The results showed that the accuracy of Transpara™ was comparable to the average of the radiologists, with Transpara AI being more accurate than over 60% of the radiologists in the study
NIJMEGEN, Netherlands (PRWEB) March 11, 2019
In the study, “Stand-Alone Artificial Intelligence for Breast Cancer Detection in Mammography: Comparison with 101 Radiologists,” researchers retrospectively compared the breast cancer detection performance of Transpara™ AI software to that of 101 breast screening radiologists. This comprehensive study, using independent data from the United States and Europe, included more than 28,000 unique radiologist readings of approximately 2,600 mammography exams (650 with cancer characteristics), acquired with systems from four different vendors.
Like a radiologist, Transpara™ AI provided a likelihood of malignancy score for each mammogram in this dataset. Using these scores, the accuracy of both Transpara™ AI and the radiologists was measured in terms of area under the receiver operating characteristic curve (AUC). The results showed that the accuracy of Transpara™ was comparable to the average of the radiologists, with Transpara AI being more accurate than over 60% of the radiologists in the study (AUC-Transpara™ AI = 0.840, AUC-Radiologists = 0.814). This result is further enhanced by the similar sensitivity performance of Transpara™ AI in comparison to the radiologists at different recall rates.
The results of this extensive study show the high accuracy and reliability of Transpara™ to detect breast cancer in mammograms. According to the researchers, “Artificial intelligence that functions at the level of an expert radiologist for breast cancer detection in mammograms might herald a change in the breast healthcare workflow, whether in a screening or in a clinical setting”. This first-of-its-kind study evaluating AI for breast cancer detection in mammography enhances ScreenPoint Medical’s leading position as a commercial provider of AI software solutions to help radiologists in the early detection of breast cancer.
“High-performing AI systems such as Transpara™ AI could become an essential tool for radiologists in the screening process, to alleviate the workload and increase performance. Given the flexibility of the software, which can also operate for 2D and 3D mammograms, radiologists can use Transpara™ AI in the way that fits their needs the best: concurrently for support, for triaging of exams, or as an independent reader. Furthermore, the robustness shown by Transpara™ AI to detect breast cancer in mammograms from different vendors and its integration with leading PACS workstations makes Transpara™ a tool suitable for every need,” said Prof. Nico Karssemeijer, founder and CEO of ScreenPoint Medical.
Utilizing state-of-the-art image analysis and revolutionary deep learning technology, Transpara™ automatically identifies soft-tissue and calcification lesions in 2D and 3D mammograms. Findings are shown to radiologists via a unique interactive decision support interface, a scientifically-proven method that boosts reading performance, as opposed to traditional mammography CAD systems. In addition, based on the findings from all views of the exam, each case is categorized using the innovative Transpara™ Score, which categorizes the exams according to their risk of harboring cancer and can be used to automatically triage exams with confidence.
Transpara™ is FDA cleared for digital mammography, has European regulatory approval (CE Mark) for use with digital mammography and digital breast tomosynthesis and is compatible with images from leading mammography systems.
About ScreenPoint Medical BV:
ScreenPoint Medical develops image analysis technology for automated reading of mammograms and digital breast tomosynthesis exams, exploiting Big Data, Deep Learning and the latest developments in Artificial Intelligence. ScreenPoint Medical was founded in 2014 by Nico Karssemeijer and Michael Brady, two experts in breast imaging, machine learning, computer vision, and computer-aided detection. The main office is in Nijmegen, The Netherlands. For more information, please contact firstname.lastname@example.org