Circulating tumor DNA profiling may allow for surgery-free detection of CNS lymphoma
Click Here to Manage Email Alerts
Circulating tumor DNA can be readily detected in patients with central nervous system lymphoma and may be a strong clinical biomarker for risk stratification, outcomes prediction and surgery-free lymphoma classification, results showed.
“There are two major challenges associated with the clinical management of CNS lymphomas,” Florian Scherer, MD, of the department of medicine at University Medical Center Freiburg in Germany, said during his presentation of the findings at the plenary session of ASH Annual Meeting and Exposition. “On one hand, outcomes following methotrexate-based therapies are highly heterogenous and patients experiencing disease progression or relapse have a very poor prognosis. However, tools for accurate risk stratification and prediction of clinical outcomes are largely missing.”
In addition, invasive neurosurgical biopsies required for CNS lymphomas carry procedural risks and often are inconclusive or unable to be performed on frail or elderly patients.
Circulating tumor DNA (ctDNA) has emerged as an important noninvasive biomarker in multiple malignancies. However, its role in CNS lymphoma has been limited because of low ctDNA concentrations in plasma — resulting in low detection rates by next-generation sequencing — and limited applicability of single-gene assays, Scherer said.
Scherer and colleagues sought to overcome these limitations with Cancer Personalized Profiling by Deep Sequencing (CAPP-Seq) and Phased Variant Enrichment and Detection Sequencing (PhasED-Seq), developed by laboratories at Stanford University. They used CAPP-Seq to genotype genomic DNA in tumor tissue and ctDNA in cerebrospinal fluid (CSF) and plasma and PhasED-Seq to monitor ctDNA in CSF and plasma.
The primary cohort included 67 patients (median age, 65 years; 55.2% male) with primary (82.1%) or isolated secondary CNS lymphomas. Sixty-five patients had diffuse large B-cell lymphoma, one had chronic lymphocytic leukemia and one had CD38+ large B-cell lymphoma. Most (77.6%) had received curative-intent treatment.
Researchers profiled DNA from samples gathered at diagnosis, during treatment and at disease progression. To evaluate specificity of their approach, they obtained samples from 44 patients with other primary brain tumors or brain metastases and 24 healthy controls.
The investigators detected somatic mutations in 100% of CNS lymphoma tumor biopsies, with a median 288 mutations per patient. The most frequently affected genes included MYD88, PIM1 and CD79B.
With the ultrasensitive PhasED-Seq, they detected ctDNA in 78% of pretreatment plasma samples with 96% specificity and in 100% of pretreatment CSF specimens with 97% specificity.
Levels of ctDNA in plasma significantly correlated with total radiographic tumor volumes, but tumor size did not correlate with ctDNA levels in CSF, Scherer said. Periventricular localization of the tumor appeared to largely influence ctDNA detection in CSF.
Investigation of the whether pretreatment plasma ctDNA could be used as a predictive biomarker showed patients with detectable ctDNA had much higher rates of disease progression within 1 year (80% vs. 31%) and death within 2 years of blood draw (71% vs. 8%) than those with undetectable ctDNA. Further analysis showed significantly inferior PFS and OS among patients with positive vs. negative ctDNA.
“While patients with detectable ctDNA and high tumor volume seemed to have a particularly poor outcome, patients with undetectable ctDNA and low tumor volume showed an exceptionally favorable 2-year overall survival of 100%,” Scherer said.
An analysis of plasma ctDNA during curative-intent therapy also showed significant associations of ctDNA positivity with clinical outcomes, including PFS and OS.
Researchers then sought to test their hypothesis that CNS lymphomas could be diagnosed without surgical procedures based upon ctDNA mutational profiles in plasma or CSF. They used an ensemble of empirical Bayesian models to define a classifier score, which they applied to an independent validation cohort of 183 samples.
“We were able to correctly classify CNS lymphoma in 59% of CSF and 25% of plasma samples,” Scherer said. “Most importantly, we did not observe any false-positive results in our non-CNS lymphoma cohort, leading to 100% specificity and 100% positive predictive value.
The researchers proposed a possible clinical path “in which noninvasive CNS lymphoma classification directly leads to further staging and systemic treatment, and if the classifier predicts non-CNS lymphoma, the patient follows the conventional diagnostic procedure including brain tumor biopsy,” Scherer said.
Collapse
Mutter JA, et al. Abstract 6 (plenary session). Presented at: ASH Annual Meeting and Exposition. Dec. 11-14, 2021; Atlanta.
Read more about
Click Here to Manage Email Alerts