New diagnostic and therapeutic options in early-stage vulvar cancer Oonk, Maaike Henrike Marije

Transcription

1 University of Groningen New diagnostic and therapeutic options in early-stage vulvar cancer Oonk, Maaike Henrike Marije IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record Publication date: 2011 Link to publication in University of Groningen/UMCG research database Citation for published version (APA): Oonk, M. H. M. (2011). New diagnostic and therapeutic options in early-stage vulvar cancer Groningen: s.n. Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons). Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Downloaded from the University of Groningen/UMCG research database (Pure): For technical reasons the number of authors shown on this cover page is limited to 10 maximum. Download date:

2 NEW DIAGNOSTIC AND THERAPEUTIC OPTIONS IN EARLY-STAGE VULVAR CANCER Maaike Henrike Marije Oonk

3 New diagnostic and therapeutic options in early-stage VULVAR CANCER Maaike Henrike Marije Oonk ISBN: Copyright 2011, M.H.M. Oonk All rights reserved. No part of this thesis may be reproduced, stored in a retrieval center of any nature, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the permission of the author. Cover design: Regina Broersma ( Lay out: Legtron Electronic Publishing, Rotterdam, the Netherlands Printing: Ipskamp Drukkers, Enschede, the Netherlands The research in this thesis (Chapter 4) was financially supported by the Dutch Cancer Society. For publication, financial support of the following institutions and companies is gratefully acknowledged: University of Groningen, University Medical Center Groningen, Dutch Cancer Society, Boehringer Ingelheim BV, Guerbet Nederland BV.

4 NEW DIAGNOSTIC AND THERAPEUTIC OPTIONS IN EARLY-STAGE VULVAR CANCER Proefschrift ter verkrijging van het doctoraat in de Medische Wetenschappen aan de Rijksuniversiteit Groningen op gezag van de Rector Magnificus, dr. E. Sterken, in het openbaar te verdedigen op woensdag 18 mei 2011 om 13:15 uur Door Maaike Henrike Marije Oonk geboren op 3 april 1977 te Almelo

5 Promotores: Prof. dr. A.G.J. van der Zee Prof. dr. H. Hollema Co-promotor: Dr. J.A. de Hullu Beoordelingscommissie: Prof. dr. I. Vergote Prof. dr. G.G. Kenter Prof. dr. L.F.A.G. Massuger

6 Contents Chapter 1 Introduction 7 Chapter 2 Prediction of lymph node metastases in vulvar cancer: a review 15 International Journal of Gynecological Cancer 2006; 16: Chapter 3 EGFR expression is associated with groin node metastases in 33 vulvar cancer, but does not improve their prediction Gynecologic Oncology 2007; 104: Chapter 4 Identification of inguinofemoral lymph node metastases by 45 methylation markers in vulvar cancer Submitted Chapter 5 Sentinel node dissection is safe in the treatment of 61 early-stage vulvar cancer Journal of Clinical Oncology 2008; 26: Chapter 6 Size of sentinel node metastasis and chances of non-sentinel node 81 involvement and survival in early-stage vulvar cancer: results from GROINSS-V, a multicentre observational study Lancet Oncology 2010; 11: Chapter 7 A comparison of quality of life between vulvar cancer patients 97 after sentinel lymph node procedure only and inguinofemoral lymphadenectomy Gynecologic Oncology 2009; 113: Chapter 8 The value of routine follow-up in patients treated for vulvar cancer 111 Cancer 2003; 98: Chapter 9 Summary and future perspectives 123

7 Chapter 10 Nederlandse Samenvatting en Toekomstvisie 133 Chapter 11 Dankwoord 143 Curriculum Vitae 147

8 Chapter 1 Introduction

9 8 Chapter 1 Introduction In the Netherlands approximately 320 women are diagnosed each year with vulvar cancer. 1 Mean age at diagnosis is about 70 years. Squamous cell carcinomas account for 80% of cases, whereas melanomas, adenocarcinomas, basal cell carcinomas and sarcomas are much less common. 2 Squamous cell carcinoma of the vulva spreads by three routes. Initial spread occurs usually to the inguinofemoral lymph nodes. The presence and the number of inguinofemoral lymph node metastases is the most important prognostic factor in vulvar cancer patients. 3 Haematogenous spread and spread by direct extension are both infrequent. Vulvar cancer has been clinically staged until Considering that palpation of the groins is inaccurate in approximately 25% of the patients, 4,5 the International Federation of Gynecology and Obstetrics (FIGO) changed the vulvar cancer staging to a surgico-pathological system in This staging system provided far better discrimination of survival between stages than the 1970 FIGO clinical staging system. 6 Different reports on prediction of lymph node metastases showed that the risk increased with a greater depth of invasion The term microinvasive vulvar cancer was postulated for lesions < 2cm in diameter and with a depth of invasion 5mm. This definition underestimated the potential of these lesions to metastasize and led to undertreatment and death of patients with normally curable disease. Hacker concluded the term microinvasive vulvar cancer should be reserved for tumors with a depth of invasion 1mm, because of the negligible risk on lymph node metastases. 11 In 1994 stage IA was added to the staging system for that reason. Table 1 and 2 show the present FIGO and TNM staging system, respectively. 12,13 FIGO and TNM classification were recently changed in order to allow for better prognostic discrimination between stages and less heterogeneity within stages. The major change is that number and morphology of the metastatic lymph nodes is taken into account now. Also former stage I and II are combined in the revised classification system. 14 Surgery is first choice in the treatment of patients with squamous cell carcinoma of the vulva. Early in the twentieth century, Basset was the first to propose en bloc dissection of the vulva and inguinofemoral lymph nodes. 15 Taussig and Way clinically applied the principles proposed by Basset and developed the radical vulvectomy with inguinofemoral lymphadenectomy en bloc (Figure 1). 16,17 This radical approach drove out simple local excision in the second half of the last century and became standard of care for a prolonged period. The rationale for this approach was the assumption that the prognosis is better after elective inguinofemoral lymphadenectomy compared to surveillance of the groins, despite the fact that only 20 30% of patients will have lymph node metastases. The morbidity of this treatment was very high. Wound breakdown, wound infections and lymphedema were of great concern and were often prolonging hospitalization. Since, many modifications of surgery have been proposed in the treatment of vulvar cancer patients, in order to reduce the complication rate: inguinofemoral lymphadenectomy through separate incisions (Figure 2), replacement of radical vulvectomy by wide local excision with excision margins of at least 1

10 Introduction 9 2 cm, abandonment of bilateral inguinofemoral lymphadenectomy in lateralized tumors 25,26 and abandonment of inguinofemoral lymphadenectomy in microinvasive tumors. 11 Table 1. FIGO Classification for Vulvar Cancer Stage Stage I Stage II Stage III Stage IV Description Tumor confined to the vulva or perineum No lymph node metastases IA: Lesions 2cm in size, with depth of invasion 1mm IB: Lesions > 2cm in size, with depth of invasion > 1mm Tumor of any size with extension to adjacent perineal structures (⅓ lower urethra, ⅓ lower vagina, anus) with negative nodes Tumor of any size with or without extension to adjacent perineal structures (⅓ lower urethra, ⅓ lower vagina, anus), with positive inguinofemoral lymph nodes: IIIA: 1 lymph node metastasis ( 5mm) or 1-2 lymph node metastases (< 5mm) IIIB: 2 lymph node metastases ( 5mm) or 3 lymph node metastases (< 5mm) IIIC: Lymph node metastases with extracapsular spread Tumor invades other regional or distant structures IVA: Tumor invades any of the following Upper urethral and/or vaginal mucosa, bladder mucosa, rectal mucosa, or fixed to the pelvic bone Fixed or ulcerated inguinofemoral lymph nodes IVB: Any distant metastases, including pelvic lymph nodes Table 2. TNM Classification for Vulvar Cancer 7 th Edition TNM T = Tumor N = Nodes M = Metastasis T1A T1B T2 T3 N0 N1a N1b N2a N2b N2c N3 M0 M1 Description Tumor confined to vulva/perineum, 2cm in size, and depth of invasion 1mm Tumor confined to vulva/perineum, > 2cm in size, or depth of invasion > 1mm Tumor extension to lower ⅓ urethra, lower ⅓ vagina or anus Tumor extension to upper ⅔ urethra, upper ⅔ vagina, bladder or rectal mucosa, or fixed to the pelvic bone No lymph nodes involved 1 2 inguinofemoral lymph node metastasis < 5mm 1 inguinofemoral lymph node metastasis 5mm 3 inguinofemoral lymph node metastases < 5mm 2 inguinofemoral lymph node metastases 5mm Inguinofemoral lymph node metastases with extracapsular spread Fixed or ulcerated inguinofemoral lymph nodes No distant metastases Any distant metastases

11 10 Chapter 1 Aim of all modifications of treatment is: to reduce morbidity of vulvar cancer treatment without compromising survival rates. The most ideal approach would be a noninvasive procedure that is able to exclude lymph node metastases with a high negative predictive value, resulting in about 70% of vulvar cancer patients who can be safely excluded for inguinofemoral lymphadenectomy. Since missing lymph node metastases is nearly always fatal, the first requirement for such a diagnostic test is a very low false-negative rate. Figure 1. Radical Vulvectomy with Bilateral Inguinofemoral Lymphadenectomy en bloc Figure 2. Radical Vulvectomy with Bilateral Inguinofemoral Lymphadenectomy Through Separate Incisions

12 Introduction 11 In Chapter 2 of this thesis a literature overview is presented on the current available noninvasive techniques for detection of lymph node metastases in vulvar cancer patients. Also the minimally invasive sentinel lymph node procedure is discussed here. Chapter 3 explores the expression of epithelial growth factor receptor (EGFR) in vulvar cancer and the relation of EGFR expression with the presence of lymph node metastases. EGFR is especially interesting, because it is known from other cancer types that EGFR expression is frequently upregulated and that aberrant EGFR activation can be associated with uncontrolled cell proliferation. Because no noninvasive techniques are available for accurate prediction of lymph node metastases in vulvar cancer patients, and until now no cell biological markers are known that show adequate correlation with the presence of lymph node metastases, the sentinel lymph node procedure was investigated in the treatment of early-stage vulvar cancer patients. The sentinel lymph node is defined as the first lymph node in a lymphatic basin that receives primary lymph flow from a tumor (Figure 3). Accurate pathologic examination of the sentinel lymph node is required since false-negative assessment of the sentinel lymph node may lead to untreatable tumor outgrowth in metastatic lymph nodes that are left behind. It is known that micrometastases are missed in a variable degree, dependent on the technique that is used to examine the lymph node. We studied the feasibility of a very sensitive method, methylation specific PCR, for detection of lymph node (micro)metastases in inguinofemoral lymph nodes of vulvar cancer patients. Chapter 4 describes the results of this pilot study. Figure 3. The Concept of the Sentinel Lymph Node Procedure

13 12 Chapter 1 Many feasibility and accuracy studies have been performed and showed that the sentinel lymph node procedure with at least the use of a radioactive tracer and eventually blue dye was highly accurate in early-stage vulvar cancer patients. The most elegant way to prove the safety and clinical utility of implementation of the sentinel lymph node procedure in early-stage vulvar cancer patients would be a clinical trial in which patients with a negative sentinel lymph node are randomized between observation and inguinofemoral lymphadenectomy. However, in such a trial at least 700 early-stage vulvar cancer patients are needed and several collaborative groups in gynecological cancer have deemed such a number of vulvar cancer patients, due to the rareness of this disease, as highly unrealistic. Chapter 5 describes the results of our international multicenter observational study GROINSS-V (GROningen INternational Study on Sentinel nodes in Vulvar cancer). Our observational study design with stopping rules had several favorable characteristics; a much lower number of patients was needed in comparison to a randomized trial, our stopping rules were based on data from literature on groin recurrence rates and in fact served as historic controls, thereby avoiding unacceptable high groin recurrence rates and finally continuous sequential analysis of events (groin recurrences) ensured relative safety for participating patients. In our view our observational study design provides at least level III evidence and should be considered as second best in rare cancers such as vulvar cancer. The accuracy of the pathologic examination of lymph nodes from vulvar cancer patients has dramatically improved by the introduction of pathologic ultrastaging in the examination of sentinel lymph nodes. However, the relevance of these by pathologic ultrastaging detected micrometastases is unknown. We reviewed all sentinel lymph node slides from the Dutch participating centers in GROINSS-V with the aim to determine the relation between size of sentinel node metastases and the presence of non-sentinel node metastases and prognosis. The results of this study are described in Chapter 6, including an update on the follow-up of these patients. The introduction of the sentinel lymph node procedure brought about a decrease in treatmentrelated morbidity as shown in Chapter 5. Our hypothesis was that this decrease in treatment-related morbidity might also improve quality of life. Therefore quality of life questionnaires were sent to patients who underwent the sentinel lymph node procedure in the near past. We compared those who underwent the sentinel lymph node procedure alone with those who underwent subsequent inguinofemoral lymphadenectomy. The results of this study are summarized in Chapter 7. The value of a strict routine follow-up program for cancer patients is often questioned. For some types of cancer, early detection leads to better treatment options and better disease-specific survival. On the contrary, for other types of cancer early detection of recurrence only shortens disease-free survival without any positive effect on disease-specific survival. Chapter 8 describes the results of our follow-up analysis in vulvar cancer patients. A summary of all results is given in Chapter 9 (Dutch version in Chapter 10). The results of the different studies are discussed with an emphasis on their clinical implications. Finally, some future perspectives are discussed here.

