The treatment schema included fluorouracil (1000 mg/m2 daily intravenous continuous infusion for 4 days), cisplatin (20 mg/m2 daily intravenous continuous infusion for 4 days), and radiotherapy (66-72 Gy at 180-200 cGy daily or 72 Gy at 120 cGy twice daily).
Tan A, Adelstein DJ, Rybicki LA, Saxton JP, Esclamado RM, Wood BG, Lorenz RR, Strome M, Carroll MA. Ability of Positron Emission Tomography to Detect Residual Neck Node Disease in Patients With Head and Neck Squamous Cell Carcinoma After Definitive Chemoradiotherapy. Arch Otolaryngol Head Neck Surg. 2007;133(5):435-440. doi:10.1001/archotol.133.5.435
Copyright 2007 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2007
To report our experience using the neck examination, computed tomography (CT), and positron emission tomography (PET) to clinically evaluate node-positive patients with head and neck squamous cell cancer for residual neck node disease after definitive chemoradiotherapy.
Retrospective review of all Cleveland Clinic patients with head and neck squamous cell cancer and N2 or N3 neck node involvement at presentation who were treated with definitive concurrent chemoradiotherapy and who underwent clinical restaging after treatment using the neck examination, CT, and PET.
Tertiary care referral institution.
Forty-eight patients with 72 positive necks at diagnosis were followed up for a median of 20 months.
Main Outcome Measures
Palpable nodes on examination, nodes larger than 1 cm, nodes with central necrosis on CT, or any hypermetabolic lymph nodes on PET were considered clinical evidence of residual nodal disease. The true rate of pathologic involvement was determined by histologic examination after planned neck dissection or if regional recurrence developed. The sensitivity, specificity, positive predictive value, negative predictive value, and accuracy were calculated for all 3 clinical assessment tools.
Planned neck dissection was performed in 33 necks and was positive for residual neck node disease in 5 necks. A delayed neck dissection was performed in 5 necks and was positive in 3 necks. The positive predictive value was low for all 3 clinical assessment tools. The addition of PET did not significantly improve the negative predictive value or positive predictive value of CT and the clinical examination.
Residual neck node disease after definitive chemoradiotherapy was infrequent and was not well predicted by PET. A positive PET finding in this setting is of little utility. Although a negative PET finding was highly predictive for control of neck disease after chemoradiotherapy, it added little to the clinical neck examination and CT.
In patients with head and neck squamous cell cancer, optimal management of the neck after definitive concurrent chemoradiotherapy is controversial. It is standard practice to perform at least a selective neck dissection in patients with residual palpable nodes or computed tomographic (CT) evidence of suspicious lymphadenopathy after completion of treatment. However, controversy arises about the role of neck dissection in the subset of patients with N2 or greater disease at presentation who have achieved a complete clinical and radiologic response to chemoradiotherapy. Based on the high response rates seen and the low risk of isolated neck failure after chemoradiotherapy, some recommend expectant management alone for these patients.1- 6 However, others advocate a planned neck dissection in all patients with N2 or N3 disease at diagnosis regardless of their clinical response to treatment.7- 13 This approach is supported by investigations7- 13 showing the presence of residual viable cancer in neck nodes even in those patients considered to have achieved a complete response, as well as by evidence of a lower rate of regional disease recurrence and improved survival in patients undergoing neck dissection. It also recognizes the lower success rate of salvage surgery when occult nodal disease becomes clinically manifest.14,15
Although the risk of neck dissection is generally considered to be low, complication rates after chemoradiotherapy, including fistula formation, chyle leak, and impaired wound healing, can be significant, particularly when the neck dissection is performed in conjunction with primary site resection.16- 18 Ideally, neck dissections should only be performed in patients with residual neck disease after definitive nonoperative treatment. The problem is how to effectively differentiate patients who will benefit from neck dissection from those who will not.
The clinical examination of the neck after chemoradiotherapy is often imprecise even when performed with the patient under anesthesia because of the body habitus and the scarring and edema that may occur after radiation. Similarly, conventional imaging using CT may be unable to distinguish residual nodal enlargement from nodal involvement and cannot reliably identify small-volume nodal disease after chemoradiation.
