Abstract
INTRODUCTION: This study aimed to externally evaluate the accuracy of four predictive scores for conversion to open surgery after rectal laparoscopic resection. None of the four scores achieved external validation previously. METHODS: This was a retrospective analysis of two prospectively maintained databases from two academic centers in France and Morocco. All consecutive patients who underwent laparoscopic resection for rectal adenocarcinoma between 2005 and 2020 were included. Logistic regression was used to assess the association between the factors present in the four scores and conversion. The accuracy of each score was assessed using the area under the curve (AUC). Observed and predicted conversion rates were compared for each score using the Chi-square goodness-of-fit test. RESULTS: Four hundred patients were included. There were 264 men (66%) with a mean age of 65.95 years (standard deviation 12.2). The median tumor height was 7 cm (quartiles 4-11) and 29% of patients had low rectal tumors. Conversion rate was 21.75%. The accuracy to predict conversion was low with an AUC lower than 0,62 for the four models. The observed conversion rates were significantly different from the predicted rates, except for one score. CONCLUSIONS: The four models had low accuracy in predicting the conversion to open surgery for laparoscopic rectal resection. There is a need for new well-designed studies, analyzing more specific variables, in a multicentric design to ensure generalizability of the results for daily surgical practice.