CONTRAST-ENHANCED 18 F-FDG-PET/CT FOR THE ASSESSMENT OF NECROTIC LYMPH NODE METASTASES

Transcription

1 ORIGINAL ARTICLE CONTRAST-ENHANCED 18 F-FDG-PET/CT FOR THE ASSESSMENT OF NECROTIC LYMPH NODE METASTASES Stephan K. Haerle, MD, 1 Klaus Strobel, MD, 2 Nader Ahmad, MD, 1 Alex Soltermann, MD, 3 Daniel T. Schmid, MD, 2 Sandro J. Stoeckli, MD 1,4 1 Department of Otolaryngology Head and Neck Surgery, University Hospital Zurich, Frauenklinikstrasse 24, CH-8091 Zurich, Switzerland. stephan.haerle@usz.ch 2 Division of Nuclear Medicine, Department of Medical Radiology, University Hospital Zurich, Raemistrasse 100, CH-8091 Zurich, Switzerland 3 Department of Pathology, University Hospital Zurich, Schmelzbergstrasse 12, CH-8091 Zurich, Switzerland 4 Department of Otolaryngology Head and Neck Surgery, Kantonsspital St Gallen, CH-9007 St Gallen, Switzerland Accepted 23 February 2010 Published online 22 July 2010 in Wiley Online Library (wileyonlinelibrary.com). DOI: /hed Abstract: Background. Cystic lymph node metastasis (CLNM) is commonly found in human papillomavirus (HPV)- associated tonsillar squamous cell carcinoma (SCC). The aim of this study was to compare the accuracy in detecting cystic lymph node metastasis from tonsillar SCC between contrastenhanced CT, 18 F- fluorodeoxyglucose-positron emission tomography (FDG-PET), non-enhanced 18 F-FDG-PET/CT, and contrast-enhanced 18 F-FDG-PET/CT. Methods. Thirty-four patients with a tonsillar SCC undergoing a pretreatment contrast-enhanced 18 F-FDG-PET/CT followed by a neck dissection as a standard of reference were included. The contrast-enhanced CT part, the 18 F-FDG-PET part, the non-enhanced 18 F-FDG-PET/CT part, and the contrast-enhanced 18 F-FDG-PET/CT were assessed separately for correct N classification and the differentiation of N0 versus Nþ. Results. Contrast-enhanced 18 F-FDG-PET/CT, non-enhanced 18 F-FDG-PET/CT, and contrast-enhanced CT are equally accurate for correct neck staging. Regarding pn0 versus pnþ, contrastenhanced CT and contrast-enhanced 18 F-FDG-PET/CT are superior to non-enhanced 18 F-FDG-PET/CT (p ¼.017). Conclusion. Contrast-enhanced CT and contrast-enhanced 18 F-FDG-PET/CT perform equally and better than nonenhanced 18 F-FDG-PET/CT in detecting CLNM in tonsillar SCC. Therefore, in patients scheduled for 18 F-FDG-PET/CT, we strongly suggest performing a contrast-enhanced 18 F-FDG- PET/CT, which is not routine in most centers. VC 2010 Wiley Periodicals, Inc. Head Neck 33: , 2011 Keywords: positron emission tomography; CT; tonsil cancer; squamous cell carcinoma; neck staging The staging of the cervical lymph nodes in patients with head and neck squamous cell carcinoma (HNSCC) is known to be the strongest prognostic factor and to considerably impact the individual treatment plan. 1 Therefore, the accurate detection of nodal metastases by the pretreatment imaging is crucial. Correspondence to: S. K. Haerle VC 2010 Wiley Periodicals, Inc. Different imaging modalities, such as ultrasound, contrast-enhanced CT, and MRI are available at most institutions. Functional imaging using 18 F-fluoro-2- deoxy-d-glucose positron emission tomography ( 18 F- FDG-PET) has gained widespread use due to the avidity of HNSCC for fluorodeoxyglucose (FDG). In order to combine the metabolic information with the anatomic orientation, 18 F-FDG-PET is now routinely combined with a non-enhanced, low-dose CT ending in a non-enhanced 18 F-FDG-PET/CT. For specific questions, contrast can be applied for the CT part of the 18 F-FDG-PET/CT giving a contrast-enhanced 18 F-FDG-PET/CT, which is not routine in most centers. Cystic lymph node metastasis is found in up to 64% of tonsillar squamous cell carcinoma (SCC). 2 The presence of central necrosis or cystic degeneration in neck nodes on imaging is considered as a reliable sign of lymph node metastasis. 