Read the Full Video Transcript

Niven Mehra: My presentation is on when to do tumor genomic profiling in advanced prostate cancer and what. So these are my disclosures.

So let's start with the guidelines—who to test, when, and what. So the EAU guidelines have been recently updated, and they are quite clear. So metastatic prostate cancer should be tested early on, preferably before onset of first-line mCRPC therapy. What to test? HRR defects, mismatch repair defects, MSI-H.

The ESMO guidelines are still from 2020. They are currently being updated. And no clear recommendations in the 2020 guidelines on early versus late testing. The genomic genes are the same as in the EAU.

And the NCCN guidelines are more clear. So test patients with metastatic prostate cancer on HRR defects, including BRCA. You can also test localized prostate cancer. But mCRPC patients should be tested for MSI, mismatch repair deficiency, and high TMB.

In the APCCC 2022 conference, we queried if you should also test for other alterations, such as PTEN, RB1, TP53, or SPOP, but no clear recommendations were seen for intensification in low-volume mHSPC if you had compound mutations in one of these three tumor suppressor genes. And also, in high-volume mHSPC, I don't think there was a clear altered treatment sequence. Also, all patients already are intensified with triplet therapy. Most of the people voted also in this, if you had a compound mutation, to intensify.

So which tissue to test on? We have many options in metachronous or synchronous disease. We can test on biopsies, which can be random, can be targeted. We can do radical prostatectomies, metastatic lymph nodes. We can do fresh biopsies in CRPC, and we can do liquid.

So the guidelines are quite clear. They say, use metastatic tissue over local tissue, over primary tissue. And only actually if you cannot do this, do liquid. And especially the NCCN guidelines is the most strict. They say, do not use liquid unless it’s unsafe or unfeasible to do biopsies. So the APCCC has not queried this in the 2022 or 2024 questions. But I think, probably, it will be in the diagnostic APCCC next year.

So retesting. The NCCN guidelines are the only guidelines which discuss retesting. And they state that tumor molecular profiles may change with subsequent treatment, and re-evaluation may be considered at time of cancer progression for treatment decision-making.

So how do I treat my patients? How do I do genomic testing influenced also by a lack of unlimited resources for repeat testing? So in my daily management, I infrequently test my patients with mHSPC. This was also seen in APCCC 2022, where it was consensus that genomic test results do not influence, at present, the results for first-line treatment.

Nevertheless, this is rapidly changing with also the onset of molecular trials in the mHSPC setting. And there are some data, but not level 1 evidence, that high-risk mutations are relevant also in metachronous disease. So level 1 data needs to come.

So mCRPC—most of my patients, I test on tissue. I rebiopsy before first-line mCRPC. I do whole-exome sequencing or whole-genome sequencing also to look at more— to less frequent alterations.

So why do we do the rebiopsy in mCRPC? So this is clear. Between different tissues, localized and metastatic, and mHSPC, mCRPC, we see enrichment in genes such as AR, of course, but also other tumor suppressor genes and HRR genes. On the left, that was performed in whole-genome sequencing, but also, on the right, by the study by Wassim Abida, that you can see enrichment across disease states.

What I also want to point out is that in localized disease, there’s intratumoral heterogeneity. So from the paper from Espiritu in 2018, we already knew that 60% of local tumors have some degree of intratumoral heterogeneity. But a very recent paper from Alexander Wyatt’s group and Piet Ost really nicely showed an mHSPC in 43 patients that one in five patients in that study had some degree of intratumoral heterogeneity.

What you see on the left side is that there is a truncal mutation in TP53, but subclonal mutations in PTEN and, on the right side, a patient with a germline BRCA mutation with a second hit, which is truncal, but many subclonal events in TP53 and PTEN. And all of these clones can metastasize to the metastatic niche. So we do not know which of these clones give rise to lethal mCRPC tissue. So that’s another reason for using mCRPC tissue for genomics.

Liquid biopsies are being used, particularly in the US. So I think liquid is a great promise, but there are also some pitfalls to be aware of in that we have false negatives due to low ctDNA fraction. Particularly under 20% to 30% ctDNA fraction, we are not detecting LOH, so not looking at this inactivation of the second allele and focal or larger biallelic inactivation.

And ctDNA fractions are still quite low. Especially in mHSPC low-volume metachronous disease, it’s sometimes undetectable. High-volume mHSPC or CRPC, we can predominantly detect ctDNA. But still, the median fractions are below the 20%, meaning that more than 50% of our patients, we cannot do comprehensive genomic analysis.

And lastly, also, we have false positives due to CHiP. So over 40%, 50% of our patients who are 70 and older have some degree of clonal hematopoiesis of indeterminate potential genes, which can give rise to false positives.

False negatives are very common. So this is a nice study published last year in the PROfound study, looking at matched liquid and tissue for BRCA1, BRCA2, and ATM, showing that, overall, there is a quite nice level of concordance. But what we are missing, and what you can see also highlighted on the right side, is that the biallelic inactivation of BRCA2 is not detected in the majority of patients, where concordance is only 27%. So be aware of what you’re missing.

So let’s go over the most important genes. So I think BRCA1 and BRCA2, these are highly concordant if you use high tumor fractions. So I think this is an early event for mismatch repair deficiency or MSI. Predominantly, these are early truncal events, but there are some cases of intratumoral heterogeneity when we’re missing these events and maybe evolution.

So if you look at the whole spectrum of genes, I think that there’s a lot of data out there. I think for some pathways or some genes, there are early drivers. But there is some evidence for retesting in certain pathways, and there is certainly evolution to be aware of. So it depends— if you retest, what are you looking for?

So liquid biopsies are easy to use. They are predominantly being used in the US at this moment. So some patients are being tested predominantly on liquid and not on tissue. So what data do we have from this series? So let me highlight one study, which was recently published. Sixteen hundred patients were queried from the PROMISE database, and 144 patients were identified with serial testing. They were queried for these actionable alterations published in JCO Precision Medicine.

What you can see is that the first NGS test, which was 70% in mHSPC patients, was predominantly done with liquid biopsies— so limited tissue. The second, third, fourth, et cetera, was predominantly liquid.

Results: so in the second test, in 11% of patients, new results were found. And also, in the third, fourth, and fifth tests, we found relevant mutations in BRCA genes, non-BRCA genes, MSI, but also high TMB. Predominantly, the high TMB is interesting. Nine out of 11 patients with high TMB were detected in serial testing. Are these actionable? I’m not sure. I think this trial, which is a ctDNA precision medicine trial where patients are retested upon progression and rerandomized, may tell a little bit about the actionability.

So let me wrap up. I think tissue is still the issue, as is also in the guidelines. So if you test on the right tissue, up to 90%—we have a successful report, and successful meaning more than 30% tumor fraction.

Retesting, I think, is relevant, especially in low tumor fractions below 20%. Use fresh biopsies. Otherwise, liquid with high tumor ctDNA fractions. And repeat the test— I’m not sure yet. I think we need more data also on actionability.

So thank you. I made it before the cowbell. So thank you for your attention.