Clinical evaluation of assistant diagnostic system for mammograms using the auto-analyzing method

Area of Science:

• Radiology
• Medical Imaging
• Artificial Intelligence in Healthcare

Background:

• Mammography is a key tool for breast cancer screening.
• Accurate detection of dense areas and cancer lesions is crucial for early diagnosis.
• Automated analysis systems can potentially enhance mammogram interpretation efficiency and accuracy.

Purpose of the Study:

• To evaluate the clinical usefulness of a novel auto-analyzing system for mammograms.
• To assess the system's ability to detect abnormal dense areas, including cancer lesions.
• To compare the system's findings with final diagnoses.

Main Methods:

• Analysis of 345 breasts, including confirmed cancers, mastopathies, fibroadenomas, cysts, lipomas, and normal cases.
• Utilizing a newly developed auto-analyzing system for mammogram interpretation.
• Comparing system-identified lesions with final pathological diagnoses.

Main Results:

• The system demonstrated high sensitivity, correctly identifying 84 out of 86 cancer lesions (97.9%).
• Cancer lesions were ranked: 64.0% at the first rank and 96.5% within the top three ranks.
• Correct diagnosis of calcifications was achieved in 68 out of 74 breasts, with 48 ranked in the top three.

Conclusions:

• The auto-analyzing system shows significant promise as an assistive tool for breast cancer screening.
• The system effectively identifies and ranks suspicious areas, including cancer and calcifications.
• Further system refinement is recommended to maximize its utility in mammogram analysis.