Future proofing radiology workflows through MR-STAT and Artificial Intelligence

Abstract

This PhD thesis investigates several interrelated bottlenecks in radiology: long MRI acquisition times, limited quantitative interpretability of conventional MRI, increasing radiologist workload, and challenges associated with emerging technologies such as AI into clinical practice. The central premise was that technical innovation alone is insufficient. Rather, meaningful clinical impact requires early evidence generation, validation in clinically relevant settings, iterative feedback with end users, and reflection on broader moral, economic and organizational implications. The results presented in Chapters 2-4 demonstrate that fast, quantitative MRI using MR-STAT can contribute to alleviating multiple technical and clinical constraints simultaneously, while Chapter 5 explores implementation scenarios for MR-STAT and Chapter 6 shows that AI adoption in radiology approached as a socio-technical challenge provides relevant insights, compared to viewing it as a purely technical one. Together, these findings support the overarching thesis that early, reflective evaluation of emerging (imaging) technology increase the likelihood for successful and responsible clinical implementation.