Relationship between ileocolic arterial stump length and lymph node yield - a retrospective study using machine learning.
Authors
Affiliations (5)
Affiliations (5)
- Universidad Nacional Autónoma de México, Mexico City, Mexico. [email protected].
- Department of Radiology and Imaging, National Institute of Medical Sciences and Nutrition "Salvador Zubirán", Mexico City, Mexico.
- Universidad Nacional Autónoma de México, Mexico City, Mexico. [email protected].
- Department of Colorectal Surgery, National Institute of Medical Sciences and Nutrition "Salvador Zubirán", Vasco de Quiroga 15, Belisario Domínguez Seccion 16, Tlalpan, Mexico City, PC 14080, Mexico.
- Department of General and Endoscopic Surgery, "Dr. Manuel GEA González" General Hospital, Mexico City, Mexico.
Abstract
In right-sided colon cancer surgery, ileocolic artery stump length may reflect the extent of mesenteric resection and lymph node harvest. This study evaluated the association between postoperative stump length measured on CT, lymph node yield, and recurrence or progression within 24 months. This single-center retrospective study included 64 patients who underwent right or extended right hemicolectomy for right-sided colon adenocarcinoma between 2016 and 2024. Right-sided tumors were defined as those located in the cecum, ascending colon, hepatic flexure, or proximal transverse colon. Only primary surgical cases were included. Ileocolic artery stump length was measured on contrast-enhanced CT scans performed approximately 12 months postoperatively (± 1 month) by two independent radiologists. A Random Forest regression model using stump length as the only predictor variable was developed to estimate lymph node yield. Model performance was assessed using 10-fold cross-validation with R² and mean squared error (MSE). Exploratory recurrence/progression analysis used logistic regression with natural cubic splines. Mean stump length was 2.03 cm (range 1.0-4.5 cm), and median lymph node yield was 17 (range 10-55). Adequate nodal evaluation (≥ 12 nodes) was achieved in 85.9% of patients. The model demonstrated a consistent inverse relationship between stump length and lymph node yield (R² = 0.78; MSE = 0.16). Exploratory analysis suggested a non-significant trend toward higher recurrence or progression with longer stumps. Longer ileocolic artery stumps were associated with reduced lymph node yield, suggesting stump length may reflect the extent of mesenteric resection. External validation is required before clinical application.