3 The likelihood of central necrosis in metastatic lymph nodes has been shown to correlate with the nodal size. Lymph node metastases larger than 1.5 cm in diameter show signs of necrosis on imaging in 56% to 63% of the cases. 4,5 Recently, the presence of human papillomavirus (HPV) infection has been proposed to be an additional factor for the development of cystic-necrotic lymph node transformation in tonsillar SCC. 6 The rising incidence of HPV-associated tonsillar 7 cancers supports the search for a reliable imaging modality for necrotic lymph node metastases. At present, according to the study by Yousem et al, 8 contrast-enhanced CT is still considered the best technique for detection of lymph node necrosis. Recently, some authors have reported on the use of 18 F-FDG-PET/CT for nodal staging, regardless of the grade of lymph node necrosis. 9 In our own subjective experience, we believe that necrotic lymph nodes are often missed by nonenhanced 18 F-FDG-PET/CT because of a lack of FDG uptake in the necrotic tissue. On the other hand, we 324 Imaging of Necrotic Lymph Node Metastases in Tonsillar Carcinoma HEAD & NECK DOI /hed March 2011

2 FIGURE 1. (A) CT images from a 77-year-old patient with squamous cell carcinoma of the right tonsil (long arrows). Axial contrast-enhanced CT shows a cystic lymph node metastasis on the right side (arrowhead). (B) This lymph node is not fluorodeoxyglucose (FDG)-active on positron emission tomography (PET) and (C) non-enhanced CT, and (D) PET/CT images. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.] to 18 F-FDG-PET/CT for exclusion of distant metastases prior to therapy decision. Furthermore, the patient cohort was selected for surgery as the firstline treatment as the histologic workup of the neck dissection specimen served as a standard of reference. The histology reports of the neck dissection specimens (tonsillar SCC) were retrospectively reviewed by 2 head and neck surgeons. At the same time, all contrast-enhanced 18 F-FDG-PET/CT images were reviewed in consensus on screen by 2 doubly board certified radiologists and nuclear medicine physicians with 10/8 years experience in CT reading and 5/4 years experience in reading combined 18 F-FDG-PET/ CT in head and neck cancer patients for the presence of lymph node metastases, presence of lymph node necrosis, SUVmax of the largest lymph node, and correct neck staging. The contrast-enhanced CT part, the 18 F-FDG-PET part, the non-enhanced 18 F-FDG- PET/CT part, and the contrast-enhanced 18 F-FDG- PET/CT part were assessed separately with a minimum time interval of 2 weeks between the readouts. The SUVmax, as assessed by non-enhanced 18 F-FDG- PET/CT, was correlated to the grade of lymph node necrosis as assessed by the histopathologic review. The reviewers of the imaging were blinded to the histology results and vice versa. This study was believe that small lymph node metastases might be missed by contrast-enhanced CT (Figure 1, Figure 2, and Figure 3). The aims of this study were (1) to compare the accuracy in detecting necrotic lymph node metastases from tonsillar SCC between contrast-enhanced CT, 18 F-FDG- PET, non-enhanced 18 F-FDG-PET/CT, and contrastenhanced 18 F-FDG-PET/CT, and (2) to assess a potential correlation between the standardized uptake value maximum (SUVmax) and the grade of lymph node necrosis. MATERIALS AND METHODS All patients with a previously untreated tonsillar SCC who underwent a pretreatment contrastenhanced 18 F-FDG-PET/CT followed by a neck dissection as part of their initial treatment at the Department of Otorhinolaryngology, Head and Neck Surgery, University Hospital Zurich, Switzerland, between January 2002 and December 2007 were included in this study. This patient cohort does not represent a consecutive group of all patients diagnosed with a tonsillar SCC within the given time frame, because, according to the institutional policy, mainly patients with advanced disease (stage III/IV) based on clinical and ultrasound findings are referred FIGURE 2. (A) This lymph node is normal sized (9 mm) and, therefore, was missed with contract-enhanced CT alone. (B,C,D) Axial images of the same patient with fluorodeoxyglucose (FDG)-active lymph node metastasis on the right side (long arrow). Contrast-enhanced FDG-positron emission tomography (FDG-PET)/CT led to the correct lymph node staging (N2b). