Perineural and Vascular Invasion in Oral Cavity Squamous Carcinoma

Transcription

1 Perineural and Vascular Invasion in Oral Cavity Squamous Carcinoma Increased Incidence on Re-review of Slides and by Using Immunohistochemical Enhancement Kevin A. Kurtz, MD; Henry T. Hoffman, MD; M. Bridget Zimmerman, PhD; Robert A. Robinson, MD, PhD Context. Perineural invasion and vascular invasion may be adverse prognostic factors in patients with oral cavity squamous cell carcinoma. However, the incidence of perineural and vascular invasion varies in the literature, and the use of immunohistochemistry to enhance their detection has not been evaluated in oral cavity squamous cell carcinomas. Objective. To determine if the previously assessed incidence of perineural and vascular invasion in cases of oral cavity squamous cell carcinoma would be increased by rereview of the original routinely hematoxylin-eosin stained sections as well as review of slides stained immunohistochemically with S00 and CD to enhance visualization of nerves and vessels. Design. Forty cases of oral cavity squamous cell carcinoma in which the status of perineural and vascular invasion had been part of the original pathology report were reviewed. All original routinely stained slides were reviewed as well as S00- and CD-stained sections of each case s tissue blocks that contained tumor. Results. Perineural invasion was identified in 0% (/ 0) of tumors in the original reports, 6% (5/0) of the authors re-review of the same slides, and 8% (/0) when cases were stained with S00. Vascular invasion was identified in 0% (/0) of tumors in the original reports, 5% (/0) of the authors re-review of the same slides, and % (7/0) when cases were stained with CD. False-positive and false-negative results were common in the original reports. The number of foci of both types of invasion was related to its discovery in the original reports. Vascular invasion, but not perineural invasion, was significantly associated with death at 5-year follow-up. Conclusions. Although careful re-review of routinely stained slides will detect a significant number of cases of perineural and vascular invasion, immunohistochemical enhancement further improves the accuracy of the determination. (Arch Pathol Lab Med. 005;9:5 59) Both perineural invasion and vascular invasion in head and neck squamous cell carcinomas (HNSCCs) have been associated with poor outcomes in some studies. Tumor recurrence and spread to the base of the skull as well as to other regional sites by HNSCCs are thought to be caused in part by invasion of and extension along perineural sheaths. The true significance of perineural invasion is not established in the literature, because conflicting results exist. This is not surprising, because the reported incidence of finding perineural invasion has ranged from approximately 5% to 50% in various studies.,7 As is true for perineural invasion, the significance of vascular invasion in HNSCC is not clear, with some believing it to be Accepted for publication October 0, 00. From the Departments of Pathology (Drs Kurtz and Robinson) and Otolaryngology Head & Neck Surgery (Dr Hoffman), the Roy J. and Lucille A. Carver College of Medicine, and the Department of Biostatistics, College of Public Health (Dr Zimmerman), University of Iowa, Iowa City. The authors have no relevant financial interest in the products or companies described in this article. Reprints: Robert A. Robinson, MD, PhD, Department of Pathology, 5 RCP, University of Iowa, Iowa City, IA 5 ( robert-a-robinson@uiowa.edu). indicative of a worse prognosis and others disagreeing.,, Our experience in searching for perineural and vascular invasion in HNSCC has been that it can be a tedious and difficult task. Brisk stromal tissue responses can make finding nerves difficult, and retraction artifact around tumor nests can mimic vascular invasion. Time pressures to finalize reports may tend to cause underreporting of perineural and vascular invasion. We wished to determine if re-review of the original hematoxylin-eosin stained (H&E) slides would increase the detection of perineural and vascular invasion in cases of oral cavity squamous cell carcinomas (OCSCCs) whose perineural and vascular invasion status had been previously included in the original pathology report. We also wished to determine if immunohistochemical enhancement of the slides with S00 and CD would increase the number of cases of OCSCC found to have perineural or vascular invasion. MATERIALS AND METHODS Case Selection Forty cases of previously untreated OCSCC were studied for the presence or absence of perineural and vascular invasion. All patients had their resections at this institution between 99 and 5 Arch Pathol Lab Med Vol 9, March 005 Perineural and Vascular Invasion in Oral Carcinoma Kurtz et al

