Prognostic value of Mandard score and nodal status for recurrence patterns and survival after multimodal treatment of oesophageal adenocarcinoma.

BACKGROUND: This study evaluated the association of pathological tumour response (tumour regression grade, TRG) and a novel scoring system, combining both TRG and nodal status (TRG-ypN score; TRG1-ypN0, TRG>1-ypN0, TRG1-ypN+ and TRG>1-ypN+), with recurrence patterns and survival after multimodal treatment of oesophageal adenocarcinoma. METHODS: This Dutch nationwide cohort study included patients treated with neoadjuvant chemoradiotherapy followed by oesophagectomy for distal oesophageal or gastro-oesophageal junctional adenocarcinoma between 2007 and 2016. The primary endpoint was the association of Mandard score and TRG-ypN score with recurrence patterns (rate, location, and time to recurrence). The secondary endpoint was overall survival. RESULTS: Among 2746 inclusions, recurrence rates increased with higher Mandard scores (TRG1 30.6%, TRG2 44.9%, TRG3 52.9%, TRG4 61.4%, TRG5 58.2%; P < 0.001). Among patients with recurrent disease, the distribution (locoregional versus distant) was the same for the different TRG groups. Patients with TRG1 developed more brain recurrences (17.7 versus 9.8%; P = 0.001) and had a longer mean overall survival (44 versus 35 months; P < 0.001) than those with TRG>1. The TRG>1-ypN+ group had the highest recurrence rate (64.9%) and worst overall survival (mean 27 months). Compared with the TRG>1-ypN0 group, patients with TRG1-ypN+ had a higher risk of recurrence (51.9 versus 39.6%; P < 0.001) and worse mean overall survival (33 versus 41 months; P < 0.001). CONCLUSION: Improved tumour response to neoadjuvant therapy was associated with lower recurrence rates and higher overall survival rates. Among patients with recurrent disease, TRG1 was associated with a higher incidence of brain recurrence than TRG>1. Residual nodal disease influenced prognosis more negatively than residual disease at the primary tumour site.

Complication Prediction after Esophagectomy with Machine Learning.

Esophageal cancer can be treated effectively with esophagectomy; however, the postoperative complication rate is high. In this paper, we study to what extent machine learning methods can predict anastomotic leakage and pneumonia up to two days in advance. We use a dataset with 417 patients who underwent esophagectomy between 2011 and 2021. The dataset contains multimodal temporal information, specifically, laboratory results, vital signs, thorax images, and preoperative patient characteristics. The best models scored mean test set AUROCs of 0.87 and 0.82 for leakage 1 and 2 days ahead, respectively. For pneumonia, this was 0.74 and 0.61 for 1 and 2 days ahead, respectively. We conclude that machine learning models can effectively predict anastomotic leakage and pneumonia after esophagectomy.