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.] Imaging of Necrotic Lymph Node Metastases in Tonsillar Carcinoma HEAD & NECK DOI /hed March

3 FIGURE 3. Representative histologic sections and cytologic smear. The primary tumor was a squamous cell carcinoma (SCC) of solid type (A), as were some of the metastases (B). The positron emission tomography (PET)-negative lymph node was cystic (C) and represented by sheets of malignant squamous cells on a background of inflammatory and necrotic debris on fine-needle aspiration (D). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.] conducted in accordance with the local guidelines established by the ethics committee for a retrospective evaluation, and written informed consent was waived for all patients. of 18 F-FDG-PET alone, contrast-enhanced CT alone, and fused non-enhanced and contrast-enhanced 18 F- FDG-PET/CT with linked cursors using the AW Workstation (GE Health Systems). Imaging Protocol. For this study, a combined 18 F- FDG-PET/CT inline system (Discovery LS, RX or Discovery STE, GE Health Systems, Milwaukee, WI) was used. This device integrates a PET scanner with a multi-slice helical CT (slice thickness, 2.5 mm) and permits the acquisition of coregistered CT and PET images in the same session. Patients fasted for at least 4 hours before their scanning, which started approximately 60 minutes after the injection of a standard dose of approximately 350 MBq of 18 F-FDG. An oral CT contrast agent (Micropaque Scanner, Guerbet AG, Aulnay-sous-bois, France) was given 15 minutes before the injection of 18 FDG. Patients were examined in the supine position. The CT scan was acquired during breath holding in the normal expiratory position. Immediately following the CT acquisition, the PET emission scan was acquired. The CT data were used for the attenuation correction, and the images were reconstructed using a standard iterative algorithm (ordered set expectation maximization) for 3-dimensional PET reconstruction. The acquired images were viewed with a software product that provides multiplanar reformatted images Histologic Workup. All 34 patients underwent surgical resection of the primary tumor and a unilateral or bilateral neck dissection within 4 weeks of the scans. The reference standard for the presence or absence of necrotic lymph node metastases was the histologic workup of the neck specimen. Complete workup of neck dissections was performed according to the guidelines of the Swiss Society of Pathology. All detectable lymph nodes were sectioned along the greatest diameter and paraffin-embedded. Whole sections were stained with hematoxylin-eosin, thus 2 surfaces per node were available. The grade of necrosis was defined using a semiquantitative necrosis score: mildly necrotic was defined as necrosis of 0% to 10% of the surface, moderately necrotic as 11% to 50% of necrosis, and severely necrotic as >50% of necrosis. This was performed in analogy to the tumor necrosis scoring system of the current Swiss SAKK trial 17/04 for malignant mesothelioma and the regression grading after neoadjuvant therapy for non-small cell lung cancer. 10 Morphologically, necrotic lymph nodes were not differentiated from the cystic lymph nodes because of their indistinguishable appearance on imaging. 326 Imaging of Necrotic Lymph Node Metastases in Tonsillar Carcinoma HEAD & NECK DOI /hed March 2011

