Adnexal Lesion Discrimination Using Deep Learning Analysis of Dynamic Contrast-enhanced US Images

Purpose To develop a multimodality deep learning model (Ovarian Cancer Network [OCNet]) using dynamic contrast-enhanced US images to classify adnexal lesions. Materials and Methods This retrospective study included patients with pathologically confirmed adnexal lesions detected at US across 14 hospitals in China between January 2018 and July 2023. Data were divided into the training set (<i>n</i> = 275), internal testing set (<i>n</i> = 57), and external testing set (<i>n</i> = 63). Two deep learning models (OCNet<sub>manual</sub> and OCNet<sub>automated</sub>) were developed and compared with Ovarian-Adnexal Reporting and Data System (O-RADS) US and the Assessment of Different Neoplasias in the Adnexa (ADNEX) model. Diagnostic performances of radiologists with and without assistance of OCNet were also assessed. Results A total of 395 female patients (median age, 43 years; IQR, 31-55 years) were included (252 benign and 143 malignant lesions). OCNet<sub>manual</sub> and OCNet<sub>automated</sub> achieved an area under the receiver operating characteristic curve (AUC) of 0.94 (95% CI: 0.89, >0.99) and 0.91 (95% CI: 0.83, 0.99), respectively, outperforming O-RADS US (AUC, 0.79; 95% CI: 0.68, 0.89; <i>P</i> = .002 and <i>P</i> = .03, respectively) and the ADNEX model (AUC, 0.86; 95% CI: 0.77, 0.95; <i>P</i> = .04 and <i>P</i> = .36, respectively). Additionally, the assistance of OCNet enhanced diagnostic performance for junior radiologists, improving the average AUC from 0.86 to 0.94 and the average specificity from 52% to 73%. Conclusion The OCNet model achieved higher performance than O-RADS US and the ADNEX model for classifying adnexal lesions and improved the diagnostic performance of junior radiologists. <b>Keywords:</b> Adnexal Lesion, Deep Learning, Contrast-enhanced US, Multimodal <i>Supplemental material is available for this article.</i> © RSNA 2025 See also commentary by Huber and Adams in this issue.

Authors

Institutions

Publication Details

Journal
Radiology Artificial Intelligence
Published
2025-11-05
DOI
https://doi.org/10.1148/ryai.240786
Citations
1
Primary Topic
Ovarian cancer diagnosis and treatment
Type
article
Field-Weighted Citation Impact
1.70
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Adnexal Lesion Discrimination Using Deep Learning Analysis of Dynamic Contrast-enhanced US Images

Xinling Zhang, Shuangyu Wu, Peter T. W. Cheng, Li Xiao et al.
1 citations
Radiology Artificial Intelligence
Ovarian cancer diagnosis and treatment
1.70
article

Adnexal Lesion Discrimination Using Deep Learning Analysis of Dynamic Contrast-enhanced US Images

Xinling Zhang, Shuangyu Wu, Peter T. W. Cheng, Li Xiao, Manting Su, Ying Chen, Tingting Liu, Manli Wu, Enze Qu, Xiaotao Meng, Yaping Wang, Ruixuan Wang, Juan Liu, Tianming Liang, Jian Gu, Hong Wen, Xiaofeng Sun, Jian–Fang Hu, Liang Mu, Hong Yang, Rui Zhang, S. Y. Chen, Qinghong Zhao, Man Zhang, Ying Wang, Ruili Wang
article en
1 citations

Abstract

Purpose To develop a multimodality deep learning model (Ovarian Cancer Network [OCNet]) using dynamic contrast-enhanced US images to classify adnexal lesions. Materials and Methods This retrospective study included patients with pathologically confirmed adnexal lesions detected at US across 14 hospitals in China between January 2018 and July 2023. Data were divided into the training set (<i>n</i> = 275), internal testing set (<i>n</i> = 57), and external testing set (<i>n</i> = 63). Two deep learning models (OCNet<sub>manual</sub> and OCNet<sub>automated</sub>) were developed and compared with Ovarian-Adnexal Reporting and Data System (O-RADS) US and the Assessment of Different Neoplasias in the Adnexa (ADNEX) model. Diagnostic performances of radiologists with and without assistance of OCNet were also assessed. Results A total of 395 female patients (median age, 43 years; IQR, 31-55 years) were included (252 benign and 143 malignant lesions). OCNet<sub>manual</sub> and OCNet<sub>automated</sub> achieved an area under the receiver operating characteristic curve (AUC) of 0.94 (95% CI: 0.89, >0.99) and 0.91 (95% CI: 0.83, 0.99), respectively, outperforming O-RADS US (AUC, 0.79; 95% CI: 0.68, 0.89; <i>P</i> = .002 and <i>P</i> = .03, respectively) and the ADNEX model (AUC, 0.86; 95% CI: 0.77, 0.95; <i>P</i> = .04 and <i>P</i> = .36, respectively). Additionally, the assistance of OCNet enhanced diagnostic performance for junior radiologists, improving the average AUC from 0.86 to 0.94 and the average specificity from 52% to 73%. Conclusion The OCNet model achieved higher performance than O-RADS US and the ADNEX model for classifying adnexal lesions and improved the diagnostic performance of junior radiologists. <b>Keywords:</b> Adnexal Lesion, Deep Learning, Contrast-enhanced US, Multimodal <i>Supplemental material is available for this article.</i> © RSNA 2025 See also commentary by Huber and Adams in this issue.

Radiology Artificial IntelligenceVol. 8(1)
Nanchang University (CN), Sun Yat-sen University (CN), Shanxi Medical University (CN), Wuhan University (CN), Chengdu Women's and Children's Central Hospital (CN), Shaanxi Provincial People's Hospital (CN), The First Affiliated Hospital of Guangxi University of Traditional Chinese Medicine (CN), Shaoxing City Women and Children Hospital (CN), Baogang Group (China) (CN), First Affiliated Hospital of GuangXi Medical University (CN), First Affiliated Hospital of Chengdu Medical College (CN), Henan Provincial People's Hospital (CN), Chengdu Fifth People's Hospital (CN), Huizhou Central People's Hospital (CN), Renmin Hospital of Wuhan University (CN), Third Affiliated Hospital of Sun Yat-sen University (CN), Shanxi Provincial Children's Hospital (CN), Nanchong Central Hospital (CN), Shanxi Provincial People’s Hospital (CN), First Affiliated Hospital of Nanchang University (CN), Huazhong University of Science and Technology (CN)
Openalex Percentile: Top 12%
Ovarian cancer diagnosis and treatment
1.70
DOISourceOpenAlex
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.