2 Table. Cases of Oral Cavity Squamous Cell Carcinoma Positive for Perineural and Vascular Invasion by Original Report, Authors Review of Original Slides, and Review of Immunohistochemical Stained Slides (S00 for Perineural Invasion and CD for Vascular Invasion)* Type of Invasion Perineural invasion Vascular invasion * H&E indicates hematoxylin-eosin. Original Report Using H&E Slides, % (Proportion) 0 (/0) 0 (/0) Authors Review Using H&E Slides, % (Proportion) 6 (5/0) 5 (/0) Immunohistochemical-Stained Slides (S00 or CD), % (Proportion) 8 (/0) (7/0) 998 and were followed a minimum of 5 years or until death. The numbers of cases originating within each subsite of the oral cavity were as follows: oral tongue, 8; floor of mouth, ; gingiva, 6; retromolar trigone, ; and buccal,. All tumors had been examined, sectioned, and embedded according to the pathology department s protocol. This protocol is essentially that of a published gross room procedure manual and follows as well a pathology reporting guideline., The cases were selected by a search of the department s laboratory information system on the primary basis that perineural and vascular invasion status (present or absent) was noted in the report. During the time period of the initial primary diagnostic review of these cases, the presence or absence of perineural and vascular invasion was routinely reported. However, on review of the reports from this time period, cases were found in which perineural or vascular space invasion was not mentioned; these reports were excluded from the study. This hospital s pathology faculty had reviewed all cases studied, and in all cases, perineural and vascular invasion had been assessed with standard H&E-stained slides. Review of Slides The authors, blinded to the status of perineural and vascular invasion according to the pathology reports, reviewed the original H&E-stained slides for the presence of both types of invasion. Perineural invasion was defined as tumor involvement of the perineural space with complete or partial encircling of the enclosed nerve. Vascular invasion was defined as tumor clearly within a vascular space, whether lymphatic space or blood vessel, and the tumor was required to be adherent to vessel endothelium or attached to thrombus in the vessel. For a case to be considered positive by the authors in this study, only focus of perineural or vascular invasion was required to be observed. The number of separate foci of perineural and vascular invasion was also counted in each case, with the number of foci defined as the number of distinct areas of tumor invading nerves or vessels in the slides. We did not count as separate foci tumor cell aggregates that were aligned in a linear pattern along a nerve, because these were considered to represent tumor along the length of a single nerve. Immunohistochemistry The entire collection of pathology materials was reviewed in each case to ascertain which slides contained tumor. All tumorcontaining tissue blocks were cut and stained with S00 to enhance the detection of nerves associated with tumor and with CD to highlight the endothelial cells of both lymphatics and blood vessels. In brief, 5- m sections of formalin-fixed, paraffinembedded tumor tissue were deparaffinized with xylene and alcohol. Endogenous peroxide was blocked by immersing the sections in methanol/h O. For S00 staining, no antigen retrieval was performed. For CD staining, the slides were boiled in citrate buffer for 0 minutes and then cooled for hour in the same buffer. Nonspecific binding was blocked by using normal goat serum for 0 minutes (:0 in phosphate-buffered saline) with % bovine serum albumin (in phosphate-buffered saline). The sections were then immunostained with monoclonal antibodies to S00 (:500; Dako Corporation, Carpinteria, Calif) and CD (:0; Dako) for 60 minutes and 0 minutes at room temperature, respectively. The secondary reactions were carried out using a streptavidin-biotin complex (DAKO LINK and DAKO LSAB; Dako) and, -diaminobenzidine tetrahydrochloride as a chromagen. The slides were rinsed and counterstained lightly with Harris hematoxylin. Internal controls of peripheral nerves and blood/lymphatic vessels were present in the tissues of interest. During interpretation of the immunohistochemical slides, the authors were unaware of the perineural or vascular invasion status according to both the original reports and their review of the H&E-stained slides. Statistical Analysis Statistical analysis between groups was carried out using Pearson. The differences between means were carried out using a -sample t test. Significance was determined to be at P.05. Cox proportional hazard regression analysis was performed for survival times. 