4 Statistical Analysis. Lymph node metastases detected by the respective imaging modality were considered true-positive if the same metastasis was confirmed by histopathological workup. A lymph node metastasis was considered a false-negative if the lymph nodes were histopathologically confirmed as SCC but not detected by imaging. The sensitivity, specificity, accuracy, and the positive and negative predictive values (PPV and NPV, respectively) were calculated accordingly for each imaging modality for the endpoints of correct neck staging and differentiation of N0 versus Nþ. For the first endpoint, a falsepositive staging was defined as overstaging the pn classification; a false-negative staging as understaging the pn classification. A receiver operating characteristic (ROC) curve was generated to determine the differences with regard to imaging performance in detecting the correct nodal classification and differentiating between N0 and Nþ. Comparison of the areas under the curve was made between the modalities using z statistics. 11 For the correlation of the SUVmax with the grade of necrosis, a linear regression analysis was performed to set a model exploring the influence and relationship of the independent variable (grade of necrosis) on the dependent variable (SUVmax). To prove the usefulness of this model in predicting SUVmax, we tested the coefficient of variation for this linear regression model. All statistical tests were carried out using SPSS 17.0 for Windows and Excel A p value of less than.05 was considered to be statistically significant. RESULTS Table 1. Tumor characteristics: T/N classification. No. of patients pn0 pn1 pn2a pn2b pn2c pn3 pt pt pt pt Total Patient and Tumor Characteristics. Between 2002 and 2007, the total number of 34 patients (28 men, 6 women) was evaluated with a contrast-enhanced 18 F-FDG-PET/CT allowing separate analysis of contrast-enhanced CT, 18 F-FDG-PET, non-enhanced 18 F-FDG-PET/CT, and contrast-enhanced 18 F-FDG- PET/CT for a previously untreated TSCC. All patients underwent surgery (resection of the primary tumor and neck dissection for locoregional control) as firstline therapy. The mean age of the patients was 58 years (range, years). The primary tumor was located in the palatine tonsil in all patients. The vast majority of patients presented with cervical lymph node involvement and were, therefore, classified as stages III and IV in 29 of 34 cases (85.3%; International Union against Cancer UICC1997). The ptn classifications of all tumors are shown in Table 1. Neck dissection was able to reveal a total number of lymph node metastases of 116 with a mean number of 4 metastases (range, 0 23) per patient. The total number of investigated lymph nodes was 934 with a mean of 28 lymph nodes (range, 7 64) per patient, respectively. The morphologic pattern of the largest lymph node metastasis found in the neck specimen was as follows: 7 of 29 metastatic lymph nodes were found to be solid, 5 lymph nodes were mildly necrotic, 10 were moderately necrotic, and 7 were severely necrotic. The mean size of the largest lymph node metastasis found in the neck specimen was 3.5 cm (range, ). Overall, 36 neck dissections were performed in 34 patients with 2 patients undergoing bilateral neck dissection due to preoperative imaging findings. Performance and Accuracy of Imaging: Neck Staging. Nodal staging of all tumors by the different imaging modalities is shown in Table 2. The sensitivity, specificity, the PPV, the NPV, and the accuracy for the different imaging modalities with regard to correct pn classification are shown in Table 3. Further statistical analysis using ROC-analysis revealed statistically significant differences only between contrastenhanced 18 F-FDG-PET/CT and 18 F-FDG-PET alone (p ¼.004). The sensitivity, specificity, the