T* Imaging Assessment of Neoadjuvant Radiation Therapy Combined With Pharmacological Ascorbate in Extremity Soft-Tissue Sarcomas: A Pilot Study.

Background: Extremity soft-tissue sarcomas (STS) are commonly treated with neoadjuvant radiation therapy followed by surgical resection. However, the pathological near-complete response rate is low (9-25%). Noninvasive imaging assessment that predicts treatment response before and during treatment is desirable to optimize treatment regimens. This pilot study aimed to investigate the application of a quantitative MRI parameter, T*, in assessing neoadjuvant radiation therapy combined with pharmacological ascorbate in extremity STS.

Methods: This prospective cohort study included seven patients diagnosed with extremity STS and scheduled to receive neoadjuvant radiation therapy combined with pharmacological ascorbate. T* maps were obtained from each patient before treatment (baseline MRI), two weeks after initiating treatment (on-treatment MRI), and before surgery (pre-surgery MRI). The T* values within the tumor region were transformed into z-scores with respect to the normal- appearing tissue region. The voxel-wise z-scores within the tumor region were thresholded to generate masks representing significantly high (z-score>1.96) and low z-score (z-score<-1.96) voxels. The means of the total z-scores and within each of the significantly high and low z-score mask were computed. Their correlations with percent necrosis from pathological examination were evaluated using Spearman's rank correlation coefficient r. A correlation was considered as moderate or strong when r is higher than 0.6 and 0.8, respectively. A correlation was considered as fair or weak when r is below 0.6.

Results: For the baseline and on-treatment MRIs, the means of the significantly high z-scores of the T* measurements showed moderate correlations with percent necrosis (r = 0.68 and 0.6; p = 0.11 and 0.24). For the pre-surgery MRI, the means of the total and significantly high z-scores showed strong correlations with percent necrosis (r = 0.8 and 0.9; p = 0.13 and 0.08). Tumor volume and baseline MRI-based percent necrosis showed fair or weak correlations (r = 0.3-0.54; p = 0.24-0.68).

Conclusion: T* measurements prior to treatment, two weeks after initiating treatment, and before surgery showed moderate to strong correlations with percent necrosis. These results support the potential for using T* mapping to predict and assess response to neoadjuvant radiation therapy combined with pharmacological ascorbate in extremity STS. .