# Evolving Paradigms in Gastric Cancer Staging: From Conventional Imaging to Advanced MRI and Artificial Intelligence

## TL;DR

This paper reviews how gastric cancer staging is evolving with new imaging techniques like MRI and AI to improve accuracy and treatment decisions.

## Contribution

The paper highlights the emerging roles of MRI, FAPI PET tracers, and AI in improving gastric cancer staging beyond conventional imaging.

## Key findings

- MRI with DWI and multiparametric protocols outperforms CT in assessing serosal invasion and nodal involvement.
- FAPI PET tracers show promise in detecting mucinous and diffuse gastric cancer where FDG is limited.
- Radiomics and deep learning models offer new quantitative biomarkers for non-invasive risk stratification.

## Abstract

Background: Accurate preoperative staging is the cornerstone of therapeutic decision-making in gastric cancer (GC), yet standard modalities often fail to capture the full extent of disease, particularly in diffuse and poorly cohesive histotypes. This review aims to provide a comprehensive update on diagnostic imaging for GC, evaluating the established roles of CT, EUS, and PET/CT alongside the emerging capabilities of Magnetic Resonance Imaging (MRI) and Artificial Intelligence (AI). Methods: A structured narrative review was conducted by searching indexed biomedical databases for studies published between 2015 and 2024. A structured literature search screening process identified 410 relevant studies focusing on T, N, and M staging accuracy, quantitative imaging biomarkers, and radiomics. Results: While Multidetector CT remains the universal first-line modality, its sensitivity declines in infiltrative tumors and low-volume peritoneal carcinomatosis. EUS retains superiority for early (T1-T2) lesions but may offer limited value in advanced stages. Conversely, MRI (leveraging diffusion-weighted imaging (DWI) and multiparametric protocols) indicates superior soft-tissue contrast, potentially outperforming CT in the assessment of serosal invasion, nodal involvement, and occult peritoneal metastases. Furthermore, emerging fibroblast activation protein inhibitor (FAPI) PET tracers show promise in overcoming the limitations of FDG in mucinous and diffuse GC. Finally, radiomics and deep learning models are providing novel quantitative biomarkers for non-invasive risk stratification. Conclusions: Contemporary GC staging requires a tailored, multimodality approach. Evidence supports the increasing integration of MRI and quantitative imaging into clinical workflows to overcome the limitations of conventional techniques and support precision oncology.