Decipher use in practice

[NotMattSpraker]

Hi all, searched the forum and didnt find a lot of direct discussion around this question. Would greatly appreciate your thoughts!

How do you all use Decipher for prostate in your practice, if at all. If you are using it for intermediate risk decision making (i.e. decision making for ADT or AS), are you doing this in favorable and unfavorable? Im curious to hear patterns of practice.

---

Members don't see this ad.

 

[medgator]

Most useful in favorable int for determining lupron or low risk for determining tx vs as. Also helpful for that pt who is reluctant to consider lupron at all

 

[Mandelin Rain]

Mainly to determine how long I give enzalutamide treatment during active surveillance. 2 years vs at time of bankruptcy?

 

[NotMattSpraker]

Most useful in favorable int for determining lupron or low risk for determining tx vs as. Also helpful for that pt who is reluctant to consider lupron at all

Awesome, my main question was actually use in favorable int so thanks for that. The literature seems to kind of just lump int risk all together so I wasnt sure if people were using it for ADT determination in those favorable patients.

In these patients, are you holding radiation until the results come back to start concurrently or going ahead and adding ADT once the results are in?

 

[medgator]

Awesome, my main question was actually use in favorable int so thanks for that. The literature seems to kind of just lump int risk all together so I wasnt sure if people were using it for ADT determination in those favorable patients.

In these patients, are you holding radiation until the results come back to start concurrently or going ahead and adding ADT once the results are in?

The latter.... Some data came out within the last couple of years regarding timing saying we can start them together

 

[cmw021]

Anyone using in the adjuvant/salvage setting?

DEFINE_ME

Or using to help guide ADT decision making in the salvage setting?

 

[communitydoc13]

Anyone using in the adjuvant/salvage setting?

DEFINE_ME

Or using to help guide ADT decision making in the salvage setting?

Not sure how this publication helps determine how to use the Decipher test? Is it better than using traditional risk factors or nomograms? Predictive value regarding biochemical failure trends with Gleason score very well.

Are you looking for discordance with traditional risk factors? Gleason 9 but low decipher and you are not adding ADT? Gleason 6 but higher decipher are you adding it?

 

[DoctwoB]

I haven’t really found that niche for genomics in prostate cancer.

The pre biopsy tests are as expensive and equivalent or worse to MRI, which also adds targeting information for biopsy, surgical planning, and node evaluation.

Post biopsy tests have prognostic data, but to my knowledge nothing shown to improve outcomes when compared to a multi factorial nomogram like UCSF-CAPRA. I occasionally get them in patients with low volume GG2 or very high volume GG1 (like all cores positive or with a PIRADs 5 lesion) considering surveillance, but even then I doubt it changes management much.

Post RP again there is prognostic information (better then nomograms?) but nothing I’ve seen that shows it would be better then monitoring and early salvage if needed.

Honestly the one I’m most excited about is urine RNA tests for prostate cancer screening, of which this is the one I’m following most closely. Initial data suggests both high sensitivity and specificity for clinically significant prostate cancers And even GG3 and above vs GG2 and below. This is in a pre-bx setting so presumably in patients with abnormal PSA/DRE, would need to be evaluated in a lower pre test probability/screening setting.

American Urological Association

Just now commercially available under the name MIR Sentinel

 

[communitydoc13]

Post biopsy tests have prognostic data, but to my knowledge nothing shown to improve outcomes when compared to a multi factorial nomogram like UCSF-CAPRA.

I feel like this is the rub with all genomic testing for risk stratification at this point. What I see are papers indicating concordance with outcomes and predictive power but very little data to indicate whether the genomic test or traditional risk assessment or a fairly refined nomogram are better in the cases of non-concordance. (This is admittedly hard to do).

It feels to me like there is a little of the "genomic medicine is the better medicine" assumption when valuing these products. In principle, particularly when talking about very multifactorial interactions of genes, environment and behaviors, a given genotype may not be more predictive than a given phenotype.

A Comparison of Local Recurrence Risk Estimates After Breast-Conserving Surgery for DCIS: DCIS Nomogram Vs Refined Oncotype DX Breast DCIS Score™

A DCIS Nomogram, integrating 10 clinicopathologic/treatment factors, and a Refined DCIS Score (RDS), incorporating a genomic assay and 3 clinicopathologic factors (Oncotype DX DCIS Score™), are available to estimate DCIS 10-year local recurrence ...

www.ncbi.nlm.nih.gov

 

[BobbyHeenan]

Aside - it drives me nuts when the pathologist at our GU conference goes over the decipher chart that "predicts" response to different agents.

is that stuff validated in any way?
LIke a high risk case where surgery vs. radiation is discussed and path is like "well the decipher says not great radiation response."

 

[Doctorer]

Aside - it drives me nuts when the pathologist at our GU conference goes over the decipher chart that "predicts" response to different agents.

is that stuff validated in any way?
LIke a high risk case where surgery vs. radiation is discussed and path is like "well the decipher says not great radiation response."

