Quantitative ex vivo assessment of target temperature and ablation duration for protocol optimization of microwave ablation procedures with mr thermometry

To quantitatively assess the influence of target temperature and ablation duration on the quality of proton resonance frequency shift (PRFS)-based MR thermometry during microwave ablation (MWA) in a controlled ex vivo model, and to identify parameter ranges associated with improved thermometry performance. Thirty-two MWAs were performed in 10 ex vivo bovine livers in a 1.5-tesla MRI system with multi-slice volumetric real-time thermometry yielding temperature and thermal dose maps. The experiments were conducted twice using all combinations of four target temperatures (60; 80; 100; 120 °C) and four ablation times (5:00; 7:30; 10:00; 15:00 min). Thermometry quality was rated on a 5‑point Likert scale. Ablation areas were compared with histopathology (hematoxylin and eosin, H&E; and nicotinamide adenine dinucleotide, NADH‑diaphorase) and correlated using Spearman coefficients. Likert scores were compared across temperatures using Kruskal-Wallis and Mann-Whitney U tests. All evaluations were performed independently by two readers. Lesion areas varied from 2.6 to 12.9 cm², increasing primarily with target temperature. Ablation areas from temperature and thermal dose maps correlated strongly with macroscopically visual necrosis (p < 0.01). Likert scores differed significantly across temperatures (p < 0.05). The highest image quality was achieved at 60 °C for 7:30 min, showing comparable scores as at 80° for 15:00 min, but significantly differing from 100 °C to 120 °C. In this controlled ex vivo setting, lower target temperatures were associated with improved MRI thermometry quality, providing more reliable visualization of ablation zones; however, ablation volumes decreased at lower temperatures. Furthermore, these empirical ex vivo observations suggest that a staged two-level approach may support a clinical workflow strategy aimed at balancing thermometry image quality and ablation volume. Given the absence of perfusion and motion effects, these findings require further validation before clinical translation.