Positron Emission Mammography Image Interpretation for Reduced Image Count Levels.

Area of Science:

• Nuclear Medicine
• Medical Imaging
• Radiology

Background:

• Positron emission mammography (PEM) utilizes (18)F-FDG for breast cancer detection.
• Optimizing imaging protocols, including injected activity and scan timing, is crucial for diagnostic accuracy.

Purpose of the Study:

• To evaluate the impact of reduced (18)F-FDG injection activity on PEM image interpretation.
• To compare image interpretation between two postinjection imaging times (60 and 120 minutes).

Main Methods:

• A receiver-operating-characteristic (ROC) study was conducted using PEM images reconstructed with varying count levels (100%, 50%, 25%) corresponding to different injected activities.
• Thirty patients underwent two PEM scans at 60 and 120 minutes postinjection, with half receiving standard activity and the other half half-standard activity.
• Eight radiologists assessed 232 PEM images for malignant lesions using a 5-point confidence scale, with data analyzed via logistic regression and ROC analysis.

Main Results:

• A trend for increased lesion detection sensitivity was observed with higher image counts (odds ratios of 2.2 and 1.9 per doubling of counts at 60 and 120 minutes, respectively).
• The area under the ROC curve (AUC) was significantly higher for 100%-count, 60-minute images (0.83) compared to 50%-count images (0.75, P=0.02).
• Lesion-to-background ratios increased significantly between 60 and 120 minutes (P < 0.001), but this did not sufficiently improve detection sensitivity with reduced counts.

Conclusions:

• Reduced (18)F-FDG activity and lower image counts decrease diagnostic accuracy in PEM.
• While later imaging (120 min) increases lesion-to-background ratio, it does not fully compensate for reduced counts to enhance detection sensitivity.
• Standard injected activity protocols are recommended for optimal diagnostic performance in PEM.