From radiomics to transformers in pancreatic cancer detection and prognosis

TL;DR

This paper reviews AI approaches for pancreatic cancer detection and prognosis, from radiomics to transformers, and highlights new ways to integrate diverse data for better clinical outcomes.

Contribution

The paper introduces a generational taxonomy of AI methods and evaluates their clinical applications while emphasizing multi-modal data integration.

Findings

A taxonomy of AI methods for pancreatic cancer spans from classical radiomics to transformer-based models.

Integration of imaging, pathology, and molecular data is critical for improving detection and prognosis.

Current models face limitations in generalizability and translational readiness.

Abstract

Pancreatic ductal adenocarcinoma (PDAC) remains one of the deadliest malignancies, primarily due to late diagnosis and poor therapeutic response. Advances in artificial intelligence (AI), particularly in medical imaging and multi-modal data integration, have created new opportunities for improving early detection and personalized prognostication.

This systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 statement. The protocol was prospectively registered with the Open Science Framework, covering studies published between 2015 and 2025.

Distinct from prior surveys that focus narrowly on specific algorithms or data types, this work introduces a generational taxonomy of AI approaches—ranging from classical radiomics-based machine learning to deep learning and contemporary transformer-based models—and maps their application to core clinical tasks such as detection, segmentation, classification, and outcome prediction. A key contribution is the integration of diverse datasets across imaging, pathology, and molecular sources; we further assess trends in availability, usage, and sample scale.

We critically evaluate limitations in generalizability, external validation, model calibration, and translational readiness, and outline recommendations for multi-center validation, standardized reporting, domain adaptation, and clinician-centered interpretability.

https://doi.org/10.17605/OSF.IO/2DVHJ.

Figures