Abstract

Objective: The present study aimed to analyze the treatment outcomes and develop a predictive model of surgical treatment for gastric cancer, taking into account the comorbidity of patients. Material and Methods. The treatment outcomes were studied in 477 (100 %) gastric cancer patients. Patients were stratifed into two groups according to the Age-Adjusted Charlson Comorbidity Index (ACCI) (0–4 points – 311 (65,2 %), more than 4 points – 166 (34,8 %) patients). Two hundred and seventy six patients (57.9 %) underwent distal or total gastrectomy, 178 (37.3 %) total gastrectomy, 19 (4 %) minimally invasive gastrectomy, and 4 (0.8 %) patients underwent extirpation of the gastric stump. The construction of models for the prognosis of surgical treatment was carried out using binary logistic regression, the comparison of models – by analyzing error curves (ROC-analysis). Results. The incidence rate of postoperative complications (grade III-V) according to the Clavien-Dindo classifcation was signifcantly higher in group 2 (9.9 %, 47/477) compared to group 1 (2.1 %, 10/477) (χ2 =64.79, p<0.001; OR 11.9 [5.82–24.3]). Postoperative mortality rate was 0.2 % (1/477) in group 1 and 1.9 % (9/477) in group 2 (χ2 =13.7, p<0.001; OR 17.8 [2.23–142]). In a comparative analysis, the prognostic model, taking into account the ACCI of the patient, allowed prediction of the development of postoperative grade III–V complications according to the Clavien-Dindo with the greatest accuracy. The area under the curve (AUC) was 0.921 ± 0.01 (95 % CI: 0,96–0,998), sensitivity – 71.9 %, specifcity – 99.8 %, diagnostic accuracy – 96.4 %. Conclusion. The inclusion of comorbidity in gastric cancer patients in the predictive concept makes it possible to improve the accuracy of prediction of postoperative complications of III–V grade according to Clavien-Dindo.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call