I note that current NCCN guidelines will admit PSMAs for as low as ~8-10% incidence.
As someone who interprets both, it has to be more sensitive because of the characteristics of the test. You can have very avid subcentimeter nodes that would never be called on CT. Happens all the time. But whether they are true or false positive is very hard to know. I’m inclined to say they are true positive, but that’s my gestalt as a rad.
It's not so much CT and bone scan per se, and why I was harping on N+; I had MRI in mind. IIRC I have seen data that the sensitivity for MRI can be better than PSMA for nodes. I think PSMA will be felt as superior for bone staging in long run.
That's interesting... I never really think of MRI as being the go-to modality for LN, except for maybe in the mesorectum. Then again, it's been a few since I treated a prostate
This is Ga68, which is not F18-PSMA
jamanetwork.com
The inquiry was not specific. PSMA vs CT, bone scanThis is Ga68, which is not F18-PSMA
Diagnostic Accuracy of 68Ga-PSMA-11 PET for Pelvic Nodal Metastasis Detection
This single-arm diagnostic imaging trial assesses the accuracy of 68Ga-PSMA-11 positron emission tomographic imaging in detecting pelvic nodal metastases compared with histopathology.
But, PPV was 74%-88% depending on which patients (all or the reasonable subsets they evaluated). Meaning of 100 PET avid lymph nodes, 74-88 of them were deemed 'real' at time of surgery.
ejnmmires.springeropen.com
Because there are so many candidate PSMA tracers, you have to be specific in which ligand/radiopharmaceutical pair you are talking about. They all have slightly different biodistributions which have impacts on diagnostic performance.The inquiry was not specific. PSMA vs CT, bone scan
As far as comparing Ga68 and F18 a few papers are in this thread
Intraindividual comparison of [68 Ga]-Ga-PSMA-11 and [18F]-F-PSMA-1007 in prostate cancer patients: a retrospective single-center analysis - EJNMMI Research
Background The analysis aimed to compare the radiotracers [68Ga]-Ga-PSMA-11 and [18F]-F-PSMA-1007 intraindividually in terms of malignant lesions, mi(molecular-imaging)TNM staging and presumable unspecific lesions retrospectively as used in routine clinical practice. Methods A retrospective...
From the conclusion
"Since it seems at least challenging for most nuclear medicine departments to provide both [18F]-F-PSMA-1007 and [68Ga]-Ga-PSMA-11, it appears reasonable to choose the PSMA radiotracer depending on local availability with attention to the greater occurrence of nonspecific bone findings with [18F]-F-PSMA-1007."
I think most of the early reports of MDT were with Ga68 but it seems clear that F18 is more likely to have false-positives especially in bone.
pubmed.ncbi.nlm.nih.gov
RT to prostate for T2a, Gl 3+3 in 1 out of 12 cores, PSA 6, s/p lung metastectomy? What says the group?
(Deescalated) active surveillance
Absolutely I would have agreed with you in that setting but I didn’t mean NCCN low risk. I meant relatively low risk in the salvage setting with a biochemical recurrence (initially IR, long disease free interval, PSA 0.5 with doubling time right at a year). After the metastatectomy PSA was undetectable so I never treated the prostate bed. Still following. I do think the PET helped here. We would have seen the met on a CT (if we imaged the chest) but it was small (maybe 6 mm) and he didn’t have any recent comparisons. It would have fallen into the indeterminate category for sure. But as I said above, this is the minority. I have far more “it’s probably nothing” or benign biopsies than this one game changing case.
You know the amazing thing about PSMA PET? Despite everything written, it is still better than anything else for staging
Agree. Bone scan and ct are terrible.
Axumin isn't bad and is more available than psma in my area
Yes the reimbursement side for the tracer is very dicey, probably why I've seen a lot fewer centers doing it now compared to a few years ago
The problem again is not the data. It is how the data is used.
Perhaps. Alternatively, a more accurate test will allow us to stage better, setting the stage for new more effective treatment paradigms. The transition may be tricky, but in the long run, knowing more is better.
with a LN sens/spec of 40%/95%, and a ~10% risk of LN positivity w/ e.g. T3a Gl3+4 PSA 5 low-decipher prostate cancer pre-RP, I am calc'ing that this +PSMA PET could have a positive predictive value worse than a coin flip. (Coin flips are cheaper than PSMA scans I think?)
You *can* sure, lol, I just did. You have to make some best guess as to the pre-test probability here. I believe the incidence of true pLN+ in post-RP PSA<0.2 and basically a Gleason 3+4 w/ low Decipher and pT3a at RP would be in the 10% range. Others are welcome to make a higher estimate which will improve the PPV.
C'mon man... can't conflate pre-treatment nomogram with a recurrence 2-3 years after surgery..with a LN sens/spec of 40%/95%, and a ~10% risk of LN positivity w/ e.g. T3a Gl3+4 PSA 5 low-decipher prostate cancer pre-RP, I am calc'ing that this +PSMA PET could have a positive predictive value worse than a coin flip. (Coin flips are cheaper than PSMA scans I think?)
