![]() Patient age, screening interval, and method of cancer detection were representative of a real-world screening population. Secondly, simulated double reading performance using AI as an independent second reader was compared to historical human double reading.Īll comparisons were determined on the same unenriched cohorts. The AI system was evaluated firstly comparing the AI system’s standalone performance to the historical first human reader, the only guaranteed independent read at all participating sites. This study aimed to evaluate whether a novel AI system could act as a reliable independent reader while automating a substantial part of the double reading workflow, and to demonstrate standalone performance compared to historical results. Such studies should evaluate model performance on images from various hardware vendors, using the most relevant screening metrics. Rigorous large-scale studies are needed to assess performance of AI in double reading on diverse cohorts of women across multiple screening sites and programmes, and on unenriched screening data representative of populations the AI will process in real-world deployments ( 28). ![]() AI and its potential to positively transform clinical practice on real-world screening populations remains to be confirmed, as also highlighted in a recent systematic review ( 27). The imaging datasets were also significantly skewed towards a single mammography hardware vendor. These included small-scale reader studies ( 22- 24) and larger-scale retrospective studies ( 24- 26) performed on artificially enriched datasets, often involving resampling, to approximate a more representative screening population. Recent studies suggest the current generation of AI-based algorithms using deep learning may interpret mammograms at least to the level of human readers ( 22- 26). Modern artificial intelligence (AI) has emerged as a promising alternative. When tested in the United Kingdom National Health Service Breast Screening Programme (UK NHSBSP) as an alternative to double reading, a traditional CAD system reduced specificity with a significant increase in recall rates ( 21). Recent studies question CAD’s benefit to screening outcomes ( 19- 20). The high cost of two expert readers to interpret every mammogram, alongside growing shortages of qualified readers, means double reading is difficult to sustain ( 15- 17).īreast radiology has experience using computer-aided detection (CAD) software to automate screening mammogram analysis, which has been adopted by over 83% of US facilities ( 18). The model is standard practice in at least 27 countries in Europe, and in Japan, Australia, the Middle East and the UK ( 11- 14). Using two readers (double reading), with arbitration, increases cancer detection rates by 6-15%, while keeping recall rates low ( 8- 10). Randomised trials and incidence-based mortality studies have demonstrated that population-based screening programs substantially reduce breast cancer mortality ( 2- 6).įull-field digital mammography (FFDM) is the most widely used imaging modality for breast cancer screening globally ( 7, 11- 14). Despite improvements in therapy, breast cancer remains the leading cause of cancer-related mortality among women worldwide, accounting for approximately 600,000 deaths annually ( 1).
0 Comments
Leave a Reply. |