Development and validation of an interpretable model integrating multimodal information for improving ovarian cancer diagnosis.

Ovarian cancer, a group of heterogeneous diseases, presents with extensive characteristics with the highest mortality among gynecological malignancies. Accurate and early diagnosis of ovarian cancer is of great significance. Here, we present OvcaFinder, an interpretable model constructed from ultrasound images-based deep learning (DL) predictions, Ovarian-Adnexal Reporting and Data System scores from radiologists, and routine clinical variables. OvcaFinder outperforms the clinical model and the DL model with area under the curves (AUCs) of 0.978, and 0.947 in the internal and external test datasets, respectively. OvcaFinder assistance led to improved AUCs of radiologists and inter-reader agreement. The average AUCs were improved from 0.927 to 0.977 and from 0.904 to 0.941, and the false positive rates were decreased by 13.4% and 8.3% in the internal and external test datasets, respectively. This highlights the potential of OvcaFinder to improve the diagnostic accuracy, and consistency of radiologists in identifying ovarian cancer.

Development and validation of ultrasound-based radiomics model to predict germline BRCA mutations in patients with breast cancer.

BACKGROUND: Identifying breast cancer (BC) patients with germline breast cancer susceptibility gene (gBRCA) mutation is important. The current criteria for germline testing for BC remain controversial. This study aimed to develop a nomogram incorporating ultrasound radiomic features and clinicopathological factors to predict gBRCA mutations in patients with BC. MATERIALS AND METHODS: In this retrospective study, 497 women with BC who underwent gBRCA genetic testing from March 2013 to May 2022 were included, including 348 for training (84 with and 264 without a gBRCA mutation) and 149 for validation(36 patients with and 113 without a gBRCA mutation). Factors associated with gBRCA mutations were identified to establish a clinicopathological model. Radiomics features were extracted from the intratumoral and peritumoral regions (3 mm and 5 mm) of each image. The least absolute shrinkage and selection operator regression algorithm was used to select the features and logistic regression analysis was used to construct three imaging models. Finally, a nomogram that combined clinicopathological and radiomics features was developed. The models were evaluated based on the area under the receiver operating characteristic curve (AUC), calibration, and clinical usefulness. RESULTS: Age at diagnosis, family history of BC, personal history of other BRCA-related cancers, and human epidermal growth factor receptor 2 status were independent predictors of the clinicopathological model. The AUC of the imaging radiomics model combining intratumoral and peritumoral 3 mm areas in the validation set was 0.783 (95% confidence interval [CI]: 0.702-0.862), which showed the best performance among three imaging models. The nomogram yielded better performance than the clinicopathological model in validation sets (AUC: 0.824 [0.755-0.894] versus 0.659 [0.563-0.755], p = 0.007). CONCLUSION: The nomogram based on ultrasound images and clinicopathological factors performs well in predicting gBRCA mutations in BC patients and may help to improve clinical decisions about genetic testing.

Deep Learning-assisted Diagnosis of Breast Lesions on US Images: A Multivendor, Multicenter Study.

Purpose: To evaluate the diagnostic performance of a deep learning (DL) model for breast US across four hospitals and assess its value to readers with different levels of experience. Materials and Methods: In this retrospective study, a dual attention-based convolutional neural network was built and validated to discriminate malignant tumors from benign tumors by using B-mode and color Doppler US images (n = 45 909, March 2011-August 2018), acquired with 42 types of US machines, of 9895 pathologic analysis-confirmed breast lesions in 8797 patients (27 men and 8770 women; mean age, 47 years ± 12 [SD]). With and without assistance from the DL model, three novice readers with less than 5 years of US experience and two experienced readers with 8 and 18 years of US experience, respectively, interpreted 1024 randomly selected lesions. Differences in the areas under the receiver operating characteristic curves (AUCs) were tested using the DeLong test. Results: The DL model using both B-mode and color Doppler US images demonstrated expert-level performance at the lesion level, with an AUC of 0.94 (95% CI: 0.92, 0.95) for the internal set. In external datasets, the AUCs were 0.92 (95% CI: 0.90, 0.94) for hospital 1, 0.91 (95% CI: 0.89, 0.94) for hospital 2, and 0.96 (95% CI: 0.94, 0.98) for hospital 3. DL assistance led to improved AUCs (P < .001) for one experienced and three novice radiologists and improved interobserver agreement. The average false-positive rate was reduced by 7.6% (P = .08). Conclusion: The DL model may help radiologists, especially novice readers, improve accuracy and interobserver agreement of breast tumor diagnosis using US.Keywords: Ultrasound, Breast, Diagnosis, Breast Cancer, Deep Learning, Ultrasonography Supplemental material is available for this article. © RSNA, 2023.