FROM TUMOR RESECTION TO PROGNOSIS: THE ROLE OF AI IN FORECASTING RECURRENCE AND SURVIVAL IN SURGICAL ONCOLOGY

Abstract

Surgical resection of solid tumors still is the first choice in curative treatment. Up to now, large part of patients have been suffered from postoperative recurrence which largely decrease long survival or affects life quality. Current prognostic strategies (mostly TNM staging and clinicopathological features) provide low accuracy because they do not reflect the biology complexity and diversity of tumor progression. Artificial intelligence is a breakthrough technology that can improve prediction of recurrence and survival for surgical oncology by evaluating sophisticated analyses of high-dimensional, multimodal data.

This narrative review provides an overview of the emerging role of AI-based prognostication after tumor resection in common types of cancer, such as colorectal, lung, breast cancer and hepatobiliary, gastroesophageal and other solid tumors. ML, following the tool provided by DL and integrative multi-omics, can achieve higher prediction performance in contrast to conventional prognostic systems through revealing nonlinear knowledge based on radiological imaging, digital pathology images, genomic characteristics as well as clinical sources. There is promising potential for AI- based tools to inform individualized surveillance intervals, adjuvant therapy decision-making, consistent patient involvement in shared decision making and re-purposing healthcare resources at the time of treatment.

Despite the promising progress, challenges remain, such as data silos, lack of generalizability raised by available models interpretability issues, workflow integration difficulties and evolving ethical and regulatory frameworks. Emerging areas—explainable AI, federated learning, multimodal data integration and real-time intraoperative risk appraisal—will likely advance clinical translation and supporting precision-guided postoperative care. Under the current evidence of AI-based prognostic tools, this review identified promising and constraints on both aspects: interdisciplinary efforts are warranted to ensure competent, safe and equitable integration into surgical oncology practice.