14 Introduction 13 References 1. IKC statistics. Available on: 2. Hacker NF: Vulvar cancer, in Berek JS, Hacker NF (eds): Practical Gynecologic Oncology (ed3). Baltimore, MD, Williams & Williams, 2000, pp Homesley HD, Bundy BN, Sedlis A, et al. Assessment of current International Federation of Gynecologic and Obstetrics staging of vulvar carcinoma relative to prognostic factors for survival (A Gynecologic Oncology Group Study). Am J Obstet Gynecol 1991;164(4): Prodratz KC, Symmonds RE, Taylor WF, Williams TJ. Carcinoma of the vulva: analysis of treatment and survival. Obstet Gynecol 1983; 61: Homesley HD, Bundy BN, Sedlis A, et al. Prognostic factors for groin node metastasis in squamous cell carcinoma of the vulva (A Gynecologic Oncology Group Study). Gynecol Oncol 1993;49: Hopkins MP, Reid GC, Johnston CM, Morley GW. A comparison of staging systems for squamous cell carcinoma of the vulva. Gynecol Oncol 1992;47(1): Magrina JF, Webb MJ, Gaffey TA, Symmonds RE. Stage I squamous cell cancer of the vulva. Am J Obstet Gynecol 1979;134(4): Andreasson B, Nyboe J. Predictive factors with reference to low-risk of metastases in squamous cell carcinoma in the vulvar region. Gynecol Oncol 1985;21: Boyce J, Fruchter RG, Kasambilides E, Nicastri AD, Sedlis A, Remy JC. Prognostic factors in carcinoma of the vulva. Gynecol Oncol 1985;20: Ross MJ, Ehrmann RL. Histologic prognosticators in stage I squamous cell carcinoma of the vulva. Obstet Gynecol 1987;70(5): Hacker NF, Berek JS, Lagasse LD, Nieberg RK, Leuchter RS. Individualization of treatment for stage I squamous cell vulvar carcinoma. Obstet Gynecol 1984;63(2): Pecorelli S. Revised FIGO staging van carcinoma of the vulva, cervix, and endometrium. Int J Gynaecol Obstet 2009;105(2): Edge SB, Byrd DR, Compton CC, et al, Eds. AJCC cancer staging manual, 7 th ed, Springer, New York. 14. Hacker NF. Revised FIGO staging for carcinoma of the vulva. Int J Gynaecol Obstet 2009 ;105(2) : Basset A. Traitement chirurgical opératoire de l épithelioma primitif du clitoris: Indications, technique, résultats. Rev Chir 1912;46: Taussig FJ. Cancer of the vulva: an analysis of 155 cases ( ). Am J Obstet Gynecol 1949;40: Way S. Carcinoma of the vulva. Am J Obstet Gynecol 1960;79: Byron RL, Lamb EJ, Yonemoto RH, Kase S. Radical inguinal node dissection in the treatment of cancer. Surg Gynecol Obstet 1962;114: Byron RL, Mishell DR, Yonemoto RH. The surgical treatment of invasive carcinoma of the vulva. Surg Gynecol Obstet 1965;121(6): Hacker NF, Leuchter RS, Berek JS, Castaldo TW, Lagasse LD. Radical vulvectomy and bilateral inguinal lymphadenectomy through separate incisions. Obstet Gynecol 1981;58: Hopkins MP, Reid GC, Morley GW. Radical vulvectomy. The decision for the incision. Cancer 1993;72(3): Heaps JM, Fu YS, Montz FJ, Hacker NF, Berek JS. Surgical-pathological variables predictive of local recurrence in squamous cell carcinoma of the vulva. Gynecol Oncol 1990;38(3):

15 14 Chapter Hacker NF, van der Velden J. Conservative management of early vulvar cancer. Cancer 1993;71(4 Suppl): De Hullu JA, Hollema H, Lolkema S, et al. Vulvar carcinoma. The price of less radical surgery. Cancer 2002;95(11): Iversen T, Aas M. Lymph drainage from the vulva. Gynecol Oncol 1983;16(2): Burger MPM, Hollema H, Bouma J. The side of groin node metastases in unilateral vulvar carcinoma. Int J Gynecol Cancer 1996;6:

16 Chapter 2 Prediction of lymph node metastases in vulvar cancer: a review Maaike H.M. Oonk Harry Hollema Joanne A. de Hullu Ate G.J. van der Zee International Journal of Gynecological Cancer 2006; 16:

17 16 Chapter 2 Abstract The aim of this study was to review the literature on currently available non- and minimally-invasive diagnostic methods and analysis of primary tumor characteristics for prediction of inguinofemoral lymph node metastases in patients with primary squamous cell carcinoma of the vulva. We used the English language literature in Pubmed and reference lists from selected articles. Search terms included vulvar carcinoma, prognosis, lymph node metastases, ultrasound, computer tomography, magnetic resonance imaging, positron emission tomography, and sentinel lymph node. No study type restrictions were imposed. Currently no noninvasive imaging techniques exist that are able to predict lymph node metastases with a high enough negative predictive value. A depth of invasion 1mm is the only histopathologic parameter that can exclude patients for inguinofemoral lymphadenectomy. No other clinicopathologic parameter allows exclusion of lymph node metastases with a high enough negative predictive value. The minimally invasive sentinel node procedure is a promising technique for selecting patients for inguinofemoral lymphadenectomy, but its safety has not been proven yet.

18 Prediction of lymph node metastases in vulvar cancer: A review 17 Introduction Vulvar carcinoma is a rare disease, mainly affecting elderly women. The majority of vulvar carcinomas are squamous cell carcinomas (SCC), which account for 90% of all vulvar carcinomas. 1 The pattern of dissemination of SCC is predominantly lymphogenic to the inguinofemoral lymph nodes. Pelvic and distant metastases are rare and mostly fatal. Radical vulvectomy with bilateral inguinofemoral lymphadenectomy en bloc has been the standard treatment for years in patients with SCC of the vulva with a depth of invasion > 1mm. The morbidity of this surgical treatment was significant: frequent wound breakdown, wound infections, lymphocysts, lymphedema and impressive psychosexual consequences, especially in younger patients. 2,3 These complications often prolong hospitalization. The most common complication is chronic lymph edema, which has been reported in up to 69% of patients. 3 In a study by de Hullu et al., it was shown that after a median follow-up of 118 months, 47% of patients treated for vulvar carcinoma still experienced either severe pain and / or severe lymphedema in the legs. 4 In the past decades, changes toward less radical treatment have been made in the standard treatment of vulvar carcinoma with depth of invasion > 1mm: wide local excision and ipsi- or bilateral inguinofemoral lymphadenectomy via separate incisions. However, also in early-stage disease (T1 or T2 tumors with clinically unsuspicious inguinofemoral lymph nodes), all patients undergo inguinofemoral lymphadenectomy based on the finding that 20 30% will have lymph node metastases. The rationale for this approach is the assumption that prognosis is superior after elective inguinofemoral lymphadenectomy compared with surveillance of the groins. This assumption is primarily based on clinical observations that, by the time groin recurrences manifest themselves, curative therapy frequently is no longer possible, resulting in poor regional control with dismal sequels for the patient. The prognosis for groin recurrences of vulvar carcinoma is very poor, most patients die of disease. 5 No randomized trials have addressed the issue of elective versus delayed inguinofemoral lymphadenectomy in patients with vulvar carcinoma and clinically unsuspicious lymph nodes. Doubt remains, however, about the therapeutic role of inguinofemoral lymphadenectomy in node-negative patients. Of all patients with clinically unsuspicious inguinofemoral lymph nodes, only 20 30% have inguinofemoral lymph node metastases. The other 70 80% will probably not benefit from the lymphadenectomy but are at risk for its significant morbidity. The morbidity of inguinofemoral lymphadenectomy and the absence of inguinofemoral lymph node metastases in the majority of patients with early-stage disease are compelling arguments to develop methods to define a low-risk group. In this low-risk group of selected vulvar carcinoma patients, inguinofemoral lymphadenectomy might be omitted. This review focuses on currently available non- and minimally invasive diagnostic methods and clinicopathologic analysis of primary tumor characteristics for prediction of inguinofemoral lymph node metastases in patients with primary SCC of the vulva.

19 18 Chapter 2 Data sources Relevant studies were identified by a computer search of English language abstracts in the Pubmed database until May We searched the medical literature using combinations of the following heading terms: vulvar carcinoma, SCC, prognosis, lymph node metastases, ultrasound, computer tomography (CT), magnetic resonance imaging (MRI), positron emission tomography (PET), and sentinel lymph node. Outcomes of interest were articles on identification / prediction of lymph node metastases. With the button related articles in Pubmed, papers were identified that were not found in the primary search. Reference lists from selected articles were used in order to identify missing relevant publications. Noninvasive methods Palpation Palpation of the inguinofemoral lymph nodes is a cheap and simple approach for determining the inguinofemoral lymph node status but is impaired by several conditions: obesity of the patients, small size of the metastases, metastases located deep in the subcutaneous fat tissue, and scar tissue due to former surgery and / or radiation. Initially, the staging system of vulvar carcinoma was based on pretreatment evaluation according to the tumor node metastasis (TNM) classification (Table 1). 6 The information of the inguinofemoral lymph node status was obtained by palpation of the groins (N0: no lymph nodes, N1: non-suspicious lymph nodes and N2: suspicious lymph nodes). In all three groups of the clinical groin classification, errors were made compared to the final pathological review of the lymph nodes after surgery. The overall error rate was nearly 25%. 2 These disappointing results were confirmed by the observation that intraoperative lymph node palpation had positive and negative predictive values of only 56% an 89%, respectively, even when done by experienced gynecologic oncologists. 7 Comparable results were recently reported by Katz et al. 8 From these observations, it appears that the accuracy of palpation is not sufficient to select patients in whom an inguinofemoral lymphadenectomy can be omitted. Ultrasound Ultrasound is safe, noninvasive, and highly acceptable to patients. Ultrasound has been used to assess nodal status in head and neck tumors, breast carcinoma, upper aerodigestive tract malignancy, and cervical carcinoma The accuracy of ultrasound ranges between 67 and 95%, depending on the different parameters examined (short axis diameter, long axis diameter, long axis / short axis ratio, shape, nodal vascular pattern, echogenic pattern, and regularity of nodal outline). 16

20 Prediction of lymph node metastases in vulvar cancer: A review 19 Table 1. Tumor Node Metastasis (TNM) Classification of Vulvar Cancer TNM T1 T2 T3 T4 N0 N1 N2 N3 M0 M1a M1b Description Tumor confined to the vulva, 2cm in largest diameter Tumor confined to the vulva, > 2cm in diameter Tumor of any size with adjacent spread to the urethra and / or vagina and / or perineum and / or anus Tumor of any size infiltrating the bladder mucosa and / or the rectal mucosa or including the upper part of the urethral mucosa and / or fixed to the bone No nodes palpable Nodes palpable in either groin, not enlarged, mobile (not clinically suspect for metastases) Nodes palpable in either or both groins, enlarged, firm and mobile (clinically suspect for metastases) Fixed or ulcerated nodes No clinical metastases Palpable deep pelvic lymph nodes Other distant metastases Mäkela et al. compared palpation and ultrasound for the detection of metastatic inguinofemoral lymph nodes in 25 patients with vulvar carcinoma. The results of ultrasound were significantly better than those of palpation, with a negative predictive value of 94% for ultrasound. 17 Comparable results were obtained for different lymph node basins in patients with cutaneous melanoma in whom ultrasound also appeared superior to palpation. 18,19 However, especially due to the low number of vulvar carcinoma patients included in studies, there is currently too little evidence for excluding lymph node metastases with ultrasound alone. Ultrasound has also been combined with ultrasound-guided fine-needle aspiration cytology (FNAC) to improve accuracy by Moskovic et al. 20 This combined technique could accurately predict nodal status in the majority of cases. Falsely negative cytology occurred when the metastatic focus was 3mm (two false-negative results out of 40 groins). Hall et al., who extended the study of Moskovic et al. to 44 patients, concluded that the combination of ultrasound and FNAC provides a sensitive and specific tool for preoperative assessment (sensitivity 93%, specificity 100%). 21 Another possibility is to combine two sonography criteria as the definition for suspicious lymph nodes (short axis diameter > 8mm and long axis / short axis ratio [L / S ratio] 2) and to sample suspicious lymph nodes by ultrasound-guided FNAC. 16 Even in the hands of experienced investigators, however, micrometastases will inevitably be missed with this method. Another drawback is that the adequacy of this technique is highly dependent on the skills of the radiologist and that radiologists with these skills are rare. For an overview of abovementioned studies, see Table 2. In conclusion, so far, the negative predictive value of ultrasound in patients with vulvar carcinoma (to exclude lymph node metastases) may be not high enough to allow its use as a criterion to omit inguinofemoral lymphadenectomy in selected patients. In combination with FNAC, ultrasound may

