In a study (INSPiRED 006) reported in the Journal of Clinical Oncology, Cai et al found that an artificial intelligence–guided shear wave elastography (AI-SWE) model provided accuracy similar to expert reading of B-mode ultrasound in the diagnosis of breast cancer in women with Breast Imaging Reporting and Data System (BI-RADS) 3 or 4 breast masses. AI-SWE was associated with a reduced false-positive rate.
Study Details
The study involved data from an international multicenter trial assessing SWE in women with BI-RADS 3 or 4 breast masses, with images from multiple study sites being used to develop an EfficientNetB1-based deep learning model and images from one separate site serving as an external validation set. Additional validation was performed in a separate institutional set using the latest SWE software.
Key Findings
The development set included 924 patients (4,026 images). The external validation sets included 194 patients (562 images) and 176 patients (188 images, using the latest SWE software).
AI-SWE was associated with receiver operating characteristic area under the curve values of 0.94 (95% confidence interval [CI] = 0.91–0.96) in the first external validation set and 0.93 (95% CI = 0.88–0.98) in the external validation set using the latest SWE software.
In the first external validation set, sensitivity in diagnosis was similar with AI-SWE vs B-mode ultrasound (97.9% vs 98.1%, P = .912), whereas AI-SWE was associated with a significantly reduced false-positive rate (20.4% vs 53.8%, P < .001). In the external validation set using the latest SWE software, sensitivity was comparable (97.8% vs 98.1%, P = .810), whereas AI-SWE was associated with a significantly reduced false-positive rate (33.3% vs 53.8%, P = .001).
The investigators concluded: “AI-SWE demonstrated accuracy comparable with human experts in malignancy detection while significantly reducing false-positive imaging findings (ie, unnecessary biopsies). Future studies should explore its integration into multimodal breast cancer diagnostics.”
Michael Golatta, MD, of the Department of Obstetrics and Gynecology, Heidelberg University Hospital, Heidelberg, Germany, is the corresponding author for the Journal of Clinical Oncology article.
Disclosure: The study was supported by the HSO 2 Fellowship in Digital Oncology by the German Cancer Research Center and the National Center for Tumor Diseases Heidelberg. For full disclosures of all study authors, visit ascopubs.org.
