December 6, 2022 | A research team from The Ohio State University and Foundation Medicine suggests that tumor fraction has the potential to serve as a pragmatic, tumor-agnostic prognostic tool. They published their results in Annals of Oncology last month.
Over the past few years, there has been growing interest in tumor fraction as a potential biomarker, explains Daniel Stover, an Associate Professor of Medicine at The Ohio State University College of Medicine and the James Cancer Hospital/Solove Research Institute.
Circulating tumor fraction (TF) is the proportion of circulating tumor DNA (ctDNA) in a blood sample. In 2018, Stover and colleagues published work in the Journal of Clinical Oncology focused on triple-negative breast cancer. They evaluated circulating DNA tumor fraction for 164 patients and found that tumor fraction ≥ 10% was associated with significantly worse survival. (DOI: 10.1200/JCO.2017.76.0033)
But the team wanted to explore whether tumor fraction could offer robust prognostic information for patients with advanced cancer across multiple tumor types.
“To date, most prognostic implications of TF quantification have been evaluated primarily within individual disease settings and not across cancer types,” they wrote in the Annals of Oncology paper (DOI: 10.1016/j.annonc.2022.09.163). So they worked with Foundation Medicine and other researchers from the University of Michigan, Flatiron Health, and the University of Bologna to test TF as a prognostic biomarker in other indications and see how representative it truly is.
The authors looked at patient data from the Flatiron Health-Foundation Medicine clinico-genomic database (CGDB), comprising de-identified data from about 280 US cancer clinics. The team considered patients with metastatic castration-resistant prostate cancer (mCRPC), metastatic breast cancer (mBC), advanced (stage IIIB-IV or progressive/recurrent) NSCLC (aNSCLC), and metastatic colorectal cancer (mCRC)—all common tumor types known to release circulating tumor DNA, Stover says.
They further narrowed the pool to patients within the CGDB with suitable liquid biopsy results from either the FoundationOne Liquid CDx assay or the FoundationOne Liquid assay, sufficient documentation of treatment start, and quantifiable tumor fraction measured within 60 days pre-therapy. Overall survival was determined via retrospective longitudinal clinical data from EHRs. A total of 1,725 patients were included across the four indications.
They found that across cancer types, elevated TF of at least 10% was strongly associated with worse outcomes in each cancer type.
“One of the impactful things about this [study] is that looking across many hundreds of patients—with totally different cancer types—this same cutoff of ctDNA that we had studied five years ago holds up,” Stover says.
Tumor fraction is now routinely returned when patients get the FoundationOne Liquid CDx assay or the FoundationOne Liquid assay along with blood tumor mutational burden (bTMB) and microsatellite instability high (MSI-H). And while TF could be a standalone test, Stover thinks the biomarker’s greatest utility will continue to be conjunction with the other test results.
“I think its greatest utility will probably be integrated with looking at specific mutations as, for example, Foundation does. Many of the other liquid assays are likely to do something similar where they report a quantification or a tumor fraction, but if they detect mutation, also report those. I really believe that it offers complementarity with the mutations and can give clinicians more data.”
TF is variable, Stover acknowledges, across tumor type and individual. For instance, renal cell carcinoma tends to shed less circulating tumor DNA than other cancers, but even for some blood cancers—myeloma, for instance—ctDNA can be a marker.
In the paper, the authors highlight that the findings could be immediately applicable in the analysis of clinical trial cohorts. “Quantification of TF and comparison to real-world cohorts could characterize whether an enrolled population is representative of the expected clinical presentation of the disease. If a phase II trial has a high rate of disease stability without tumor response, measurement of ctDNA TF could identify whether these favorable outcomes may be due to the enrollment of patients with low ctDNA shed,” they write.
Stover also points out another interesting avenue for research to Diagnostics World. “Across tumor types, if you see that a tumor has a low tumor fraction—so low shed of circulating tumor DNA—those patients generally do better. Can we use this as a way to start to think about giving some patients less therapy, or less aggressive therapy, because we think that they’re going to have outcomes just based on this?” he asks. “If we’re starting to think of ways we can tune or tailer therapy, I think that clinical setting offers potential utility for this idea.”
TF also varies across individual patients, and it’s not yet proven that a patient’s TF will always increase with disease progression, he adds. The dynamic nature of the biomarker—when and how it responds to disease progress or regression—is of great interest to Stover and his team.
“If it changes from high to low when a patient is on treatment, or low to high, is that reliable enough to help us direct therapy? That’s being investigated by us and many other groups. That’s sort of a next wave.”