John Hermans

Radboud UMC

In short, John Hermans's research entails the following:

Artificial Intelligence for Opportunistic Screening of Pancreatic Cancer on Medical Images

Pancreatic ductal adenocarcinoma, the most common type of pancreatic cancer, is expected to become the second leading cause of cancer-related deaths in Western countries by 2030. Due to the lack of early, disease-specific symptoms, 80–85% of patients are diagnosed only at an advanced stage. However, patients with stage I have a significantly better prognosis than those with stage IV (median survival: 26 versus 4.8 months), making early detection the most effective strategy to improve outcomes.

Contrast-enhanced CT scans are the most common type of CT imaging and are the first choice for diagnosing pancreatic cancer. Studies show that in 16%–84% of diagnosed patients, secondary signs of cancer, such as dilation of the pancreatic duct and pancreatic atrophy, are already visible on pre-diagnostic CT scans 3 to 36 months before clinical diagnosis. However, these features and early subtle focal abnormalities can be missed in clinical practice, leading to delayed diagnosis and reduced survival.

This project aims to develop and validate an AI algorithm that can be implemented in routine clinical care to detect pancreatic cancer at an early stage, significantly improving patients' survival rates and quality of life.

Short summary of progress/results February 2026

The project progresses according to plan, with near completion of data collection and AI inference for the retrospective study. Early results indicate that the current AI algorithm can identify pre-diagnostic pancreatic cancer cases on routine CECT scans. Ongoing labelling and upcoming analyses will further clarify clinical utility.

The establishment of international collaborations and preparation for a prospective  clinical trial position the project well to deliver significant scientific and clinical impact in the coming years.