P14.08.a prognostic value of csf il-10 at the 2-month assessment after the start of induction chemotherapy in primary cns lymphomas. a loc network study

Abstract BACKGROUND Despite a high rate of objective response at first evaluation of ongoing high dose methotrexate (hd MTX) based induction chemotherapy, the risk of early relapse remains high and unpredictable in patients suffering from primary CNS lymphomas (PCSNL). Since cerebrospinal fluid (CSF) IL-10 level has been shown prognostic in PCNSL, both before and at the end of the treatment, it is used in routine in some centers. The objective of the present study was to assess the prognostic value of early CSF IL-10 level (e-IL-10), performed 2 months after the beginning of the induction chemotherapy. MATERIAL AND METHODS We retrospectively selected the immunocompetent patients treated at Pitié-Salpêtrière hospital and Institut Curie for a PCNSL diagnosed between January 2019 and July 2022 according to the following criteria : histological diagnosis of diffuse large B-cell lymphoma ; hd MTX-based chemotherapy as induction in first line treatment; complete response (CR) or partial response (PR) according to the IPCG criteria at the 2-month evaluation; availibility of dosage of e-IL-10. CSF IL-10 was classified as detectable (≥2.5 pg/ml) or undetectable (<2.5 pg/ml) RESULTS Thirty patients (median age 63 years, median KPS 70, 17 men) met the selection criteria. At initial diagnosis, all patients presented with brain lymphoma, associated with CSF (N=10) and/or intraocular involvement (N=2). Before treatment, 22 of them had a dosage of CSF IL-10, which was detectable in 21 cases (median 27.5 pg/mL, range 3-1844 pg/mL), with a ratio IL-10/IL-6>1 in 21 cases. At the 2-month evaluation after the beginning of the treatment, 25 patients were in PR and 5 were in CR and e-IL-10 was undetectable in 22 patients and detectable in 8 patients, with the following values: 105, 30, 20, 16, 14, 10, 6, 3 pg/mL . For the patients with detectable e-IL-10, e-IL-10 was superior to initial IL-10 level in 1 case and inferior in 5 cases (no data at baseline for 2 patients) and the ratio CSF e-IL-10/e-IL-6 was superior to 1 in 8/8 cases. At the end of induction treatment, all but one patient (the one with an e-IL-10 value of 3 pg/ml) had experienced progressive disease in the group with detectable e-IL-10 (2 progression in the brain, 5 in the CSF) vs 4/22 patients in the group with undetectable e-IL-10 (p=0.001). The median follow-up was 16 months. Median PFS was 5.8 months in the group with detectable e-IL-10 and 28.6 months in the group with undetectable e-IL-10 (p=0.0001). 1-year OS was 71% in the group with detectable e-IL-10 and 86% in the group with undetectable e-IL-10 (p=0.36). CONCLUSION Our results suggest that detectable CSF IL-10 at 2 months after the start of first line induction chemotherapy in PCNSL is associated with a high risk of early relapse and that closer follow-up of these patients is warranted. Further studies are needed to confirm these results and to assess whether the therapeutic strategy should be adapted accordingly in these cases.