PPV, the NPV, and the accuracy for the different imaging modalities with regard to pn0 versus pnþ are shown in Table 3. The statistical analysis using ROC-analysis showed a statistically significant difference between contrast-enhanced 18 F-FDG-PET/CT and non-enhanced 18 F-FDG-PET/CT (p ¼.017), between contrastenhanced 18 F-FDG-PET/CT and 18 F-FDG-PET alone (p ¼.017), between contrast-enhanced CT and nonenhanced 18 F-FDG-PET/CT (p ¼.017), and between contrast-enhanced CT and 18 F-FDG-PET (p ¼.017) alone, respectively. There was, however, no significant difference between contrast-enhanced CT and contrastenhanced 18 F-FDG-PET/CT. Correlation between SUVmax and the Grade of Necrosis. The mean SUVmax of the largest metastatic lymph nodes was 8.50 (SD 4.02; range, ). In a simple linear regression, we could detect a Table 2. Nodal (N) classification by different imaging modalities. No. of patients cn0 cn1 cn2a cn2b cn2c cn3 Contrast-enhanced CT Contrast-enhanced F-FDG-PET/CT Non-enhanced 18 F-FDG-PET/CT F-FDG-PET Abbreviation: FDG-PET, fluorodeoxyglucose-positron emission tomography. Imaging of Necrotic Lymph Node Metastases in Tonsillar Carcinoma HEAD & NECK DOI /hed March

5 Table 3. Correct pn classification and pn0 versus pnþ. Sensitivity, % Specificity, % PPV, % NPV, % Accuracy, % Contrast-enhanced CT Correct pn-classification pn0 vs pnþ F-FDG-PET Correct pn- classification pn0 vs pnþ Non-enhanced 18 F-FDG-PET/CT Correct pn- classification pn0 vs pnþ Contrast-enhanced 18 F-FDG-PET/CT Correct pn- classification pn0 vs pnþ Abbreviations: PPV, positive predictive value; NPV, negative predictive value; FDG-PET, fluorodeoxyglucose-positron emission tomography. negative, statistically significant correlation between SUVmax and the grade of necrosis (p ¼.011; r ¼ 0.44; Figure 4). DISCUSSION The development of cystic lymph node metastases in the neck is considered to be a peculiarity of a subset of HNSCC originating from the Waldeyer s ring. 2,12 Recent reports have demonstrated an increasing number of HPV-associated cancers with a strong predilection for the palatine and lingual tonsils. 13,14 Goldenberg et al 6 were able to reveal a strong association of HPV-related tonsillar SCC with cystic cervical lymph node involvement. Though nodal staging is accepted as the most important prognosticator in HNSCC, the debate on what imaging modality to use is still ongoing. In most cases, the modality that seems most appropriate for the primary tumor is also applied to the neck. With regard to the rising numbers of HPV-related tonsillar SCC, the accurate detection of centrally necrotic lymph node metastases is gaining more and more importance. Yousem et al 8 found contrast-enhanced CT outperforming MRI for this specific question. Conversely, Curtin et al 15 reported a comparable performance of contrastenhanced CT and MRI for the detection of metastatic lymph nodes. Due to the avidity of HNSCC for FDG, 18 F-FDG-PET/CT has been introduced for the diagnosis of HNSCC with promising results. In order to combine the metabolic information with the anatomic orientation for better localization of pathologic findings and reduction of false-positive lesions, the 18 F- FDG-PET is coregistered with a low-dose nonenhanced CT. In our subjective experience, cystic lymph node metastases were often missed by nonenhanced 18 F-FDG-PET/CT due to the lack of glucose metabolism in the necrotic tissue. On the other hand, the non-enhanced CT part was not able to detect the central necrosis and peripheral enhancement of these lymph nodes due to the missing contrast. Rodrigues et al 9 reported their advocacy of contrast-enhanced 18 F-FDG-PET/CT for the initial detection of nodal disease in patients with HNSCC. Therefore, the aim of this study was to assess which part of the contrastenhanced 18 F-FDG-PET/CT was the most indispensable