QVCAD System

The QVCAD System aims to improve breast cancer detection in women with dense breast tissue by serving as a computer-aided detection tool for 3D Automated Breast Ultrasound (ABUS). It employs neural network-based feature extraction on volumetric data from the Somo.v device, analyzing tissue patterns in the coronal plane to highlight suspicious lesions for radiologists.

The QVCAD System is a sophisticated computer-aided detection (CAD) software specifically developed to enhance breast cancer screening in women with dense breast tissue using 3D Automated Breast Ultrasound (ABUS). Its primary function is to process volumetric ultrasound data acquired from the Somo.v ABUS device. The system meticulously scans the coronal plane of the breast ultrasound volumes, employing advanced tissue pattern analysis algorithms to identify and highlight potentially suspicious regions. These regions are then marked on-screen, providing visual cues to radiologists during their review. Underpinning QVCAD's detection capabilities is a robust science centered on neural-network-based feature extraction. This neural network is meticulously trained with a comprehensive dataset comprising examples of both benign and malignant breast tissues, enabling it to distinguish subtle pathological patterns. Clinical evidence suggests that QVCAD significantly improves the detection rate of lesions that are typically not visible or are obscured on traditional mammograms, particularly in dense breasts, without causing an undue increase in false positive findings. Consequently, QVCAD serves as a valuable adjunct tool, augmenting the radiologist's interpretive skills and supporting more effective breast cancer screening protocols for this specific patient population.

AI Regulatory Concerns

Powered by

Jev reads the full FDA letter and returns calibrated concern probabilities per regulatory category. These are AI estimates, not regulatory advice.

• Q1. What specific sensitivity and specificity metrics does the QVCAD system achieve in detecting mammographically occult cancers in dense breasts when used with the Somo.v device?; How diverse was the training dataset for the neural network in terms of lesion types, sizes, and patient demographics, and how does this affect its generalizability?; What is the learning curve for radiologists integrating QVCAD into their existing ABUS workflow, and how does it impact reading times?; Does the QVCAD system demonstrate differential performance in detecting specific histological subtypes of breast cancer (e.g., invasive ductal vs. invasive lobular carcinoma) in dense tissue?; How does the performance of QVCAD-assisted ABUS compare to other supplemental screening modalities, such as breast MRI, in women with dense breasts and an elevated risk?