21 20 Chapter 2 become of more importance in the near future, possibly in selecting patients for sentinel node detection or in the follow-up of vulvar carcinoma patients after a negative sentinel node. Table 2. Overview of Studies on Determining Inguinal Nodal Status with Ultrasound in Patients with Vulvar Cancer Author Number of patients Method Sensitivity Specificity Makela Ultrasound 82% 87% Preoperative palpation 9% 100% Intraoperative palpation 55% 90% Abang Mohammed Ultrasound 83% 88% Moskovic Ultrasound 85% 83% Cytology 58% 100% Combined 83% 82% Hall Ultrasound 86% 96% Cytology 75% 100% Combined 93% 100% CT and MRI The only experience with CT in patients with vulvar carcinoma was the measurement of the distance in centimeters between the skin and the underlying inguinofemoral lymph nodes for planning of groin radiation. 22,23 No other literature is available on the diagnostic value of CT for detection of inguinofemoral lymph node metastases in patients with vulvar carcinoma. MRI is being used more often in the treatment of gynecological malignancies because MRI offers an opportunity to stage both the primary tumor and the regional lymph nodes. Recently, two studies on identification of inguinal lymph node metastases in vulvar carcinoma by MRI were published. MRI was highly specific for the detection of nodal involvement in 22 patients with vulvar carcinoma (for superficial nodes a specificity of 97% and for deep inguinal nodes 100%). However, the sensitivity for both groups of inguinal nodes was low (respectively, 40% and 50%), which means a high falsenegative rate. 24 Another pilot study showed a specificity of 91% and a sensitivity of 89% for detection of inguinal lymph node metastases by MRI in ten patients with vulvar carcinoma. 25 An explanation for the difference in sensitivity between both studies might be that the parameters used to define a lymph node as abnormal were more extensive in the second study. These preliminary data indicate that MRI may have some value for selecting women in whom the risk of nodal metastases is very

22 Prediction of lymph node metastases in vulvar cancer: A review 21 low from those in whom the risk is relatively high, but further experience is needed before firm conclusions can be drawn. Positron Emission Tomography The clinical use of PET for detection and staging of malignant tumors is rapidly increasing in different malignancies. In the future, PET may obtain a role in diagnostic imaging as well as in monitoring of therapeutic interventions in specific malignancies. 26,27 In general, limitations exist with regard to false-positive results. Acute or chronic inflammation as well as aspecific reactions following radiotherapy may mimic tumor tissue. 28 In cutaneous melanoma, the accuracy of detecting lymph node metastases was studied in 56 lymph node basins: PET detected 100% of metastases > 10mm; 83% of metastases 6 10mm; and 23% of metastases 5mm. 29 In stage I and II melanoma, PET was positive in only one of eight patients with a positive sentinel node, yielding a sensitivity of 13%. 30 Comparable results (sensitivity 15%) were obtained in melanoma patients with Breslow thickness > 1.0mm and no palpable lymph nodes. 31 Apparently, PET cannot detect subclinical microscopic disease with acceptable sensitivity. In SCC of the vulva, sensitivity and specificity of detection of inguinofemoral lymph node metastases by PET using L-[11C]-tyrosine as a tracer were 62% and 89% per groin. 32 Another study with fluorodeoxyglucose showed similar results (sensitivity 67%, specificity 95%). 33 Both studies concluded that PET is not able to adequately predict or exclude presence of inguinofemoral lymph node metastases in patients with SCC of the vulva. Minimally invasive methods Sentinel lymph node procedure The lack of accurate noninvasive techniques to detect inguinofemoral lymph node metastases encouraged the use of the minimally invasive sentinel lymph node procedure. The sentinel lymph node is defined as the first draining lymph node, and the pathology of the sentinel lymph node is considered to be representative for the non-sentinel lymph nodes. It implies that a negative sentinel lymph node predicts the absence of tumor metastases in the other non-sentinel lymph nodes. 34 The sentinel node procedure is already generally accepted as an accurate method of staging in patients with cutaneous melanoma and breast carcinoma, but large randomized trials are still under way to prove its safety and clinical utility. In 1994, the sentinel node procedure with only blue dye was first described in patients with vulvar carcinoma. 38 In subsequent pilot studies, the sentinel node procedure with the combined technique (preoperative lymphoscintigram with 99m Technetium-labeled nanocolloid and blue dye) proved to be feasible in patients with primary vulvar carcinoma

23 22 Chapter 2 In 1999, Ansink et al. published a report on the first large series of patients studied to investigate the negative predictive value of the sentinel lymph node procedure in vulvar carcinoma. 42 In this multicenter study, only blue dye was used. Shortly after this publication, DeCicco et al. reported no false-negative sentinel lymph node in 37 patients with preoperative lymphoscintigram and gamma probe-guided surgery. 43 De Hullu et al. used the combined technique and reported no false-negative sentinel lymph node in 59 patients. 44 See Table 3 for an overview of the studies investigating the negative predictive value of a negative sentinel lymph node Only studies with more than 20 patients are included. In all these studies, the sentinel node procedure was followed by inguinofemoral lymphadenectomy. The results of these studies show an identification rate of nearly 100% when preoperative lymphoscintigram and gamma probe-guided surgery have been performed. Table 3. Studies on the Accuracy of the Sentinel Lymph Node Procedure in Vulvar Cancer Author Number of patients Blue dye Tracer Lympho- Scintigram Identification rate False negative patients Ansink Yes No No 56% 2 DeCicco No Yes Yes 100% 0 De Hullu Yes Yes Yes 100% 0 Levenback Yes No No 88% 0 Sliutz Yes # Yes Yes 100% 0 Moore Yes Yes Yes 100% 0 Puig-Tintore Yes Yes Yes 96% 0 # in only 8 of the 26 patients, blue dye was used An extended study is ongoing in the United States (GOG-173) to further determine the accuracy of the sentinel node procedure in vulvar carcinoma, by performing the sentinel node procedure with isosulfan blue, followed by standard complete lymphadenectomy. Pre- or intraoperative lymphoscintigraphy is optional. In conclusion, based on 169 patients in five studies, the sentinel node procedure with at least the preoperative use of a radioactive tracer, eventually combined with blue dye, is highly accurate in predicting lymph node metastases, with a negative predictive value of a negative sentinel node of nearly 100%. 43,44,46-48 However, safety and clinical utility need still to be proven in large clinical trials, in which full lymphadenectomy is no longer performed after a negative sentinel lymph node.

24 Prediction of lymph node metastases in vulvar cancer: A review 23 Primary tumor characteristics Histopathologic parameters In the past, many studies on histopathologic parameters of the primary tumor have been performed in vulvar carcinoma with the aim to define a group with low risk on lymph node metastases. Initially, several authors considered, by analogy with cervical carcinoma, depth of invasion 5mm as microinvasive, with an extremely low risk on inguinofemoral lymph node metastases, but several more recent studies showed that the percentage of patients with inguinofemoral lymph node metastases rises steadily with each millimeter of invasion See Table 4 for an overview of five studies, which registered the percentage of inguinofemoral lymph node metastases related to depth of invasion per millimeter in patients with T1 ( 2cm) tumors. 50,52,55-57 Table 4. Lymph Node Status in Patients with T1 Tumors in Relation to Depth of Invasion Depth of invasion Number of patients Number of lymph node metastases 1mm (0%) 1.1 2mm 92 5 (5.4%) 2.1 3mm 79 7 (8.9%) 3.1 4mm (25%) 4.1 5mm 25 7 (28%) > 5mm (32%) Total (11.1%) Only in patients with SCC of the vulva with 1mm of invasion (stage IA) 54 is it now generally accepted to omit inguinofemoral lymphadenectomy based on the low risk of inguinofemoral lymph node metastases, as shown in Table 4. Only occasional reports have been published of patients with stage IA vulvar carcinoma with inguinofemoral lymph node metastases Moreover, the long-term results of these patients who underwent only local treatment without inguinofemoral lymphadenectomy are excellent. 61 Univariate analyses of other histopathologic variables, such as tumor diameter, presence of lymphovascular invasion, grade of the tumor, growth pattern (pushing versus spray-like), concurrent vulvar intraepithelial neoplasia, site of the lesion, and uni- or multifocality of the tumor, have been performed in different studies, with conflicting results regarding their ability to predict inguinofemoral lymph node metastases. Several authors performed multivariate analyses to define which clinicopathologic factors, apart from depth of invasion, are independently related to inguinofemoral

25 24 Chapter 2 lymph node metastases. See Table 5 for an overview of these studies. 56,57,62,63 However, it appears to be impossible to define a real low-risk group based on these factors. Apart from this inability, it remains to be decided what exactly low risk means. When debating a possible acceptable falsenegative rate for sentinel node detection in vulvar carcinoma, the Gynecologic Cancer Group of the European Organization for Research and Treatment of Cancer (EORTC) decided that in light of the significant decrease in morbidity by omitting elective lymphadenectomy, a maximum increase of groin recurrences (i.e., a false negative rate) of 6% might be acceptable. In analog to this discussion, this would imply that primary tumor characteristics should be able to define a group of patients with a risk of metastases < 6%. Within this point of view, inguinofemoral lymphadenectomy could also be omitted in patients with vulvar carcinoma with a depth of invasion of mm (a risk on lymph node metastases of 5.4%; see Table 4). In conclusion, based on conventional histopathologic primary tumor characteristics, there is no possibility of defining low-risk groups of patients with invasive SCC for whom it is safe to omit inguinofemoral lymphadenectomy, other than patients with tumors with a depth of invasion 1mm. Table 5. Overview of Studies with Multivariate Analyses Concerning Relation Between Clinicopathologic Factors and Inguinofemoral Lymph Node Metastases Author Number of patients Variables independently related to lymph node metastases Homesley Clinical node status, lymphovascular invasion, tumor differentiation (GOG), age, depth of invasion Andreasson Tumor site, depth of invasion, tumor size Boyce Clinical node status, depth of invasion, growth pattern Smyczek-Gargya Clinical node status, tumor size, depth of invasion, tumor grading Cell biologic parameters Apart from the more conventional clinical and histopathologic parameters described above, more recently discovered cell biologic parameters have also been evaluated for their possible value in predicting lymph node metastases in vulvar carcinoma. In the following paragraphs, we will summarize those cell biologic parameters that were analyzed for their usefulness in predicting lymph node metastases. DNA ploidy status is a prognostic factor for survival in vulvar melanoma 64 but appears to have no prognostic value in SCC of the vulva Tumors with positive inguinal lymph nodes appeared to be associated with a higher rate of aneuploidy compared to tumors with negative inguinal nodes, but the difference was not significant. 67 However, a significantly higher S-phase fraction (proliferation

26 Prediction of lymph node metastases in vulvar cancer: A review 25 index) was found in tumors from patients with lymph node metastases when compared to tumors from patients with negative lymph nodes (median value in patients with lymph node metastases [n=13] 18.2, in patients with negative lymph nodes [n=26] 8.9). 68 In 1976 the role of human papillomavirus (HPV) was first proposed in the development of carcinoma of the female genital tract. 69 The virally (HPV) encoded oncoproteins E6 and E7 can form a complex with protein products of tumor suppressor genes, E6 binding to the p53 protein and E7 binding to the retinoblastoma gene product. This may lead to loss of the normal negative growth control. The relation between HPV infection and vulvar carcinoma is not as clear as it is in cervical carcinoma. About one-third of SCC of the vulva is associated with high-risk HPV infection, particularly HPV types 16 and Data on the relation between the HPV status and the presence of lymph node metastases are conflicting. One study with a small number of patients reported that the rate of metastases in HPV-positive vulvar carcinomas (3 / 7, 43%) was higher than for the HPV-negative carcinomas (4 / 16, 25%). 71 Another larger study, however, found no relation between the presence of HPV and lymph node metastases. 72 Epidermal growth factor receptor (EGFR) activation plays a key role in cell adhesion, cell locomotion, cell survival, invasion, and angiogenesis, which results in modulation of tumor progression. EGFR expression was analyzed in benign vulvar epithelium, primary vulvar carcinoma, and in groin node metastases. There was a progressive increase in EGFR expression from benign vulvar epithelium to primary malignant vulvar tissue to metastatic lesions within the same patient. Increased expression of EGFR in the primary vulvar malignancy was significantly associated with lymph node metastases and decreased patient survival in a study of 61 patients (an increased EGFR level in tumor tissue identified 71% of lymph node metastases). 73 Twenty-nine percent of the lymph node metastases could, however, not be identified by an increased EGFR level, so EGFR level analysis is not sensitive enough to exclude patients for complete lymphadenectomy. Overexpression of the tumor suppressor gene p53 in squamous cell vulvar carcinomas varies from 39% to 61% in different studies No relation was found between overexpression of p53 in the primary tumor and lymph node metastases in patients with vulvar carcinoma. 70,76,77 The tumor suppressor gene retinoblastoma (Rb) has also been investigated in vulvar carcinoma. Loss of Rb immunostaining was observed in about 20% of vulvar SCC but did not seem to play any role for prognosis. 79 The loss of Rb expression increased from FIGO stage I to IV, suggesting that Rb may play a role in tumor progression. 81 Another study showed progressive decrease of Rb2 / p130 (a retinoblastoma-related protein) expression from non-neoplastic epithelial alterations through intraepithelial neoplasia to invasive vulvar carcinoma, suggesting a role for this tumor suppressor gene in vulvar carcinogenesis. There was no correlation between Rb2 / p130 expression and the presence of lymph node metastases. 82 The presence of the HER-2 / neu immunopositivity was found to be associated with lymph node metastases in early SCC of the vulva as 69% of the node-positive tumors was HER-2 / neu positive

