DeepBT Detector-Plus

DeepBT Detector-Plus is an AI software system that assists trained medical professionals in radiation therapy treatment planning by automatically generating initial contours of brain tumors (brain metastases, meningiomas, and acoustic neuromas) on MRI images. The system accepts both T1-weighted contrast-enhanced and T2-weighted MRI images, outputs results in DICOM format for viewing and editing in treatment planning systems, and requires medical professionals to confirm or modify the contours before clinical use.

The primary difference from the predicate device (VBrain) is that DeepBT Detector-Plus can process both single-parametric (T1W+C only) and bi-parametric (T1W+C and T2W) MRI images simultaneously, whereas VBrain supports only single-parametric T1W+C images. Both use deep learning neural networks, operate on Linux, integrate with PACS networks, and target the same three tumor types for adult patients. The bi-parametric capability provides improved tumor boundary identification for certain tumors like cystic lesions.

IEC 62304:2006/A1:2016 (Medical device software life cycle processes) and ISO 14971:2019 (Risk management for medical devices). Software verification and validation followed FDA's June 14, 2023 guidance on 'Content of Premarket Submissions for Device Software Functions' at the Enhanced Documentation Level.

DeepBT Detector-Plus is substantially equivalent to VBrain because both are AI-based software systems for brain tumor contouring in radiation therapy with identical intended clinical use, same tumor types, same patient population (adults), and comparable performance metrics. Although DeepBT supports an additional bi-parametric imaging capability that improves accuracy, both products require physician finalization of contours and do not alter clinical workflow, and the performance data demonstrates that DeepBT's results are comparable to the predicate across multiple evaluation metrics.