Utility Of Post Therapy Brain Surveillance Imaging In The Detection Of Primary CNS Lymphoma (PCNSL) Relapse

The optimal follow-up strategy for primary CNS lymphoma (PCNSL) patients in remission after first line therapy is not clear. The goal of this study is to determine the utility of planned brain surveillance imaging in the detection of relapse in a large cohort of PCNSL patients.

Patients were from consecutive PCNSL cases (N=209) included in Leon Berard Cancer Centre registry (Lyon, France), from 1987 to 2011 (date of diagnosis). Patients were all treated by chemotherapy, 92% of them by high-dose methotrexate containing chemotherapy followed by brain radiotherapy for 107 patients (51%). All patients were followed for relapse, retreatment and death. Patient clinical records were reviewed for details at relapse and relationship to planned follow-up visits and brain surveillance imaging.

Among the 209 PCNSL patients, 28 (13%) presented toxic death, one patient died from another reason and 41 patients (20%) had a progressive disease during first-line therapy. The remaining 139 patients (66%) entered in post-treatment observation, 128 of them in complete remission (92%) and 11 in partial remission (8%). The median follow up was 36 months for the patients who entered in post-treatment observation. Among these 139 patients, 7 (5%) were lost of follow-up, 62 (45%) patients are still in remission and 70 (50%) relapsed. Among these 70 relapses, 15 (21%) were detected by planned brain surveillance imaging but 53 (76%) patients were symptomatic and presented earlier than a planned follow-up visit; two patients (3%) had no information at time of relapse. If we consider only patients in complete remission who entered in post-treatment observation, 13 (20%) relapses were detected by brain surveillance imaging and 50 (80%) patients presented symptoms between planned visits and imaging. Among the 7/11 patients considered in partial remission after initial treatment who relapsed, two (29%) relapses were detected by brain surveillance imaging and five (71%) by symptoms between planned visits and imaging. Among the 53 symptomatic patients at relapse, 41 (77%) presented a brain tumor relapse, six had an isolated leptomeningeal relapse, one a spinal cord localization and five an extra-cerebral relapse (three abdominal nodes, one soft-tissue mass, one testis). The most common symptoms at relapse were cognitive troubles, motor or sensitive deficits, epilepsy, and alteration of performance status. We did not observe any difference between asymptomatic relapse patterns before and after 2 years with 54% relapses before the two years of follow-up (brain imaging mainly every 4 months) and 46% relapses after 2 years of follow-up (brain imaging mainly every 6 months).

This study showed that PCNSL relapses frequently occurred outside of planned follow-up visits and were notably detected by symptoms between two brain surveillance imaging. Even during the two first years of follow-up with a closed brain imaging monitoring, planned surveillance imaging seemed not efficient.

No relevant conflicts of interest to declare.