Related Experiment Video
Updated: Apr 12, 2026
Tracking the Mammary Architectural Features and Detecting Breast Cancer with Magnetic Resonance Diffusion Tensor Imaging
Published on: December 15, 2014
Breast Cancer: Computer-aided Detection with Digital Breast Tomosynthesis
Lia Morra1, Daniela Sacchetto1, Manuela Durando1
1From the Department of Research and Development, im3D, Via Lessolo 3, 10153 Turin, Italy (L.M., D.S., S.A., S.D., D.P., A.B.); Department of Radiology, University of Turin, Turin, Italy (M.D., G.M., P.F.); Department of Diagnostic Imaging and Radiation Therapy, Radiology University of Torino, Azienda Ospedaliero Universitaria Città della Salute e della Scienza di Torino, Turin, Italy (M.D., G.M., P.F.); Unità di Radiologia, IRCCS Policlinico S. Donato, Milan, Italy (L.C.); C.d.C. Paideia, Rome, Italy (B.P.); and Department of Radiology, Sant'Anna Hospital, Turin, Italy (V.M.).
A computer-aided detection (CAD) system for digital breast tomosynthesis demonstrated high sensitivity in detecting breast cancers, including masses and microcalcifications. This CAD system achieved an 89% detection rate with an acceptable false-positive rate, aiding in cancer diagnosis.
Area of Science:
- Radiology
- Medical Imaging
- Oncology
Background:
- Digital breast tomosynthesis (DBT) is an advanced mammography technique.
- Computer-aided detection (CAD) systems aim to improve diagnostic accuracy in medical imaging.
- Independent validation of commercial CAD systems is crucial for clinical adoption.
Purpose of the Study:
- To independently evaluate the performance of a commercial tomosynthesis computer-aided detection (CAD) system.
- To assess CAD system accuracy using a multicenter dataset of DBT examinations.
Main Methods:
- A multicenter dataset of 175 diagnostic and screening tomosynthesis mammographic examinations was analyzed.
- The dataset included 123 patients with 132 biopsy-proven cancers and 52 negative examinations.
- CAD performance was evaluated based on per-lesion sensitivity and false-positive rates.
Main Results:
- The CAD system achieved an 89% per-lesion sensitivity for detecting 111 lesions (masses, microcalcifications, architectural distortions).
- Specific sensitivities included 95% for microcalcification clusters and 89% for masses.
- The system demonstrated a false-positive rate of 2.7 per view.
Conclusions:
- A digital breast tomosynthesis CAD system effectively detects a significant proportion of breast cancers.
- The system shows promise for aiding radiologists in identifying masses and microcalcification clusters.
- Further research with larger, multi-vendor datasets is recommended to confirm findings and explore radiologist-CAD interaction.
