Residual Diffusion-Weighted Imaging Hyperintense Signal in Primary Central Nervous System Lymphoma Can Predict Early Recurrence

The treatment response of primary central nervous system lymphomas (PCNSLs) is mainly evaluated using postcontrast T1-weighted imaging (T1WI). Because poorly enhanced lesions may contain residual tumors, the combination of evaluation methods will potentially improve the accuracy of determining treatment effectiveness. In this study, we evaluated the usefulness of diffusion-weighted imaging (DWI) in predicting recurrence among patients with PCNSL who achieved complete response (CR)/unconfirmed CR (CRu). Fifty-four patients newly diagnosed with PCNSL who were treated at our institution and achieved CR/CRu at the end of treatment were included in this study. The patients were divided into two groups according to the presence or absence of residual DWI hyperintense signal at the tumor site at the end of treatment. Kaplan–Meier analysis was performed to analyze the median overall survival (OS) and progression-free survival (PFS). The mean age of the 54 patients was 66.4 ± 13.3 years. The induction therapies were HD-MTX in 20 patients, R-MPV in 29 patients, and other chemotherapies in five patients. Radiotherapy was performed in 35 patients, high-dose cytarabine therapy in 14 patients, and autologous hematopoietic stem cell transplantation in one patient, and of the 54 patients, 10 had no consolidation therapy. The residual DWI hyperintense signal sign was observed in 18 patients. The R-MPV regimen was statistically associated with a lower rate of residual DWI hyperintense signal (p = 0.0453). The median PFS was statistically shorter in the residual DWI hyperintense signal group than in the non-residual DWI hyperintense signal group (14.0 months vs. 85.1 months) (p < 0.0001, log-rank test). A residual DWI hyperintense signal at the end of treatment was statistically associated with shorter PFS. Among patients who achieved CR/CRu evaluated based on postcontrast T1WI, DWI could be a valuable additional sequence to predict the early recurrence of PCNSL.