Diagnostic Significance of Hyperintense Lesion in the Pons on Brain Magnetic Resonance Imaging in Patients with Intravascular Lymphoma

Introduction: Neurological symptoms related to the involvement of the central nervous system (CNS) have been commonly observed at diagnosis and at relapse in intravascular large B-cell lymphoma (IVLBCL). Although various patterns of abnormal findings on brain magnetic resonance imaging (MRI) in patients with IVLBCL have been reported, most of them were from case reports or small case series of selected patients.Hence, we aimed to investigate the prevalence and clinical value of abnormal findings detected using brain MRI in patients with IVLBCL regarding diagnosis and prognosis. Methods: A total of 33 consecutive patients diagnosed with IVLBCL who underwent treatment at Kameda Medical Center between 1998 and 2017 were available for data of routine pretreatment brain MRI.The diagnosis of IVLBCL was pathologically made by an expert hematopathologist (KT) in all patients. Brain MRI was performed as previously reported, and the abnormalities were classified into the following 4patterns by 2 neuroradiologists in consensus (Figure 1): (A) hyperintense lesion in the pons on T2-weighted imaging (T2WI), (B) nonspecific white matter lesions, (C) infarct-like lesions, and (D) meningeal thickening and/or enhancement. We subsequently identified 77 consecutive patients (52 patients at initial presentation and 25 patients at relapse) with pathologically diagnosed diffuse large B-cell lymphoma (DLBCL) without IVLBCL and 41 patients who received random-skin biopsy on the suspicion of IVLBCL but were found to be negative, as control groups for the presence or absence of hyperintense lesions in the pons. Results: Pretreatment brain MRI revealed abnormal findings in 29 (87.9%) patients. Hyperintense lesions in the pons on T2WI was the most common abnormal finding and was detected in 19 (65.5%) patients. Among them, 10 (52.6%) patients did not have impaired consciousness (Figure 2). Among the 7 patients in whom hyperintense lesions in the pons on T2WI was the sole abnormality, 5 patients (71.4%) did not have impaired consciousness. Infarct-like lesions were detected in 8 (27.6%) patients, and impaired consciousness was more frequent in patients with this pattern than in those without (87.5% vs. 28.0%; P=0.005). Nonspecific white matter lesions and meningeal thickening and/or enhancement were detected in 14 (48.3%) and 4 (13.8%) patients, respectively. No significant difference in overall survival (OS) was detected between patients with and without hyperintense lesions in the pons on T2WI (Figure 3). Patients with nonspecific white matter lesions had relatively shorter OS than those without the finding, although the difference was not statistically significant (median OS, 19.8 months vs. not achieved; P=0.057). Infarct-like lesions were associated with unfavorable survival (median OS, 12.5 months vs. not achieved; P=0.030). Follow-up brain MRI revealed improvements in abnormal findings in most of the patients who responded to chemotherapy (Figure 4). Furthermore, postmortem examinations revealed pathological changes in the brain related to the lymphoma lesions, indicating that these MRI findings might represent these lesions of the brain. Next, we reviewed findings on brain MRI in 77 control patients with DLBCL without IVLBCL. Among them, 16 (20.8%) patients had concomitant CNS involvement of lymphoma. No patients harbored hyperintense lesions in the pons, in contrast with the patients with IVLBCL (P<0.001). This finding was detected in no patient also among those who received random-skin biopsy on the suspicion of IVLBCL but were found to be negative (P<0.001). Conclusions: Our findings revealed that most patients with IVLBCL presented abnormal findings on pretreatment brain MRI, even if they exhibited no neurological symptoms. In particular, hyperintense lesions in the pons on T2WI were frequently observed in patients with IVLBCL, irrespective of the presence or absence of impaired consciousness, and were highly specific in IVLBCL compared to those in control groups, suggesting that this pattern may be pathognomonic and valuable for the timely diagnosis of IVLBCL. Improvements in all types of abnormal findings on follow-up brain MRI indicated that these findings might reflect structural changes associated with IVLBCL and might be useful for confirmation of the therapeutic effect. Further longitudinal studies are required to validate our findings and determine their clinical implications. Disclosures No relevant conflicts of interest to declare.