Refinement of primary central nervous system lymphoma prognostication and response assessment using 3-dimensional MRI.

BACKGROUND

Primary central nervous system lymphoma (PCNSL) is an aggressive lymphoma restricted to the CNS in which outcomes cannot be reliably predicted. The International PCNSL Collaborative Group developed standardized response assessment utilizing 2-dimensional (2D) Magnetic Resonance Imaging (MRI) tumor measurements. Considerable challenges of this approach exist due to many reasons. Recent glioblastoma and PCNSL data demonstrated that radiological assessment of baseline 3-dimensional volume (3DV) as well as 3DV reduction (3DVR) may be a sensitive prognostic parameter.

METHODS

Our multicentre retrospective study evaluated semiautomated 3DV in 74 PNCSL patients undergoing curative-intent chemoimmunotherapy.

RESULTS

Baseline tumor 3DV was not associated with survival. Compared to 3DVR < 58% (ROC-determined threshold based on our cohort), both interim and End-of-Treatment (EOT) 3DVR ≥ 58% in responding patients were associated with statistically significant prolonged 2-year progression-free survival (PFS) (interim: 73% (95%CI 57-83) versus 22% (95%CI 3-51),  = 0.005; EOT: 75% (95%CI 59-85) versus 0%,  = 0.002) and 2-year OS (interim: 83% (95%CI 68-91) versus 38% (95%CI 9-67),  = 0.02; EOT: 86% (95%CI 70-93) versus 0%,  = 0.0002). However, no significant differences in PFS or OS were observed in patients achieving standard 2D complete response (CR) compared to partial response (PR).

CONCLUSION

Although PCNSL tumor 3DV at baseline is not associated with survival outcomes, 3DVR of ≥58% in interim and EOT confers superior PFS and OS. Whereas, no difference in survival was observed using standard 2D CR versus PR response assessment at the same time-points. 3DV calculations may offer a sensitive method of response assessment for PCNSL. We are currently validating this in clinical trials.