5 RESULTS Patient Data Upon review of the 0 cases of oral cavity tumors, all were squamous cell carcinomas. No cases of verrucous carcinoma, basaloid squamous carcinoma, or other subtypes of squamous carcinoma were present. The patients ages at diagnosis ranged from 5 to 8 years, with a mean of 58 years. There were 9 men and women. The stages of the tumors were as follows: were stage I, were stage II, 8 were stage III, and were stage IV. There were 8 well-differentiated tumors, 7 moderately differentiated tumors, and 5 poorly differentiated tumors. Seven tumors were T, tumors were T, 9 tumors were T, and tumors were T. 6 The majority of patients (/0) received postoperative irradiation, most likely because of the large number of stage III and IV malignancies included in this study. It is uncertain whether some patients were selected to receive irradiation based solely on the presence of perineural or vascular invasion. Perineural Invasion According to the surgical pathology reports, perineural invasion was indicated to be present in 0% (/0) of the cases (Table ). When the original H&E-stained slides were reviewed by the authors, 6% (5/0) of the cases were interpreted to contain perineural invasion. Examination of the S00-stained slides found 8% (/0) of the cases to be positive for perineural invasion. Comparing the results of the original reports with the S00-stained slides, there were no false-positive cases. There were, however, false-negative reports of perineural invasion. Compared with the S00-stained slides, the reviewers (authors) had false-positive case of perineural invasion when reviewing the original slides (non immunohistochemical-stained slides) and 9 false-negative cases. Using the S00-stained slides as the standard, there were no significant differences between patients with tumors that exhibited perineural invasion and those whose tu- Arch Pathol Lab Med Vol 9, March 005 Perineural and Vascular Invasion in Oral Carcinoma Kurtz et al 55

3 Table. Characteristic Site of tumor Oral tongue Floor of mouth Gingiva Retromolar trigone Buccal Nodal metastases Present Absent Pathologic stage I II III IV Tumor differentiation Well Moderate Poor Clinical and Pathologic Features of 0 Patients With Oral Cavity Squamous Cell Carcinoma With and Without Perineural and Vascular Invasion Perineural Invasion Positive (n ) Negative (n 7) P Vascular Invasion Positive (n 7) Negative (n ) P mors did not in regard to subsite of tumor, presence or absence of lymph node metastases, pathologic stage, or tumor differentiation (Table ). Of the positive perineural invasion cases, the number of foci of perineural invasion ranged from to 5 separate foci, with a mean of. foci. Of the cases that were originally reported positive for perineural invasion in the original report, the mean number of foci was 7 foci per case. Within the false-negative cases, the mean number of foci of perineural invasion was 9.7 foci per case. The difference between these means was significant (P.06). Of the authors false-negative cases of perineural invasion, the mean number of foci of invasion was. per case. Vascular Invasion According to the surgical pathology reports, vascular invasion was reported in 0% (/0) of the cases. When the original slides were reviewed by the authors, 5% (/ 0) of the cases were interpreted as containing vascular invasion. Examination of the CD-stained slides revealed vascular invasion in % (7/0) of the cases (Table ). Our most important finding in the reassessment of vascular invasion was the high incidence of false-positive and false-negative reporting. Using the CD-stained slides as a reference and standard, there were 6 false-positive and false-negative cases of vascular invasion in the original reports. The authors on re-review had false-positive case of vascular invasion and false-negative cases. We found no significant differences between those patients whose tumors exhibited vascular invasion and those whose tumors were negative in regard to subsite of tumor, presence or absence of lymph node metastases, pathologic stage, or tumor differentiation (Table ). In cases with vascular invasion, the number of foci ranged from to 9 separate foci, with a mean of.6 foci per case. Of the 6 cases that were correctly reported positive in the original reports, the