(P 382) METHOD FOR VOLUMETRIC ASSESMENT OF TUMOR RESPONSE FOLLOWING NEOADJUVANT RADIATION IN PRIMARY MYXOFIBROSARCOMA

Objective: Myxofibrosarcoma (MFS) is an aggressive subtype of soft tissue sarcoma typically treated with neoadjuvant radiotherapy (NRT) and surgical resection. Tumor response to NRT is defined by RECIST criteria which classify response based on changes in the greatest dimension of the tumor. MFS tumors are often irregularly shaped with perilesional stranding which may obscure measurements and complicate response assessment. Three-dimensional volumetric analysis has been used to measure other tumor types. We propose a volumetric method to assess MFS tumors and compare accuracy and reliability to unidimensional RECIST measurements.

Methods: Patients aged 18-100 years diagnosed with MFS and treated with NRT at KUMC from 2002-2021 were included. Pre- and post-NRT MRIs were measured in their greatest dimension 3 times by 3 reviewers. Each MRI stack was uploaded to 3D-Slicer and 3 tumor models per reviewer created. Tumor size changes were compared pre- vs post-NRT. Tumors were grouped by location: upper extremity (UE), lower extremity (LE) and torso. Intra-rater and inter-rater reliability was assessed using interclass correlations. Paired t-tests and Pearson correlations were used to compare changes in tumor size between methods.

Results: Inter- and intra-rater reliability for both methods showed no statistically significant difference (p=0.092) and no significant correlation between methods (r=0.199). On average, volumetric measurements showed 24% shrink in UE, 29% growth in LE and 45% growth in torso tumors. Unidimensional measurements showed 21% shrink, 9% growth and 2% shrink, respectively. Individually, RECIST classified all UE and LE tumors, as well as 4/6 torso tumors as stable disease. Volumetrically, all UE tumors shrank while three LE tumors shrank and five grew. All “stable disease” torso tumors grew, while the “partial response” tumor shrank and the “progressive disease” tumor grew.

Conclusion: Volumetric analysis demonstrated similar reliability to RECIST but may provide more accurate classification of tumor response. Differences are likely due to more complete and sensitive inclusion of MFS tumor morphology. Further development of volumetric MFS analyses may be a valuable tool in evaluating treatment response.