EUVIMEDEUVIMED entdecken
EditorialKommentar

Beyond the stopwatch: establishing diagnostic trust in AI-accelerated prostate MRI

Arnaldo Stanzione, Jeries Paolo Zawaideh, Tristan Barrett

European Radiology · 28.04.2026 · 36 (8) · 6450-6452

DOI: 10.1007/s00330-026-12575-8 ↗

Kurz gefasst

Editorial in der Zeitschrift European Radiology, online erschienen am 28.04.2026. Verfasst von Arnaldo Stanzione und 2 weiteren Autor:innen. Beteiligte Einrichtung: Department of Advanced Biomedical Sciences, University of Naples Federico II, Naples, Italy. arnaldo.stanzione@unina.it. Die Arbeit nennt 9 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The global burden of prostate cancer (PCa) continues to rise, necessitating scalable and ef fi cient diagnostic pathways. While multiparametric MRI (mpMRI) has revolutionized the detection and risk strati fi cation of clinically signi fi cant prostate cancer (csPCa), its implementation as a fi rst-line screening tool remains hindered by long acquisition times, high costs, and limited throughput. Standard protocols typically require 25 – 30 min of room time, creating a bottleneck in healthcare systems facing increasing demand. Where do we stand today? Although abbreviated biparametric protocols (bpMRI) have reduced scan times, they often still exceed 15 min, which is insuf fi cient for high-volume screening. Consequently, the pertinent question is no longer whether we can accelerate prostate MRI, but how far we can push acceleration factors using AI before diagnostic reliability is compromised. The source article by van Lohuizen et al [1] contributes a robust evaluation of this question through a retrospective, multi-reader, multi-case study. The authors assessed a state-of-the-art AI reconstruction model, vSHARP, trained on multi-coil k-space data. By testing acceleration factors of R = 3 and R = 6, they demonstrated that a sixfold acceleration of T2-weighted (T2w) sequences, reducing scan times to as little as 33 seconds, resulted in no statistically signi fi cant reduction in csPCa detection compared to fully sampled scans (AUROC 0.80 vs 0.86; p = 0.08). A critical added value of this work is its scale: interpreted by eight experienced radiologists and validated against histopathology, it provides high-level evidence

Autor:innen und Institutionen

  • Department of Advanced Biomedical Sciences, University of Naples Federico II, Naples, Italy. arnaldo.stanzione@unina.it.
  • Department of General and Interventional Radiology, IRCCS Azienda Ospedaliera Metropolitana, Genova, Italy.
  • Department of Radiology, University of Cambridge, Cambridge, UK.
  • Department of Radiology, Cambridge University Hospitals NHS Foundation Trust Addenbrooke's Hospital, Cambridge, UK.

Zitieren

APA

Stanzione, A., Zawaideh, J. P., & Barrett, T. (2026). Beyond the stopwatch: establishing diagnostic trust in AI-accelerated prostate MRI. European Radiology, 36(8), 6450-6452. https://doi.org/10.1007/s00330-026-12575-8

Vancouver

Stanzione A, Zawaideh JP, Barrett T. Beyond the stopwatch: establishing diagnostic trust in AI-accelerated prostate MRI. European Radiology. 2026;36(8):6450-6452. doi:10.1007/s00330-026-12575-8

BibTeX
@article{Stanzione2026,
  author = {Stanzione, Arnaldo and Zawaideh, Jeries Paolo and Barrett, Tristan},
  title = {Beyond the stopwatch: establishing diagnostic trust in AI-accelerated prostate MRI},
  journal = {European Radiology},
  year = {2026},
  volume = {36},
  number = {8},
  pages = {6450--6452},
  doi = {10.1007/s00330-026-12575-8}
}
Kostenlos starten →