Background: Neoadjuvant chemoradiotherapy before surgery can improve survival in patients with potentially curable esophageal cancer, but not all patients respond. Fluorodeoxyglucose positron emission tomography (FDG-PET) has been proposed to identify nonresponders early during neoadjuvant chemoradiotherapy. The aim of the present study was to determine whether FDG-PET could differentiate between responding and nonresponding esophageal tumors early in the course of neoadjuvant chemoradiotherapy.
Methods: This clinical trial comprised serial FDG-PET before and 14 days after start of chemoradiotherapy in patients with potentially curable esophageal carcinoma. Histopathologic responders were defined as patients with no or less than 10% viable tumor cells (Mandard score on resection specimen). PET response was measured using the standardized uptake value (SUV). Receiver operating characteristic analysis was used to evaluate the ability of SUV in distinguishing between histopathologic responders and nonresponders.
Results: In 100 included patients, 64 were histopathologic responders. The median SUV decrease 14 days after the start of therapy was 30.9% for histopathologic responders and 1.7% for nonresponders (P = 0.001). In receiver operating characteristic analysis, the area under the curve was 0.71 (95% CI = 0.60-0.82). Using a 0% SUV decrease cutoff value, PET correctly identified 58 of 64 responders (sensitivity 91%) and 18 of 36 nonresponders (specificity 50%). The corresponding positive and negative predictive values were 76% and 75%, respectively.
Conclusions: SUV decrease 14 days after the start of chemoradiotherapy was significantly associated with histopathologic tumor response, but its accuracy in detecting nonresponders was too low to justify the clinical use of FDG-PET for early discontinuation of neoadjuvant chemoradiotherapy in patients with potentially curable esophageal cancer.