27 26 Chapter 2 against 36% of the node-negative tumors (see Table 6). 69 Assessment of the proliferation-associated markers Ki-67 and Ag-NOR did not contribute to the prediction of lymph node metastases. 83 Table 6. Overview of Sensitivity, Specificity, Positive Predictive Value (PPV), and Negative Predictive Value (NPV) of Cell Biologic Parameters that Showed a Significant Correlation with Lymph Node Metastases Sensitivity (%) Specificity (%) PPV (%) NPV (%) HER-2 / neu (N=39) HPV (N=23) EGF-R (N=61) Studies on pro- and anti-apoptotic parameters in vulvar carcinoma are limited. While Bcl-2 overexpression leads to blocking of apoptosis, overexpression of Bax promotes apoptosis. In a study of 72 vulvar carcinoma patients, immunohistological staining for Bcl-2 showed that Bcl-2 positivity is associated with lymph node metastases (p=0.03). 84 In contrast, in a preliminary study of 28 patients by Hefler et al. no correlation was found between Bcl-2 expression and lymph node involvement. 85 Cyclooxygenase-2 (COX-2) is one of the two isoforms of the key enzyme in the biosynthesis of prostaglandins and (in critical steps of tumor progression) overexpression has been associated with increased metastatic potential. The use of the ratio between COX-2 in the tumor cells and COX-2 in the stroma seems to correlate more effectively with vulvar tumor extension and metastatic lymph node involvement than the levels of COX-2 in tumor cells. In a study of 36 patients with invasive vulvar carcinoma, a higher tumor / stroma COX-2 ratio was observed in cases with lymph node metastases compared to cases without lymph node metastases (p=0.037). It remains to be verified whether assessment of COX-2 status can be useful in the prognostic characterization of vulvar carcinoma. 86 Comparative genomic hybridization (CGH) is a powerful tool in the study of carcinoma that has quickly gained recognition for its ability to scan the entire genome for DNA copy number changes that are mapped to chromosome regions. Chromosome regions gained or lost may harbor oncogenes or tumor suppressor genes, respectively. Loss of 4p13-pter, and concomitant loss of 3p and gain in 3q, is the most frequent genetic alteration in SCC of the vulva. 87 Another study found that gain of 3q was more common in HPV-positive carcinomas compared to HPV-negative carcinomas, while chromosome 8q was more frequently gained in HPV-negative carcinomas. Chromosome arms 3p and 11q were frequently lost in both categories of vulvar carcinoma. 88 The different genetic changes indicate that there are at least two different types of vulvar carcinoma. However, all studies on CGH included only a limited number of patients. To our knowledge, no articles are available that study the relationship between lymph node metastases and chromosome regions gained or lost.

28 Prediction of lymph node metastases in vulvar cancer: A review 27 Discussion In SCC of the vulva, in particular, a very high negative predictive value for a diagnostic method is necessary to safely omit inguinofemoral lymphadenectomy. With respect to the groin, uni- or bilateral inguinofemoral lymphadenectomy is currently the standard treatment. Because of the high morbidity and complication rate of this surgical procedure, there is a trend towards less radical treatment. In 50 60% of the patients, the inguinofemoral lymphadenectomy may only be a diagnostic method to exclude lymph node metastases. It has to be kept in mind that missing lymph node metastases has significant consequences because of the high mortality. This points to the clinical relevance of a reliable non- or minimally invasive technique for detection of lymph node metastases. No noninvasive procedure (palpation of the groins, ultrasound, MRI, CT, PET) is available that can predict lymph node metastases with a negative predictive value that is high enough to safely omit inguinofemoral lymphadenectomy. Ultrasound combined with fine-needle aspiration shows promising results, but small metastases will eventually be missed by this procedure. However, this method may be helpful preoperatively to exclude patients from the sentinel lymph node procedure by detecting lymph node metastases, leading to immediate inguinofemoral lymphadenectomy. Studies on different histopathologic parameters showed that inguinofemoral lymphadenectomy can be omitted in tumors with a depth of invasion less than 1mm. No other histopathologic parameters are able to define low-risk groups of patients with invasive SCC of the vulva. Data on cell biologic parameters are inconclusive, conflicting and the number of patients analyzed is often too small. Some cell biologic parameters showed significant correlation with lymph node metastases: S-phase fraction, HER-2 / neu, HPV positivity, EGFR overexpression, Bcl-2, and COX-2. We calculated the sensitivity, specificity, negative predictive value, and positive predictive value of the parameters of which the necessary data were available in literature (HER-2 / neu, HPV and EGFR; see Table 6). The only parameter with a high enough negative predictive value to exclude lymph node metastases in patients with vulvar carcinoma was EGFR overexpression. However, this parameter had a very low specificity, so very few patients without lymph node metastases will be identified by this technique. Also, the number of patients in these three studies was small (respectively 39, 23, and 61 patients). Further research is necessary to define the consequences of these positive results. Until now, the cell biologic parameters investigated have not been sensitive enough to exclude patients for inguinofemoral lymphadenectomy. At present, the minimally invasive sentinel lymph node procedure is the most promising diagnostic tool for assessment of lymph node status, but its safety is still to be proven. Extended studies are ongoing to further determine the accuracy and safety of the sentinel node procedure in vulvar carcinoma. For now, wide local excision with uni- or bilateral inguinofemoral lymphadenectomy is still the standard treatment for patients with SCC of the vulva with a depth of invasion > 1mm.

29 28 Chapter 2 References 1. Hacker NF. Vulvar Cancer. In: Berek JS, Hacker NF, eds. Practical Gynecologic Oncology. 3 rd ed. Baltimore: Williams and Wilkins, 2000: Prodratz KC, Symmonds RE, Taylor WF, Williams TJ. Carcinoma of the vulva: analysis of treatment and survival. Obstet Gynecol 1983;61: Andersen BL, Hacker NF. Psychosexual adjustment after vulvar surgery. Obstet Gynecol 1983;62: De Hullu JA, Ansink AC, Tymstra T, Van der Zee AGJ. What doctors and patients think about false-negative sentinel lymph nodes in vulvar cancer. J Psychosom Obstet Gynaecol 2001;22: Stehman FB, Bundy BN, Ball H, Clarke-Pearson DL. Sites of failure and times to failure in carcinoma of the vulva treated concervatively: a Gynecologic Oncology Group study. Am J Obstet Gynecol 1996;174: Greene FL, Page DL, Fleming I, et al. Vulva. In: American Joint Committee on Cancer: AJCC cancer staging manual, 6 th edn. New York: Springer, 2002; Arango HA, Hoffman MS, Roberts WS, et al. Accuracy of lymph node palpation to determine need for lymphadenectomy in gynecologic malignancies. Obstet Gynecol 2000;95: Katz A, Eifel PJ, Jhingran A, Levenback CF. The role of radiation therapy in preventing regional recurrences of invasive squamous cell carcinoma of the vulva. Int J Radiat Oncol Biol Phys 2003;57: Ahuja A, Ying M, Yang WT, et al. The use of sonography in differentiating cervical lymphomatous lymph nodes from cervical metastatic lymph nodes. Clin Radiol 1996;51: Griffith JF, Chan AC, Ahuja AT, et al. Neck ultrasound in staging squamous oesophagal carcinoma a high yield technique. Clin Radiol 2000;55: Vijaykumar DK, Prahlad S, Chaturvedi HK. Ultrasound evaluation of pelvic nodes in cervical carcinoma. Eur J Gynaecol Oncol 1995;16: Cheung TH, Lo WK, Yu MY, et al. Extended experience in the use of laparoscopic ultrasound to detect pelvic nodal metastasis in patients with cervical carcinoma. Gynecol Oncol 2004;92: Couto D, Dias M, Goncalo M, et al. Diagnostic value of ultrasound and color Doppler in identifying axillary lymph node metastases in patients with breast cancer. Eur J Gynaecol Oncol 2004;25: Selinko VL, Middleton LP, Dempsey PJ. Role of sonography in diagnosing and staging invasive lobular carcinoma. J Clin Ultrasound 2004;32: Vaidya JS, Vyas JJ, Thakur MH, et al. Role of ultrasonography to detect axillary node involvement in operable breast cancer. Eur J Surg Oncol 1996;22: Abang Mohammed DK, Uberoi R, de B Lopes A, Monaghan JM. Inguinal node status by ultrasound in vulva cancer. Gynecol Oncol 2000;77: Makela PJ, Leminen A, Kaariainen M, Lehtovirta P. Pretreatment sonographic evaluation of inguinal lymph nodes in patients with vulvar malignancy. J Ultrasound Med 1993;12: Binder M, Kittler H, Steiner A, et al. Lymph node sonography versus palpation for detecting recurrent disease in patients with malignant melanoma. Eur J Cancer 1997;33: Blum A, Schlagenhauff B, Stroebel W, et al. Ultrasound examination of regional lymph nodes significantly improves early detection of locoregional metastases during the follow-up of patients with cutaneous melanoma. Cancer 2000;88: Moskovic EC, Shephard JH, Barton DPJ, et al. The role of high resolution ultrasound with guided cytology of groin lymph nodes in the management of squamous cell carcinoma of the vulva: a pilot study. Br J Obstet Gynaecol 1999;106: Hall TB, Barton DP, Trott PA, et al. The role of ultrasound-guided cytology of groin lymph nodes in the management of squamous cell carcinoma of the vulva: 5-year experience in 44 patients. Clin Radiol 2003;58:

30 Prediction of lymph node metastases in vulvar cancer: A review McCall AR, Olson MC, Potkul RK. The variation of inguinal lymph node depth in adult women and its importance in planning elective irradiation for vulvar cancer. Cancer 1995;75: Koh WJ, Chiu M, Stelzer KJ, et al. Femoral vessel depth and the implications for groin node radiation. Int J Radiat Oncol Biol Phys 1993;15: Sohaib SA, Richards PS, Ind T, et al. MR imaging of carcinoma of the vulva. Am J Roentgenol 2002;178: Hawnaur JM, Reynolds K, Wilson G, et al. Identification of inguinal lymph node metastases from vulval carcinoma by magnetic resonance imaging: an initial report. Clin Radiol 2002;57: Weber WA, Avril N, Schwaiger M. Relevance of positron emission tomography (PET) in oncology. Strahlenther Onkol 1999;175: Greven KM, Williams DW 3 rd, McGuirt WF, et al. Serial positron emission tomography scans following radiation therapy of patients with head and neck cancer. Head Neck 2001;23: Strauss LG. Sensitivity and specificity of positron emission tomography (PET) for the diagnosis of lymph node metastases. Recent Results Cancer Res 2000;157: Crippa F, Leutner M, Belli F, et al. Which kinds of lymph node metastases can FDG PET detect? A clinical study in melanoma. J Nucl Med 2000;41: Fink AM, Holle-Robatsch S, Herzog N, et al. Positron emission tomography is not useful in detecting metastasis in the sentinel lymph node in patients with primary malignant melanoma stage I and II. Melanoma Res 2004;14: Havenga K, Cobben DCP, Oyen WJG, et al. Fluorodeoxyglucose-positron emission tomography and sentinel lymph node biopsy in staging primary cutaneous melanoma. Eur J Surg Oncol 2003;29: De Hullu JA, Pruim J, Qué TH, et al. Noninvasive detection of inguinofemoral lymph node metastases in squamous cell cancer of the vulva by L-[1-11 C]-tyrosine positron emission tomography. Int J Gynecol Cancer 1999;9: Cohn DE, Dehdashti F, Gibb RK, et al. Prospective evaluation of positron emission tomography for the detection of groin node metastases from vulvar cancer. Gynecol Oncol 2002;85: Morton DL, Wen DR, Wong JH, et al. Technical details of intraoperative lymphatic mapping for early-stage melanoma. Arch Surg 1992;127: Gershenwald JE, Colome MI, Lee JE, et al. Patterns of recurrence following a negative sentinel lymph node biopsy in 243 patients with stage I or II melanoma. J Clin Oncol 1998;16: Roumen RM, Kuijt GP, Liem IH, van Beek MW. Treatment of 100 patients with sentinel node-negative breast cancer without further axillary dissection. Br J Surg 2001;88: Cody HS 3rd. Current surgical management of breast cancer. Curr Opin Obstet Gynecol 2002;14: Levenback C, Burke TW, Gershenwald DM, et al. Intraoperative lymphatic mapping for vulvar cancer. Obstet Gynecol 1994;84: Levenback C, Burke TW, Morris M, et al. Potential allpications of intraoperative lymphatic mapping in vulvar cancer. Gynecol Oncol 1995;59: De Hullu JA, Doting E, Piers DA, et al. Sentinel lymph node identification with 99m Technetium labeled nanocolloid in squamous cell cancer of the vulva. J Nucl Med 1998;39: Rodier JF, Janser JC, Routiot T, et al. Sentinel node biopsy in vulvar malignancies: A preliminary feasibility study. Oncol Rep 1999;6: Ansink AC, Sie-Go DM, van der Velden J, et al. Identification of sentinel lymph nodes in vulvar carcinoma patients with the aid of a patent blue V injection. Cancer 1999;86: De Cicco C, Sideri M, Bartolomei M, et al. Sentinel node biopsy in early vulvar cancer. Brit J Cancer 2000;82:

