EUVIMEDEUVIMED entdecken
VergleichsstudieOpen Access · CC BY 4.0

From ACR O‐RADS 2022 to Explainable Deep Learning

Ali Abbasian Ardakani iD, Afshin Mohammadi, Alisa Mohebbi iD, Anushya Vijayananthan, Sook Sam Leong iD, Yi Ting Lim iD, Mohd Kamil Bin Mohamad Fabell, Gernot Hudelist iD, Balint Balogh iD, Marlina Tanty Ramli Hamid, U Rajendra Acharya iD, Sepideh Hatamikia iD

Journal of Ultrasound in Medicine · 2026-12 · 45 (12) · 2871-2885

DOI: 10.1002/jum.70401 ↗

Kurz gefasst

Vergleichsstudie in der Zeitschrift Journal of Ultrasound in Medicine, online erschienen am 09.08.2026. Verfasst von Ali Abbasian Ardakani und 11 weiteren Autor:innen. Beteiligte Einrichtung: Department of Medicine, Faculty of Medicine and Dentistry, Danube Private University, Krems an der Donau, Austria. Hauptthemen laut MEDLINE: Deep Learning, Ovarian Neoplasms, Image Interpretation, Computer-Assisted, Radiology Information Systems. Die Arbeit nennt 29 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Objectives

The 2022 update of the Ovarian‐Adnexal Reporting and Data System (O‐RADS) improves risk stratification of adnexal lesions; however, radiologist interpretation remains subject to inter‐observer variability and conservative diagnostic thresholds. Concurrently, deep learning (DL) models demonstrated promise in ovarian mass characterization. This study evaluates radiologist performance applying O‐RADS version 2022 (v2022), compares it to convolutional neural network (CNN) and vision transformer (ViT) models, and investigates diagnostic gains from hybrid human–artificial intelligence (AI) frameworks with emphasis on explainable DL approaches that could enhance clinical applicability.

Methods

In this retrospective study, a total of 512 ultrasound images from 227 patients (110 with at least 1 malignant lesion) were analyzed. Sixteen DL models, including DenseNets, EfficientNets, ResNets, VGGs, Xception, and ViTs were trained and validated. For each model, a hybrid framework integrating radiologist‐assigned O‐RADS scores with DL‐predicted malignancy probabilities was constructed.

Results

Radiologist‐only O‐RADS assessment achieved an area under the curve (AUC) of 0.683 and an accuracy of 68.0%. CNN models yielded AUCs of 0.620–0.908 and accuracies of 59.2–86.4%, while ViT16‐384 reached the best performance, with an AUC of 0.941 and an accuracy of 87.4%. Hybrid human–AI frameworks significantly enhanced most CNNs (9 out of 12 CNNs, p < .05) and ViTs (3 out of 4 ViTs, p < .05).

Conclusions

DL models outperform radiologist‐only O‐RADS v2022 assessment. The integration of expert radiologist scores with AI yields the highest accuracy, supporting hybrid human–AI paradigms as a promising approach to standardize ultrasound interpretation, reduce false‐positive diagnoses, and improve identification of high‐risk ovarian lesions.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Deep LearningOvarian NeoplasmsImage Interpretation, Computer-AssistedRadiology Information SystemsFemaleHumansConvolutional Neural NetworksUltrasonographyRetrospective StudiesRisk AssessmentRadiologistsReproducibility of Results

Stichwörter der Autor:innen

O‐RADS v2022deep learningovarian cancerultrasoundvision transformer

Autor:innen und Institutionen

  • Department of Medicine, Faculty of Medicine and Dentistry, Danube Private University, Krems an der Donau, Austria.
  • Department of Radiology, Faculty of Medicine, Urmia University of Medical Science, Urmia, Iran.
  • Department of Biomedical Imaging, Universiti Malaya Research Imaging Centre, Faculty of Medicine, Universiti Malaya, Kuala Lumpur, Malaysia.
  • Centre for Medical Imaging Studies, Faculty of Health Sciences, Universiti Teknologi MARA Selangor, Shah Alam, Malaysia.
  • Department of Gynaecology, Hospital St John of God, Vienna, Austria.
  • Dpt. of Gynaecology, Jaghiellonian University, Krakow, Poland.

ORCID-Profile: Ali Abbasian Ardakani, Alisa Mohebbi, Sook Sam Leong, Yi Ting Lim, Gernot Hudelist, Balint Balogh, U Rajendra Acharya, Sepideh Hatamikia

Zitieren

APA

Ardakani, A. A., Mohammadi, A., Mohebbi, A., Vijayananthan, A., Leong, S. S., Lim, Y. T., Fabell, M. K. B. M., Hudelist, G., Balogh, B., Hamid, M. T. R., Acharya, U. R., & Hatamikia, S. (2026). From ACR O‐RADS 2022 to Explainable Deep Learning. Journal of Ultrasound in Medicine, 45(12), 2871-2885. https://doi.org/10.1002/jum.70401

Vancouver

Ardakani AA, Mohammadi A, Mohebbi A, Vijayananthan A, Leong SS, Lim YT, et al. From ACR O‐RADS 2022 to Explainable Deep Learning. Journal of Ultrasound in Medicine. 2026;45(12):2871-2885. doi:10.1002/jum.70401

BibTeX
@article{Ardakani2026,
  author = {Ardakani, Ali Abbasian and Mohammadi, Afshin and Mohebbi, Alisa and Vijayananthan, Anushya and Leong, Sook Sam and Lim, Yi Ting and Fabell, Mohd Kamil Bin Mohamad and Hudelist, Gernot and Balogh, Balint and Hamid, Marlina Tanty Ramli and Acharya, U Rajendra and Hatamikia, Sepideh},
  title = {From ACR O‐RADS 2022 to Explainable Deep Learning},
  journal = {Journal of Ultrasound in Medicine},
  year = {2026},
  volume = {45},
  number = {12},
  pages = {2871--2885},
  doi = {10.1002/jum.70401}
}
Kostenlos starten →