Development and Validation of a Predictive Model of Therapeutic Effect in Patients with Esophageal Squamous Cell Carcinoma Who Received Neoadjuvant Treatment: A Nationwide Retrospective Study in Japan.

Abstract

Neoadjuvant treatment has become the standard of care for patients with advanced esophageal cancer. However, few studies have elucidated the clinical factors that predict response to neoadjuvant therapy in a large multicenter cohort. We aimed to develop a predictive model of therapeutic effect in patients with esophageal squamous cell carcinoma (SCC) who received neoadjuvant treatment. This nationwide study included 4078 patients from 85 institutions. Patients who received neoadjuvant treatment followed by surgery for esophageal SCC were eligible. We developed a logistic regression model to predict good pathological therapeutic effects, and a predictive nomogram was generated by applying the logistic regression formula. Among neoadjuvant regimens, cisplatin plus 5-fluorouracil (CF) was the most frequently used (60.2%), followed by docetaxel plus CF (DCF, 27.4%), CF with radiotherapy (CF-RT, 4.5%), adriamycin plus CF (3.6%), nedaplatin plus 5-fluorouracil (0.9%), and DCF-RT (0.5%). Multivariable analysis revealed that male sex, advanced cT category, and increased pretherapeutic SCC antigen level were independently associated with not achieving a good therapeutic effect. Moreover, intensified neoadjuvant regimens were independently associated with favorable therapeutic effects; DCF-RT elicited the best therapeutic effect, followed by CF-RT and DCF. A predictive model including nine commonly measured preoperative variables was generated, and the area under the curve was 0.679 (95% confidence interval: 0.658-0.700). This nomogram was also adequately validated internally. The model developed in this study was validated and predicts the therapeutic effect in patients with esophageal SCC who received neoadjuvant treatment. This model might contribute to individualized treatment strategies.