31 30 Chapter De Hullu JA, Hollema H, Piers DA, et al. Sentinel lymph node procedure is highly accurate in squamous cell carcinoma of the vulva. J Clin Oncol 2000; 18: Levenback C, Coleman RL, Burke TW, et al. Intraoperative lymphatic mapping and sentinel node identification with blue dye in patients with vulvar cancer. Gynecol Oncol 2001;83: Sliutz G, Reinthaller A, Lantzsch T, et al. Lymphatic mapping of sentinel nodes in early vulvar cancer. Gynecol Oncol 2002;84: Moore RG, DePasquale SE, Steinhoff MM, et al. Sentinel node identification and the ability to detect metastatic tumor to inguinal lymph nodes in squamous cell cancer of the vulva. Gynecol Oncol 2003;89: Puig-Tintoré LM, Ordi J, Vidal-Sicart S, et al. Further data on the usefulness of sentinel lymph node identification and ultrastaging in vulvar squamous cell carcinoma. Gynecol Oncol 2003;88: Hacker NF, Nieberg RK, Berek JS, et al. Superficially invasive vulvar cancer with nodal metastases. Gynecol Oncol 1983;15: Hacker NF, Berek JS, Lagasse LD, et al. Individualization of treatment for stage I squamous cell vulvar carcinoma. Obstet Gynecol 1984;63: Sedlis A, Homesley H, Bundy BN, et al. Positive groin lymph nodes in superficial squamous cell vulvar cancer. Am J Obstet Gynecol 1987;156: Magrina JF, Webb MJ, Gaffey TA, Symmonds RE. Stage I squamous cell cancer of the vulva. Am J Obstet Gynecol 1979;134: Buscema J, Stern JL, Woodruff JD. Early invasive carcinoma of the vulva. Am J Obstet Gynecol 1981;140: Shephard JH. Cervical and vulva cancer: changes in FIGO definitions of staging. Br J Obstet Gynaecol 1996;103: Ross MJ, Ehrmann RL. Histologic prognosticators in stage I squamous cell carcinoma of the vulva. Obstet Gynecol 1987;70: Boyce J, Fruchter RG, Kasambilides E, et al. Prognostic factors in carcinoma of the vulva. Gynecol Oncol 1985;20: Andreasson B, Nyboe J. Predictive factors with reference to low-risk of metastases in squamous carcinoma of the vulvar region. Gynecol Oncol 1985;21: Hicks ML, Hempling RE, Piver MS. Vulvar carcinoma with 0.5 mm of invasion and associated inguinal lymph node metastasis. J Surg Oncol 1993;54: Atamdede F, Hoogerland D. Regional lymph node recurrence following local excision for microinvasive vulvar carcinoma. Gynecol Oncol 1989;34: Van der Velden J, Kooyman CD, Van Lindert AC, Heintz AP. A stage Ia vulvar carcinoma with an inguinal lymph node recurrence after local excision. Int J Gynecol Cancer 1992;2: Magrina JF, Gonzalez-Bosquet J, Weaver AL, et al. Squamous cell carcinoma of the vulva stage IA: long-term results. Gynecol Oncol 2000;76: Homesley HD, Bundy BN, Sedlis A, et al. Prognostic factors for groin node metastasis in squamous cell carcinoma of the vulva. Gynecol Oncol 1993;49: Smyczek-Gargya B, Volz B, Geppert M, Dietl J. A multivariate analysis of clinical and morphological prognostic factors in squamous cell carcinoma of the vulva. Gynecol Obstet Invest 1997;43: Scheistroen M, Trope C, Koern J, et al. Malignant melanoma of the vulva. Evaluation of prognostic factors with emphasis on DNA ploidy in 75 patients. Cancer 1995;75: Ballouk F, Ambros RA, Malfetano JH, Ross JS. Evaluation of prognostic indicators in squamous carcinoma of the vulva including nuclear DNA content. Mod Path 1993;6:

32 Prediction of lymph node metastases in vulvar cancer: A review Dolan JR, McCall AR, Gooneratne S, et al. DNA ploidy, proliferation index, grade, and stage as prognostic factors for vulvar squamous cell carcinomas. Gynecol Oncol 1993;48: Mariani L, Conti L, Atlante G, et al. Vulvar squamous carcinoma: prognostic role of DNA content. Gynecol Oncol 1998;71: Gordinier ME, Steinhoff MM, Hogan JW, et al. S-Phase fraction, p53, and HER-2/neu status as predictors of nodal metastasis in early vulvar cancer. Gynecol Oncol 1997;67: Zur Hausen H. Condylomata accumminata and human genital cancer. Cancer Res 1976;36: Fox H, Wells M. Recent advances in the pathology of the vulva. Histopathology 2003;42: Nuovo GJ, Delvenne P, MacConnell P, et al. Correlation of histology and detection of human papillomavirus DNA in vulvar cancers. Gynecol Oncol 1991;43: Monk BJ, Burger RA, Lin F, et al. Prognostic significance of human papillomavirus DNA in vulvar carcinoma. Obstet Gynecol 1995;85: Johnson GA, Mannel R, Khalifa M, et al. Epidermal growth factor receptor in vulvar malignancies and its relationship to metastasis and patient survival. Gynecol Oncol 1997;65: Milde-Langosch K, Albrecht K, Joram S, et al. Presence and persistence of HPV infection an p53 mutation in cancer of the cervix uteri and the vulva. Int J Cancer 1995;63: Kagie MJ, Kenter GG, Tollenaar RAEM, et al. p53 protein overexpression is common and independent of human papillomavirus infection in squamous cell carcinoma of the vulva. Cancer 1997;80: Scheistroen M, Trope C, Pettersen EO, Nesland JM. p53 protein expression in squamous cell carcinoma of the vulva. Cancer 1999;85: Emanuels AG, Koudstraal J, Burger MP, Hollema H. In squamous cell carcinoma of the vulva, overexpression of p53 is a late event and neither p53 nor mdm2 expression is a useful marker to predict lymph node metastases. Br J Cancer 1999;80: Ngan HY, Cheung AN, Liu SS, et al. Abnormal expression or mutation of TP53 and HPV in vulvar cancer. Eur J Cancer 1999;35: Lerma E, Matias-Guiu X, Lee SJ, Prat J. Squamous cell carcinoma of the vulva : study of ploidy, HPV, p53, and prb. Int J Gynecol Pathol 1999;18: Koyamatsu Y, Yokoyama M, Nakao Y, et al. A comparative analysis of human papillomavirus types 16 and 18 and expression of p53 gene and Ki-67 in cervical, vaginal, and vulvar carcinomas. Gynecol Oncol 2003;90: Chan MK, Cheung TH, Chung TK, et al. Expression of p16ink4 and retinoblastoma protein Rb in vulvar lesions of Chinese women. Gynecol Oncol 1998;68: Zamparelli A, Masciullo V, Bovicelli A, et al. Expression of cell-cycle-associated proteins prb2/p130 and p27kip in vulvar squamous cell carcinomas. Hum Pathol 2001;32: Emanuels AG, Burger MP, Hollema H, Koudstraal J. Quantitation of proliferation-associated markers Ag-NOR and Ki-67 does not contribute to the prediction of lymph node metastases in squamous cell carcinoma of the vulva. Hum Pathol 1996;27: Hantschmann P, Kurzl R. Regulation of apoptosis in squamous cell carcinoma of the vulva. J Reprod Med 2000;45: Hefler L, Tempfer C, Kainz C. Expression of bcl-2 in vulvar cancer tissue. Gynecol Oncol 1999;73: Ferrandina G, Ranelletti FO, Salutari V, et al. Expression of cyclooxygenase-2 (COX-2) in non-neoplastic and neoplastic vulvar epithelial lesions. Gynecol Oncol 2004;92: Jee KJ, Kim YK, Kim KR, et al. Loss in 3p and 4p and gain of 3q are concomitant aberrations in squamous cell carcinoma of the vulva. Mod Pathol 2001;14: Allen DG, Hutchins AM, Hammet F, et al. Genetic aberrations detected by comparative genomic hybridization in vulvar cancers. Brit J Cancer 2002;86:

33

34 Chapter 3 EGFR expression is associated with groin node metastases in vulvar cancer, but does not improve their prediction Maaike H.M. Oonk Geertruida H. de Bock David J. van der Veen Klaske A. ten Hoor Joanne A. de Hullu Harry Hollema Ate G.J. van der Zee Gynecologic Oncology 2007; 104:

35 34 Chapter 3 Abstract Objectives High morbidity of elective inguinofemoral lymphadenectomy in early-stage vulvar cancer patients urges the need for defining a group of low-risk patients in whom inguinofemoral lymphadenectomy can be safely omitted. Aim of the study was to evaluate whether in addition to classic clinicopathological factors determination of EGFR expression in vulvar cancer can be helpful in defining such a low-risk group. Methods Formalin-fixed paraffin-embedded tumor tissue samples of 197 surgically treated T1 / 2 patients were collected in a Tissue Micro Array (TMA). On this TMA, immunohistochemistry for EGFR was performed. Logistic regression analyses were performed including histopathological characteristics with the presence of nodal metastases as outcome. A predictive model was constructed, and absolute risks were calculated. Results EGFR expression was present in 68% of the vulvar tumors and related to the presence of lymph node metastases (OR 2.12, 95% CI ). Our predictive model with only clinicopathological factors was able to define a group of patients with a likelihood of absence of lymph node metastases of 13% (95% CI 5 36), which could be decreased to 6% (95% CI 0 29) after inclusion of EGFR expression (p=0.07). Conclusions EGFR expression is present in the majority of vulvar tumors and is associated with groin node metastases in vulvar cancer. Current classic clinicopathological predictive factors for inguinofemoral lymph node metastases with or without EGFR analysis are not strong enough for identification of sufficiently low risk T1 / 2 vulvar cancer patients. Our predictive model approach however is excellent for evaluation of new cell biological parameters, associated with clinical outcome.

