High-Risk Features of Esophageal Adenocarcinoma Following Neoadjuvant Chemoradiation: Patients for Whom Surgery Should Not Be Delayed

Objective: Clinical predictors of pathological complete response have not reliably identified patients for whom an organ-sparing approach following neoadjuvant chemoradiation be undertaken for esophageal cancer patients. We sought to identify high-risk predictors of residual carcinoma that may preclude patients from a selective surgical approach. Background: Patients treated with neoadjuvant chemoradiation followed by esophagectomy for esophageal adenocarcinoma were identified. Patients and Methods: Correlation between clinical and pathologic complete responses were examined. Regression models and recursive partitioning were utilized to identify features associated with residual carcinoma. External validation of these high-risk factors was performed on a data set from an independent institution. Results: A total of 326 patients were identified, in whom clinical complete response was noted in 104/326 (32%). Pathologic complete response was noted in only 33/104 (32%) of these clinical complete responders. Multivariable analysis identified that the presence of stricture (P=0.011), positive biopsy (P=0.010), and signet ring cell histology (P=0.019) were associated with residual cancer. Recursive partitioning corroborated a 94% probability of residual disease, or greater, for each of these features. The positive predictive value was >90% for these characteristics. A SUVmax >5.4 at the esophageal primary in the absence of esophagitis was also a high-risk factor for residual carcinoma. External validation confirmed these high-risk factors to be implicated in the finding of residual carcinoma. Conclusions: Clinical parameters of response are poor predictors of complete pathologic response leading to challenges in selecting candidates for active surveillance. However, we characterize several high-risk features for residual carcinoma which indicate that esophagectomy should not be delayed.