ARRS 2022 Abstracts

RETURN TO ABSTRACT LISTING


1965. Dynamic Contrast Enhancement Processing Comparison for Determining True Progression from Pseudoprogression in High Grade Glioma
Authors * Denotes Presenting Author
  1. Ahmad Amer *; Chicago Medical School at Rosalind Franklin University of Medicine and Science; The University of Texas MD Anderson Cancer Center
  2. Swapnil Khose; The University of Texas MD Anderson Cancer Center; The University of Texas Memorial Hermann
  3. Halyna Pokhylevych; The University of Texas MD Anderson Cancer Center
  4. Susana Calle; The University of Texas MD Anderson Cancer Center
  5. Ho-Ling Liu; The University of Texas MD Anderson Cancer Center
  6. Jason Johnson; The University of Texas MD Anderson Cancer Center
Objective:
Glioblastoma is a devastating brain tumor with mean survival of 14 months from diagnosis. A variety of treatment-related changes occur on brain tissues related to the effects of radiation and temozolomide. Delayed radiation or treatment-induced enhancement (pseudo-progression) can mimic a tumor on MR imaging. The MR technique of dynamic contrast enhancement (DCE) is commonly performed to acquire further physiologic information about suspicious enhancement. The purpose of this study is to compare the performance of DCE processing methods for determination of true progression from pseudoprogression in high-grade gliomas.

Materials and Methods:
We identified a group of 67 patients (39 male, 57 +/- 12 years old) with high-grade glioma treated with surgery and chemoradiation who had DCE perfusion imaging to assess new or increasing enhancement. Each of these patients had confirmatory surgery or biopsy within 3 months from the date of the DCE imaging, with pathologic confirmation of progressive disease (PD) or pseudo-progression (PsP). DCE imaging was performed at 3T and data were analyzed using nordicICE (NordicNeuroLab). The MCA, SSS, and Parker models were each processed using standardized methodology to create k-trans maps. The k-trans maps were then reviewed by experienced neuroradiologists, and three equivalent sized region of interest (ROIs) were placed at sites of peak enhancement within the lesion. Patient ROI data for each processing method was then processed for the mean and the max ROI k-trans value. A student's t-test was then performed to assess for statistical significance between the group of patients with pathology-confirmed PD and PsP.

Results:
Sixty of the patients had pathology confirmation of PD and 7 patients had pathology-proven PsP. The data from 67 patients were successfully processed in nordicICE. For the Parker method, the mean and peak values for the PsP group was 0.093 +/- 0.034 and 0.104 +/- 0.036, respectively, with the PD values of 0.150 +/- 0.091 and 0.169 +/- 0.104. This difference was statistically significant for the mean value at a P = 0.05. For the MCA method, the mean and peak values for the PsP group was 0.070 +/- 0.029 and 0.081 +/- 0.034, respectively, with the PD values of 0.100 +/- 0.094 and 0.113 +/- 0.115. For the SSS method, the mean and peak values for the PsP group was 0.049 +/- 0.021 and 0.055 +/- 0.023, respectively, with the PD values of 0.059 +/- 0.042 and 0.066 +/- 0.048, respectively. Values for the MCA and SSS methods were not statistically different.

Conclusion:
Our results suggest that the Parker method is a preferred processing method for differentiating PD from PsP in a group of patients with high-grade glioma treated with surgery and chemoradiation and subsequent abnormal enhancement. The MCA and SSS processing methods did not separate out the two groups. Further research will be performed to assess if this performance difference is seen within other categories of disease including patients followed with clinical observation instead of surgical confirmation.