Controlling Signal Artifact With Software Threshold Imaging for Magnetic Resonance-Guided Laser Interstitial Thermal Therapy.

Background: Magnetic resonance-guided laser interstitial thermal therapy (MRgLITT) uses intraoperative temperature mapping and thermal damage estimates to guide ablations of intracranial targets. In select cases, signal artifact presents at the target site and impairs intraprocedural decision-making by obscuring the visualization of both temperature imaging and the thermal damage estimate calculation. To date, the etiology and impact of signal artifact are unknown. However, user-selected MRgLITT software settings may play a role in generating artifact.

Objective: To assess the effect of the thresholding feature in MRgLITT software on signal artifact generation during intracranial ablations.

Methods: Ablations were performed with the Visualase MRI-guided Laser Ablation System (Medtronic). For each LITT procedure, raw thermal data were extracted at a reference threshold of 40 and reprocessed at 5 additional threshold values ranging from 35 to 60. Artifact growth rates relative to threshold values were derived using simple linear regressions and then assessed within the context of laser power and duration using Pearson correlations.

Results: A total of 33 patients were included, with 28 artifact-containing and 5 artifact-free cases. For artifact-containing cases, a 13% increase in artifact area occurred for every 1-point increase in threshold (R2 > 0.99). Artifact growth rates were not correlated with laser power (r = 0.15, P = .44) or duration (r = 0.0049, P = .98). One of the 5 artifact-free cases developed artifact at a threshold of 60.

Conclusion: Artifact generation is likely multifactorial involving tissue properties and software settings. Operators can minimize software-introduced artifact by reducing threshold values.