ASH 2023: ctDNA in lymphoma management

Chimeric antigen T-cell (CAR-T) therapy has improved survival outcomes in patients but there is a gap in knowledge on why some patients do not improve after CAR T-cell therapy.¹ Circulating tumor DNA (ctDNA) has become an emerging biomarker in many cancers used to predict CAR T-cell resistance and stratify the patients possible survival outcomes earlier.¹

During the 65th American Society of Hematology (ASH) Annual Meeting and Exposition updates from three key trials assessing ctDNA as a tool to predict treatment response in patients with diffuse large B-cell lymphoma (DLBCL) were presented. Here, we summarize these updates; Foerster presented the association between ctDNA levels and CAR T-cell therapy,¹ Jardin presented ctDNA analysis from the POLARIX trial (NCT03274492),² and Alizadeh presented ctDNA analysis from the TRANSFORM trial (NCT03575351).³

AVENIO ctDNA analysis workflow¹

Foerster presented a trial exploring the value of ctDNA in patients with R/R DLBCL treated with CAR T-cell therapy and the association between ctDNA and known clinical risk factors. In total 16 patients were included with 48 samples collected. ctDNA concentration was increased in patients with:

• elevated LDH compared with normal LDH levels;
• transformed DLBCL compared with de novo;
• more prior lines of therapy (≥3 compared with 1–2 lines); and
• lower progression-free and overall survival.

ctDNA concentration was decreased in patients with complete and partial metabolic response after bridging therapy but before CAR T-cell therapy. After treatment, patients with complete and partial metabolic responses showed a significant reduction in ctDNA levels, whereas those with stable or progressive disease had no significant ctDNA reduction.

Presenter’s conclusions¹

Overall, pre-treatment ctDNA was associated with high-risk features and inferior progression-free survival (PFS) and overall survival in patients with DLBCL. ctDNA decreased one week after treatment in most cases and early ctDNA assessment after one week corresponds with early radiologic response after 4–6 weeks.

CtDNA: Results from the POLARIX trial²

The Lymphoma Hub previously reported the primary results from the POLARIX trial. Jardin presented a substudy of the POLARIX trial assessing the mutational landscape and subtyping of DLBCL by plasma ctDNA (using AVENIO assay) vs tumor tissue (using whole genome sequencing assay). At baseline, 443 patients had tissue DNA and ctDNA data:

• Mutated genes identified were consistent between tumor tissue and ctDNA assays (96% vs 71%, respectively)
• Among the top 20 mutated genes, 14 were present in both tumor tissue and ctDNA detection assays
• Overall, 11 genes showed statistical differences in prevalence in tumor tissue vs ctDNA, and the majority were expressed lower in tumor tissue compared with ctDNA (9.1% vs 90.9%, respectively)
• LymphGen results using tissue single nucleotide variants as a reference showed that 75% of cases have the same molecular subtype by tissue vs ctDNA, but 25% had different subtypes
• In the POLARIX trial, progression-free survival (PFS) was similar when assessed by ctDNA vs tumor tissue (Table 1)

Table 1. 2-year PFS estimates of subtypes determined by tumor tissue vs ctDNA*

PFS in patients by subtype, %	Tumor tissue DNA	ctDNA
ctDNA, circulating tumor DNA, EZB, EZH2 mutations and BCL2 translocations; MCD, MYD88L265P and CD79B mutations; PFS, progression-free survival. *Adapted from Jardin.2
MCD	75	77
EZB	80	76
Undetermined	73	74

Presenters conclusions²

This data shows that the mutational landscape of tissue DNA determined by whole exome sequencing is captured when using ctDNA by AVENIO assay. Additionally, similar PFS outcomes were reported in patients with molecular subtypes defined by tissue or ctDNA; therefore, supporting the use of plasma ctDNA as an alternative to tumor tissue typing in patients with DLBCL. Consistency between tissue and ctDNA single nucleotide variants was 75% which demonstrated that further enhancement of ctDNA sequencing is required for future study.

CtDNA: Results from the TRANSFORM trial³

Alizadeh presented data of a substudy from the TRANSFORM trial, previously summarized by the Lymphoma Hub. This trial met its primary endpoint with lisocabtagene maraleucel demonstrating higher event-free survival over the standard of care (52.6% vs 20.8%, respectively). Here, ctDNA-minimal residual disease levels were assessed in these patients to demonstrate surrogacy for disease burden and evaluate prognostic value (N = 184). Overall, prerandomization ctDNA levels correlated with:

• disease burden indices including secondary age-adjusted International Prognostic Index (p < 0.0001) and Ann Arbor stage (p = 0.024);
• radiographic burden (p < 0.0001); and
• lactate dehydrogenase (LDH) (p < 0.0001).

Prerandomization ctDNA levels did not correlate with relapsed/refractory disease status or histology and cell or origin. After treatment:

• Rapid reduction of ctDNA levels was seen in patients with complete response (CR) as early as Day 15
  • Association of ctDNA clearance and durable outcome was most significant at Month 3 (p < 0.0001)
• Half the number of patients with CR achieved ctDNA negativity by Day 15 with continued durable ctDNA clearance over time
• Patients with progressive disease and detectable ctDNA at Day 15 with lower ctDNA clearance was 89%
• Undetectable ctDNA after treatment was associated with durable benefit (p = 0.011), this was most significant at Month 3
• Detectable ctDNA in patients with CR at Month 1 and 12 identified those with a risk of progressive disease

Presenter’s conclusions³

Overall, this trial demonstrated the value of evaluating ctDNA levels in patients with non-Hodgkin lymphoma for disease surveillance as well as prediction of CR, durable clinical benefit, and risk of relapse.