Positron emission tomography (PET) with 18F-fluorodeoxyglucose is being increasingly used in the staging and restaging of many types of malignant neoplasms. Uptake of the radiotracer 18F-fluorodeoxyglucose, an analogue of glucose, is substantially increased in most types of cancer cells compared with its uptake in healthy tissues, reflecting the increase in glucose metabolism by tumor cells.19- 21 The high sensitivity and specificity of PET in the detection of malignancy before definitive treatment is well documented.22 This has prompted investigations evaluating the use of PET as a restaging tool for patients with head and neck cancer to help identify residual nodal disease after chemoradiotherapy so that appropriate intervention can proceed.
The objective of this retrospective study was to assess the ability of PET to detect residual neck node disease in patients after definitive chemoradiotherapy. We detail our experience using clinical neck examination, CT, and PET and compare these results with the pathologic persistence or recurrence of regional disease.
In this study, we retrospectively reviewed the medical records of all patients presenting to our institution with N2-N3, M0 head and neck squamous cell cancer who were treated with definitive chemoradiotherapy and who underwent posttreatment restaging, including PET. These patients were identified from our prospectively collected Cleveland Clinic institutional review board–approved Head and Neck Cancer Chemoradiotherapy Registry.
At the Cleveland Clinic, patients with newly diagnosed, locoregionally advanced squamous cell head and neck cancer are offered definitive chemoradiotherapy as a treatment option unless there is an absolute indication for surgery (eg, bone or significant laryngeal cartilage involvement) or a contraindication to chemotherapy. Our chemoradiotherapy schedule consists of cisplatin and fluorouracil chemotherapy concurrent with external beam radiation (Figure). Radiotherapy is generated by a 6-MV linear accelerator using opposed lateral fields. An electron beam boost is given to selected nodal regions as indicated. A total dose of 66 to 72 Gy in once-daily fractions of 180 to 200 cGy, or 72 Gy in twice-daily fractions of 120 cGy, is given without planned interruption. During the first and fourth weeks of radiotherapy, patients also receive 4-day continuous intravenous infusions of cisplatin (20 mg/m2 per day) and fluorouracil (1000 mg/m2 per day). Clinical restaging is then performed 8 to 12 weeks after completion of treatment, and consideration is given to a neck dissection and to primary site salvage surgery if necessary.
This posttreatment restaging includes a careful neck examination and CT. Since 2002, we have also gradually incorporated PET in this clinical evaluation. Patients are considered to have persistent clinical evidence of neck disease if there are palpable lymph nodes on neck examination. Computed tomography criteria for clinically persistent disease include residual nodes larger than 1 cm, marginal enhancement following intravenous contrast administration, or central nodal necrosis. Positron emission tomography criteria include any nodes with any residual hypermetabolic uptake. No standard uptake value (SUV) cutoff was used given the lack of evidence that a particular cutoff improves the diagnostic accuracy of PET.23- 25 For the purposes of this study, patients with clinical evidence of bilateral neck node disease were assigned a separate N designation for each positive neck.
Pathologic evidence of neck node involvement was only confirmed after histologic examination of neck specimens obtained at the time of planned neck dissection or after neck dissection performed for regional recurrence identified during follow-up. Although neck dissection was recommended for all patients with N2 or N3 disease at presentation, the decision to proceed with neck surgery was based on the judgment of the responsible head and neck surgeon. Neck procedures performed included radical neck dissection, modified radical neck dissection, selective neck dissection, or excisional lymph node biopsy.
The χ2 test was used to determine whether findings from the clinical examination, CT, or posttreatment PET were predictive of neck node positivity. The sensitivity, specificity, negative predictive value (NPV), positive predictive value (PPV), and accuracy were also calculated for all 3 clinical assessment tools and their combinations. Sensitivity is defined as the percentage of patients with residual or recurrent neck node disease who have a positive test result, and specificity is the percentage of patients without residual or recurrent neck node disease who have a negative test result. Negative predictive value is the percentage of patients who are truly negative for disease in the neck given that the test result is negative, while the PPV is the percentage of patients who are truly positive for disease in the neck given that the test result is positive. Accuracy is the extent to which the results of PET reflect true disease status in the neck. Follow-up time was calculated from initiation of treatment until the time of death or last contact.
The study included 48 patients with N2 or N3 disease at presentation. These patients had a total of 72 necks positive for neck node disease. Patient characteristics are given in Table 1. All patients had stage IV disease at presentation. Oropharyngeal cancer was the most common primary site diagnosis. Among 72 necks, the nodal disease status was 5 N0 necks (7%), 20 N1 necks (28%), 13 N2a necks (18%), 23 N2b necks (32%), and 11 N3 necks (15%). The 5 N0 necks reflect 5 patients with ipsilateral N2 or N3 disease who were considered to have false-positive PET findings in the contralateral neck. The mean SUV in neck nodes in patients with a positive pretreatment PET finding in the neck was 12.1, with a median of 10 (range, 2.9-30.3). Restaging PET was performed at a median of 2.5 months (range, 1.1-5.3 months) after completion of definitive chemoradiation. Patients had been followed up for a median of 20 months (range, 9.9-54.8 months) as of September 24, 2006. Four patients had died (2 from unrelated causes while free of disease).
After chemoradiotherapy, 33 planned neck dissections were performed in 27 patients. Pathologic evidence of residual neck node disease was found in 5 necks. Recurrent primary site or neck node disease prompted 5 delayed neck dissections in 4 patients. Three necks were positive for residual neck node disease. No neck dissection was performed in 34 necks. In total, residual or recurrent disease was identified in 8 (11%) of 72 necks after chemoradiotherapy. The types of surgical procedures performed (of 38 procedures) were 3 radical neck dissections (8%), 6 modified radical neck dissections (16%), 25 selective neck dissections (66%), and 4 excisional lymph node biopsies (11%).
Table 2 summarizes the results of clinical restaging, CT, and PET vs the pathologic persistence or recurrence of disease. The clinical examination proved to be a better predictor of pathologic positivity than CT. Positron emission tomography was of little value in distinguishing patients with and without residual disease.
Table 3 summarizes these results as sensitivity, specificity, PPV, NPV, and accuracy. Despite the low PPVs of all 3 clinical staging modalities and their combinations, the NPVs were high. This presumably reflects the infrequency of residual or recurrent neck node disease. There was no SUV cutoff for PET positivity; any increased uptake was considered positive. The mean SUV in neck nodes in patients with a positive restaging PET finding was 4.0, with a median of 3.4 (range, 2.6-8.7).
The findings from this study suggest that no single staging modality or combination of modalities is sufficiently sensitive and specific in the patient reassessment for residual neck node disease after chemoradiotherapy. All modalities, including PET, have a high NPV, similar to results reported by others.26- 28 Given the infrequency of residual or recurrent regional disease after chemoradiotherapy, this is not surprising. It is the disappointingly low PPV and sensitivity of these modalities that point out our limited clinical ability to identify those patients in need of a neck dissection.
Does PET help the clinician at all with this decision? We address this question by assessing the contribution of PET to the combination of the neck examination and CT in the restaging of the neck after chemoradiotherapy. Table 4 summarizes the common clinical scenarios faced by the surgeon. Although the subgroups are small, the assessments are illustrative.
Few would argue with the need for a neck dissection if both the clinical examination and CT suggest residual disease or even if only one of them is positive. Here, PET adds little. The PPV of PET in this setting is low, and the NPV is not high enough to allow the clinician to avoid surgery. The more difficult question is whether PET is of value when both the restaging clinical examination and CT are negative in a patient with N2 or N3 disease at presentation. Once again, the study adds little. The PPV is low, and the negative PET finding only slightly improves the NPV of negative CT and clinical examination findings from 93% to 95%.
This suggests that little is gained by adding PET to our clinical reassessment. Several explanations can be offered. It seems that concurrent chemoradiotherapy is effective in sterilizing neck node disease. In this series, residual or recurrent disease was identified in 8 (11%) of 72 necks. A 31% incidence of residual or recurrent neck node disease in patients with N2 or N3 disease was previously reported using this chemoradiotherapy regimen.11 The difference observed in this more recent cohort is incompletely understood. However, it may reflect the greater use of hyperfractionated radiotherapy schedules, the increasing incidence of tumors originating in the oropharynx, and the shorter follow-up period.
