Dr. Agarwal on genomic testing, AI, and other developments in prostate cancer

In a recent interview with our sister site CancerNetwork, Neeraj Agarwal, MD, discussed the significance of genomic testing, the emergence of artificial intelligence (AI), and the evolving treatment paradigm in prostate cancer.

Neeraj Agarwal, MD

Agarwal is a professor of medicine, presidential endowed chair of cancer research, and director of both the Genitourinary Oncology (GU) Program and the Center of Investigational Therapeutics at the Huntsman Cancer Institute (HCI) of the University of Utah.

What factors do you take into consideration when determining which patients should receive genetic testing?

All patients with metastatic prostate cancer should undergo 2 types of testing as far as genomic testing is concerned. Number one is germline testing, where patients undergo testing to look for one of the mutations present in their germline DNA, which means they have inherited this mutation from their parents and may share these mutations with their siblings and their children. If they have germline mutations in HRR pathways, they are a candidate for treatment with PARP inhibitors.

The second type of testing is tumor genomic testing or somatic testing. There are multiple reasons why we should be doing the tumor testing for those genomic mutations. This tumor testing can be solid tumor testing, meaning you obtain the biopsy or the tumor tissue and order a comprehensive genomic profiling, or you can even do with circulating tumor DNA.

The reason we need to do that is because we have PARP inhibitors either as monotherapy or in combination with AR pathway inhibitors approved for patients who have HRR mutations. But this testing has implications beyond HRR mutation testing. We know there are several trials going on right now that are recruiting patients based specifically on certain types of mutations. We also know that many of these mutations—especially if you have RB, PTEN, or p53 loss—cause these patients to have a more aggressive clinical phenotype of prostate cancer, and they may require more close monitoring as far as how often we see them in the clinic.

We also know that if patients have all 3 of these mutations, they may have an even more aggressive type of prostate cancer, and they are more likely to have what we call de-differentiated prostate cancer, where [the disease], after a certain time of being treated with conventional therapies, develop phenotypes which comprise a non-secretory [subtype]. In this case, there is not a sufficient level of prostate-specific antigen [PSA] to reflect the tumor burden. In those patients, we cannot rely on PSA monitoring. These patients will require more frequent imaging studies. These patients are also more likely to not respond to standard AR pathway inhibitor therapies. If you look at the first-line, metastatic CRPC setting, enzalutamide is usually able to control the disease progression for 20 months if you look at median progression-free survival [PFS]. In patients who have many of these mutations, enzalutamide or novel hormonal therapies may not be as effective. They may be failed by these treatment options earlier. That may provide the rationale for offering clinical trials to these patients or offering combination therapies such as enzalutamide plus talazoparib if they have HRR mutations.

As I said, tumor genomic testing and germline testing do not just have implications on their family members and how we select therapies for these patients. These testing results have other implications on how we monitor these patients, how often we should do scans, and how we counsel these patients about their survival? These tests, germline and tumor somatic testing, are extremely important and can no longer be ignored in the clinic when we see these patients.

What are your thoughts about the current use and potential future applications of AI in prostate cancer?

AI is becoming more and more important. With each passing month, we see new papers in different cancer settings, including in prostate cancer, which continue to highlight the importance of utilizing AI in improving our diagnosis of these patients and improving prognostication. In the near future, we will also be using AI in frequent selection of these patients.

I don’t think AI-based tools are going to replace clinicians. However, clinicians who do not use AI-based tools are going to be replaced by clinicians who are using AI-based tools. That’s how important AI will be in our clinics in the very near future. We have already seen some wonderful studies being done in the localized prostate cancer setting in the context of radiation therapy.Multiple teams have reported that AI-based tools, which read the pathology slides, were able to provide a better prognostication than the conventional algorithms that used tumor and patient characteristics as far as recurrence is concerned. I think this is just the beginning, and we will see AI playing a much bigger role in treatment selection, diagnosis, prognostication, and drug development.

What treatments are on the horizon that clinicians should be aware of?

I remain optimistic about many options that are going to be approved for patients with prostate cancer in the very near future. The drug development in prostate cancer has been unprecedented. We have never seen so much momentum in drug development in prostate cancer in the recent past as we’re seeing now.

I don’t think we are far away from a day when we will be able to assure our patients that they will not die of their prostate cancer; they will die with their prostate cancer. Some of the therapies I would like to highlight, which I think are close to being approved, include AR-targeted therapies—novel ways to target AR. I remain optimistic about the product technology, the AR degraders, and the antibody-drug conjugates that are using novel ways to deliver chemotherapy in prostate cancer. I also remain optimistic about immunotherapy beyond immune checkpoint inhibitors such as CAR T cells and bispecific T cells that are being developed by different companies. B7-H3 targeting is another novel tool that I hope will see some more momentum in the phase 3 trials.

We also know that the ¹⁷⁷Lu-PSMA-617 [Pluvicto] approval¹ was just the start of delivering radionuclide therapy in precise fashion in the prostate cancer cells. We are going to see an antibody, rosopatamab [J591] that will deliver ¹⁷⁷Lu-PSMA-617 by utilizing an antibody to deliver the beta particles into the prostate cancer cells. Antibodies can carry a bigger payload of radiation particles, so hopefully we’ll use fewer doses to achieve the same results.

Reference

1. Novartis Pluvicto approved by FDA as first targeted radioligand therapy for treatment of progressive, PSMA positive metastatic castration-resistant prostate cancer. News release. Novartis. March 23, 2022. Accessed September 27, 2023. https://bit.ly/36lu4Ct