Adaptive statistical iterative reconstruction for computed tomography of the spine

Area of Science:

• Radiology
• Medical Imaging
• Image Processing

Background:

• Sagittal views of CT spine scans are crucial but often suffer from noise, degrading image quality.
• Iterative reconstruction (IR) techniques reduce noise but can impact spatial resolution.

Purpose of the Study:

• To evaluate the effectiveness of adaptive statistical iterative reconstruction (ASiR-V) for improving sagittal CT spine images compared to filtered back projection (FBP).

Main Methods:

• Subjective assessment of image quality (spatial resolution, noise, overall quality) by three radiologists on 25 patients' clinical images.
• Objective assessment using z-direction modulation transfer function (z-MTF) and z-axis noise power spectrum (z-NPS) with phantoms.
• ASiR-V with 50% and 100% blending levels (ASiR-V50, ASiR-V100) were compared to FBP.

Main Results:

• Subjectively, ASiR-V100 yielded the highest overall image quality, though with lower spatial resolution scores.
• Objectively, ASiR-V50 and ASiR-V100 showed slightly degraded low-contrast z-MTF compared to FBP.

Conclusions:

• ASiR-V effectively enhances sagittal CT spine image quality compared to FBP.
• Overall image quality, not solely high resolution, is key for reviewing sagittal CT spinal reformats.