Clinically missed cancer: how effectively can radiologists use computer-aided detection?

Area of Science:

• Radiology
• Medical Imaging
• Oncology

Background:

• Screening mammography is crucial for early breast cancer detection.
• Missed cancers on screening mammograms can lead to delayed diagnosis and treatment.
• Computer-aided detection (CADe) systems aim to assist radiologists in identifying subtle abnormalities.

Purpose of the Study:

• To evaluate the effectiveness of computer-aided detection (CADe) in improving radiologists' ability to detect breast cancers missed during screening mammography.
• To quantify the impact of CADe on radiologist sensitivity and recall rates.

Main Methods:

• An observer study involving eight radiologists reading 300 analog mammographic examinations.
• Mammograms were analyzed with and without CADe assistance.
• Data analyzed using jackknife alternative free-response receiver operating characteristic (FROC) analysis.

Main Results:

• CADe use resulted in a statistically significant increase in radiologist sensitivity of 9.9% (p < 0.001).
• The callback rate also increased by 12.1% (p < 0.001) with CADe.
• Radiologists failed to utilize correct computer prompts in 71% of cases where cancer was missed.

Conclusions:

• Computer-aided detection (CADe) offers a modest improvement in radiologist sensitivity for detecting missed breast cancers.
• There is substantial room for improving CADe's clinical impact by enhancing radiologist interaction with system prompts.
• Further research and system optimization are needed to maximize CADe's benefits in breast cancer screening.