mean number of foci was 6.8, and of the false-negative cases, the mean number of foci of vascular invasion was.8. The difference between these means was significant (P.0). Patient Outcomes The patients were followed for at least 5 years or until death. Follow-up times ranged from to 6 months, with a median of 6 months for all patients. Of the 0 patients, 9 were dead and were alive. The median survival for those dead was months. Of the patients alive at the end of the follow-up, 7 had received postoperative irradiation. Of the 9 patients dead at the conclusion of the study, 7 had also received postoperative irradiation. Six of the 0 patients developed local recurrences, and developed distant metastases. Although all of these patients had tumors with perineural invasion, the finding did not achieve statistical significance (P.09). All patients with distant metastases also had vascular invasion. One patient with vascular invasion had a local recurrence. Five patients with vascular invasion did not develop metastases, nor did they suffer local recurrence. Of those patients dying, 7 had perineural invasion, and did not. Of those alive, 6 had perineural invasion, and 5 did not. Twelve patients with vascular invasion were dead, whereas 5 patients with vascular invasion were alive. Sixteen patients without vascular invasion were alive, and 7 were dead. Survival analysis was performed based on perineural and vascular invasion status, with patients stratified into low-stage (stages I and II) and high-stage (stages III and IV) categories. There was no significant effect of the presence of perineural invasion on mortality of our patients (P.5; hazard ratio of.; 95% confidence internal: ; data not shown). However, survival analysis did show that the presence of vascular invasion was a significant predictor of mortality (P.006; hazard ratio of.0; 95% confidence interval,.5 0.8; Figure, A and B). COMMENT In some studies, perineural invasion has been associated with an increased risk of local recurrence, a higher rate of metastasis, and decreased survival in patients with HNSCC. 6, In contrast, a significant number of other studies failed to detect an association between the pres- 56 Arch Pathol Lab Med Vol 9, March 005 Perineural and Vascular Invasion in Oral Carcinoma Kurtz et al

4 Figure. Model-based estimate of the survival distribution comparing vascular invasion positive cases (VI ) with vascular invasion negative cases (VI ) for stages I and II (A) and stages III and IV (B) cancer patients. There was a significant effect of the presence of vascular invasion on mortality in oral cavity squamous cell carcinoma patients (P.006; hazard ratio of.0; 95% confidence interval,.5 0.8). ence of perineural invasion and adverse outcome. 7,8,,7 Likewise, the significance of vascular invasion in the prognosis of HNSCC is not clear in the literature, with some data supporting an increased risk of lymph node metastases or worse prognosis and other data showing no association.,, The discrepancies in outcome reported in these studies may be attributable in part to differing sensitivities of detecting perineural and vascular invasion. Upon review of the H&E-stained slides in our study of 0 cases of OCSCC, we detected perineural invasion in 6% of the cases, whereas the presence of perineural invasion was documented in only 0% of the original reports. To our knowledge, this is the highest prevalence of perineural invasion in HNSCC reported in the literature. We found vascular invasion in 5% of cases in our review of the H&E-stained slides, whereas the original reports described vascular invasion in 0% of the cases. The discrepancy between the originally reported perineural or vascular invasion status and our review of the slides most likely results from the focused nature of this study. We concentrated solely on the detection of perineural and vascular invasion but not on other factors such as margin status or extent of tumor. During the original sign-out of these cases, the presence of perineural or vascular invasion may have been missed because of attention to these other important histologic findings. Although immunohistochemical stains have been used to enhance detection of perineural and vascular invasion in gastric and colorectal carcinomas, 8,9 none of the studies attempting to correlate perineural or vascular invasion with outcome in HNSCC has described the use of immunohistochemistry. In our series, 8% of OCSCCs contained perineural invasion when the slides were stained for S00 to highlight peripheral nerves, compared with the detection rate of 6% when we reviewed the H&E-stained Arch Pathol Lab Med Vol 9, March 005 Perineural and Vascular Invasion in Oral Carcinoma Kurtz et al 57

