Background: Large experiences have reviewed the risk of central nervous system (CNS) involvement in diffuse large B-cell lymphoma (DLBCL), but there are limited data on CNS involvement by peripheral T cell lymphomas (PTCL). We characterized the incidence of CNS involvement, risk factors and outcome in a large single institution dataset of PTCL. Methods: Retrospective review of the T-cell lymphoma database at Memorial Sloan Kettering Cancer Center. We identified 232 patients with any subtype of PTCL between 1994-2011 with a minimum 6 months of follow-up or an event defined as relapse or death. We excluded indolent forms of cutaneous T cell lymphoma. Results: Histologies included PTCL-NOS (31%), angioimmunoblastic (16.8%), anaplastic (ALCL), ALK negative (12%), ALCL, ALK positive (6%), extranodal NK/T cell lymphoma (7.3%), adult T cell leukemia/lymphoma (ATLL) (7.3%), and transformed MF (8.6%). Median age was 58 years with 59.9% men. CNS disease was found in 17 patients (7.32%). 8 (47%) had pathologic confirmation and 7 (41.2%) were clinically diagnosed. Two had other diagnoses at biopsy: DLBCL and glioblastoma. Median time to CNS involvement was 2.33 months (range, 0.16 to 103.1). CNS prophylaxis was given to 24 (10.34%), primarily intrathecal methotrexate. There was no difference in CNS involvement in patients who received prophylaxis vs. those who did not: 3/24 (12.5%) vs. 12/208 (5.77%) (p=0.192) respectively. Univariate analysis identified: stage III-IV (p=0.03), bone marrow involvement (p=0.018), >1 extranodal site (p<0.001), and ATLL vs. all other subtypes, 23.5% vs. 6.4% (p=0.003) as risk factors for CNS disease. On multivariate analysis, >1 extranodal site (p=0.004) and high intermediate (H-I) and high (H) IPI (IPI 3-5 & 4-5) were predictive for CNS involvement (p<0.05). The median survival of patients with CNS involvement was 2.628 months. Conclusions: Despite high relapse rates, PTCL carries a low risk of CNS involvement other than the ATLL subtype. As with other aggressive lymphomas, survival of patients with CNS involvement is poor and risk factors include: >1 extra nodal site and H-I-H IPI. In this dataset, prophylactic intrathecal chemotherapy does not appear to reduce the risk of CNS disease.