Academic Journal
Fluent but not fail-safe: machine-translated radiology reports still need expert review.
| Τίτλος: | Fluent but not fail-safe: machine-translated radiology reports still need expert review. |
|---|---|
| Συγγραφείς: | Seibold C; Clinic for Diagnostic and Interventional Radiology, University Hospital Heidelberg, Heidelberg, Germany., Leichenich T; Clinic for Diagnostic and Interventional Radiology, University Hospital Heidelberg, Heidelberg, Germany., Stein P; Clinic for Diagnostic and Interventional Radiology, University Hospital Heidelberg, Heidelberg, Germany., Atsiatorme E; Clinic for Diagnostic and Interventional Radiology, University Hospital Heidelberg, Heidelberg, Germany., Kliesch P; Clinic for Diagnostic and Interventional Radiology, University Hospital Heidelberg, Heidelberg, Germany., Sevim GA; Clinic for Diagnostic and Interventional Radiology, University Hospital Heidelberg, Heidelberg, Germany., Trauth P; Clinic for Diagnostic and Interventional Radiology, University Hospital Heidelberg, Heidelberg, Germany., Kremer A; Clinic for Diagnostic and Interventional Radiology, University Hospital Heidelberg, Heidelberg, Germany., Bischoff A; Clinic for Diagnostic and Interventional Radiology, University Hospital Heidelberg, Heidelberg, Germany., Kauczor HU; Clinic for Diagnostic and Interventional Radiology, University Hospital Heidelberg, Heidelberg, Germany., Fink MA; Clinic for Diagnostic and Interventional Radiology, University Hospital Heidelberg, Heidelberg, Germany. matthias.fink@uni-heidelberg.de. |
| Πηγή: | European radiology experimental [Eur Radiol Exp] 2026 Oct 14; Vol. 10 (1). Date of Electronic Publication: 2026 Oct 14. |
| Τύπος έκδοσης: | Journal Article |
| Γλώσσα: | English |
| Στοιχεία περιοδικού: | Publisher: SpringerOpen Country of Publication: England NLM ID: 101721752 Publication Model: Electronic Cited Medium: Internet ISSN: 2509-9280 (Electronic) Linking ISSN: 25099280 NLM ISO Abbreviation: Eur Radiol Exp Subsets: MEDLINE |
| Imprint Name(s): | Original Publication: [London, United Kingdom] : SpringerOpen, [2017]- |
| Ιατρικοί όροι (MeSH): | Translating* , Radiology Information Systems* , Radiology*, Retrospective Studies ; Humans ; Germany ; Large Language Models |
| Περίληψη: | Objectives: Cross-border care, teleradiology, and international referral pathways increasingly require radiology report translation, but isolated mistranslations may alter clinical meaning. We evaluated German-to-English machine translation of radiology reports and correction-based localization of text requiring review. Materials and Methods: In this retrospective study, 310 de-identified German radiology reports from June and July 2025 were translated into English with DeepL. Eight radiologists assigned 5-point quality scores and corrected perceived errors. Thirty reports were reviewed by all eight readers, and 280 reports were assigned for single-reader evaluation, yielding 520 evaluations. Automated metrics and large language model judges were tested for report-level identification of translations requiring review; a correction-based segment-level edit-prediction model was evaluated on held-out shared reports. Results: Most translations were rated good or excellent (339/520, 65.2%), but clinically important mistranslations involved diagnosis, anatomy, and radiologic terminology. Agreement was low for ordinal quality scores (Krippendorff α = 0.078, p = 0.006) but high for expert corrections (pairwise "bilingual evaluation understudy"-BLEU = 83.0, p < 0.001). In the conservative primary screen of 30 shared reports, no automated method discriminated translations requiring review (area under receiver operating characteristic curve [AUC] = 0.51-0.65; p = 0.100-0.920). The final edit-prediction model achieved a concordance index of 0.697 and an AUC of 0.731 for segments edited by at least six of eight readers. Conclusion: Machine-translated radiology reports were often acceptable on average, but translations containing errors requiring expert review persisted. Expert corrections were more reproducible than ordinal quality scores and supported localization of text requiring human review. Key Points: Question Cross-border radiology workflows increasingly require report translation, but fluent machine translations may still contain meaning-changing errors requiring expert review for safe care. Findings Most machine translations were rated good or excellent, yet clinically important mistranslations occurred, and automated report-level metrics did not reliably identify reports needing review. Relevance statement Machine translation may support cross-border radiology workflows, but clinically important mistranslations in otherwise fluent reports require human-in-the-loop oversight and tools that localize text requiring expert review. (© 2026. The Author(s).) |
| Competing Interests: | Declarations. Ethics approval and consent to participate: This retrospective study used de-identified radiology reports and was approved by the institutional review board of University Hospital Heidelberg (S-083/2018). Informed consent was waived because the study was retrospective and all reports were de-identified before analysis. Consent for publication: Not applicable. This manuscript does not contain identifiable person data, images, or videos. All radiology report excerpts were de-identified before analysis and presentation. Competing interests: The authors declare that they have no competing interests. |
| References: | Eur Radiol Exp. 2025 Jan 15;9(1):7. (PMID: 39812924) Eur Radiol Exp. 2026 Feb 23;10(1):. (PMID: 41729370) BMJ. 2014 Dec 15;349:g7392. (PMID: 25512386) Eur Radiol. 2024 May;34(5):2817-2825. (PMID: 37794249) Eur Radiol Exp. 2024 Oct 30;8(1):124. (PMID: 39477904) BMJ. 2011 Jan 17;342:d296. (PMID: 21242212) Radiology. 2010 Oct;257(1):14-7. (PMID: 20851935) Radiology. 2023 Jun;307(5):e230877. (PMID: 37310247) Med Image Anal. 2023 May;86:102798. (PMID: 36989850) Eur Radiol Exp. 2025 Feb 18;9(1):24. (PMID: 39966263) J Biomed Inform. 2018 Sep;85:56-67. (PMID: 30031857) Insights Imaging. 2016 Aug;7(4):463-79. (PMID: 27188379) Eur Radiol Exp. 2025 Jun 19;9(1):61. (PMID: 40536631) |
| Contributed Indexing: | Keywords: Artificial intelligence; Language translation; Medical reports; Patient safety; Radiology |
| Entry Date(s): | Date Created: 20261014 Date Completed: 20261014 Latest Revision: 20261017 |
| Update Code: | 20261017 |
| PubMed Central ID: | PMC13669263 |
| DOI: | 10.1186/s41747-026-00827-7 |
| PMID: | 42873404 |
| Βάση Δεδομένων: | MEDLINE |
| ISSN: | 2509-9280 |
|---|---|
| DOI: | 10.1186/s41747-026-00827-7 |