for the detection of cystic lymph node metastases. To answer this question, the contrast-enhanced CT part, the 18 F-FDG-PET part, the non-enhanced 18 F-FDG-PET/ CT part, and the contrast-enhanced 18 F-FDG-PET/CT part were analyzed separately for nodal disease in a cohort of patients with TSCC. The histologic evaluation of the neck specimen served as a standard of reference. In our series, 75.9% of the investigated neck specimens showed either mildly necrotic (5 of 29), moderately necrotic (10 of 29), or severely necrotic (7 of 29) lymph node metastases confirming the high prevalence in tonsillar SCC. For the correct pn classification in our patient cohort, contrast-enhanced 18 F- FDG-PET/CT and contrast-enhanced CT achieved comparable results with a good accuracy of 75.6% and 73.9%, respectively. Contrast-enhanced 18 F-FDG-PET/ CT showed a trend to perform better (p ¼.085) than non-enhanced 18 F-FDG-PET/CT and a significantly FIGURE 4. Correlation between standardized uptake value (SUV) maximum and the grade of metastatic necrosis. 328 Imaging of Necrotic Lymph Node Metastases in Tonsillar Carcinoma HEAD & NECK DOI /hed March 2011

6 better performance (p ¼.004) than 18 F-FDG-PET alone. Nevertheless, the prediction of the correct N classification was significantly less accurate than the differentiation between N0 and Nþ. For this second endpoint, contrast-enhanced 18 F-FDG-PET/CT (91.9%) and contrast-enhanced CT (91.9%) reached the highest accuracies performing significantly better than non-enhanced 18 F-FDG-PET/CT and 18 F-FDG-PET alone. Therefore, it is important to state that imaging by either modality will understage and overstage nodal disease in a considerable number of patients. Contrast-enhanced CT and contrast-enhanced 18 F- FDG-PET/CT perform equally well suggesting that the addition of contrast to the CT is more important for detecting cystic metastases than the addition of FDG. The most accurate staging for the neck remains to be the histologic workup of the neck dissection specimen. In nonsurgical patients, the ultimate nodal staging remains unknown, which renders the comparison of surgical and nonsurgical therapy strategies difficult. For the neck staging, some authors have advocated ultrasound with fine-needle aspiration cytology to be most appropriate. 16,17 In order to be able to answer the clinically important question, whether a neck is nodally negative or positive, contrast-enhanced CT and contrast-enhanced 18 F-FDG-PET/CT both show excellent results superior to non-enhanced 18 F-FDG-PET/CT. Again, the contrast added to the CT seems to be decisive for the detection of cystic lymph nodes. Our findings with regard to the identification rate for necrosis by contrast-enhanced CT have been confirmed by other authors with sensitivities ranging from 74% to 91% and specificities from 93% to 94%. 3,18 Rodrigues et al 9 recently described an accuracy rate for detection of nodal necrosis by contrast-enhanced CT of 81%. In our study, we were able to reveal a negative correlation between the grade of lymph node necrosis and the SUVmax. The larger and more necrotic the lymph nodes are, the lower the glucose metabolism. In our opinion, this is the reason why nonenhanced 18 F-FDG-PET/CT and even more 18 F-FDG- PET alone are prone to miss these nodes. The contrast added to the CT highlights the vital periphery of the cystic lymph nodes in contrast to the dark necrotic center and makes these nodes more easily detectable. Overall, this study shows that non-enhanced 18 F-FDG-PET/CT, as routinely performed in most institutions, bears a considerable risk of missing necrotic neck metastases in tonsillar SCC. In patients with tonsillar SCC scheduled for 18 F-FDG-PET/CT, we suggest a contrast-enhanced 18 F-FDG-PET/CT to be performed. CONCLUSION The prevalence of cystic lymph node metastases in tonsillar