PET/CT in the evaluation of treatment response to neoadjuvant chemoradiotherapy and prognostication in patients with locally advanced esophageal squamous cell carcinoma

OBJECTIVE: To investigate the role of fluorine-18-fluorodeoxyglucose PET/computed tomography for the prognostication and evaluation of neoadjuvant chemoradiotherapy response in locally advanced esophageal squamous cell carcinoma. METHODS: All consecutive biopsy-proven esophageal squamous cell carcinoma patients with PET/computed tomography at baseline (PET0) and 1 month after the completion of neoadjuvant chemoradiotherapy (PET1) between January 2008 and December 2013, followed by esophagectomy, were included. Maximum and mean standard uptake values (SUVmax and SUVmean), metabolic tumor volume, and total lesion glycolysis of all lesions at PET0 and PET1 were analyzed. Logistic and Cox regressions were used to identify factors predictive of pathological complete remission (pCR), overall survival, and recurrence-free survival. Cut-offs were identified using leave-one-out cross-validation adjusted receiver operator curve-based methods. A Kaplan-Meier model was adopted to compare survivals between groups using log-rank tests. RESULTS: Of a total of 52 patients (45 men, age 21-78 years), pCR was achieved in 21 (40.4%). SUVmax of primary tumor at PET1 was independently predictive of pCR [P=0.013, odds ratio=0.736, 95% confidence interval (CI): 0.578-0.937]; using a leave-one-out cross-validation-adjusted cut-off of 2.7, pCR could be predicted with a sensitivity of 71.0%, a specificity of 66.7%, a positive predictive value of 75.9%, and a negative predictive value of 60.9%. In the subset of 40 patients with standardized treatment included in survival analysis, total lesion glycolysis (P=0.002, hazard ratio: 1.029, 95% CI: 1.01-1.048) and SUVmax (P=0.003, hazard ratio: 1.167, 95% CI: 1.055-1.290) of nodal metastasis at PET0 were independently predictive of overall survival and recurrence-free survival, respectively. CONCLUSION: Baseline total lesion glycolysis and SUVmax of nodal metastases are significant independent predictors of survival, whereas post-treatment SUVmax of the primary tumor is predictive of pCR. However, the predictive value of the latter is modest, which may limit its clinical utility.