Purpose: Brain tumors are typically heterogeneous, and may contain different grades of tumor cells, different types of tumor cells, edema and/or abnormal vascular structures. Anatomical imaging alone can be limited in the evaluation of tumor heterogeneity, especially in those tumors that demonstrate little to no enhancement. While there are physiologic MR tools available in daily clinical practice such as perfusion or diffusion, we wanted to develop a more powerful, sensitive sequence for the characterization of tumor heterogeneity. We propose that tumor heterogeneity imaging (THI) can provide quantitative distributions of different grades of tumor cells and capillary blood perfusion within the tumor in a single clinical imaging scan with more accuracy than previously reported with traditional diffusion techniques.
Materials & Methods: 11 adult patients with known or suspected brain gliomas that were non-enhancing or had substantial non-enhancing regions (>50%) underwent simultaneous 3, 4-dihydroxy-6-[18F] fluoro-L- phenylalanine (18FFDOPA) PET/ MRI prior to planned standard-of-care surgical resection and/or stereotactic biopsy. Of these, 7 patients also underwent THI, a new diffusion MRI protocol, microstructure modeling, and inverse computation technique. The THI maps were then compared to the 18FDOPA and coordinate-guided biopsy or surgical resection results. Perfusion maps extracted from THI were calculated. ADC cut-offs for tumor grade based on the THI data were then determined and tumor grade maps created.
Results: Grade 4 tumors ADC cutoff was 0.3-0.5×10−3 mm2/s. Grade 3 was 0.5-0.8×10−3 mm2/s. Grade 1 and 2 was 0.81.5×10−3 mm2/s. Table 1 summarizes the subject’s demographic characteristics and well as the correlation between 18F-FDOPA and THI maps. We found that in 7/7 patients (100%) THI maps correlated with tumor grade on pathological evaluation. Interestingly, 18F-FDOPA was negative on subject S7, whereas THI correctly identified not only the tumor, but the tumor grade at the region of stereotactic biopsy sample.
Conclusion: This preliminary study demonstrated the capability of a new diffusion MRI method, THI, to noninvasively characterize the structural heterogeneity in brain tumors, including various grades of tumor cells and capillary blood perfusion within the tumors, consistent with pathology assessment on biopsy tissues. Although our preliminary data suggest THI is a promising multiparametric imaging technique to accurately measure cellularity and tumor grade, larger studies will be needed before definitive conclusions can be made about the role of this technique. This preliminary study also suggested the unmet need to develop new generation of MRI technique that is capable to provide direct pathophysiological measures for tumor characterization.
Note: This work is partly presented at 2nd Edition of International Conference on Clinical Oncology and Molecular Diagnostics, June 11- 13, 2018 held at Dublin, Ireland