SCC is high. Due to the rising number of HPV-related tonsillar SCC detection of cystic lymph nodes is gaining more importance. For the correct N classification and the differentiation between N0 and Nþ, contrast-enhanced CT and contrast-enhanced 18 F-FDG-PET/CT perform equally well and better than non-enhanced 18 F-FDG-PET/CT or 18 F-FDG- PET alone. Due to the missing glucose metabolism in necrotic lymph nodes, FDG is of less value than the contrast added to the CT for diagnosis. In patients with TSCC scheduled for 18 F-FDG- PET/CT, we strongly suggest that a contrastenhanced 18 F-FDG-PET/CT should be performed, which is not routinely done in most centers. REFERENCES 1. Layland MK, Sessions DG, Lenox J. The influence of lymph node metastasis in the treatment of squamous cell carcinoma of the oral cavity, oropharynx, larynx, and hypopharynx: N0 versus Nþ. Laryngoscope 2005;115: Thompson LD, Heffner DK. The clinical importance of cystic squamous cell carcinomas in the neck: a study of 136 cases. Cancer 1998;82: King AD, Tse GM, Ahuja AT, et al. Necrosis in metastatic neck nodes: diagnostic accuracy of CT, MR imaging, and US. Radiology 2004;230: Friedman M, Roberts N, Kirshenbaum GL, Colombo J. Nodal size of metastatic squamous cell carcinoma of the neck. Laryngoscope 1993;103: Don DM, Anzai Y, Lufkin RB, Fu YS, Calcaterra TC. Evaluation of cervical lymph node metastasis in squamous cell carcinoma of the head and neck. Laryngoscope 1995;105(7 Pt 1): Goldenberg D, Begum S, Westra WH, et al. Cystic lymph node metastasis in patients with head and neck cancer: an HPV-associated phenomenon. Head Neck 2008;30: Psyrri A, Gouveris P, Vermorken JB. Human papillomavirusrelated head and neck tumors: clinical and research implication. Curr Opin Oncol 2009;21: Yousem DM, Som PM, Hackney DB, Schwaibold F, Hendrix RA. Central nodal necrosis and extracapsular neoplastic spread in cervical lymph nodes: MR imaging versus CT. Radiology 1992; 182: Rodrigues RS, Bozza FA, Christian PE, et al. Comparison of whole-body PET/CT, dedicated high-resolution head and neck PET/CT, and contrast-enhanced CT in preoperative staging of clinically M0 squamous cell carcinoma of the head and neck. J Nucl Med 2009;50: Schmücking M, Baum RP, Bonnet R, Junker K, Müller KM. Correlation of histologic results with PET findings for tumor regression and survival in locally advanced non-small cell lung cancer after neoadjuvant treatment. [Article in German] Pathologe 2005;26: DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988;44: Regauer S, Mannweiler S, Anderhuber W, et al. Cystic lymph node metastases of squamous cell carcinoma of Waldeyer s ring origin. Br J Cancer 1999;79: Gillison ML, Koch WM, Capone RB, et al. Evidence for a causal association between human papillomavirus and a subset of head and neck cancers. J Natl Cancer Inst 2000;92: Mellin H, Dahlgren L, Munck-Wikland E, et al. Human papillomavirus type 16 is episomal and a high viral load may be correlated to better prognosis in tonsillar cancer. Int J Cancer 2002;102: Curtin HD, Ishwaran H, Mancuso AA, Dalley RW, Caudry DJ, McNeil BJ. Comparison of CT and MR imaging in staging of neck metastases. Radiology 1998;207: van den Brekel MW, Castelijns JA, Stel HV, Golding RP, Meyer CJ, Snow GB. Modern imaging techniques and ultrasoundguided aspiration cytology for the assessment of neck node metastases: a prospective comparative study. Eur Arch Otorhinolaryngol 1993;250: de Bondt RB, Nelemans PJ, Hofman PA, et al. Detection of lymph node metastases in head and neck cancer: a meta-analysis comparing US, USgFNAC, CT and MR imaging. Eur J Radiol 2007;64: van den Brekel MW, Stel HV, Castelijins JA, et al. Cervical lymph node metastasis: assessment of radiologic criteria. Radiology 1990;177: Imaging of Necrotic Lymph Node Metastases in Tonsillar Carcinoma HEAD & NECK DOI /hed March