Interim Circulating Tumor DNA As a Prognostic Biomarker in the Setting of Interim PET-Based Adaptive Therapy for DLBCL

Background: While the majority of diffuse large B-cell lymphoma (DLBCL) patients are cured with R-CHOP immunochemotherapy, a significant proportion of patients still experience disease relapse. Studies using interim PET (iPET) to select patients for therapy intensification have failed to improve survival, at least partially, due to imperfect risk stratification. Circulating tumor DNA (ctDNA) is an emerging biomarker in lymphoma and ctDNA dynamics early in therapy have been shown to predict treatment outcomes in DLBCL (Kurtz, JCO 2018). Here, we assess the utility of interim ctDNA after 3 cycles of front-line therapy, in the context of standardized interim PET/CT imaging and PET-driven adaptive therapy.

Methods: We quantified ctDNA levels using Cancer Personalized Profiling by Deep Sequencing (CAPP-Seq) in plasma samples collected before treatment and prior to cycle 4 in 39 patients with de novo DLBCL. 28 patients were from the NHL21 trial of the Australasian Leukaemia Lymphoma Group, in which patients initially received R-CHOP14 as frontline therapy (Hertzberg, Haem 2017). Patients with positive iPET after cycle 4 were escalated to R-ICE (rituximab, ifosfamide, carboplatin, etoposide) followed by 90Y-ibritumomab tiuxetan-BEAM (BCNU, etoposide, cytarabine, melphalan) and autologous stem cell transplant (ASCT). iPET negative patients continued R-CHOP14 for a total of 6 cycles. iPET was evaluated centrally by 2 nuclear imaging specialists according to the International Harmonization Project (IHP) criteria. An additional group of 11 patients received 6 cycles of R-CHOP as standard therapy at Stanford University. All survival analyses are calculated from date of diagnosis. Matched germline DNA was sequenced in 28/39 cases and all samples were uniformly processed at Stanford.

Results: Median follow-up for progression-free survival (PFS) for the patients from the NHL21 cohort was 2.6 years [95%CI 2.23; 2.97]. Within the 28 NHL21 patients evaluated, mean baseline metabolic tumor volume was 898.76 cm³(+/- 112.01). Of the 28 patients with interim PET data, 10 were iPET-positive. Agnostic to ctDNA detection, the 2 year mean progression free survival (PFS) for these 10 patients was 60.0% [95% CI 36.2; 99.5%] as compared to 66.7% [95% CI 48.1%; 92.4%] within the iPET-negative group (Hazard Ratio (HR) 1.649 [95% CI 0.50; 5.4], p=0.4, Fig 1A). Within the 39 patients monitored across both cohorts, 17 patients possessed detectable disease by ctDNA at C4D1 with a mean disease burden of 3.59 haploid genome equivalents per milliliter of plasma (hGE/mL, +/- 1.23). Plasma genotyping within the 28 patients from the NHL21 cohort discovered an average of 112.9 single nucleotide variants (SNVs) which were then used to monitor plasma samples at the C4D1 timepoint, comparable to a broader cohort of NHL cases previously described (Kurtz, JCO2018). 2-year PFS of the 22 patients with undetectable disease as assessed by ctDNA at C4D1 was 81.82 [95% CI 67.2; 99.6] compared to 47.1% within the 17 patients with detectable ctDNA [95% CI 28.4; 77.9] (HR 4.42 [95% CI 1.32; 13.76],p=0.006, Fig. 1B). Detection of ctDNA at C4D1 also has a prognostic value for overall survival (p=0.02, Fig. 1C). iPET concordance with with C4D1 ctDNA-positivity was 50% and within the 18 patients who were iPET-negative as part of the NHL21 study and did not receive therapy escalation, 9 patients with undetectable disease by ctDNA demonstrated 88.9% 2-year PFS [95% CI 70.6; 1.00] compared to 44.4% [95% CI 21.4; 92.3] for the 9 patients with detectable ctDNA (p=0.03, Fig. 1D).

Conclusions: Late in the course of DLBCL therapy, ctDNA carries promising value as a biomarker for stratifying predicted patient response to therapy as evidenced by additional detection of additional 50% detection of relapsing cases not categorized as iPET-positive. Data from additional patients and relationships between ctDNA response as measured by EMR (C2D1), MMR (C3D1), and C4D1 will be presented at the meeting.

View largeDownload slide

View largeDownload slide

Close modal