What is the risk of LN+ in a patient like this in the PSA recurrence <0.2 situation. I had to make some guess! Actually nomographically it's in the ~4% range pre-RP so I thought doubling the risk to ~10% was reasonable. But as I said I wanted some other guess but I didn't know where to look exactly. Dr. Collins has a patient with a rising PSA, but as mentioned it's in the <0.2 range (which I think is important), and it seems he's going to wholly rely on the PSMA scan to offer salvage RT. I.e., in Dr Collins' mind, and for the patient's treatment, the LN+ PPV for this scan was 100%. I know it's not 100%, though, and would like to know the probability people like Dr Collins (and I, once I start PSMA scanning) are making the wrong tx decision.
If we can’t make some educated guess about what his pre test LN+ probability is, a positive PSMA PET is almost uninterpretable.I do not know if data exists for BCRPC sensitivity and speciality. The closest I can find is this study which uses disease detection rate in men with biochemically recurrent disease and negative conventional imaging.
I don’t think it logically makes sense to assume the same percentage of nodal metastasis at baseline vs recurrence. This guy has maybe recurred (assuming PSA is actually elevated). Now he’s in the 10% population which presumably had metastatic disease at baseline.
Diagnostic Performance of 18F-DCFPyL-PET/CT in Men with Biochemically Recurrent Prostate Cancer: Results from the CONDOR Phase III, Multicenter Study
Purpose: Current FDA-approved imaging modalities are inadequate for localizing prostate cancer biochemical recurrence (BCR). 18F-DCFPyL is a highly selective, small-molecule prostate-specific membrane antigen–targeted PET radiotracer. CONDOR was a prospective study designed to determine the...clincancerres.aacrjournals.org
Agree but the numbers you are using make you sound like Evicore.
Mathematically none of this answered the question. I want to find mets, obviously. No test is perfect. If the pretest probabilities are infaust for a PSMA scan, and I think they might be for this patient, then the NPV of an MRI will be greater than the PPV of a PSMA scan in this scenario. For any diagnostic test, patient selection is key. Was this patient properly selected. Still no one has given one speck of a data based guess about the pretest LN+ here. It’s the KEY issue. Is a 10% isolated single LN+ pretest probability THAT bad of a guess for a guy with primary Gleason 3 and a low decipher and a sub 0.2 PSA.
Mathematically the data you want doesn't exist. Unless there is a nomogram for biochemically recurrent patients X years after definitive therapy stratified by tumor, Gleason, etc, it is unknown what the incidence is.
Well that’s something at least. No one has said that out loud; I don’t know how to use a test clinically unless I have an idea of its PPV and NPV. And to not be wordy I’ll just say I want those for Bayesian purposes. I don’t know how a sensitivity and specificity were calc’d for LN testing via PSMA PET if the (stratified) incidence(s) was unknown but so be it. I guess then the PPV of LN+ by PSMA scan is always … perfect?…
Writing “yes” in green increases the PPV for a PSMA positive LN by how much
There is no data in the space you are looking for, nor will there ever be. We do not have any historical data in a situation of somebody who already had RALP with those exact (or even FIR findings), had elevated PSA suggestive of recurrence, and then did a RPLND to see what percentage of patients now had LN disease. That is not a thing that happened.
I'm causing agita over wondering how likely a positive PSMA scan is truly positive. I had no idea the general consensus would be "Who knows." One of the most important things we do in medicine is when ordering a diagnostic test, have an idea of the likelihood of the condition before the test and an idea of the likelihood of the condition after the test. Many, many (statistically pontificating) articles have been written on this and it's sort of med school 101. I *do not* want to be the rad onc that orders PSMA PETs and sees things, at least not in certain scenarios. If the urologist in my group orders a PSMA PET with the first rise in PSA and sees a spot and tells me "irradiate the spot," whereas that irradiation would not have been considered standard of care without the PSMA PET... yeah, I got questions. I don't want to be inappropriately seduced.
“Why are you so distrustful of PSMA?” If I could have data that would inform me about the PPV of a +PSMA PET I could answer your question. At the end of the day no one is showing any likely reason to disbelieve that the PPV is about 50% here.
The problem is that the "endpoint" to date associated with "research" of these tests have been very weak ones like "changes management decisions", PPV or NPV.
It’s a new test with no clinical track record that doctors are using to determine if and where people have cancer in these gray zone situations and hasn’t been externally validated, and furthermore clinicians appear resigned to call it un-validatable in certain situations versus accepted gold standards. Given these uncertainties, it’s still being used to deviate treatments from the standard of care… and this appears to be its selling point.
Is it “the bottom line”? How many times has a positive PET in the chest for lung cancer borne no fruit pathologically. For me, many times. Or positive PETs under the armpit in breast.
I have an idea of PET sensitivity and specificity for nodes in lung cancer. The quoted sensitivity for nodes in PSMA PET is 40%. That’s not great! It means what you’re testing for should be rather likely (and the specificity needs to be A+ high) before you run the test or you will have a >50% chance of a false positive. People are “anti” PSMA in prostate too; but depends on risk stratification. Looking at “appropriate” vs “inappropriate” for PSMA PET, it can mean the difference between one less positive core in some patients.
No one wants to hear me say this but I will anyways. In a population where you have the god like knowledge of knowing almost everyone has the condition (“known prostate lesions” as you say), the rate of false negatives approaches 1 minus sensitivity; so when sensitivity is low (as it is in PSMA PET), false negatives can be more common than to what we might ordinarily be accustomed.
