Changes in cancer detection and false-positive recall in mammography using artificial intelligence: a retrospective, multireader study

TLDR

Mammography remains the standard modality for breast cancer screening. The study aimed to develop an AI algorithm for breast cancer diagnosis on mammograms and assess whether it could enhance radiologists’ diagnostic accuracy. The AI was trained and validated on 170,230 mammograms from five institutions, then evaluated in a multicentre, observer‑blinded reader study where 14 radiologists assessed 320 images with and without AI assistance. AI alone achieved an AUROC of 0.959 overall, outperforming radiologists (0.810) and improving radiologists’ performance to 0.881 when aided, with higher sensitivity for masses, distortions, T1, and node‑negative cancers. The significant performance gains support AI as a diagnostic aid, and the study was funded by Lunit.

Abstract

BackgroundMammography is the current standard for breast cancer screening. This study aimed to develop an artificial intelligence (AI) algorithm for diagnosis of breast cancer in mammography, and explore whether it could benefit radiologists by improving accuracy of diagnosis.MethodsIn this retrospective study, an AI algorithm was developed and validated with 170 230 mammography examinations collected from five institutions in South Korea, the USA, and the UK, including 36 468 cancer positive confirmed by biopsy, 59 544 benign confirmed by biopsy (8827 mammograms) or follow-up imaging (50 717 mammograms), and 74 218 normal. For the multicentre, observer-blinded, reader study, 320 mammograms (160 cancer positive, 64 benign, 96 normal) were independently obtained from two institutions. 14 radiologists participated as readers and assessed each mammogram in terms of likelihood of malignancy (LOM), location of malignancy, and necessity to recall the patient, first without and then with assistance of the AI algorithm. The performance of AI and radiologists was evaluated in terms of LOM-based area under the receiver operating characteristic curve (AUROC) and recall-based sensitivity and specificity.FindingsThe AI standalone performance was AUROC 0·959 (95% CI 0·952–0·966) overall, and 0·970 (0·963–0·978) in the South Korea dataset, 0·953 (0·938–0·968) in the USA dataset, and 0·938 (0·918–0·958) in the UK dataset. In the reader study, the performance level of AI was 0·940 (0·915–0·965), significantly higher than that of the radiologists without AI assistance (0·810, 95% CI 0·770–0·850; p<0·0001). With the assistance of AI, radiologists' performance was improved to 0·881 (0·850–0·911; p<0·0001). AI was more sensitive to detect cancers with mass (53 [90%] vs 46 [78%] of 59 cancers detected; p=0·044) or distortion or asymmetry (18 [90%] vs ten [50%] of 20 cancers detected; p=0·023) than radiologists. AI was better in detection of T1 cancers (73 [91%] vs 59 [74%] of 80; p=0·0039) or node-negative cancers (104 [87%] vs 88 [74%] of 119; p=0·0025) than radiologists.InterpretationThe AI algorithm developed with large-scale mammography data showed better diagnostic performance in breast cancer detection compared with radiologists. The significant improvement in radiologists' performance when aided by AI supports application of AI to mammograms as a diagnostic support tool.FundingLunit.