We anticipate that ctDNA will have broad utility for dissecting tumor heterogeneity within and between patients with lymphomas and other cancer types, with applications for the identification of adverse risk groups, the discovery of resistance mechanisms to diverse therapies, and the development of risk-adapted therapeutics.— Florian Scherer, MD, and colleagues
Tweet this quote
In a study reported in Science Translational Medicine, Florian Scherer, MD, David M. Kurtz, MD (Conquer Cancer Foundation Young Investigator), Aaron M. Newman, PhD, and colleagues from Stanford University found that analysis of circulating tumor DNA (ctDNA) permits identification of patterns of clonal evolution and classification of tumor subtypes in patients with diffuse large B-cell lymphoma (DLBCL).1 Levels of ctDNA at diagnosis were associated with clinical measures and independently associated with outcome.
Study Details
The investigators hypothesized that characterization of mutational heterogeneity and genomic evolution using ctDNA profiling could identify molecular determinants of adverse outcomes. In the study, CAPP-Seq (Cancer Personalized Profiling by Deep Sequencing) was used to perform deep molecular profiling of lymphoma tissue biopsies and cell-free DNA samples from 92 lymphoma patients and 24 healthy subjects, in order to define features predictive of clinical outcomes.
Clinical Indices and Outcomes
At diagnosis, levels of ctDNA levels were correlated with serum lactate dehydrogenase (r = 0.94, P < .0001). Pretreatment ctDNA levels were also correlated with metabolic tumor volume on 18F-fluorodeoxyglucose positron-emission/computed tomography scans (r = 0.67, P < .0001), Ann Arbor stage (P = .0003), and International Prognostic Index (P < .0001). In multivariate analyses including clinical parameters, higher ctDNA levels were continuously and independently correlated with poorer progression-free survival.
David M. Kurtz, MD
Aaron M. Newman, PhD
Minimal Residual Disease and Early Relapse
Profiling of plasma samples at the time of radiographic complete response showed that ctDNA was identified in all patients at the time of clinical relapse and was also detectable as minimal residual disease prior to relapse in at least 1 plasma sample in 8 (73%) of 11 patients. The mean time between first ctDNA-positive measurement and clinical relapse was 188 days, with the longest lead time being 2.5 years before radiographic recurrence.
Compared with immunoglobulin sequencing (Ig high-throughput sequencing), ctDNA measurements detected minimal residual disease in twice as many patients, with a mean lead time of more than 2 months. ctDNA was never detected in plasma samples from 10 patients who were disease-free for at least 24 months after therapy or in 24 healthy adult subjects.
Progression-free survival was significantly longer in patients never positive for ctDNA vs ever positive after treatment (P = .0003). A borderline significant trend for overall survival was also observed (P = .056). The investigators noted: “[T]hese results illustrate the promise of ctDNA profiling by targeted sequencing for improved MRD assessment and early relapse detection.”
Emergent Resistance Mutations
The technique also permitted noninvasive identification of emergent resistance mutations. For example, noninvasive genotyping of three patients with progressive disease receiving ibrutinib (Imbruvica) identified emergent resistance mutations in BTK, which exhibited distinct clonal dynamics in two of them.
In one patient with a BTK C481S substitution, two adjacent mutations in the corresponding codon were never observed within the same ctDNA molecule—a finding indicating convergent evolution of independent resistant subclones. To the investigators’ knowledge, BTK C481S had not been previously described in DLBCL.
They stated: “These results suggest that tumor genotyping from plasma can facilitate monitoring of BTK-targeted therapy, regardless of histology. Thus, ctDNA profiling with CAPP-Seq has utility for real-time assessment of dynamic tumor processes, including clonal evolution and the acquisition of molecular resistance.”
Cell-of-Origin Classification
Multiplexed somatic mutation profiling was used to develop a technique for cell-of-origin classification from tumor or pretreatment plasma. In the group of 76 tumor biopsies, 44 patients were predicted to have germinal center B-cell–like (GCB) and 32 to have non-GCB DLBCL.
Circulating Tumor DNA and Diffuse Large B-Cell Lymphoma
- ctDNA analysis allowed identification of patterns of clonal evolution and classification of tumor subtypes in patients with DLBCL.
- ctDNA levels at diagnosis were correlated with clinical indices and independently predictive of patient outcome.
Comparison of these results with blinded, centralized immunohistochemical classification using the Hans algorithm (current clinical standard) showed concordance of approximately 80%. GCB was associated with significantly better progression-free survival vs non-GCB (P = .02), and cell-of-origin classifier scores were continuously associated with improved progression-free survival (P = .003).
In comparison, the Hans algorithm failed to stratify clinical outcomes among patients analyzed by both immunohistochemistry and DNA genotyping. Testing of the cell-of-origin classifier in pretreatment plasma (n = 41) showed 88% concordance with predictions from tumor tissue, and the DLBCL molecular subtypes predicted from plasma were significantly associated with progression-free survival (P = .02).
Clonal Evolution
The investigators also identified distinct patterns of clonal evolution distinguishing indolent follicular lymphomas from those that transformed into DLBCL, potentially permitting noninvasive prediction of histologic transformation. Use of CAPP-Seq in three groups of paired tumor samples—ie, diagnostic follicular lymphomas vs transformed follicular lymphomas, diagnostic follicular lymphomas vs nontransformed follicular lymphomas, and diagnostic de novo DLBCL vs relapsed or refractory DLBCL—showed the greatest genomic evolutionary distance among tumor pairs associated with histologic transformation.
It was found that genomic divergence was independent of time to progression or transformation and number of prior therapies. Analysis of tumor biopsies and follow-up plasma from patients with indolent lymphomas experiencing transformation, progression without transformation, or relapsed or refractory DLBCL showed that plasma genotyping was largely consistent with that in analysis of sequential tumors; a higher fraction of emergent variants distinguished transformed follicular lymphomas from other histologies. In addition, it was found that higher ctDNA levels distinguished transformed follicular lymphomas and relapsed or refractory DLBCL from nontransformed follicular lymphomas.
The latter findings suggest that aggressive lymphomas have similar tumor cell proliferation and turnover kinetics despite separate origins. Overall, noninvasive classification was able to distinguish transformed follicular lymphomas from nontransformed follicular lymphomas with 83% sensitivity and 89% specificity.
The investigators concluded: “Collectively, our results demonstrate that ctDNA analysis reveals previously undescribed biological factors underlying lymphoma clinical outcomes and could facilitate individualized therapy.” They noted: “[N]oninvasive genotyping and serial ctDNA monitoring represent promising approaches for uncovering biology and improving patient management. We anticipate that ctDNA will have broad utility for dissecting tumor heterogeneity within and between patients with lymphomas and other cancer types, with applications for the identification of adverse risk groups, the discovery of resistance mechanisms to diverse therapies, and the development of risk-adapted therapeutics.” ■
Disclosure: This work was supported by the Damon Runyon Cancer Research Foundation, American Society of Hematology, V Foundation for Cancer Research, German Research Foundation, Stanford TRAM Pilot Grant, ASCO, National Cancer Institute, National Institutes of Health Director’s New Innovator Award Program, and Ludwig Institute for Cancer Research. For full disclosures of the study authors, visit stm.sciencemag.org.
Reference