5 Figure. A, Hematoxylin-eosin stained section of oral cavity squamous cell carcinoma. The arrows point to foci that may represent perineural invasion. The larger arrow shows a concentric rimming pattern often seen in perineural invasion. It is not clear if the focus identified by the smaller arrow is nerve, muscle, or collagen (original magnification 00). B, Corresponding section stained with S00. Only the focus at the smaller arrow demonstrates perineural invasion (original magnification 00). Figure. A, Hematoxylin-eosin stained section of oral cavity squamous cell carcinoma showing a focus of tumor surrounded by a clear space, equivocal for vascular invasion (original magnification 00). B, Corresponding section stained with CD, demonstrating a layer of endothelial cells at the periphery of the clear space (arrow), confirming the presence of vascular invasion (original magnification 00). slides. We found that the S00-stained slides were especially useful for finding nerves with small diameters embedded within the tumor and for differentiating nerves from the deceptively similar-appearing desmoplastic stroma (Figure, A and B). In comparison with the close examination of H&E slides for perineural invasion, reviewing S00-stained sections was much faster. The primary reason for the enhanced speed was the sharp contrast between S00-stained nerves and the hematoxylin counterstain, allowing low-power detection of perineural invasion. The enhanced detection of vascular invasion by the use of immunohistochemistry was not as pronounced as the increased detection rate demonstrated for perineural invasion. In the CD-stained slides, vascular invasion was identified in % of the cases, compared with 5% determined by reviewing the H&E-stained slides. One primary advantage of the CD-stained slides was the highlighting of thrombus in vessels associated with tumor. Another major advantage of using CD-stained slides was the differentiation between endothelial-lined vascular spaces and retraction artifact (Figure, A and B). The CD-stained slides demonstrated that vascular invasion had been overcalled in a substantial number of cases reported to be positive according to the original pathology reports. Using data derived from examination of the immunohistochemically stained slides, we found a significant correlation between the presence of vascular invasion and decreased survival (P.006). However, we did not find statistical significance correlating perineural invasion status to ultimate outcome. This is not surprising, because our study found such a high incidence of perineural invasion, with 8% of the cases being positive. Also, we did not assess the diameter of nerves involved by perineural invasion or attempt to correlate survival to the number of foci of perineural invasion. It is possible that perineural invasion of a few small nerves embedded within the tumor is not as biologically significant as perineural invasion of many large-diameter nerves. If the presence of perineural or vascular invasion is un- 58 Arch Pathol Lab Med Vol 9, March 005 Perineural and Vascular Invasion in Oral Carcinoma Kurtz et al

6 derreported or overreported, accurate statistical measurement of the prognostic significance is impossible. It appears from our study that it is more likely for pathologists to diagnose perineural and vascular invasion if the quantity of foci of either is increased. Although this concept makes intuitive sense, we were able to show a significant difference in the ability of routine pathology review to identify perineural or vascular invasion based on the number of foci of invasion. In addition, our data demonstrated that the use of immunohistochemical stains to highlight nerves and vessels can further enhance the detection of perineural and vascular invasion compared with careful review of H&E-stained slides. Having the most accurate data with which to determine prognosis or on which to base treatment decisions is important. Because the presence of perineural and vascular invasion has been reported to have purported prognostic significance and may influence treatment decisions, further studies are warranted to determine if immunohistochemical enhancement of tissues for perineural and vascular invasion is a more precise determinant of prognosis in HNSCC. This work was supported in part by the Frederic W. Stamler Professorship to Dr Robinson. References. Brown B, Barnes L, Mazariegos J, Taylor F, Johnson J, Wagner RL. Prognostic factors in mobile tongue and floor of mouth carcinoma. Cancer. 