A Prospective Study of Clonal Evolution in Follicular Lymphoma: Circulating Tumor DNA Correlates with Overall Tumor Burden and Fluctuates over Time without Therapy

Background: Follicular lymphoma (FL) shows marked variation in clinical course including spontaneous regression and histologic transformation (HT). Watchful waiting (W&W) is routinely applied to pts with newly diagnosed FL, but monitoring strategies are not standardized. Pts with progression within 1-2y of diagnosis have worse outcomes, but the biologic basis is unclear and biologic-based classifiers are not routinely applied at diagnosis. Circulating tumor DNA (ctDNA) is a highly tumor-specific biomarker that is prognostic in aggressive B-cell lymphomas, but its ability to serially monitor FL remains undefined. We applied a next-generation sequencing assay to identify tumor clonotypes for serial monitoring of peripheral blood in pts with untreated FL as part of an ongoing prospective clinical trial [NCT03190928].

Methods: Pts with grade I-II or 3A FL are eligible if evaluable disease on CT or FDG-PET, age ≥18, ECOG ≤2, no evidence of HT, and no prior systemic therapy. Pts undergo W&W until they meet uniform protocol-defined treatment criteria and remain on study until second-line therapy. Baseline testing includes labs, peripheral blood flow cytometry, BM biopsy/aspirate, CT and FDG-PET scans, and research biopsy. Pt have clinic visits every 4m for 2y, every 6m in years 3-5, then annually. CT scans are every 8m for 2y, then annually. FDG-PET scans are at baseline, at 2y, and any time of suspected progression. Peripheral blood samples including Streck tubes (plasma) and PBMCs are drawn at each clinic visit and stored. For ctDNA analysis, tumor DNA was amplified from FFPE using locus-specific primer sets for the Ig heavy-chain and light-chain loci along with BCL1/BCL2 translocations (Adaptive Biotechnologies). Amplified products were sequenced and tumor clonotypes were identified in plasma and PBMCs. Serial tracking of ctDNA was done in plasma and blinded to clinical outcomes.

Results: 77 pts enrolled between July 2017 and July 2021. Median age was 57 (range 24-83) including 14 (18%) low-risk, 29 (38%) intermediate-risk, and 34 (44%) high-risk by FLIPI. Fourteen (18%) pts had stage I-II disease. Forty-three (56%) pts had monoclonal B-cells on peripheral blood flow cytometry. Twenty-nine (38%) pts progressed requiring frontline therapy including 7 (9%) pts with HT. Twenty-five (32%) pts were monitored ≥2y with no progression including 10 (13%) pts with evidence of at least some spontaneous regression by CT. Twenty (26%) pts were on study for <2y and 3 (4%) pts were unevaluable. Of 36 pts with FFPE analyzed, 35 (97%) had ≥1 dominant tumor clonotype identified for tracking. Of these, 32 (91%) had ≥1 dominant clonotype identified in plasma suitable for tracking. Three pts without available FFPE had the dominant clonotype identified from plasma that was used for tracking. In 28 pts, plasma was compared to PBMCs, and ≥1 clonotype was identified in both compartments in 21 (75%), plasma only in 5 (18%), and neither in 2 (7%). The median overall levels of baseline plasma ctDNA was 104.7 counts/mL (range 4-2096); progressors had median levels of 136.0 counts/mL (range 4-1493) compared to 40.9 count/mL (32-2096)(p=0.85) for non-progressors (Figure 1A). Baseline ctDNA correlated with the total metabolic tumor volume (TMTV) on FDG-PET scan (r=0.49, p=<0.006)(Figure 1B). Pts who required immediate treatment often had very high baseline levels of ctDNA (black dots, Figure 1C) and serial monitoring prior to progression demonstrated patterns of sharp increases in ctDNA prior to progression or relatively stable levels in others (Figure 1C). Notably, significant fluctuations were noted in ctDNA for non-progressing pts with an overall trend for decreasing values including four pts with evidence of spontaneous regression by CT with corresponding decreases in ctDNA over time (Figure 1D). As a notable examples, FL_14 had baseline ctDNA level of 2096 counts/mL with decreases to 101 counts/mL after 4 months, 14 counts/mL at 16 months, and no detectable ctDNA at 2 years.

Conclusions: ctDNA quantified from plasma in FL mirrors TMTV. Serial monitoring of ctDNA in patients without therapy demonstrated various patterns of fluctuation, including some patients in which ctDNA became undetectable coincident with spontaneous clinical regressions. ctDNA thus provides a non-invasive platform to monitor the natural history of FL, enabling future studies of tumor immune surveillance in this disease.

Figure 1

View largeDownload slide

Figure 1

View largeDownload slide

Close modal