Circulating tumor DNA reliably measures lymphoma disease burden

ATLANTA — Circulating tumor DNA appeared to be a reliable biomarker in aggressive lymphomas and could be used to refine clinical estimates of disease burden among patients with diffuse large B-cell lymphoma, according to an international study presented at the ASH Annual Meeting and Exposition.

“The performance of circulating tumor DNA detection across diverse patients from diverse medical centers had not yet been extensively supported, and that was one of the goals of this study,” David M. Kurtz, MD, postdoctoral medical fellow of oncology at Stanford University Medical Center, said during his presentation. “Specifically, we wanted to ask how circulating tumor DNA varies with clinical and preanalytical factors, including the choice between plasma versus serum samples; clinical risk factors, including stage and the International Prognostic Index [IPI]; and metabolic tumors volume.”

Pretreatment tumor burden is a known prognostic factor in DLBCL. IPI — which includes lactate dehydrogenase and clinical stage — is used to capture tumor burden but lacks specificity.

Previous research by Kurtz and colleagues described circulating tumor DNA (ctDNA) levels as a novel quantitative biomarker for disease burden in DLBCL.

In this study, researchers assessed pretreatment ctDNA levels among 183 patients with DLBCL from six centers around the world.

Kurtz and colleagues performed Cancer Personalized Profiling by Deep Sequencing (CAPP-Seq) on pretreatment cell-free and germline DNA from patients undergoing intent-to-cure anthracycline-based therapy. Researchers identified somatic mutations and determined variant allele fractions. They used the mean tumor allele fraction and total concentration of cell-free DNA to determine ctDNA levels.

Researchers then correlated levels of ctDNA with clinical factors that included stage, IPI, pretreatment metabolic tumor volume and patient outcomes.

Patients enrolled in the study represented Stanford University (n = 52); University of Eastern Piedmont in Novara, Italy (n = 36); The University of Texas MD Anderson Cancer Center (n = 21); Hospital Le Bocage in Dijon, France (n = 25); the NCI (n = 33); and University Hospital Essen in Germany (n = 16).

Researchers detected ctDNA prior to therapy among 97% of patients, with at least 90% of patients from each center having detectable ctDNA. Concentration of ctDNA— expressed as haploid genome equivalents per mL (hGE/mL) — did not significantly differ across cohorts, with a median of 295 hGE/mL.

“In a multivariate analysis considering ctDNA concentration, IPI and site of origin, only ctDNA remained significantly prognostic for outcomes, which suggests ctDNA is actually the strongest at predicting risk factors that we have today,” Kurtz said.

PAGE BREAK

Without knowledge of tumor genotype, all but five patients had mutations identifiable via noninvasive genotyping. Among 88 patients with tumor samples available, researchers confirmed a median of 81% (interquartile range, 34-92) of variants identified through noninvasive genotyping by tumor sequencing.

Additionally, researchers found a significant difference in ctDNA levels between patients in different IPI risk groups; patients with higher-stage disease had higher ctDNA levels.

ctDNA levels also correlated with lactose dehydrogenase.

“ctDNA measurement by CAPP-Seq was robust and detectable in 97% of our subjects,” Kurtz said. “Mutations identified in plasma and tumors were highly concordant at about 80%; plasma appeared better than serum for finding mutations in patients; and pretreatment ctDNA was correlated with stage, IPI, and tumor volume. It was prognostic for EFS, potentially guiding our approaches in the future.”

The study confirmed a robust correlation between metabolic tumor volume and ctDNA level that was consistent across all three cohorts with metabolic tumor volume available (Essen, P = .0005; r = 0.8; Dijon, P < .0001; r = 0.73; Stanford, P = .006; r = 0.5).

Pretreatment ctDNA concentration appeared continuously associated with EFS (P < .0001) and OS (P = .0014). Patients with high ctDNA levels had significantly inferior EFS compared with those with low ctDNA (HR = 2.7; 95% CI, 1.4-4.6).

“We can look at it as pretreatment ctDNA measuring tumor burden, and posttreatment ctDNA measuring chemotherapy responsiveness,” Kurtz said. “Since these are independent predictors, we’re now thinking about ways we can combine pretreatment and dynamic risk factors to make a better mouse trap, a better prognostic tool.” – by Chuck Gormley

Reference:

Kurtz DM, et al. Abstract 310. Presented at: ASH Annual Meeting and Exposition; Dec. 9-12, 2017; Atlanta.

Disclosures: Kurtz reports no relevant financial disclosures. Please see the study for all other authors’ relevant financial disclosures.