Determining the Optimal Angular Range of the X-Ray Source Motion in Tomosynthesis Using Virtual Clinical Trials

Area of Science:

• Medical Imaging
• Radiology
• Computational Imaging

Background:

• Digital breast tomosynthesis (DBT) utilizes limited angular data, potentially compromising the detection of small lesions like microcalcifications.
• Under-sampled datasets in DBT can hinder the visibility of crucial calcifications in reconstructed images.

Purpose of the Study:

• To evaluate the impact of varying angular acquisition ranges on calcification detection in DBT.
• To assess the performance of computer simulations and virtual readers in a virtual clinical trial (VCT) for DBT analysis.

Main Methods:

• A virtual clinical trial (VCT) approach using simulated breast phantoms and images.
• Multiple-reader multiple-case (MRMC) receiver operating characteristic (ROC) analyses with channelized Hotelling observers (CHOs).
• Simulation of continuous and step-and-shoot x-ray motion with angular ranges of ±7.5°, ±15°, and ±25°.

Main Results:

• Calcification detection performance is significantly affected by the angular motion path, especially the ±25° range.
• A wider angular range (±25°) combined with continuous motion and larger voxel sizes reduced detection performance (AUC difference: -0.030 to -0.067).
• No significant difference in calcification detection was observed between ±7.5° and ±15° angular ranges.

Conclusions:

• The angular range is a critical factor influencing calcification detection in DBT.
• Wider angular acquisitions in DBT may pose limitations for detecting microcalcifications, an area requiring further research and optimization.