36 EGFR expression in vulvar cancer 35 Introduction Carcinoma of the vulva is a rare disease, which predominantly affects elderly women. The most common histological type is squamous cell carcinoma which accounts for 90% of all vulvar carcinomas. 1 Standard treatment for squamous cell carcinoma of the vulva is wide local excision with elective unior bilateral superficial and deep inguinofemoral lymphadenectomy via separate incisions. Although effective with respect to cure, the morbidity of this treatment is high and especially related to the lymphadenectomy with frequent wound breakdown, infections, and lymphedema of the legs as major complications. 2,3 About 20 30% of patients with early-stage squamous cell carcinoma of the vulva (T1 ( 2 cm) or T2 (> 2cm) tumor, without suspicious groins) will have inguinofemoral lymph node metastases. In retrospect, 70 80% of these patients will therefore probably not benefit from inguinofemoral lymphadenectomy but are at risk for its major complications. Moreover, it has to be kept in mind that missing inguinofemoral lymph node metastases may have major impact for the patients because of the high mortality associated with groin recurrences, especially in an undissected groin. 4 In the light of these considerations, gynecologic oncologists throughout the world regard an elective inguinofemoral lymphadenectomy as standard of care in early-stage vulvar cancer patients with depth of invasion > 1mm. To obviate the morbidity of elective inguinofemoral lymphadenectomy, while maintaining a very low number of groin recurrences, the demands for non- or minimally invasive techniques for exclusion of inguinofemoral lymph node metastases are extremely high. In a recent review, 5 it was shown that currently available noninvasive diagnostic tests (palpation, ultrasound (with or without FNA cytology), CT, MRI, PET) are not able to exclude inguinofemoral lymph node metastases with a certainty that is high enough to safely omit inguinofemoral lymphadenectomy, while the safety of the sentinel node technique is currently being evaluated in phase II studies. 6-8 Studies on different histopathological parameters of the primary vulvar tumor showed a very low (< 1%) risk for inguinofemoral metastases in tumors with a depth of invasion less than 1mm No other classic histopathological parameters are able to define comparable low risk groups of patients with invasive squamous cell carcinoma of the vulva. In addition, a variety of cell biological parameters, more or less associated with metastatic behavior of the primary tumor, have been evaluated for their predictive value, but overall these data are inconclusive and conflicting, especially because the number of patients analyzed in these studies is often too small. So far, the only parameter with potentially high enough negative predictive value to exclude lymph node metastases in patients with vulvar carcinoma was the absence of epidermal growth factor receptor (EGFR) overexpression. EGFR activation plays a key role in cell adhesion, cell locomotion, cell survival, invasion, and angiogenesis, which results in modulation of tumor progression, e.g. metastases. 14

37 36 Chapter 3 In this study, we analyzed the relation between expression of EGFR with inguinofemoral lymph node metastases in a large group of patients with early-stage squamous cell carcinoma of the vulva. As the focus for the clinician is to find a way to avoid an elective inguinofemoral lymphadenectomy, we then constructed a predictive model including classic clinicopathological and immunostaining parameters to identify a group of patients with a low likelihood of presence of inguinofemoral lymph node metastases. Patients and methods Patients Since 1984, clinicopathological and follow-up data of all patients referred to the Department of Gynecologic Oncology of the University Medical Center Groningen, the Netherlands are prospectively collected during standard treatment and follow-up and stored in a computerized registration database, which is managed in accordance with the hospitals regulations. For the current study, all consecutive squamous cell vulvar cancer patients with T1 2 tumors treated from March 1984 until January 2000 were selected. Patients who did not undergo inguinofemoral lymphadenectomy were excluded (n=25); seven because of FIGO stage IA disease and 18 because of bad general health. Until 1993, standard treatment consisted of radical vulvectomy with en bloc uni- or bilateral inguinofemoral lymphadenectomy. Since 1993, this treatment was modified to wide local excision with uni- or bilateral inguinofemoral lymphadenectomy via separate incisions. Staging was performed according to the surgicopathologic FIGO classification 15 and the AJCC TNM classification. 16 IRB approval For the present study, all relevant data were retrieved from our previously mentioned, larger, computerized database into a separate, anonymous database. In this separate, password protected database, patient identity was protected by study-specific, unique patient numbers, which codes were only known to two dedicated datamanagers, who also have daily responsibility for the larger database. In case of uncertainties with respect to clinicopathologic and follow-up data, the larger databases could only be checked through the datamanagers, thereby ascertaining the protection of patients identity. Due to these procedures according to Dutch law, no further patient or IRB approval was needed. Histology Results of the histological diagnosis of the biopsy and surgical specimen were taken together. Biopsies examined at regional pathology laboratories were requested for review. The tumors were examined for tumor thickness, degree of differentiation, vascular invasion, and multifocality. Using

38 EGFR expression in vulvar cancer 37 a calibrated eyepiece micrometer, tumor thickness was measured from the tumor surface or, in the case of superficial keratosis, from the level of the stratum granulosum downward to the deepest point of the tumor; the thickness of the keratin layer, if present, was therefore neglected. The degree of differentiation in the most prominent part of the tumor was assessed according to the criteria described elsewhere. 17 Capillary-lymphatic infiltration was defined as the presence of tumor emboli in endothelial-lined spaces. The inguinofemoral lymph nodes from each side were examined (one section per cm node) for the presence of intranodal or extranodal metastatic tumor growth. Immunohistochemistry Formalin-fixed paraffin-embedded tumor tissue samples from all patients were available for constructing Tissue Micro Arrays (TMAs). Morphologically representative areas of the primary tumor were marked on haematoxylin- and eosin-stained sections. Areas of necrosis or severe leukocyte infiltration were avoided. Three cores of 0.6mm in diameter were taken from the marked areas out of the corresponding tissue block and were placed in pre-defined array locations in a recipient blank paraffin block, using a precision instrument (Beecher Instruments, Silver Spring, Maryland). After inserting the cores, the blocks were placed in an oven of 37 C for 30 minutes in order to attach the cores to the surrounding paraffin. In total, three arrays were constructed, each containing three cores per tumor. Cores of histologically normal vulvar tissue and vulvar cancer tissue originating from the same donor blocks were incorporated in all three arrays to serve as internal control for intra-run variability and to help to orientate oneself in the TMA. Sections of 4μm were cut from the arrays and were transferred to adhesive coated slides. Immunohistochemistry for EGFR was performed on these TMAs with primary mouse anti-human EGFR monoclonal antibody (clone 31G7, Zymed, #28-005). Slides were cleared in xylene and rehydrated through a graded ethanol series in distilled water. They were subjected to antigen retrieval using a 0.1% Trypsin (0.1% Calcium Chloride) solution at 37 C for 10 minutes. Immunostaining was performed using the DAKO autostainer. Diaminobenzidine (DAB) was used as chromogen to visualize the antibody. The nuclei were counterstained in Mayer s haematoxylin, dehydrated in graded ethanol, dried and coverslipped. For EGFR staining, sections from an ovarian cancer specimen with well established EGFR overexpression were used as positive controls. Isotype IgG antibodies replacing the primary antibody served as negative control. The staining intensity for EGFR observed in each core was scored on a four point scale for EGFR (0=negative, 1=weakly positive, 2=positive, 3=strongly positive). Scoring was independently performed by two of the authors (KA ten Hoor and DJ van der Veen), and a concordance of more than 95% was found. The discordant cases were reviewed, and scores were reassigned on consensus of opinion. For our statistical analyses, we used the average intensity score of three cores.

39 38 Chapter 3 Statistics Statistical analysis was carried out using Statistical Package for the Social Sciences (SPSS) version To determine whether single predictive factors differed significantly between patients with and without nodal metastases at the moment of diagnosis, odds ratios were estimated by univariate logistic regression with the presence of nodal metastases as dependent variable. Then, a multivariate logistic regression analysis was performed with again the presence of nodal metastases as dependent variable, including the variables that contributed statistically significantly in the univariate analysis. Based on this, two predictive models were constructed: one with only classic predictive factors, and one with inclusion of EGFR. The values of the β-coefficients of the variables included in the multivariate logistic regression equation were used to generate two scoring formulas, one including classic predictive factors, and one including classic predictive factors as well as EGFR. The predicted probability for the presence of nodal metastases in both models was calculated for each patient. In the following, it was verified how many subjects were identified correctly by using different cut off values of the scoring formulas. The observed values according to the value of the scoring formulas were grouped, and the observed numbers of patients in total as well as the observed numbers of patients with nodal metastases were counted. All tests were two-sided and P values of < 0.05 were considered significant. Results Patients Between March 1984 and January 2000, 197 patients with primary T1 or T2 squamous cell tumors of the vulva were surgically treated at the University Medical Center Groningen, the Netherlands. Median age was 73 years (range: 34 94). Forty-six (23.4%) patients had a T1 tumor and 151 (76.6%) a T2 tumor. The primary vulvar tumor was unilateral in 111 patients (56.3%) and bilateral in 85 patients (43.2%) (one missing value). Multifocality of the primary tumor was seen in 30 patients (15.2%). One or more inguinofemoral lymph node metastases were present in 71 / 197 patients (36%). Thirty patients had one positive inguinofemoral lymph node, 19 patients two, and 11 patients three or more positive inguinofemoral lymph nodes. Follow-up data were available for all patients and have been published previously, 18 but are not further reported here. Immunostaining Positive staining for EGFR was observed in 68% of the tumors. Positive staining for EGFR was associated with a depth of invasion > 5mm (p=0.001) and with good to moderate grade of differentiation (p=0.01). There was no relation with age, T-status and vascular invasion. Chi-square tests for EGFR staining in relation to different tumor characteristics are shown in Table 1.

40 EGFR expression in vulvar cancer 39 Table 1. Patient and Disease Characteristics Related to the EGFR (n (%); χ 2 test; p-value) Characteristics EGFR positive (moderate/strong) χ2 test Age Under age and over T-status T1 ( 2cm) T2 (> 2cm) Depth of infiltration 5mm > 5mm Grade of differentiation Good Moderate Poor Vascular invasion No Yes 59 (60.2) 74 (74.7) 27 (58.7) 106 (70.2) 42 (54.5) 91 (76.5) 51 (65.4) 69 (75.8) 13 (46.4) 111 (65.7) 54 (78.6) χ 2 =4.75 p=0.03 χ 2 =2.13 p=0.15 χ 2 =10.3 p=0.001 χ 2 =8.70 p=0.01 χ 2 =1.82 p=0.18 Clinicopathological and immunostaining parameters in relation to inguinofemoral lymph node metastases For an overview of patient and tumor characteristics related to the presence of lymph node metastases and the univariate odds ratios by logistic regression, see Table 2. T-status (OR 2.86, 95% CI ), depth of infiltration (OR 2.93, 95% CI ), grade of differentiation (good versus poor: OR 3.35, 95% CI ), and EGFR status (OR 2.12, 95% CI ) were all associated with the presence of lymph node metastases. The two multivariate logistic regression models, one with and one without EGFR, are shown in Table 3. EGFR did not retain its predictive value in the multivariate analysis. EGFR staining intensity was positive (moderate strong) in 78 / 126 (61.9%) of the patients without nodal metastases and in 55 / 71 (77.5%) of the patients with nodal metastases (OR: 2.1, 95% CI: ). A predictive model for inguinofemoral lymph node metastases The values of the β-coefficients of the tumor characteristics included in the multivariate logistic regression equation were used in the predictive models shown in Table 4. Our predictive model based on clinicopathological parameters without EGFR allowed identification of T1 / 2 patients with a likelihood of absence of lymph node metastases of 13% (95% CI 5 36; 30 [15%] patients), which could be decreased to 6% (95% CI 0 29) after inclusion of EGFR expression (p=0.07; 17 [9%] patients).

41 40 Chapter 3 Table 2. Patient and Disease Characteristics Total and Related to the Presence of Nodal Metastasis at the Moment of Diagnosis of Primary Tumor (N (%); Univariate OR, 95% CI) Characteristics Total (N = 197) Presence of nodal metastasis (N = 71) Univariate OR (95% CI) Age Under age and over 98 (49.7) 99 (50.3) 32 (45.1) 39 (44.9) ( ) T-status T1 ( 2cm) T2 (> 2cm) 46 (23.4) 151 (76.6) 9 (12.7) 62 (87.3) ( ) Depth of infiltration* 5 mm > 5 mm 77 (39.1) 119 (60.4) 17 (23.9) 54 (76.1) ( ) Grade of differentiation Good Moderate Poor 78 (39.6) 91 (46.2) 28 (14.2) 20 (28.2) 36 (50.7) 15 (21.1) ( ) 3.35 ( ) Vascular invasion No Yes 169 (85.8) 28 (14.2) 58 (81.7) 13 (18.3) ( ) EGFR Negative (Moderate/strong) positive 64 (32.5) 133 (67.5) 16 (22.5) 55 (77.5) ( ) * Could not be assessed for one patient Table 3. Patient and Disease Characteristics Related to the Presence of Nodal Metastasis (n=71) at the Moment of Diagnosis of Primary Tumor: a Model Excluding and a Model Including EGFR (multivariate OR, 95% C.I.) Characteristics A model excluding EGFR A model including EGFR T-status T2 (> 2cm) 1.81 ( ) 1.75 ( ) Depth of infiltration > 5mm 2.53 ( ) 2.26 ( ) Grade of differentiation Moderate Poor 1.81 ( ) 3.33 ( ) 1.75 ( ) 3.94 ( ) EGFR Positive ( )

42 EGFR expression in vulvar cancer 41 Table 4. Number of Patients with Different Score Levels (Divided into Five Groups) in Relation to the Probability of the Absence of Nodal Metastases at the Moment of Diagnosis of Primary Tumor Scores A Model Including EGFR # A Model Excluding EGFR $ Number of patients with score # Number (%) without nodal metastases 95% CI Number of patients with score $ Number (%) without nodal metastases 95% CI (38%) (49%) (51%) (58%) (69%) (77%) (81%) (87%) (94%) # Score = 1.75 (if T1-status is present) (if depth of infiltration 5mm) (if good grade of differentiation is present) (if moderate grade of differentiation is present) (if EGFR is negative) $ Score = 1.81 (if T1-status is present) (if depth of infiltration 5mm) (if good grade of differentiation is present) (if moderate grade of differentiation is present) Discussion Our study shows that EGFR expression is present in the majority of primary squamous cell carcinomas of the vulva and that it is associated with the presence of inguinofemoral lymph node metastases. However, it has no value in the prediction of lymph node metastases. So far, expression of EGFR in vulvar cancer has been evaluated only to a very limited extent. Berchuck et al. showed in 34 squamous carcinomas of the cervix, vulva, and vagina that EGFR expression was moderate to heavy in all malignant cells. 19 Another small study on EGFR expression in vulvar cancer showed progressive increase in EGFR expression from benign vulvar epithelial to primary malignant tissue to metastatic lesions within the same patient. Increased expression of EFGR in the primary vulvar tumor was also significantly associated with lymph node metastases, but its predictive value was not properly addressed. 14 Studies on EGFR expression in other gynecologic malignancies showed that EGFR overexpression is associated with biological aggressiveness and poor prognosis in cervical cancer Multivariate analysis of risk factors for metastatic disease showed that EGFR overexpression in endometrial cancer is an independent predictor for the presence of metastases. 23 Regarding primary squamous cell carcinoma of the vulva, there is a wealth of studies reporting associations between a variety of cell biological parameters and patients characteristics such as presence of lymph node metastases. However, most of these studies give the decision maker no indication of the absolute number of people who might benefit from adding these biological parameters to routine diagnostics. In the presented study here, after calculating the relative risks

