Supporting scientific research
UMC Utrecht / TU Eindhoven
In short, the fellowship of Ruben Lucassen entails the following:
Pathologists are faced with increasing workload due to a rising number of cases for examination, the need for more comprehensive diagnoses, and a worsening shortage of pathologists. These factors increase the risk of delays and misdiagnoses, which could ultimately affect patient care. Artificial intelligence (AI) for pathology image analysis has advanced tremendously in recent years and is regarded as a promising solution to improve diagnostic efficiency and accuracy, which could alleviate workload, save on costs, and lead to better patient outcomes.
Despite the growing number of published studies demonstrating strong predictive performance of AI models on retrospective datasets, only very few of these models reach the point of clinical use. Several factors likely contribute to this translation gap: (1) the limited emphasis at the problem identification stage on estimating the required model performance to justify the associated costs; (2) the technical challenges associated with integrating AI models into the digital pathology workflow, which is often necessary for prospective validation; and (3) the need for careful assessment of the impact of the AI models on the workflow after implementation.
Ruben Lucassen, who is expected to finish his PhD on “computational pathology for cutaneous melanocytic lesion assessment” in 2026, plans to learn more about healthcare simulation, AI implementation, and impact assessment from experts in the respective fields though several research visits during his Fellowship. With the gained knowledge and experience, he intends to start bridging the gap between AI model development and clinical benefit. Affiliation (supervisor):
Planned research visits (supervisor):