Automated registration of diagnostic to prediagnostic x-ray mammograms: evaluation and comparison to radiologists' accuracy.

PURPOSE: To compare and evaluate intensity-based registration methods for computation of serial x-ray mammogram correspondence. METHODS: X-ray mammograms were simulated from MRIs of 20 women using finite element methods for modeling breast compressions and employing a MRI/x-ray appearance change model. The parameter configurations of three registration methods, affine, fluid, and free-form deformation (FFD), were optimized for registering x-ray mammograms on these simulated images. Five mammography film readers independently identified landmarks (tumor, nipple, and usually two other normal features) on pairs of diagnostic and corresponding prediagnostic digitized images from 52 breast cancer cases. Landmarks were independently reidentified by each reader. Target registration errors were calculated to compare the three registration methods using the reader landmarks as a gold standard. Data were analyzed using multilevel methods. RESULTS: Between-reader variability varied with landmark (p < 0.01) and screen (p = 0.03), with between-reader mean distance (mm) in point location on the diagnostic/prediagnostic images of 2.50 (95% CI 1.95, 3.15)/2.84 (2.24, 3.55) for nipples and 4.26 (3.43, 5.24)/4.76 (3.85, 5.84) for tumors. Registration accuracy was sensitive to the type of landmark and the amount of breast density. For dense breasts (> or = 40%), the affine and fluid methods outperformed FFD. For breasts with lower density, the affine registration surpassed both fluid and FFD. Mean accuracy (mm) of the affine registration varied between 3.16 (95% CI 2.56, 3.90) for nipple points in breasts with density 20%-39% and 5.73 (4.80, 6.84) for tumor points in breasts with density < 20%. CONCLUSIONS: Affine registration accuracy was comparable to that between independent film readers. More advanced two-dimensional nonrigid registration algorithms were incapable of increasing the accuracy of image alignment when compared to affine registration.

Computer-aided detection in the United Kingdom National Breast Screening Programme: prospective study.

PURPOSE: To evaluate prospectively the recall and cancer detection rates with and without computer-aided detection (CAD) in the United Kingdom National Health Service Breast Screening Programme. MATERIALS AND METHODS: The study had appropriate ethics committee approval. Informed consent was not required; however, patients were informed that their mammograms might be used in research efforts, and all patients agreed to participate. Mammograms obtained in 6111 women (mean age, 58.4 years) undergoing routine screening every 3 years were analyzed with a CAD system. Mammograms were independently double read. Twelve readers participated. Readers recorded an initial evaluation, viewed the CAD prompts, and recorded a final evaluation. Recall to assessment was decided after arbitration. Sensitivities were calculated for single reading, single reading with CAD, and double reading, as a proportion of the total number of cancers detected by using double reading with CAD. RESULTS: A total of 62 cancers were detected in 61 women. CAD prompted 51 (84%) of 61 radiographically detected cancers. Of 12 cancers missed on single reading, nine were correctly prompted; however, seven prompts were overruled by the reader. Sensitivity of single reading was 90.2% (95% confidence interval [CI]: 83.0%, 95.0%), single reading with CAD was 91.5% (95% CI: 85.0%, 96.0%), and double reading without CAD was 98.4% (95% CI: 91.0%, 100%). Cancer detection rate was 1%. Recall to assessment rate was 6.1%, with an increase of 5.8% because of CAD. Average time required, per reader, to read a case was 25 seconds without CAD and 45 seconds with CAD. CONCLUSION: CAD increases sensitivity of single reading by 1.3%, whereas double reading increases sensitivity by 8.2%.