Differences in Preoperative Chemotherapy Regimens Influence the Criteria and Content of Primary Tumor Effects Assessed by Computed Tomography

Background In the treatment of esophageal cancer (EC), it is important to evaluate the tumor effect of neoadjuvant chemotherapy (NAC) in order to predict prognosis. However, although it is fully expected that the evaluation criteria for anti-tumor effect and the characteristics of effect may differ depending on the NAC regimen, there are few reports that have examined these. In EC patients who underwent NAC followed by surgery, we compared the differences in the optimal cutoff value for primary tumor effect and the characteristics of recurrence rates after stratification between different NAC regimens. Methods We included 199 consecutive EC patients treated with a triplet regimen of NAC (DCF:docetaxel+cisplatin+5FU &FAP:5FU + adriamycin+cisplatin) followed by surgery between April 2004 and March 2020. ROC curves were used to determine optimal cut-off values, including the reduction rate of the product of the shorter and longer diameter (SLD), and responders were defined as cases exceeding each cut-off value. We compared to recurrence-free survival (RFS) curves between responders and non-responders using log-rank test. In addition, the recurrence rates for responders/non-responders were examined in the FAP and DCF groups. Results The DCF regimen tended to be superior to the FAP regimen. The optimal cutoff value for primary tumor effect was 37.8% for all cases, 30.8% for the FAP group, and 37.8% for the DCF group, with each being significantly stratified by Kaplan–Meier analysis. The recurrence rates for responders/non-responders after stratification were as follows: In the FAP regimen, 45.7% of responders had recurrence and 54.3% did not, while 80.0% of non-responders had recurrence and 20.0% did not. In the DCF regimen, 25.8% of responders had recurrence and 74.2% did not, while 58.5% of non-responders had recurrence and 41.5% did not. Conclusion The antitumor effect of DCF tended to be stronger than that of FAP, and reflecting this, the optimal cutoff value was higher in the DCF regimen than in the FAP regimen. Additionally, the prognostic prediction accuracy of responder cases in the FAP regimen was low, while the prediction accuracy of non-responder cases in the DCF regimen was low. From the above, it was inferred that the more tumor-controlling the regimen, the higher the prognostic prediction accuracy for responders. However, the prognostic prediction accuracy for non-responders was lower, indicating that further investigation into non-responders is necessary.