Free at Last, (Cell) Free at Last?

Cell-free DNA analysis may improve the scalability of genomics for prostate cancer.  The management of CRPC has long been plagued by the inability of clinicians to get an accurate sense of the underlying biology of the tumor that they are treating. 

Those of us with some familiarity with cell lines and a general sense of prostate cancer pathology have known that PTEN loss, P53 mutation and Pi3Kinase mutation are common events in the disease writ large, but we’ve never been able to tell which, if any, of these aberrations are occurring in the patient we are talking to in the clinic at or near the time that we are talking to him. Furthermore, biopsies of bone metastases, requiring meticulous decalcification, histological evaluation and DNA extraction has been relegated to large research centers. Not to diminish the importance of this work, which has taught us a lot about the disease, but it suffers from scalability problems. 

And so, emerging data with cell-free DNA technology is particularly encouraging and has my vote for the most disruptive technology to emerge in the management of advanced prostate cancer in the last couple of years. It faces some barriers of validation, but is worth watching for sure. 

Kim Chi and Alex Wyatt, both at the BC Cancer Agency in Vancouver, British Columbia, demonstrated an effective potential use of this technology in Dr Chi’s presentation earlier this summer at ASCO ( and a forthcoming publication).  In it, they were able to demonstrate differences in patient outcome based on circulating tumor DNA Analysis. 

For those who are not familiar with this technology and its scope, a brief primer: First, our blood streams are filled with circulating DNA fragments, so-called ‘cell-free’ DNA or CF DNA.

Virtually all patients with advanced CRPC will have detectable CF DNA, and in a prostate cancer patient, a significant amount of the CF DNA is circulating tumor DNA, or CTDNA. In the study presented by Dr. Chi, about three-quarters of patients had measurable CTDNA fraction. In advanced disease patients for example, a majority of the circulating DNA is in fact tumor DNA. Further, these CT DNA fragments are similar to the DNA sequences that one would see following a biopsy. Mutations found in the CT DNA replicate those detected in the biopsies of the tumor. A rich resource indeed, and easily accessible.

With this analysis method, the Vancouver team was able to analyze 73 genes that are known contributors to the pathogenesis of CRPC, among them cancer drivers (e.g. AR, SPOP, NKX3.1, FOXA1), cell cycle regulators (e.g. TP53, RB1, CDKN1B, CDKN2A), mediators of DNA repair (e.g. BRCA1/2, FANC family genes, ATM, MSH2/6) and the PI3K pathway (e.g. PIK3CA, PTEN, AKT1). There is a lot of biology, prognostic and even predictive significance in these genes, especially when you consider the potential significance of measuring their interacting effects ( for example, Ana Aparicio and colleagues at MD Anderson have shown that combined aberrancies in P53, PTEN and RB are associated with a very poor prognosis).

Aberrations of the androgen receptor are of particular interest, as they are associated with resistance to enzalutamide, abiraterone, and presumably the other drugs targeting this axis that are in development. In a paper a couple of years ago, AR copy number gain was observed in CT DNA in association with Enzalutamide resistance, and mutations within Exon 8 of the Androgen receptor ( the region associated with ligand binding) harbored many mutations.

There are some challenges in this process, and yes it needs to go through a few rounds of validation and improved scalability before we can be doing it at the bedside (or, I should say, in the clinic) on a routine basis. One major limitation is that, while the amount of CT DNA is high ( where by quantity it is prognostic) in a patient with advanced disease, it is low in a patient with a low burden of disease or localized disease, a fact that may limit its application to earlier stage prostate cancer.  Also, most of the panels that are run on CT DNA aim to detect short DNA segments, so large structural rearrangements within the genome may be missed. The intrapatient variability is also something that needs to be rectified. Do blood draws taken from the same patient on consecutive days, or consecutive hours, give the same result?

Nevertheless, the ease of use of a standard blood test and the conformity to what biopsies show suggest that this is a technology that may become very useful over time as a research tool for sure, but also potentially as a treatment predictor and prognostic test.

When will these results hit the clinic on a regular day? The easy answer is when it tells us something positive ( like what to do), as opposed to something negative ( like that our treatment isn’t working, which we probably already knew). While biomarkers of treatment resistance are important in drug development and for raising research questions, they generally don’t guide the typical clinician.  My guess is that they will become standard when they are linked to a therapy that is given as a direct result of the test, for example a parp inhibitor given when coupled to a CT DNA Test showing a BRCA 2 or ATM mutation. Many of the large registration studies with parp inhibitors are including these analyses in their protocols, with a rigorous plan to evaluate the outcome of the testing strategy in addition to the outcome of the pharmacologic therapy being tested. 

Watch this space. 

Written by: Charles Ryan, MD, B.J. Kennedy Chair in Clinical Medical Oncology, Director and Professor of Medicine, Division of Hematology, Oncology and Transplantation, University of Minnesota, Minneapolis, Minnesota

Published Date: August 24th, 2017