Quantitative metal artifact reduction algorithm for spectral CT thermometry
Leening P Liu1, Kevin M Brown2, Amy E Perkins2
1Department of Bioengineering and Department of Radiology, Perelman School of Medicine, Philadelphia, PA, USA.
Conference Proceedings. International Conference on Image Formation in X-Ray Computed Tomography
|February 10, 2025
Summary
A new Spectrally Obtained Needle Artifact Reduction (SONAR) algorithm improves metal artifact reduction in spectral CT thermometry. This enhances temperature map accuracy for non-invasive monitoring of thermal ablations, aiming to reduce tumor recurrence.
Area of Science:
- Medical Imaging
- Interventional Radiology
- Biomedical Engineering
Background:
- Spectral CT thermometry offers non-invasive internal temperature monitoring for thermal ablations.
- Accurate temperature mapping is crucial for reducing local tumor recurrences due to insufficient treatment.
- Existing metal artifact reduction methods lack the quantitative accuracy needed for clinical translation.
Purpose of the Study:
- To develop and evaluate a novel metal artifact reduction algorithm, Spectrally Obtained Needle Artifact Reduction (SONAR), for spectral CT thermometry.
- To improve the quantitative accuracy of temperature maps generated by spectral CT thermometry in the presence of ablation applicators.
- To validate the effectiveness of SONAR in correcting metal artifacts and ensuring precise temperature measurements near the applicator.
Main Methods:
- The SONAR algorithm was developed using spectral CT's material decomposition to isolate applicator projections.
- Modeled projections of an angled cylinder replaced projections with significant metal artifacts.
- A phantom study with a liver-mimicking material, ablation applicator, and thermometers was conducted using dual-layer spectral CT at 35 °C and 80 °C.
Main Results:
- SONAR significantly reduced streak artifacts along the applicator's axis.
- The algorithm corrected underestimated temperatures near the applicator and overestimated temperatures in the periphery.
- Temperature map accuracy improved, with corrected temperatures closer to expected values, especially within 1-3 cm of the applicator.
Conclusions:
- SONAR effectively minimizes metal artifacts in spectral CT thermometry, leading to quantitatively accurate temperature maps.
- Improved temperature map accuracy facilitates in vivo evaluation of spectral CT thermometry for non-invasive monitoring of thermal ablations.
- This advancement holds potential for reducing local tumor recurrences by ensuring adequate treatment margins.