Yup, this is my feeling on many of the new genetic tests as well. I feel they're being increasingly used to make treatment decisions that they're in no way validated for. People seem to love ordering them though!

 

[RSAOaky]

Hi all, searched the forum and didnt find a lot of direct discussion around this question. Would greatly appreciate your thoughts!

How do you all use Decipher for prostate in your practice, if at all. If you are using it for intermediate risk decision making (i.e. decision making for ADT or AS), are you doing this in favorable and unfavorable? Im curious to hear patterns of practice.

The two major NRG studies open right now evaluating DECIPHER are GU-009 and GU-010. I'm sure there are other studies as well. I've put the study schemas below. So if patient is willing to enroll in these trials and you offer/believe in them, that's a good approach. I definitely wouldn't escalate someone's treatment off trial, nor do I offer de-escalation, but I think a DECIPHER can be useful for a patient that doesn't know what they want and is a marginal case.

Some Examples:
Unfavorable intermediate risk:
1) Gleason 3+4 in 3/12 cores, PSA 11
2) Gleason 3+3 in 6/12 cores (high volume), PSA 7
3) Gleason 3+3 in 4 cores, 4+3 in 5% of a single core, PSA of 7
These are 3 different situations where patients are "unfavorable," but on the cusp. I mean, in #1, does something magically happen when the PSA goes from 9 to 11? In #2, does something magically happen with 6 cores instead of 5? In #3, is a single small core of "unfavorable" intermediate risk disease really enough to commit someone to ADT?

I think this is the value of genomic testing. Now, to be clear, all of the above patients I not only offer, but recommend ADT to. However, if it's a borderline "unfavorable" case and the patient tells me that they REALLY don't want hormones, I would give them the following spiel. I'd tell them it is absolutely what is recommended, I'd show them the NCCN guidelines, tell them it's standard of care, etc., and say that if it's really important to them then let's get a tiebreaker. If DECIPHER is high, then we absolutely should do ADT. If it's low, then we STILL should do ADT, but it's possible that there's less of a benefit.


GU-009

GU-010

 

[BobbyHeenan]

The two major NRG studies open right now evaluating DECIPHER are GU-009 and GU-010. I'm sure there are other studies as well. I've put the study schemas below. So if patient is willing to enroll in these trials and you offer/believe in them, that's a good approach. I definitely wouldn't escalate someone's treatment off trial, nor do I offer de-escalation, but I think a DECIPHER can be useful for a patient that doesn't know what they want and is a marginal case.

Some Examples:
Unfavorable intermediate risk:
1) Gleason 3+4 in 3/12 cores, PSA 11
2) Gleason 3+3 in 6/12 cores (high volume), PSA 7
3) Gleason 3+3 in 4 cores, 4+3 in 5% of a single core, PSA of 7
These are 3 different situations where patients are "unfavorable," but on the cusp. I mean, in #1, does something magically happen when the PSA goes from 9 to 11? In #2, does something magically happen with 6 cores instead of 5? In #3, is a single small core of "unfavorable" intermediate risk disease really enough to commit someone to ADT?

I think this is the value of genomic testing. Now, to be clear, all of the above patients I not only offer, but recommend ADT to. However, if it's a borderline "unfavorable" case and the patient tells me that they REALLY don't want hormones, I would give them the following spiel. I'd tell them it is absolutely what is recommended, I'd show them the NCCN guidelines, tell them it's standard of care, etc., and say that if it's really important to them then let's get a tiebreaker. If DECIPHER is high, then we absolutely should do ADT. If it's low, then we STILL should do ADT, but it's possible that there's less of a benefit.


GU-009

View attachment 359020

GU-010
View attachment 359021

Good post. I have one of those trials open.

I might be wrong here, but 6/12 cores (assuming all cores are G3+3/GG1) I would say is still low risk like example two you gave. My reading of NCCN risk stratification is that to be in int risk group you have to have at least one int risk factor (the cores % positive is a sub classification you can use only if you have G7, PSA>10, or cT2B, T2c.)

 

[SneakyBooger]

I do not use Decipher.

DCISionRT needs a similar evaluation as the NRG is pursuing with Decipher.

 

[madchemist89]

Good post. I have one of those trials open.

I might be wrong here, but 6/12 cores (assuming all cores are G3+3/GG1) I would say is still low risk like example two you gave. My reading of NCCN risk stratification is that to be in int risk group you have to have at least one int risk factor (the cores % positive is a sub classification you can use only if you have G7, PSA>10, or cT2B, T2c.)

I agree with this.

 

[RSAOaky]

Good post. I have one of those trials open.