43 42 Chapter 3 for the absence of groin node metastases in vulvar cancer, the absolute risks were calculated for the absence of these metastases. Based on these absolute risks we were able to conclude that, despite the fact that high EGFR expression is associated with inguinofemoral lymph node metastases, its expression analysis allows us to define only a small group of patients (9%) with a likelihood of absence of lymph node metastases of still 6% with a relatively wide 95% CI (0 29). These figures would implicate a false-negative rate for prediction of inguinofemoral lymph node metastases of at least 6%. The high morbidity of elective inguinofemoral lymphadenectomy in early-stage vulvar cancer patients urges the need for defining a group of low-risk patients in whom inguinofemoral lymphadenectomy can be safely omitted. When debating a possible acceptable false-negative rate for sentinel node detection in vulvar cancer the Gynecologic Oncology Group of the European Organization for Research and Treatment of Cancer (EORTC) decided that in the light of the significant decrease in morbidity by omitting elective lymphadenectomy a maximum increase of groin recurrences (i.e. a false-negative rate) of 6% might be acceptable. Our study showed that based on the currently available clinicopathological data with or without inclusion of EGFR expression analysis we could not identify accurately a cohort of T1 / 2 patients with such a low likelihood of inguinofemoral lymph node metastases. The upper level of our 95% CI was higher than the required 6%. The predictive model approach that we used for the present study however seems to be quite feasible for the evaluation of the predictive value of future cell biological parameters possibly associated with metastatic behavior. Currently, no noninvasive technique is available for accurate prediction of inguinofemoral lymph node status. At present, the sentinel node technique seems to be the most promising minimally-invasive technique for defining a group of vulvar cancer patients in whom inguinofemoral lymphadenectomy can be safely omitted. Further research is required in order to find more powerful predictive markers for inguinofemoral lymph node metastases. Our predictive model and the availability of TMAs of early-stage vulvar cancer will allow rapid and accurate evaluation of these new markers.

44 EGFR expression in vulvar cancer 43 References 1. Hacker NF. Vulvar Cancer. In: Berek JS, Hacker NF, eds. Practical Gynecologic Oncology. 3 rd ed. Baltimore: Williams and Wilkins, p Prodratz KC, Symmonds RE, Taylor WF, Williams TJ. Carcinoma of the vulva: analysis of treatment and survival. Obstet Gynecol 1983;61: Andersen BL, Hacker NF. Psychosexual adjustment after vulvar surgery. Obstet Gynecol 1983;62: Stehman FB, Bundy BN, Ball H, Clarke-Pearson DL. Sites of failure and times to failure in carcinoma of the vulva treated conservatively: a Gynecologic Oncology Group study. Am J Obstet Gynecol 1996;174: Oonk MHM, Hollema H, de Hullu JA, van der Zee AGJ. Prediction of lymph node metastases in vulvar cancer, a review. Int J Gynaecol Cancer, in press. 6. De Hullu JA, Hollema H, Piers DA, et al. Sentinel lymph node procedure is highly accurate in squamous cell carcinoma of the vulva. J Clin Oncol 2000;18: Moore RG, DePasquale SE, Steinhoff MM, et al. Sentinel node identification and the ability to detect metastatic tumor to inguinal lymph nodes in squamous cell cancer of the vulva. Gynecol Oncol 2003;89: Puig-Tintoré LM, Ordi J, Vidal-Sicart S, et al. Further data on the usefulness of sentinel lymph node identification and ultrastaging in vulvar squamous cell carcinoma. Gynecol Oncol 2003;88: Hacker NF, Berek JS, Lagasse LD, et al. Individualization of treatment for stage I squamous cell vulvar carcinoma. Obstet Gynecol 1984;63: Magrina JF, Webb MJ, Gaffey TA, Symmonds RE. Stage I squamous cell cancer of the vulva. Am J Obstet Gynecol 1979;134: Ross MJ, Ehrmann RL. Histologic prognosticators in stage I squamous cell carcinoma of the vulva. Obstet Gynecol 1987;70: Boyce J, Fruchter RG, Kasambilides E, et al. Prognostic factors in carcinoma of the vulva. Gynecol Oncol 1985;20: Andreasson B, Nyboe J. Predictive factors with reference to low-risk of metastases in squamous carcinoma of the vulvar region. Gynecol Oncol 1985;21: Johnson GA, Mannel RM, Khalifa M, et al. Epidermal growth factor receptor in vulvar malignancies and its relationship to metastasis and patient survival. Gynecol Oncol 1997;65: Shepherd JH. Cervical and vulvar cancer: changes in FIGO definitions of staging. Br J Obstet Gynaecol 1996;103: Vulva. In: American Joint Committee on Cancer: AJCC Cancer Staging Manual. 5 th edition. Philadelphia: Lippincott-Raven, p Fu YS, Reagan JW. Pathology of the uterine cervix, vagina, and vulva. Saunders, Philadelphia;1989: De Hullu JA, Hollema H, Lolkema S, et al. Vulvar carcinoma. The price of less radical surgery. Cancer 2002;11: Berchuck A, Rodriguez G, Kamel A, et al. Expression of epidermal growth factor receptor and HER-2/neu in normal and neoplastic cervix, vulva, and vagina. Obstet Gynecol 1990;76: Pfeiffer D, Stellwag B, Pfeiffer A, et al. Clinical implications of the epidermal growth factor receptor in the squamous cell carcinoma of the uterine cervix. Gynecol Oncol 1989;33: Hale RJ, Buckley CH, Gullick WJ, et al. Prognostic value of epidermal growth factor receptor expression in cervical carcinoma. J Clin Pathol 1993;46:

45 44 Chapter Kersemaekers AM, Fleuren GJ, Kenter GG, et al. Oncogene alterations in carcinomas of the uterine cervix: overexpression of the epidermal growth factor receptor is associated with poor prognosis. Clin Cancer Res 1999;5: Khalifa MA, Abdoh AA, Mannel RS, et al. Prognostic utility of epidermal growth factor receptor overexpression in endometrial adenocarcinoma. Cancer 1994;73:

46 Chapter 4 Identification of inguinofemoral lymph node metastases by methylation markers in vulvar cancer Maaike H.M. Oonk Jasper J.H. Eijsink Haukeline H. Volders Harry Hollema G. Bea A. Wisman Ed Schuuring Ate G.J. van der Zee Submitted

47 46 Chapter 4 Abstract Histopathologic ultrastaging of sentinel nodes (SNs) in vulvar cancer is time consuming and micrometastases can be missed. Therefore, molecular techniques, such as DNA methylation assay, might improve SN assessment. In this study, we selected methylation markers for vulvar cancer and determined if these methylation markers were suitable for SN assessment. We performed methylation specific PCR (MSP) on DNA isolated from the primary tumor, metastatic lymph nodes, and negative lymph nodes from twenty vulvar cancer patients using the following genes: P16INK4a, MGMT, TWIST1, CADM1, TERT, and TFPI2. For P16INK4a and MGMT immunohistochemistry was performed on the primary tumors and metastatic lymph nodes in order to explore intratumor heterogeneity in gene expression patterns. TERT was methylated in all primary vulvar cancers, P16INK4a in 13/20, TFPI2 in 12/20, CADM1 in 11/20, MGMT in 9/20, and TWIST1 in 7/20. A panel of three methylation markers (P16INK4a, TERT and TFPI2) reached a sensitivity of 67% and a specificity of 100% for detection of metastatic lymph nodes. Immunohistochemistry showed intratumor heterogeneity for expression of P16INK4a and MGMT in respectively 55% and 45% of the primary tumors. Our study shows methylation for one or more methylation markers in all primary vulvar cancers. A panel of three methylation markers had a moderate sensitivity, but 100% specificity for metastatic lymph node detection. Significant intratumor heterogeneity for expression of P16INK4a and MGMT may reflect intratumor heterogeneity for methylation patterns and thereby in general explain the moderate sensitivity of our marker panel for detection of metastases.

48 Identification of lymph node metastases by methylation markers 47 Introduction Vulvar cancer has an annual incidence of 2-3 per 100,000 women and represents a rare disease that accounts for 3-5% of all gynecological malignancies. 1,2 About 90% are of the squamous cell type and the primary way of spread is lymphogenic to the inguinofemoral lymph nodes. For years, standard treatment consisted of radical vulvectomy or wide local excision combined with elective uni- or bilateral inguinofemoral lymphadenectomy. Recently, the sentinel lymph node (SN) procedure was introduced and validated in the management of patients with early-stage squamous cell cancer of the vulva. An international, multicenter observational study showed less morbidity without compromising groin recurrence and survival rates in patients with a negative SN. 3 Pathologic examination of the SN includes ultrastaging for the presence of (micro)metastases (multiple sectioning and immunohistochemistry) and needs to be done with high precision since a false-negative assessment of a SN may lead to untreatable tumor outgrowth in other metastatic lymph nodes that are left behind. Immunohistochemistry with cytokeratin AE1/AE3 detects micrometastases in 23 42% of the histologically negative lymph nodes in vulvar cancer patients. 4,5 Our observational study revealed that indeed ultrastaging increased the number of detected metastatic SNs (17% of negative SNs were found positive at ultrastaging). 3 More importantly, our study also showed that in 2 / 8 patients with groin recurrence after a negative SN, metastases in the SN initially had been missed and were only observed at pathologic review. Review of SNs and size of SN metastases revealed that even in patients with isolated tumor cells in their SN, 4 5% also had non-sn involvement. 6 Therefore, a more sensitive and robust technique might help in the detection of micrometastases in SNs and therefore in better selection of patients for additional treatment. Epigenetic changes, such as methylation of CpG islands in the gene promoter area, can contribute to the development of cancer by gene silencing of especially tumor suppressor genes. Many tumor suppressor genes are currently known which can be methylated in different types of cancer, resulting in the ability of cells to proliferate uncontrollably, invade tissues and avoid apoptosis. In vulvar cancer, little is known about the methylation status of different genes. At present, only five studies have been published on methylation in vulvar cancer. P16INK4a, a cyclin dependent kinase inhibitor, was shown to be methylated in 0 70% of the vulvar cancers 7-10 and sigma, a p53-regulated inhibitor of G2/M progression showed hypermethylation in 53% of vulvar cancers. 11 Recently, a study in 30 vulvar cancers showed methylation of RASSF1A, RASSF2A, TSP-1 and MGMT in respectively 30.0, 60.0, 36.7 and 36.7% of the cases. 10 Methylation of a gene promoter can easily be detected by Methylation Specific PCR (MSP). 12 A recent study by Brock et al in stage I non-small-cell lung cancer patients found that molecular examination of lymph nodes by means of MSP improved the assessment of risk recurrence. The methylation of specific genes in histological normal regional lymph nodes probably indicated the presence of microscopically undetectable micrometastases. 13 Studies in lung cancer, colorectal cancer and gastric cancer also showed that methylation analysis of

49 48 Chapter 4 specific genes might be useful as a marker for the identification of micrometastases in regional lymph nodes Therefore, we hypothesized that a methylation assay might be an attractive alternative for pathologic ultrastaging in examination of SNs in vulvar cancer patients, since it might be more sensitive and less time consuming. The aim of the present study was to identify methylation markers for vulvar cancer in patients with histological confirmed lymph node metastases and to determine whether these methylation markers might be suitable in SN assessment. For this purpose, we investigated DNA methylation patterns of six genes in primary tumors, lymph nodes with a histological proven metastasis and negative lymph nodes from the same 20 patients with squamous cell vulvar cancer. For two genes (P16INK4a and MGMT) we also performed immunohistochemistry on the primary tumors and metastatic lymph nodes in order to explore intratumor heterogeneity in gene expression patterns, which might also reflect heterogeneity in methylation in the primary tumor and thereby explain possible discrepancies between methylation patterns in primary tumors and their metastases. Material and Methods Patients and tissue samples Twenty patients with squamous cell cancer of the vulva with lymph node metastases who were treated by radical excision and inguinofemoral lymphadenectomy between 1997 and 2006 were selected from a large, well-documented, consecutive and anonymized database. Pathologic examination of the lymph nodes in these patients was initially only performed with haematoxylin and eosin (H&E) staining. Based on revision of these H&E slides, formalin-fixed paraffin embedded tissue samples from the primary tumor, a metastatic lymph node and a non-metastatic lymph node from the same patient were selected. When more than one metastatic lymph node was available, the one with the largest metastasis was selected. A 4 µm section was obtained from each paraffin block and was stained with H&E to confirm the presence of tumor tissue in the primary tumor and tumor positive lymph node sections and to confirm absence of tumor in negative lymph node sections. Tumors with < 70% tumor cells were macrodissected to enrich for tumor cells. Five to eight 4 µm sections were used for DNA isolation. IRB approval For the present study, all relevant data were retrieved from our previously mentioned, larger, computerized database into a separate, anonymous database. In this separate database, patient identity was protected by study-specific, unique patient numbers, which codes were only known to two dedicated datamanagers, who also have daily responsibility for the larger database. Due to these procedures according to Dutch law, no further patient or IRB approval was needed.

