Resumen de Multivariate prognostic models for patients with stages i and ii colon carcinoma: A strobe-compliant retrospective cohort study

Luis F. Oñate Acuña, Roberto Herrera Goepfert, Alejandro Avilés Salas, Carlo C. Cortés González, Sagrario González Trejo, Jose F. Carrillo, Erika Ruiz García, Francisco J. Ochoa Carrillo, Vincenzo Aiello Crocifoglio, Claudia M. García Cuellar

• Colorectal cancer is the most frequent gastrointestinal malignancy worldwide. The value of adjuvant treatment is controversial in Stages I and II. Objective: The aim of this study was to construct post-operative prognostic models applicable to patients with stages I-II colon carcinoma (CC). Methods: This is a retrospective cohort study of patients with Stage I-II CC treated over a 25-year period. Exposure was defined as clinical, histopathological, and immunohistochemical factors (including CDX2 and MUC2 expression). Patients were randomly allocated to either a “modeling set” or a “validation set”. Factors associated with recurrence, disease-free survival (DFS), and overall survival (OS) were defined in the “modeling set”. Their performances were tested in the “validation set”. Results: From a total of 556 recruited patients, 339 (61%) were allocated to the “modeling set” and 217 (39%) to the “validation set”. Three models explaining recurrence, DFS, and OS were described. Tumor location in the left colon (Hazards ratio [HR] = 1.57; 95% Confidence interval [CI] 0.99-2.48), lymphocyte (HR = 0.46; 96% CI 0.27-0.88) and monocyte (HR = 0.99; 95% CI 0.99-1) counts, neutrophil/platelet ratio (HR = 1.3; 95% CI 0.74-2.3, and HR = 2.3; 95% CI 1.3-4.1; for second and third category, respectively), albumin/monocyte ratio (HR = 0.43; 95% CI 0.21-0.87), and microscopic residual disease after surgery (HR = 8.7; 95% CI 3.1-24) were independently associated with OS. T classification and expression of CDX2 and/or MUC2 were not independently associated with recurrence or prognosis. Conclusion: These models are simple and readily available, and distinguish the risk and prognosis in patients with CC stages I and II; these models require cheaper processes than the use of more sophisticated molecular biology techniques. They may guide either the need for adjuvant therapy versus post-operative surveillance only, as well as aid in the design of clinical trials. (REV INVEST CLIN. 2023;75(5):259-71)