Cell-Free DNA Shows Potential in Identifying Transformed Castration-Resistant Neuroendocrine Prostate Cancer

Himisha Beltran, MD, an associate professor of medicine at Lank Center for Genitourinary Oncology and Division of Molecular and Cellular Oncology at the Dana-Farber Cancer Institute and Harvard Medical School

Himisha Beltran, MD

Cell-free DNA (cfDNA) can potentially be utilized as an approach to identify patients with prostate adenocarcinoma that has transformed to castration-resistant neuroendocrine prostate cancer (CRPC-NE), according to results of a study that have been published in the Journal of Clinical Investigation.

In the analysis, researchers conducted whole-exome sequencing and whole-genome bisulfite sequencing of cfDNA and of matched metastatic tumor biopsies from patients with metastatic prostate adenocarcinoma and CRPC-NE. Results showed that CRPC-NE—associated genomic and epigenomic alterations were detectable in the circulation and showed high concordance with matched biopsy.

Serial sampling of patients, along with deeper analysis of subclonal alterations, revealed that CRPC-NE changes could sometimes occur early, even before a clinical manifestation of CRPC-NE. When looking across the genome, there was overall higher concordance between cfDNA and alterations found in tissue biopsy in patients with CRPC-NE versus those with castration-resistant adenocarcinoma. These findings suggest that there is less heterogeneity across metastases within an individual harboring the neuroendocrine subtype. These analyses provide insights into how cancers evolve and might help in the treatment of patients as they progress on treatment, according to lead investigator, Himisha Beltran, MD.

“We can detect neuroendocrine prostate cancer molecular alterations in the blood of patients, including changes in cfDNA methylation,” Beltran explained. “This is paving the way for using a noninvasive biomarker to identify these patients.”

In an interview with OncLive, Beltran, an associate professor of medicine at Lank Center for Genitourinary Oncology and Division of Molecular and Cellular Oncology at the Dana-Farber Cancer Institute and Harvard Medical School, discussed research evaluating cfDNA to identify therapy resistance in prostate cancer.

OncLive: Could you provide some background on CRPC-NE?

Beltran: Most prostate cancers are classified as prostate adenocarcinoma when first diagnosed. As patients are treated with drugs in the advanced setting, a subset of cancers can transform to a small cell neuroendocrine carcinoma. This is typically diagnosed by a metastatic tumor biopsy that shows features that appear more like small cell lung cancer (SCLC) and less like prostate adenocarcinoma. It is important to recognize CRPC-NE, as these patients do not tend to respond well to drugs that target the androgen receptor (AR) and they may not progress with a rising prostate-specific antigen (PSA).

As tumors evolve their phenotype toward CRPC-NE, there are times where you see mixed or overlapping features. This process, sometimes referred to as ‘lineage plasticity,’ is just one mechanism by which prostate cancers can evade some of the drugs that we use, particularly the hormonal-based agents.

How is CRPC-NE typically diagnosed and what are some of the challenges that exist with current therapeutic approaches?

Typically, if we suspect that someone might develop CRPC-NE, based on low-PSA progression, atypical spread, or aggressive clinical features, a biopsy would be performed. The biopsy will be examined via a microscope to look for small cell carcinoma morphology. Immunohistochemical markers for prostate and neuroendocrine markers may help support the diagnosis. We use regimens that are similar to small cell carcinoma or SCLC [to treat these patients] and there are a number of emerging therapeutic targets and trials in development for this specific disease subgroup.

One challenge is that metastatic biopsies are invasive. Given the fact that CRPC-NE tends to evolve from a preexisting adenocarcinoma, there is a concern that there might be heterogeneity and what you are getting is not representative of what is going on within a patient.

What was the rationale to use whole-exome sequencing and whole-genome bisulfite sequencing of cfDNA to detect CRPC-NE features?

Given the challenges of metastatic biopsy, we wanted to study if molecular features in the blood of patients could serve as a noninvasive approach to detect CRPC-NE. In order to identify a blood-based biomarker, we first had to understand what molecular features define CRPC-NE. In prior research, we extensively evaluated tumor biopsies from men with prostate adenocarcinoma and CRPC-NE and identified distinguishing features in both DNA as well as DNA methylation in CRPC-NE. The next step was to see if we could detect these changes the blood using circulating tumor DNA (ctDNA).

This was important to us, as this approach could help avoid a metastatic biopsy or even make the diagnosis of CRPC-NE more precise by not having to rely on pathologic features. That was the rationale of the study; to see if we can detect CRPC-NE molecular features while looking at cfDNA . We also wanted to use cfDNA as a tool to understand how and when CRPC-NE associated changes occur as patients were progressing on therapy.

What was the design of the study?

We evaluated patients with advanced prostate cancer across the disease spectrum—both adenocarcinoma and CRPC-NE. All patients underwent a blood test and a tissue biopsy. We performed whole-exome sequencing and whole-genome bisulfite sequencing of tumors and cfDNA and determined whether what we saw in the blood was reflective of what we see in the biopsy. In general, consistent with a number of other studies, we saw a high concordance. There were a number of shared genomic alterations between adenocarcinoma and CRPC-NE. When we looked at the prognostic value of common prostate cancer alterations between the two subtypes, we did see differences, suggesting that disease context may be important to consider when applying genomic results in the clinic.

There was also information in the blood we cannot get from tumor biopsies, such as being able to identify emergence of differences outside of common aberrations that might reflect subclonal lesions. Analyses of these lesions, as well as their evolutionary patterns on therapy and during progression, provided insights into resistance patterns.

Were there any surprising results?

I was impressed by the robust DNA methylation signal that we observed in cfDNA and how distinct this was in CRPC-NE compared with adenocarcinoma. As opposed to their very similar genomics, suggesting a prostate adenocarcinoma cell of origin of CRPC-NE, the CRPC-NE subtype is characterized by a very different epigenome.

DNA methylation is a way that cancer cells can turn on and off genes. We found that DNA methylation in tumors can not only identify be reliably detected in cfDNA, it seems to be a relatively stable biomarker. Analysis of regions that are differentially methylated in CRPC-NE, combined with transcriptome analysis, is providing new clues into genes and pathways that are dysregulated in CRPC-NE and pointing to new therapeutic targets. This extends what we can learn from the blood of patients beyond traditional ctDNA platforms that are used today.

What should your colleagues take away from this study?

CRPC-NE -associated genomic and epigenomic alterations are detectable in the circulation. This research will hopefully inform testing in future studies to identify when and how often CRPC-NE develops in patients, and for trial selection as we think about sub-segmenting advanced prostate cancer and treating patients based on their underlying molecular features.

Are you excited about any other trials examining the use of cfDNA?

There is a lot of exciting prostate cancer research studying liquid biopsies, both cfDNA and circulating tumor cells, as noninvasive means to look for actionable targets and emerging resistance mechanisms. I am confident that these tests will eventually replace metastatic biopsies, at least for some indications. I look forward to future work in the field that will help refine how best to use these molecular tests for diagnosis, early detection of treatment resistance, disease monitoring, and selection of therapy.