Limbic encephalitis (LE) is a high-risk paraneoplastic phenotype; however, the majority of cases do not have paraneoplastic etiology. Other than high-risk paraneoplastic autoantibodies, distinctive features for paraneoplastic etiology are not well elucidated.

148 (71, females; median age 60 years [range11-84]) LE cases were identified. Among these 22 (15%), 11 (7%) and 27 (18%) met definite, probable, and possible paraneoplastic neurological syndrome diagnostic criteria, respectively. The remaining 88 (60%) cases were characterized as nPLE. The majority of cases were neural autoantibody seropositive (118, 80%). The most common autoantibody among PLE (definite/probable PNS) and nPLE cases were ANNA1 (9, 27%) and LGI1 (44, 50%) respectively. High-risk paraneoplastic autoantibodies were more common in PLE (p<0.001). Cognitive, behavioral, and sleep problems were similar between groups. Co-existing brainstem dysfunction (p=0.007), cranial neuropathy (p<0.001), and peripheral nerve involvement (p=0.028) were seen more commonly in PLE. Weight loss and supernumerary CSF oligoclonal bands were more common in PLE cases (p<0.001). MRI brain in PLE more commonly demonstrated extra-temporal involvement (p=0.002). At disease nadir, PLE had more severe disability compared to nPLE (median Modified Rankin Scale, mRS [range] 4 [3-5] vs 3 [3-4], p=0.002). Progression/relapse free survival was significantly worse in PLE (median survival 11 [2-26] vs 36 [12-70.5] months, p<0.001). At last follow-up, PLE had higher frequency of intractable seizures (p=0.011), severe cognitive impairment (p=0.023), and more disability (median mRS 3 [2-6] vs 2 [1-3], p=0.002).  

Along with high-risk paraneoplastic autoantibodies, co-existing multifocal neurologic deficits, weight loss and extra-temporal MRI brain changes can help distinguish PLE from nPLE and prompt early cancer search and immunotherapy escalation to prevent poor outcomes.