50 Identification of lymph node metastases by methylation markers 49 DNA isolation protocol DNA was extracted from the formalin-fixed paraffin embedded tissue sections using standard saltchloroform extraction and ethanol precipitation for high molecular DNA and dissolved in 150 µl TE-4 buffer (10mM Tris; 1mM EDTA (ph 8.0). 17 For quality control, genomic DNA was amplified in a multiplex PCR according to the BIOMED-2 protocol. 18 Bisulfite treatment and MSP MSP was performed after bisulfite treatment on genomic DNA as described previously. 19 One microgram of genomic DNA was used for bisulfite treatment using the EZ DNA methylation kit according to the manufacturer s protocol (Zymogen, BaseClear, Leiden, The Netherlands). For MSP, 15ng of the modified DNA was used. MSP was carried out for 40 cycles after the first denaturation at 95 C for 10 minutes (95 C for 1 minute, 60 C for 1 minute, 72 C for 1 minute), a final 10-minute extension was carried out at 72 C. PCR products were separated on a 2.5% agarose gel, stained with ethidium bromide and visualized by UV illumination. They were visually scored, as methylated or unmethylated according to the presence or absence of a PCR product. As a positive control, in vitro methylated genomic DNA by Sss I DNA methyltransferase from leucocytes was used and as a negative control a pool of leukocyte DNA from healthy women. Based on previous experience in squamous cell cancers the following genes were selected for MSP: P16INK4a, MGMT, TWIST1, CADM1, TERT, and TFPI2. Primer sequences are shown in Table 1. All samples were run in duplicate, and repeated when PCR results were inconsistent. A DNA sample was considered methylated when two PCR were scored positive using methylation-specific primers. Immunohistochemistry Since the negative lymph nodes were selected based on H&E staining only, we performed immunohistochemistry (IHC) with cytokeratin 1% AE1:AE3 antikeratin solution (Boehringer Mannheim, Mannheim, Germany) to confirm absence of tumor cells in the negative lymph nodes. After every three slides for DNA isolation, one slide was made for IHC. Only cytokeratin-positive cells with a nucleus were considered metastatic tumor cells. Anucluate cytokeratin-positive structures were not considered positive. In order to explore intratumor heterogeneity in gene expression patterns, which might be caused by heterogeneity in methylation in the primary tumor and thereby be responsible for discrepancies between methylation patterns in primary tumors and their metastases, we performed IHC on the primary tumors for P16INK4a and MGMT for which antibodies were readily available. Paraffin-embedded formalin-fixed sections of the primary tumor (4 µm) were deparaffinized in xylene and endogenous peroxidase was blocked by incubation in 0.3% hydrogen peroxide for 30 minutes. Antigen retrieval for P16INK4a and MGMT was performed by overnight incubation at 80 in Tris/HCl buffer and by microwave treatment in 10mM TrisEDTA buffer, respectively. The primary antibodies were diluted in phosphate buffered saline (PBS) containing 1%

51 50 Chapter 4 bovine serum albumin (BSA) and added to the deparaffinized tumor material for one hour at room temperature. Primary antibodies and dilutions were for P16INK4a mouse monoclonal antibody (1:50, clone G , Pharmingen) and for MGMT mouse monoclonal antibody (1:100, clone MT3.1, Chemicon). 3,3 -diaminobenzidine (DAB) was used as chromogen, and the slides were counterstained in haematoxylin. For the negative control the primary antibody was replaced by mouse IgG1. For P16INK4a protein expression, strong nuclear, as well as cytoplasmic staining was considered positive. P16INK4a immunostaining was scored, according to Klaes et al, 24 as negative when <1% of cells were positive, (+) when < 5% were positive, (++) when focal clusters of cells representing <25% were positive, and (+++) when sheets or large clusters of contiguous positive cells representing >25% of cells were observed. For MGMT protein expression, specific staining in the tumor cell nucleus in more than 10% of the tumor cells was defined as positive staining. Less than 10% staining was considered negative as reported previously by Sawhney. 25 When expression in the primary tumor was only focal, or only present in the core or outer periphery of the tumor mass, we considered staining heterogeneous. Table 1. Primer sequences Gene Forward primers (5 3 ) Reverse primers (3 5 ) P16INK4a 12 M TTATTAGAGGGTGGGGCGGATCGC U TTATTAGAGGGTGGGGTGGATTGT MGMT 20 M TTTCGACGTTCGTAGGTTTTCGC U TTTGTGTTTTGATGTTTGTAGGTTTTTGT TWIST1 19,21 M GAAGGAAATCGTTTCGCGTTC U GGGAAGGAAATTGTTTGTGTTTG CADM1 22 M GAAATTTGTAACGTTTGGTTCG U TGAGTGATGGAAATTTGTAATG TERT M GGTTTCGATAGCGTAGTTGTTTC* U GGTTTGCTTTTGATAGTGTAGTTGTTTT* TFPI2 23 M GCGGTTTTTTGTTTTAGGC U TGGTTTTTTGTTTTAGGTGGTT* M = methylated sequence U = unmethylated sequence * = own design GACCCCGAACCGCGACCGTAA CAACCCCAAACCACAACCATAA GCACTCTTCCGAAAACGAAACG AACTCCACACTCTTCCAAAAACAAAACA AAACGACCTAACCCGAACG CCTAACCCAAACAACCAACC CGCTATATCAAACCGACG ACAATCACTATATCAAACCAACA CTACACCCTAAAAACGCGAAC* TCCCTACACCCTAAAAACACAAAC* GACGAAAATCGACCGAACGC CCCCAACAAAAATCAACCAAACA*

52 Identification of lymph node metastases by methylation markers 51 Results Patients and tumor characteristics All twenty vulvar cancer patients had uni- or bilateral groin node metastases. Local treatment consisted of radical vulvectomy in 13 patients (65%) and wide local excision in 7 patients (35%). In 18/20 patients (90%) bilateral inguinofemoral lymphadenectomy was performed. The primary tumors had a mean diameter of 50 mm (range: mm) and a mean depth of invasion of 11 mm (range: 2 27 mm). The clinicopathological characteristics of the patients in this study are summarized in Table 2. Table 2. Clinicopathological characteristics of patients (N=20) Characteristic Number of patients (%) Age (years; median, range) 74 (52 88) Local treatment Radical vulvectomy Wide local excision Groin treatment Inguinofemoral lymphadenectomy en bloc Inguinofemoral lymphadenectomy separate incisions 13 (65) 7 (35) 6 (30) 14 (70) Diameter primary tumor (mm; median, range) 45 (20 120) Depth of invasion (mm; median, range) 10 (2 27) Grade of differentiation Good Moderate Poor 6 (30) 11 (55) 3 (15) Number of metastatic lymph nodes (per patient; median, range) 2.6 (1 6) Extracapsular tumorgrowth Absent Present 11 (55) 9 (45) Methylation markers in squamous cell cancer of the vulva In all 20 primary squamous cell carcinomas of the vulva, one of the following methylation markers was positive: P16INK4a, MGMT, TWIST1, CADM1, TERT, and TFPI2. TERT was the only marker that was methylated in all primary tumors. P16INK4a was methylated in 13/20 vulvar tumors (65%), followed by TFPI2 in 12/20 (60%), CADM1 in 11/20 (55%), MGMT in 9/20 (45%) and TWIST1 in 7/20 (35%). For an overview of these results, see Table 3.

53 52 Chapter 4 Table 3. Methylation markers in primary squamous cell cancer of the vulva (n = 20) P16INK4a MGMT TWIST1 CADM1 TERT TFPI2 VC1 VC2 VC3 VC4 VC5 VC6 VC7 VC8 VC9 VC10 VC11 VC12 VC13 VC14 VC15 VC16 VC17 VC18 VC19 VC20 VC = vulvar cancer Black squares denote methylation, white squares denote no methylation. Methylation markers in matching lymph nodes The negative lymph nodes were all negative for the following methylation markers: P16INK4a, TWIST1, TERT, and TFPI2. Four negative lymph nodes showed methylation for MGMT. Since CADM1 was methylated in > 50% of the negative lymph nodes, this gene was excluded for further analysis. P16INK4a was methylated in 9/20 (45%) metastatic lymph nodes, MGMT in 7/20 (35%), TWIST1 in 4/20 (20%), TERT in 11/20 (55%), and TFPI2 in 4/20 (20%). Concordance between hypermethylation in primary tumor and metastatic lymph nodes was observed for P16INK4a in 16/20 cases (80%), for MGMT in 18/20 cases (90%), for TWIST1 in 15/20 cases (75%), for TERT in 11/20 cases (55%), and

54 Identification of lymph node metastases by methylation markers 53 for TFPI2 in 12/20 cases (60%). For an overview of these results, see Table 4. IHC with AE1/AE3 of the negative lymph nodes sections revealed a micrometastasis (maximum diameter 1.4 mm) in one sample (VC6). Table 4. Methylation markers in primary tumor and metastatic lymph nodes VC1 < 5 VC2 < 5 VC3 50 VC4 75 VC5 50 VC6 50 VC7 75 VC8 < 5 VC9 25 VC10 75 VC11 75 VC12 50 VC13 50 VC14 25 VC15 75 VC16 25 VC17 25 VC18 50 VC19 50 VC20 50 % P16INK4a MGMT TWIST1 TERT TFPI2 T M N T M N T M N T M N T M N % = percentage of tumor load in the examined metastatic lymph node T = primary tumor M = metastatic lymph node N = negative lymph node Black squares denote methylation, white squares denote no methylation. P16INK4a and MGMT immunostaining patterns For P16INK4a heterogeneous expression was observed in 11/20 primary tumors (55%) and for MGMT in 8/19 primary tumors (42%) (Table 5, Figure 1). For P16INK4a inconsistent methylation

55 54 Chapter 4 patterns in tumor and metastatic lymph node were observed in 4/20 cases (VC5, VC14, VC17 and VC20 [methylation in primary tumor, no methylation in metastatic lymph node]). Heterogeneity for P16INK4a expression was observed in 3 of these 4 inconsistent cases (VC14, VC17 and VC20). For MGMT inconsistent methylation patterns in tumor and metastatic lymph node were observed in 2/20 cases (VC18 and VC20 [methylation in primary tumor, no methylation in metastatic lymph node]), heterogeneity for MGMT expression was observed in one of these two cases (VC20). Heterogeneity might therefore explain 4 of the 6 cases with inconsistent methylation patterns for P16INK4a and/ or MGMT. A discrepancy in immunostaining between the primary tumors and its metastasis was observed in 7/18 cases for P16INK4a (38.9%) and in 4/19 cases for MGMT (21.1%). For a summary of immunostaining results, see Table 5. Table 5. Immunohistochemistry and methylation patterns in primary tumor and metastatic lymph node for P16INK4a and MGMT P16INK4a MGMT Primary tumor Metastatic lymph node Primary tumor Metastatic lymph node IHC Methylation IHC Methylation IHC methylation IHC methylation VC1 - M - M + U + U VC2 + U - U + U + U VC3 ++ U ++ U - U + U VC4 +++ U + U + U - U VC5 - M + U - U + U VC6 ++ M + M + U + U VC7 + M n.d. M n.d. U Weak U VC8 - M + M - M - M VC9 + U - U - U - U VC10 + M + M - U + U VC11 + M - M + U + U VC12 + M + M + U + U VC13 - M n.d. M - M - M VC14 + M + U + M + M VC15 ++ M + M + M + M VC16 - U ++ U - M - M VC17 + M - U - M - M VC18 + U ++ U + M + U VC19 ++ U + U - M - M VC20 ++ M + U + M + U For P16INK4a: < 1% is negative (-), < 5% +, < 25% ++, > 25% +++. For MGMT: < 10% is negative (-), > 10% is positive (+) = heterogeneous staining, M = methylated, U = unmethylated

56 Identification of lymph node metastases by methylation markers 55 Figure 1. Immunohistochemistry for P16INK4a and MGMT in primary tumors; examples of heterogeneous staining pattern A + B show heterogeneous staining of P16INK4a respectively with x10 and x40 enlargement C + D show heterogeneous staining of MGMT respectively with x10 and x40 enlargement