P1156: in the absence of symptoms or cerebrospinal fluid involvement, mri staging of the central nervous system (cns) in patients with systemic diffuse large b-cell lymphoma rarely detects disease

Topic: 19. Aggressive Non-Hodgkin lymphoma - Clinical Background: Synchronous CNS involvement in systemic DLBCL is rare but portends a poor prognosis. Screening for CNS involvement by cerebrospinal fluid (CSF) analysis (by cytology [CC] and flow cytometry [MFC]), and cerebral magnetic resonance imaging (MRI) is recommended for at-risk patients (pts). Most early CNS relapses are thought to derive from undetected synchronous CNS involvement at diagnosis. Resource constraints may limit ability to perform MRI. While occult CSF involvement by MFC has been described in asymptomatic individuals, the utility of MRI is uncertain. Aims: We sought to describe institutional patterns of and outcomes following CNS staging for at-risk DLBCL, and to assess the merit of CNS imaging as an adjunct to CSF analysis in patients without symptoms concerning for synchronous CNS involvement. Methods: We retrospectively analysed consecutively diagnosed pts with DLBCL Jan 2011 – Nov 2022, identified from institutional databases. Patients with DLBCL or high-grade B-cell lymphoma, including transformation of indolent lymphoma, were included for analysis if one or more risk factors were present: IPI ≥3; ≥2 extra-nodal sites, or ≥1 extra-nodal site and elevated LDH; renal, adrenal, breast or testicular involvement; double/triple-hit by FISH as defined by the WHO 2016 classification. Patients without these risk factors, or with insufficient CSF or MRI brain +/- spine data available, were excluded. CSF was considered positive (pos) if either CC or MFC or both were pos. By report, any equivocal CSF findings, if negative (neg) on repeat, were deemed negative, while recurrently suspicious CC were deemed positive. Imaging reported as suspicious or equivocal was interpreted after contemporary multidisciplinary review of each case. Results: 540 pts were identified, of whom 167 met inclusion criteria (Fig. 1). Overall, 139 (83%) underwent CNS staging by CSF analysis (n=133, 79.6%), MRI brain (n=74, 44.3%), or both (n=68, 40.7%). Six pts had CNS involvement identified, of whom five had neurological symptoms; four had pos MRI and CSF findings, and one had MRI involvement but neg CSF. The remaining asymptomatic pt had pos CSF (CC), but no MRI was performed. Five of 16 symptomatic pts (31.3%) demonstrated CNS involvement, while MRI revealed extra-cranial cause of symptoms for four pts. Both MRI (33.3% vs. 0%, Fisher’s exact test p=0.0002) and CSF (23.1% vs. 1.0%, p=0.0028) analyses were more likely to be positive in symptomatic vs asymptomatic pts. Furthermore, of 117 asymptomatic patients with negative CSF analyses, 55 (47%) had MRI performed, with none identifying CNS involvement. Excluding known baseline synchronous CNS involvement, overall 8 pts (5.0%) experienced CNS relapse after a median 9.6 months (6.2-43.6 months) from diagnosis, including six relapses in asymptomatic pts with neg CSF (5.1%) of whom had three baseline MRI, all of which were negative. High-dose intravenous and/or intrathecal methotrexate had been administered as CNS prophylaxis for seven of these eight patients with subsequent CNS relapse. Summary/Conclusion: For at-risk pts with systemic DLBCL, the presence of neurological symptoms appears sensitive but not specific for synchronous CNS involvement, and justifies both CSF analysis and MRI staging. When CSF analysis is negative in the absence of symptoms, the utility of adjunctive MRI of the CNS is very limited. These data support the selective use of MRI in the staging of patients with DLBCL, and highlight the need for more sensitive methods (such as CSF ctDNA) to detect clinically occult CNS disease.Keywords: MRI, Lymphoma, DLBCL