Positron emission tomography has limited sensitivity.29,30 Small or microscopic foci of tumor cannot be identified using PET. Unfortunately, the clinician needs the most help identifying precisely those patients with the small or microscopic foci of residual disease. In addition, the uptake of this radiotracer is not tumor specific and can reflect increased glucose use in any hypermetabolic tissue or process, including infection and inflammation.31- 33 This makes interpretation of PET particularly challenging in the postradiation setting owing to false-positive results caused by treatment-induced inflammation. Adding to this false positivity is the fact that, to our knowledge, no clear SUV cutoff has been defined to maximize the diagnostic accuracy of PET.23- 25 Further work will be needed to establish an appropriate SUV that is considered normal in the postchemoradiation setting.
The optimal PET timing after chemoradiation also remains undefined. If the study is performed too early after completion of treatment, false-positive results are more frequent because of insufficient time for the resolution of postchemoradiotherapy inflammatory changes in the neck. False-negative rates are also increased because the residual disease may not have reached the resolution threshold detectable by PET.34- 37
The low PPV of PET in detecting residual disease in the neck after definitive therapy is well documented.26,29,30,38,39 Although a high NPV has also been reported for PET, there is little suggestion that PET provides additional information to conventional evaluation using CT and the neck examination.39- 42 Therefore, current evidence would argue that the inclusion of PET in the restaging evaluation of patients after chemoradiotherapy does not seem to change the management of the neck, making it difficult to justify the additional cost of obtaining routine PET in this setting.
The potential for PET/CT to improve the results achieved using PET alone has been described.42- 44 In our series, some patients had separate PET and CT, while others had a combined study. In a combined study, the results of each test affect the interpretation of the other, imparting a degree of subjectivity to the study. Herein, we chose to independently assess the results from PET and CT. It is unclear how the interpretation of combined PET and CT would have affected the results.
This review has several inherent limitations. Although the patient population was treated using the same concurrent chemoradiotherapy regimen, there was variability in the posttreatment timing of PET. Most restaging PET was performed at least 8 weeks after chemoradiation, timing that is consistent with what has been recommended in other studies.29,34- 37 In addition, no SUV cutoff was used in this study, and PET was considered positive if any increased metabolic uptake was noted, potentially increasing our false-positive rates and reducing our PPV. In this study, we attempted to use PET to predict the presence of residual neck node disease (found in 5 patients) and the development of regional recurrence (found in 3 patients). This is likely not an entirely reasonable expectation for this test.
In conclusion, in patients with N2 or N3 head and neck squamous cell cancer, residual or recurrent neck node disease after definitive chemoradiotherapy is infrequent and is not well predicted by PET. A positive PET finding in this setting is of little utility. Although a negative PET finding was predictive for control of neck disease after chemoradiotherapy, it added little to the clinical examination and CT.
Correspondence: Ann Tan, MD, c/o David J. Adelstein, MD, Taussig Cancer Center and Department of Solid Tumor Oncology, Cleveland Clinic, Desk R-35, 9500 Euclid Ave, Cleveland, OH 44195 (firstname.lastname@example.org).
Submitted for Publication: August 21, 2006; final revision received October 19, 2006; accepted November 12, 2006.
Author Contributions: Drs Tan, Adelstein, Esclamado, Wood, and Lorenz and Ms Rybicki had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Tan, Adelstein, Saxton, Esclamado, Wood, Strome, and Carroll. Acquisition of data: Tan, Adelstein, Esclamado, and Wood. Analysis and interpretation of data: Tan, Adelstein, Rybicki, Esclamado, and Lorenz. Drafting of the manuscript: Tan, Adelstein, and Rybicki. Critical revision of the manuscript for important intellectual content: Tan, Adelstein, Rybicki, Saxton, Esclamado, Wood, Lorenz, Strome, and Carroll. Statistical analysis: Rybicki. Obtained funding: Adelstein. Administrative, technical, and material support: Tan, Adelstein, and Esclamado. Study supervision: Adelstein and Lorenz.
Financial Disclosure: None reported.
Funding/Support: This study was supported in part by a grant from the Taussig Cancer Center.
Previous Presentation: This study was presented in part at the American Head and Neck Society 2006 Annual Meeting and Research Workshop on the Biology, Prevention, and Treatment of Head and Neck Cancer; August 19, 2006; Chicago, Ill.