Studying patients’ cancer genomes with blood samples

Reporting in Nature Communications, the team demonstrates that nearly 90% of a tumour’s genetic features can be detected in blood samples using whole-exome sequencing, and that the method can be effectively applied in up to 49% of patients with advanced cancer — a number likely to increase as sequencing becomes cheaper. This less-invasive tumour sampling, which can provide a window into the cancer’s genome, has a range of potential applications.

“Our ultimate hope is to use blood biopsies to exhaustively search for and characterise even the smallest remnants of tumours,” explains Viktor Adalsteinsson, co-first author on the paper and group leader at the Broad Institute, where he leads the Blood Biopsy Team.

“And, as tumours evolve in more advanced stages of cancer, developing resistance or becoming metastatic, we might access timepoints that could be pivotal in deciding which therapies are right for that patient.”

This ability to detect and analyse cancer DNA from a patient’s blood sample is emerging as a promising alternative to invasive surgical biopsies, which can be difficult, painful, and costly — especially when tumours have appeared in locations that are challenging to access.

Blood biopsies (also called liquid biopsies) are poised to overcome many of these issues. They have the potential to allow doctors to track the progress of disease and treatment in real-time and to help researchers understand how tumours resist treatment with far greater resolution.

Cells in the body, including tumour cells, regularly expel fragments of DNA into the bloodstream when they die. With blood biopsies, clinicians collect this “cell-free” DNA from a blood draw and then detect and comprehensively profile the fragments originating from cancer cells.

Tracking these data could make it possible to monitor cancer recurrence, a patient’s response to treatment, and other clinically important features, all from blood samples.

The research and development arena for blood biopsies is frequented by both academic and industry players, but with scalable whole-exome sequencing, a team led by Adalsteinsson and colleagues Gavin Ha, Sam Freeman, Matthew Meyerson, J. Christopher Love, and Gad Getz is taking the field in a new and innovative direction. Love is a Broad associate member, associate professor of chemical engineering at MIT, and a member of the Koch Institute for Integrative Cancer Research at MIT.

Compiling a whole exome from DNA fragments currently requires at least 10% tumour DNA in a blood sample, but the fraction of tumour DNA in the blood can vary wildly from patient to patient. Because of this variation, the team first desired an unbiased approach for detecting and measuring levels of cancer DNA before attempting whole-exome sequencing.

Across the field, many blood biopsy methods detect tumour DNA by screening for mutations in known cancer-related genes, but this targeted sequencing misses cancers without those mutations.

Co-first author Ha, a postdoc at the Broad Institute and Dana-Farber Cancer Institute (DFCI), led the development of a tool called ichorCNA that can analyse DNA fragments for mutation patterns nearly universal in cancer genomes, and as a result capture cancers with both known and unknown mutations. Ha focused on detecting stretches of DNA that have either fewer or greater copies in cancer cells, in contrast to healthy cells.

The research team tested and refined ichorCNA on 1,439 blood samples collected prospectively from 520 metastatic breast or prostate cancer patients at DFCI (a significant effort championed by medical oncologists Atish Choudhury, Daniel Stover, Heather Parsons, Nikhil Wagle, and colleagues).

Using this approach, the researchers found that in 33 to 49% of the metastatic breast and prostate cancer patients, depending upon whether one or multiple blood samples were examined, tumour DNA made up greater than 10% of the cell-free DNA in their blood — enough to make whole-exome sequencing of cell-free DNA feasible.

Then, to determine whether this sequencing of cell-free tumour DNA could offer the same level of insight into cancer genetics as a tissue biopsy could, the team compared surgically obtained tumour biopsies to data collected from whole-exome sequencing of cell-free DNA from a group of 41 patients.

The researchers found that genetic data from blood whole-exome sequencing and tissue biopsies matched significantly across a number of genetic features, such as clonal somatic mutations (88% match) and copy number alterations (80% match).

These results support cell-free DNA whole-exome sequencing, from blood samples, as a potential substitute for metastatic tumour biopsy sequencing for many patients.

“Our study has demonstrated that we can get a cancer whole exome reliably, from blood; that it reflects the matched tumour biopsy; and that it can be done for a significant fraction of patients with metastatic cancer,” says Adalsteinsson, who was a postdoc in Love’s lab before joining the Broad Institute.

“This validation suggests that we can use blood biopsies for large-scale genomic characterisation of disease in patients with metastatic cancer.”

“It unlocks the potential for a lot of studies that we couldn’t do before,” adds Getz, institute member and director of the Cancer Genome Computational Analysis group at Broad and associate professor of Pathology and director of Bioinformatics at the Massachusetts General Hospital Cancer Center and Department of Pathology.

“The technology will allow us to track the dynamics of cancer and understand the evolution of drug resistance, or the development of the metastatic state, in a way that isn’t possible through surgical biopsies.”

The new study improves the analysis pipeline for blood biopsies and allows it to be performed at expanded scale. The researchers are actively applying their work to thousands of patients with metastatic cancer who may otherwise not have their tumours biopsied.

“With this work, we now have a framework for the precise measurement and quality control of tumour DNA in the plasma, enabling the genomic analysis of blood biopsies with high technical accuracy,” explains Meyerson, institute member at Broad and professor of pathology at DFCI and Harvard Medical School.