Whole-genome methylation shows promise as multicancer early-detection test
Click Here to Manage Email Alerts
Evaluation of multiple classifiers, both alone and in combination, supports the decision to focus on cell-free DNA methylation patterns to detect cancer and predict the cancer signal origin, according to a presentation at ESMO Congress 2021.
“In addition, analysis of the paired tissue and blood samples from the participants with cancer enabled determination of the circulating tumor allele fraction (cTAF), which informs a clinical limit of detection and makes it possible to assess detection performance against cell-free DNA shedding behavior,” Minetta C. Liu, MD, professor and research chair in the department of oncology and consultant in the department of laboratory medicine and pathology at Mayo Clinic, in Rochester, Minnesota, told Healio.
Liu presented results of the first substudy of the observational case-control Circulating Cell-free Genome Atlas (CCGA) study, designed to develop and validate a cell-free DNA multicancer early-detection test.
Lui and colleagues pursued the research after identifying a clear need for improvement and expansion of screening approaches, given that population-scale cancer screening paradigms are limited to breast, cervical, colorectal, lung and prostate cancer and that detection of circulating tumor DNA (ctDNA) in the peripheral blood was “a natural fit” in early detection for multiple cancers.
“Circulating cancer biomarkers were historically limited to tumor profiling for drug selection in advanced malignancies,” Lui said. “But recent technologic and scientific advances have improved the sensitivity of ctDNA detection to the degree that molecular residual disease monitoring and cancer screening through blood are realities.”
The substudy analysis included 2,261 participants: 1,414 in the independent training set (cancer, n = 854; non-cancer, n = 560) and 847 in the validation set (cancer, n = 485; non-cancer, n = 362).
Liu and colleagues collected and sequenced plasma and matched white blood cells from participants and, when available, also sequenced tumor biopsies. Investigators used whole-genome methylation data from whole-genome bisulfite sequencing, small somatic variant data from error-corrected targeted sequencing and somatic copy-number alterations, fragment length, fragment endpoint and allelic imbalance data from whole-genome sequencing.
After splitting samples into the training and validation sets, investigators trained 10 classifiers to detect solid cancer (carcinomas, sarcomas and lymphomas). They then assessed the classifiers for cancer detection and clinical limit of detection, the latter estimated as the probability of detecting cancer as a function of circulating tumor fraction using matched tumor biopsies.
Using whole-genome methylation, targeted sequencing and somatic copy-number alterations, investigators trained three additional classifiers to predict cancer signal origin and assess for accuracy.
Results showed that whole-genome methylation predicted cancer signal origin with 75% accuracy and, thus, was significantly more accurate than single-nucleotide variant-white blood cell (35%) or somatic copy-number alteration (41%) classifiers. Additionally, the clinical limit of detection with whole-genome methylation was more than 1.5-fold lower than any whole-genome sequencing or targeted sequencing classifier. Cancer signal origin prediction was more than 1.8-fold more accurate using whole-genome methylation than targeted sequencing or somatic copy-number alterations, Liu and colleagues wrote.
Other highlights of whole genome methylation, Liu said during her presentation, were that it does not require removal of biological background from paired white blood cells, unlike single-nucleotide variant-white blood cell and pan feature, and that it has significant potential for further improvement in achieving higher performance through a targeted methylation assay approach.
“These results reflect ongoing work designed to improve our understanding of the biology related to ctDNA detection,” Liu told Healio.
Liu and colleagues concluded that the whole-genome methylation assay’s outperformance of the other tests, without needing additional white blood cell sequencing for clonal hematopoiesis, and the data that resulted inform the design of a “significantly improved” target methylation, multicancer early detection test for further CCGA studies to support clinical use.
“Clinical limit of detection is a useful benchmark by which to assess cell-free DNA-based test performance,” Liu said in the presentation. “Whole-genome methylation provided the most promising cell-free DNA approach for a blood-based multicancer early-detection test.”
Perspective
Back to Top
Christian Rolfo, MD, PhD, MBA
This is impressive research in multi-omics to confirm this methylation approach is the most promising in multicancer early detection. This approach seems to be less complex and avoids confounding factors, and it has a high specificity in the validation. FDA approval for this technology is still needed.
Obviously, we are missing some earlier-stage and some high-risk populations. We don’t know if those heterogenous populations can affect the performance that is necessary in stage IV disease. There is missing data in nodule discrimination for some diseases, like lung cancer. And there are differences in histology that still need to be clarified.
So, we are passing from multi-omics to whole-genome methylation in cell-free DNA, but maybe in the future we need to go for something that includes the multi-omics beyond the cell-free DNA.
Collapse
Liu MC, et al. Abstract 1123O. Presented at: European Society for Medical Oncology Congress (virtual meeting); Sept. 17-21, 2021.
Read more about
Click Here to Manage Email Alerts