Beyond Post hoc Explanations: A Comprehensive Framework for Accountable AI in Medical Imaging Through Transparency, Interpretability, and Explainability
Bioengineering · 15.08.2025 · 12 (8) · 879
DOI: 10.3390/bioengineering12080879 ↗Kurz gefasst
Übersichtsarbeit in der Zeitschrift Bioengineering, online erschienen am 15.08.2025. Verfasst von Yashbir Singh und 10 weiteren Autor:innen. Beteiligte Einrichtung: Radiology, Mayo Clinic, Rochester, MN 55905, USA. Die Arbeit wurde bisher 76-mal zitiert (Crossref) und nennt 38 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
The integration of artificial intelligence (AI) in medical imaging has revolutionized diagnostic capabilities, yet the black-box nature of deep learning models poses significant challenges for clinical adoption. Current explainable AI (XAI) approaches, including SHAP, LIME, and Grad-CAM, predominantly focus on post hoc explanations that may inadvertently undermine clinical decision-making by providing misleading confidence in AI outputs. This paper presents a systematic review and meta-analysis of 67 studies (covering 23 radiology, 19 pathology, and 25 ophthalmology applications) evaluating XAI fidelity, stability, and performance trade-offs across medical imaging modalities. Our meta-analysis of 847 initially identified studies reveals that LIME achieves superior fidelity (0.81, 95% CI: 0.78–0.84) compared to SHAP (0.38, 95% CI: 0.35–0.41) and Grad-CAM (0.54, 95% CI: 0.51–0.57) across all modalities. Post hoc explanations demonstrated poor stability under noise perturbation, with SHAP showing 53% degradation in ophthalmology applications (ρ = 0.42 at 10% noise) compared to 11% in radiology (ρ = 0.89). We demonstrate a consistent 5–7% AUC performance penalty for interpretable models but identify modality-specific stability patterns suggesting that tailored XAI approaches are necessary. Based on these empirical findings, we propose a comprehensive three-pillar accountability framework that prioritizes transparency in model development, interpretability in architecture design, and a cautious deployment of post hoc explanations with explicit uncertainty quantification. This approach offers a pathway toward genuinely accountable AI systems that enhance rather than compromise clinical decision-making quality and patient safety.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Radiology, Mayo Clinic, Rochester, MN 55905, USA.
- Department of Radiology, University of Pennsylvania, Philadelphia, PA 19104, USA.
- Division of Gastroenterology & Hepatology, Mayo Clinic, Jacksonville, MN 32224, USA.
- Department of Surgery, Mayo Clinic, Rochester, MN 55905, USA.
- Sheikh Shakhbout Medical City (SSMC), Abu Dhabi P.O. Box 11001, United Arab Emirates.
- Department of Radiology, Università di Padova, 35122 Padua, Italy.
ORCID-Profile: Yashbir Singh, Quincy A Hathaway, Sara Salehi, Yujia Wei, Diana V Vera-Garcia, Emilio Quaia, Jesper B Andersen
Zitieren
APA
Singh, Y., Hathaway, Q. A., Keishing, V., Salehi, S., Wei, Y., Horvat, N., Vera-Garcia, D. V., Choudhary, A., Mula Kh, A., Quaia, E., & Andersen, J. B. (2025). Beyond Post hoc Explanations: A Comprehensive Framework for Accountable AI in Medical Imaging Through Transparency, Interpretability, and Explainability. Bioengineering, 12(8), 879. https://doi.org/10.3390/bioengineering12080879
Vancouver
Singh Y, Hathaway QA, Keishing V, Salehi S, Wei Y, Horvat N, et al. Beyond Post hoc Explanations: A Comprehensive Framework for Accountable AI in Medical Imaging Through Transparency, Interpretability, and Explainability. Bioengineering. 2025;12(8):879. doi:10.3390/bioengineering12080879
BibTeX
@article{Singh2025,
author = {Singh, Yashbir and Hathaway, Quincy A and Keishing, Varekan and Salehi, Sara and Wei, Yujia and Horvat, Natally and Vera-Garcia, Diana V and Choudhary, Ashok and Mula Kh, Almurtadha and Quaia, Emilio and Andersen, Jesper B},
title = {Beyond Post hoc Explanations: A Comprehensive Framework for Accountable AI in Medical Imaging Through Transparency, Interpretability, and Explainability},
journal = {Bioengineering},
year = {2025},
volume = {12},
number = {8},
pages = {879},
doi = {10.3390/bioengineering12080879}
}
EUVIMED