PATIENT EXPOSURE OPTIMISATION THROUGH TASK-BASED ASSESSMENT OF A NEW MODEL-BASED ITERATIVE RECONSTRUCTION TECHNIQUE

Area of Science:

• Medical Imaging
• Radiology
• Image Reconstruction

Background:

• Iterative reconstruction algorithms are crucial for enhancing image quality in computed tomography (CT).
• Evaluating the performance of these algorithms, especially for low-contrast detection, is essential for optimizing radiation dose and diagnostic accuracy.

Purpose of the Study:

• To investigate the performance of a novel iterative reconstruction algorithm using a model observer.
• To compare the algorithm's efficacy against human observers in a low-contrast phantom study.

Main Methods:

• A Siemens SOMATOM Force CT scanner was used to acquire images of a low-contrast phantom at varying radiation doses (CTDIvol levels).
• Images were reconstructed using the ADMIRE algorithm and evaluated by three human observers using a forced-choice experiment.
• A channelized Hotelling observer model was applied to the same image set for quantitative comparison.

Main Results:

• A strong agreement was observed between the human observers and the channelized Hotelling observer model.
• Low-contrast detection performance improved significantly when increasing ADMIRE strength from 1 to 3.
• Effective low-contrast detection was achieved even for challenging targets, indicating potential for dose reduction.

Conclusions:

• The validated iterative reconstruction algorithm demonstrates high performance comparable to human observers.
• The findings support the optimization of patient radiation dose in CT examinations without compromising diagnostic image quality.
• Further research can explore dose reduction strategies based on these performance metrics.