I might be wrong here, but 6/12 cores (assuming all cores are G3+3/GG1) I would say is still low risk like example two you gave. My reading of NCCN risk stratification is that to be in int risk group you have to have at least one int risk factor (the cores % positive is a sub classification you can use only if you have G7, PSA>10, or cT2B, T2c.)

You are 100% right, thanks for correcting me!

 

[communitydoc13]

The two major NRG studies open right now evaluating DECIPHER are GU-009 and GU-010. I'm sure there are other studies as well. I've put the study schemas below. So if patient is willing to enroll in these trials and you offer/believe in them, that's a good approach. I definitely wouldn't escalate someone's treatment off trial, nor do I offer de-escalation, but I think a DECIPHER can be useful for a patient that doesn't know what they want and is a marginal case.

Some Examples:
Unfavorable intermediate risk:
1) Gleason 3+4 in 3/12 cores, PSA 11
2) Gleason 3+3 in 6/12 cores (high volume), PSA 7
3) Gleason 3+3 in 4 cores, 4+3 in 5% of a single core, PSA of 7
These are 3 different situations where patients are "unfavorable," but on the cusp. I mean, in #1, does something magically happen when the PSA goes from 9 to 11? In #2, does something magically happen with 6 cores instead of 5? In #3, is a single small core of "unfavorable" intermediate risk disease really enough to commit someone to ADT?

I think this is the value of genomic testing. Now, to be clear, all of the above patients I not only offer, but recommend ADT to. However, if it's a borderline "unfavorable" case and the patient tells me that they REALLY don't want hormones, I would give them the following spiel. I'd tell them it is absolutely what is recommended, I'd show them the NCCN guidelines, tell them it's standard of care, etc., and say that if it's really important to them then let's get a tiebreaker. If DECIPHER is high, then we absolutely should do ADT. If it's low, then we STILL should do ADT, but it's possible that there's less of a benefit.


GU-009

View attachment 359020

GU-010
View attachment 359021

I'm familiar with both of these trials and the design absolutely drives me crazy.

If I recall correctly, in it's first incarnation GU-009 included 2 drugs for the therapeutic escalation arm (abi and apa). They then revised to a single drug, picking the more expensive agent of course. This in a space where we are unclear "how low to go" with abiraterone (now a cheap drug) for meaningful clinical benefit.

In addition, neither of these trials studies the efficacy of Decipher. It is a sort of "we will establish Decipher as SOC by adding it's inclusion in these trials" strategy.

What possible sort of results in these trials could either validate or invalidate the use of Decipher?

 

[RSAOaky]

In addition, neither of these trials studies the efficacy of Decipher. It is a sort of "we will establish Decipher as SOC by adding it's inclusion in these trials" strategy.

What possible sort of results in these trials could either validate or invalidate the use of Decipher?

I'm not scientist, but I do think it could validate the use.

Consider the 4 patient groups:
1) Low Decipher -> RT alone
2) Low Decipher -> Receives SOC (ADT + RT)
3) High Decipher -> Receives SOC (ADT + RT)
4) High Decipher -> Receives Escalation (ADT + RT + api/aba)

If group 2 has an inferior prognosis to group 3, that's direct validation of Decipher. It says that two groups with the same treatment and the same historical risk assessment do differently based on their genomic score. If 2 = 3 though, I think it will get muddy to interpret!

 

[communitydoc13]

I'm not scientist, but I do think it could validate the use.

Consider the 4 patient groups:
1) Low Decipher -> RT alone
2) Low Decipher -> Receives SOC (ADT + RT)
3) High Decipher -> Receives SOC (ADT + RT)
4) High Decipher -> Receives Escalation (ADT + RT + api/aba)

If group 2 has an inferior prognosis to group 3, that's direct validation of Decipher. It says that two groups with the same treatment and the same historical risk assessment do differently based on their genomic score. If 2 = 3 though, I think it will get muddy to interpret!

You're right, there is some validation (of decipher as a prognostic tool) if group 2 does better than group 3, although it hinges on these groups being evenly balanced in terms of the risk factors that qualify them for their risk grouping. (presumably this will occur with ~ 500 pts per arm). But, we already know decipher is a decent prognostic tool.

What bothers me is that many possible outcomes leave open questions on best care and do not validate decipher as a clinical decision making tool.

1. If Low decipher benefits from ADT and high decipher benefits from adding apa, do we know that low decipher won't benefit from adding apa?
2. If low decipher doesn't benefit from ADT but high decipher benefits from adding apa, do we know that low decipher won't benefit from adding apa?
3. If high decipher doesn't benefit from apa, do we know that low decipher won't? (It's intuitive that it won't but that's another thing).

I guess we can comfortably exclude ADT in a subset of traditionally unfavorable intermediate risk patients who have low decipher if the de-escalation arm works out with equivalent outcomes.

 

[evilbooyaa]

I do not use Decipher. If I had access to the trials evaluating the utility of them I would consider it, but I do not see any benefit of using it off protocol at the current time...