Nov. 12, 2024 — iCAD, Inc., a provider of clinically proven AI-powered cancer detection solutions, recently announced its ProFound Detection Version 4.0 for Digital Breast Tomosynthesis (DBT) has received clearance from the U.S. Food and Drug Administration (FDA). This next-generation AI solution, trained using advanced deep learning convolutional neural networks (CNN), offers advancements in cancer detection and specificity, achieving a 6.3% improved area under the receiver operating characteristic curve (AUC) over prior version, significantly improving the identification of hard-to-find and aggressive cancers while reducing false positives. This version also introduces an option for clinicians to incorporate a prior exam into a current exam’s ProFound Detection analysis.
With today’s FDA clearance, ProFound Detection Version 4.0 is now available for implementation in the U.S. Additional global and vendor-specific regulatory expansions are anticipated through 2025.
Compared to previous version 3, ProFound Detection Version 4.0 provides a 22% overall improvement in detecting some of the most challenging and aggressive cancer subtypes, including:
- 50% improvement in identifying cancers within dense breast tissue,
- 60% improvement in identifying invasive lobular cancers,
- 21% improvement in detecting invasive cancers, and
- 38% improvement in identifying cancers smaller than 1 cm
“With FDA clearance of ProFound Detection Version 4.0, iCAD continues to set new benchmarks in cancer detection, especially in the most challenging cases where accurate and early detection is critical,” said Dana Brown, President and CEO, iCAD. This "not only enhances detection for cancers feared most by even the expert fellowship-trained breast radiologists, but also reduces the burden of potential false positives, thereby providing clinicians with a highly precise and efficient AI concurrent-reader solution.”
“ProFound Detection Version 4.0 will dramatically improve breast cancer detection,” said Dr. Chirag Parghi, MD, MBA, Chief Medical Officer at Solis Mammography. “The newest version’s ability to detect subtle invasive cancers, especially within dense breast tissue, addresses one of the biggest challenges in breast imaging today. The improved precision in ProFound Detection’s lesion marking combined with the option to incorporate prior exams allows for improved clinical accuracy and more efficient reading workflows. Collectively, these upgrades will enhance the AI-enriched interpretation experience leading to meaningful adoption of AI."
Along with improved cancer detection capabilities, ProFound Detection Version 4.0 delivers more precise lesion marking, with an 18% improvement in cases with no marks, reducing potential false positives by enhanced specificity performance. This includes:
- 20% fewer marks related to vascular calcifications, and
- 51% fewer marks related to non-vascular calcifications.
ProFound Detection Version 4.0 also enables clinicians to incorporate a prior exam into its AI analysis and case-score/lesion assessment on the current case, emulating the approach radiologists take when interpreting current screening exams with historical context. By integrating prior imaging data, the software provides greater precision and insight, further enhancing clinical decision-making.
iCAD will showcase this next-generation AI at the upcoming 2024 Radiological Society of North America (RSNA) annual meeting in Chicago, Ill.
Visit iCAD’s RSNA event page to learn more and book a live demonstration: https://www.icadmed.com/about/news-events/upcoming-tradeshows-and-meetings/rsna-2024
August 29, 2024 