989;6: Yilmaz T, Hosal AS, Gedikoglu G, Onerci M, Gursel B. Prognostic significance of vascular and perineural invasion in cancer of the larynx. Am J Otolaryngol. 998;9: Martinez-Gimeno C, Rodriguez EM, Vila CN, Varela CL. Squamous cell carcinoma of the oral cavity: a clinicopathologic scoring system for evaluating risk of cervical lymph node metastasis. Laryngoscope. 995;05: Gilbert RW, Lundgren JA, van Nostrand AW, Keane TJ. TN0M0 glottic carcinoma: a pathologic analysis of patients treated surgically following radiotherapy. Clin Otolaryngol. 988;: Conte CC, Ergin MT, Ricci A Jr, Deckers PJ. Clinical and pathologic prognostic variables in oropharyngeal squamous cell carcinoma. Am J Surg. 989; 57: Woolgar JA, Scott J. Prediction of cervical lymph node metastasis in squamous cell carcinoma of the tongue/floor of mouth. Head Neck. 995;7: Ozdek A, Sarac S, Akyol MU, Unal OF, Sungur A. Histopathological predictors of occult lymph node metastases in supraglottic squamous cell carcinomas. Eur Arch Otorhinolaryngol. 000;57: Close LG, Burns DK, Reisch J, Schaefer SD. Microvascular invasion in cancer of the oral cavity and oropharynx. Arch Otolaryngol Head Neck Surg. 987; : Olsen KD, Caruso M, Foote RL, et al. Primary head and neck cancer: histopathologic predictors of recurrence after neck dissection in patients with lymph node involvement. Arch Otolaryngol Head Neck Surg. 99;0: Poleksic S, Kalwaic HJ. Prognostic value of vascular invasion in squamous cell carcinoma of the head and neck. Plast Reconstr Surg. 978;6: 0.. Crissman JD, Liu WY, Gluckman JL, Cummings G. Prognostic value of histopathologic parameters in squamous cell carcinoma of the oropharynx. Cancer. 98;5: Fagan JJ, Collins B, Barnes L, D Amico F, Myers EN, Johnson JT. Perineural invasion in squamous cell carcinoma of the head and neck. Arch Otolaryngol Head Neck Surg. 998;: Parsons JT, Mendenhall WM, Stringer SP, Cassisi NJ, Million RR. An analysis of factors influencing the outcome of postoperative irradiation for squamous cell carcinoma of the oral cavity. Int J Radiat Oncol Biol Phys. 997;9:7 8.. Frierson HF Jr, Cooper PH. Prognostic factors in squamous cell carcinoma of the lower lip. Hum Pathol. 986;7: Borges AM, Shrikhande SS, Ganesh B. Surgical pathology of squamous carcinoma of the oral cavity: its impact on management. Semin Surg Oncol. 989;5: Lydiatt DD, Robbins KT, Byers RM, Wolf PF. Treatment of stage I and II oral tongue cancer. Head Neck. 99;5: O Brien CJ, Lahr CJ, Soong SJ, et al. Surgical treatment of early-stage carcinoma of the oral tongue: would adjuvant treatment be beneficial? Head Neck Surg. 986;8: Maddox WA, Urist MM. Histopathological prognostic factors of certain primary oral cavity cancers. Oncology (Huntingt). 990;:9 ; discussion, Soo KC, Carter RL, O Brien CJ, Barr L, Bliss JM, Shaw HJ. Prognostic implications of perineural spread in squamous carcinomas of the head and neck. Laryngoscope. 986;96: Vural E, Hutcheson J, Korourian S, Kechelava S, Hanna E. Correlation of neural cell adhesion molecules with perineural spread of squamous cell carcinoma of the head and neck. Otolaryngol Head Neck Surg. 000;: Yuen PW, Lam KY, Chan AC, Wei WI, Lam LK. Clinicopathological analysis of local spread of carcinoma of the tongue. Am J Surg. 998;75:.. Magnano M, Bongioannini G, Lerda W, et al. Lymph node metastasis in head and neck squamous cell carcinoma: multivariate analysis of prognostic variables. J Exp Clin Cancer Res. 999;8: Rosai J. Manual of Surgical Pathology Gross Room Procedures. Minneapolis, Minn: University of Minnesota Press; 98.. Zarbo RJ, Barnes L, Crissman JD, Gnepp DR, Mills SE, and the Association of Directors of Anatomic and Surgical Pathology. Recommendations for the reporting of specimens containing oral cavity and oropharynx neoplasms. Mod Pathol. 000;: Schmoor C, Sauerbrei W, Schumacher M. Sample size considerations for the evaluation of prognostic factors in survival analysis. Stat Med. 000;9: Greene FL, Page DL, Fleming ID, et al, eds. AJCC Cancer Staging Manual. 6th ed. Chicago, Ill: American Joint Committee on Cancer; Alvi A, Johnson JT. Development of distant metastasis after treatment of advanced-stage head and neck cancer. Head Neck. 997;9: Mori M, Adachi Y, Kamakura T, Ikeda Y, Maehara Y, Sugimachi K. Neural invasion in gastric carcinoma. J Clin Pathol. 995;8:7. 9. Bellis D, Marci V, Monga G. Light microscopic and immunohistochemical evaluation of vascular and neural invasion in colorectal cancer. Pathol Res Pract. 99;89: 7. Arch Pathol Lab Med Vol 9, March 005 Perineural and Vascular Invasion in Oral Carcinoma Kurtz et al 59