Dynamic Response Assesment Combining Liquid Biopsy MRD and PET/CT in Follicular Lymphoma Patients Including CAR-T Cell Therapy

INTRODUCTION: Follicular lymphoma (FL) is considered an indolent disorder with a relatively favorable course. However, some patients have a more aggressive course and a poorer outcome. Chimeric antigen receptor (CAR) T-cell therapy has improved outcomes for R/R non-Hodgkin lymphomas (NHL). However, nearly 60% of patients will either fail to respond or relapse after CAR T-cell therapy. Currently, PET/CT is used to assess response in FL. Circulating tumor DNA (ctDNA), as a minimal residual disease (MRD) biomarker, has been validated in other NHL, but never in FL. METHODS: A total of 58 FL patients were included. Of them, 37 received first-line therapy, 14 a non-CAR T-cell salvage treatment, and seven salvage CAR-T cell therapy. By PET/CT, 42 patients achieved CR, five PR and 11 EE or progression. Genomic profiling to identify somatic mutations suitable for liquid Biopsy MRD monitoring (LiqBioMRD) was performed at baseline on DNA from FFPE lymph node biopsies and/or cfDNA from peripheral blood plasma samples. Then, the previously described NGS-MRD method (Onecha, E et al. Hematologica 2019) was applied to 175 cfDNA follow-up samples. Of them, 27 were screened at mid-induction and 49 at the end of treatment (EOT) and presented a paired PET/CT scan (month_3 and month_6 for CAR-T patients). A mean of 81.7 ng, (rank 3.5 - 943.8 ng) were analyzed with the LiqBioMRD test with a cut-off of 1E-4. In CAR-T patients, cfDNA was obtained on day_7 (n=4), day_28 (n=6), day_90 (n=6), day_180 (n=6) and day_365 (n=4). For those patients, PET/CT examinations were performed on day_90, day_180, and day_365. PET/CT was considered positive when DS was 4 or 5. RESULTS: The most frequently mutated genes at baseline were CREBBP (78%) KMT2D (69%), BCL2 (43%) and TNFRSF14 (37%). The mean number of MRD biomarkers analyzed per patient was 3.8 (rank 1 -12). After a median follow-up of 26 months, 16 patients relapsed with a median time to relapse of 10 months. A higher risk of progression was observed with the LiqBioMRD test (HR 14.6, 95% CI 5.1-42.2 p < 0.001) and PET/CT (HR 5.4, 95% CI 2.0-14.2, p = 0.001). The sensitivity (SE) of the LiqBioMRD test for the last TP available was 71%, with a specificity (SP) of 98%, being SE=62% and SP=86% for PET/CT. When both tests were combined, progression in less than 24 months was identified with 80% SE and 100% SP. Regarding the 15 patients with discordant results (26%), 11 were incorrectly classified by PET/CT, and only 4 were by LiqBio-MRD (Figure panel A). More importantly, of the 42 patients with concordant results, only two were misclassified. These two patients progressed despite being negative on both tests during the first six months after treatment. The first patient progressed after 12 months from CAR T-cell infusion. At this time point, both PET and liqBio-MRD were positive (Figure panel B). The second patient negative by both tests becomes positive in a sequential cfDNA sample obtained under rituximab maintenance after RCHOP, five months before progression. Of note, this patient continued maintenance therapy, achieving a complete response a few months later. Considering only CAR T-cell patients, all patients achieved CR by day_90 PET/TC evaluation. After a median follow-up of 21 months, only 1 of 7 patients progressed 12 months after infusion, presenting an MRD positive test at this point. The other six patients remain in CR. Regarding the 24 samples obtained from those patients after CAR-T infusion, at day_7, 4 of 4 samples were LiqBio-MRD positive, at day_28 only 1 of 6 remained positive. All samples at day_90 and day_180 (n=12) presented a negative LiqBioMRD result. Only one sample was positive at day_365, corresponding to the time point of progression of this patient (Figure panel B). CONCLUSIONS: Our preliminary results suggest that the combination of imaging with the LiqBio-MRD test during the first six months of treatment is a robust approach to identifying patients with a high risk of relapse independently of the therapy line. This combination presented an unprecedented sensitivity and specificity, and should be considered in future response-adapted clinical trials. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal