Combined Diffusion and Perfusion MRI As Biomarkers of Prognosis in Immunocompetent Patients with Primary CNS Lymphoma

The prospect for advances in the treatment and improvement in the prognosis of patients with primary central nervous system lymphoma (PCNSL) is likely dependent on the systemic evaluation of its pathobiology and identification of biomarkers that can guide and monitor response to therapy. The validation of noninvasive imaging-based biomarkers of prognosis and therapeutic response may have significant impact on the management of patients with PCNSL. Recent work suggests that apparent diffusion coefficient (ADC) measurements derived from diffusion-weighted magnetic resonance imaging (DW-MRI) may predict prognosis in immunocompetent patients with PCNSL. The purpose of our study was to validate ADC measurement as a prognostic biomarker in patients with PCNSL and to evaluate whether relative cerebral blood volume (rCBV) measurement derived from dynamic susceptibility-weighted contrast-enhanced (DSC) perfusion MRI may also serve as predictor of outcome.

We studied pre-therapy imaging of 26 immunocompetent patients with PCNSL that were treated homogeneously with methotrexate-based chemotherapy. We measured anatomic tumor volume and peri-tumoral edema volume from post-contrast T1-weighted and T2-weighted images, respectively. ADC map and rCBV map were generated from DW-MRI and DSC perfusion MRI, respectively. Post-contrast T1-weighted images containing regions of interest (ROIs) outlining contrast enhancing tumor were co-registered with the ADC and rCBV maps and minimum and mean ADC and rCBV values were measured. Progression-free survival (PFS) and overall survival (OS) were measured. The Welch t-test assessed differences between groups. Multivariate Cox survival analysis was performed.

Anatomic tumor volume or peri-tumoral edema volume at initial diagnosis did not correlate with OS or PFS. However, baseline ADC and rCBV measurements of tumor had significant association with clinical outcome measurements: patients with low minimum ADC (ADC_min) or low mean rCBV (rCBV_mean) had worse OS and PFS. The high ADC_min-high rCBV_mean group had the best and low ADC_min-low rCBV_mean group had the worst PFS and OS. Multivariate Cox survival analysis identified no other imaging-based variables besides ADC and rCBV that significantly affected survival.

Our study further supports the potential of non-invasive ADC measurement as a valid and reliable prognostic biomarker in patients with PCNSL. In addition, our study is the first evaluation of perfusion MRI as a prognostic biomarker in PCNSL. We have generated evidence that reduced perfusion, suggestive of hypoxic tumor microenvironment, is associated with adverse prognosis in patients with PCNSL. In addition, our data suggests that perfusion-based parameters enhance the prognostic utility of ADC measurements in the evaluation of newly diagnosed patients with CNS lymphoma.

Supported by NIH RO1CA139-83-01A1 and Pathways to Careers in Clinical and Translational Research (PACCTR) Program (FEV).

No relevant conflicts of interest to declare.