high PSA: follow with MRI vs biopsy

 

A Swedish study found that following men with a high PSA with an MRI seemed to be a reasonable approach for risk-stratifying patients, decreasing false positive results and without leading to significant delays in diagnosis of significantprostate cancer, in the Gothenburg-2 trial, a large, population-based, screening trial (see psa MRI followup NEJM2024 in dropbox, or DOI:10.1056/NEJMoa2406050). Special thanks to Brittany Berk (a really smart urologist and, it just so happens, my daughter-in-law) for some comments and suggestions.....

 

Details:

-- about 80,000 men who were 50 to 60 years of age in Gothenburg (aka “Goteborg”), Sweden were eligible for participation in the trial, from which 38,316 men were randomly selected and invited for screening, a population-based trial that started in 2015

-- those who accepted the invitation were assigned in a 1:1:1 manner to 1 of 3 groups:

    -- those with low PSA levels of 1.8-3.0 ng/mL.  these men were followed but were not targeted in the current report, only the next two categories were systematically evaluated

    -- those with PSA at least 3.0 ng/mL who were in the MRI-targeted biopsy group: 6575 men

    -- those with PSA at least 3.0 ng per mL who were in the systematic biopsy group (ie a prescribed set of multiple biopsies throughout the gland, but not specifically targeted): 6578 men

-- those in the systematic biopsy group underwent systematic biopsies and, if suspicious lesions were found on MRI, an MRI-targeted biopsy

-- those in the MRI-targeted biopsy group had MRI targeted biopsy only if the MRI detected a suspicious abnormality

-- men were invited for repeated screenings at 2, 4, or 8 years later, depending on their PSA level

-- prostate biopsies were categorized per the international Society of Urological Pathology (ISUP)

 

-- the primary outcome: detection of prostate cancer classified as clinically insignificant (ISUP grade 1); detection of clinically significant cancer (ISUP grade at least 2), and the detection of clinically advanced or high risk cancer (ISUP grade 4 or 5)

-- median follow-up of 3.9 years (approximately 26,000 person-years each group)

 

Results:

-- percentage of men who had a PSA level of at least 3 ng/mL in first screening round:

    -- MRI-targeted biopsy group: 6.8%

        -- 2.8% had both an elevated PSA and a suspicious lesion on MRI

    -- systematic biopsy group: 6.9%

        -- per the protocol, all 6.9% had systematic biopsies

 

-- prostate cancer detection rate, in an intention-to-treat analyses:

    -- overall findings:

        -- MRI-targeted biopsy group: 185 of 6575 men (2.8%)

        -- systematic biopsy group: 298 of 6578 men (4.5%)

            -- 38% lower in MRI-targeted group, RR 0.62 (0.52-0.74)

    -- on first screening:

            -- MRI-targeted biopsy group: 108 men (1.6%)

            -- systematic biopsy group: 159 men (2.4%)

                -- 32% lower in the MRI-targeted group, RR 0.68 (0.53-0.87)

    -- on screening round at least 2:

            -- MRI-targeted biopsy group: 60 men (2.7%)

            -- systematic biopsy group: 114 men (5.3%)

                -- 49% lower in the MRI-targeted group, RR 0.51 (0.38-0.70)

 

-- Clinically insignificant cancer rate, in intention-to-treat analyses:

    -- overall findings:

        -- MRI-targeted biopsy group: 68 of 6575 men (1.0%)

        -- systematic biopsy group: 159 of 6578 men (2.4%)

            -- 57% lower in MRI-targeted group, RR 0.43 (0.32-0.57), p<0.001

                -- biopsy was not performed in 20 men in the MRI-targeted group and in 116 men in the systematic biopsy group at their final visit (declined, despite recommendations to get biopsies). multiple imputations of these missing outcomes, a statistical approach to replace missing data with estimated values as if these men were getting the biopsies, actually found an even larger 62% lowering of the findings of clinically insignificant cancers in the MRI-targeted biopsy group, with RR 0.38 (0.29-0.50)

    -- on first screening:

            -- MRI-targeted biopsy group: 39 men (1.6%)

            -- systematic biopsy group: 80 men (2.4%)

                -- 51% lower in the MRI-targeted group, RR 0.49  (0.33-0.71)

    -- on screening round at least 2:

            -- MRI-targeted biopsy group: 18 men (0.8%)

            -- systematic biopsy group: 69 men (3.2%)

                -- 75% lower in the MRI-targeted group, RR 0.25 (0.15-0.42)

  

-- Clinically significant cancer rate in intention-to-treat analyses:

    -- overall findings:

        -- MRI-targeted biopsy group: 117 of 6575 men (1.8%)

        -- systematic biopsy group: 139 of 6578 men (2.1%)

            -- 16% lower in MRI-targeted group, RR 0.84 (0.66-1.07), not statistically significant

    -- on first screening:

            -- MRI-targeted biopsy group: 69 men (1.0%)

            -- systematic biopsy group: 79 men (1.2%)

                -- 13% lower in the MRI-targeted group, RR 0.87 (0.63-1.20), not statistically significant

    -- on screening round at least 2:

            -- MRI-targeted biopsy group: 42 men (1.9%)

            -- systematic biopsy group: 45 men (2.1%)

                -- 8% lower in the MRI-targeted group, RR 0.91 (0.60-1.37)

 

    --Interval cancers in intention-to-treat analyses

        -- any prostate cancer:

            -- MRI-targeted biopsy group: 17/6467 (0.3%)

            -- systematic biopsy group: 25/6419 (0.4%)

                -- lower in MRI-targeted group, RR 0.67 (0.37-1.24), not statistically significant

        -- clinically insignificant cancer:

            -- MRI-targeted group: 11 men (0.2%)

            -- systematic biopsy group: 10 men (0.2%)

                -- higher in MRI-targeted group, RR 1.09 (0.48-2.51), not statistically significant

        -- clinically significant cancer:

            -- MRI-targeted group: 6 men (0.1%)

            -- systematic biopsy group: 15 men (0.2%)

                -- 60% lower in MRI-targeted group, RR 0.40 (0.19-0.99), statistically significant

               

 -- here are a series of graphs of the cumulative incidence of clinically insignificant prostate cancer (A, B, and C) by different definitions of “insignificant”, all showing a significant divergence of the curves over time and favoring the MI-targeted biopsy approach;  and the clinically significant prostate cancers (graphs D, E, and F) with different definitions of the levels of “significant”, all showing large overlap of the confidence intervals and not reaching statistical significance: 

 

  

-- at screening rounds 2 through 4:

    -- MRI-targeted biopsy group: 98 of 2005 men (4.4%) had a biopsy indication at least once

    -- systematic biopsy group: 371 of 2147 men (17.3%) had a biopsy indication at least once

        -- though the screening intervals depended on the PSA levels, however these were similar between the groups

 

-- adherence rate to getting biopsy:

    -- MRI-targeted biopsy group: 176 of 186 men (94.6%) at visit one

        -- 88 of 98 men left front 89.8%) at repeat screenings

    -- systematic biopsy group: 384 of 456 men (84.2%) at visit one

        -- 309 of 371 men (83.3%) at repeat screenings

 

-- analysis of advanced prostate cancer detected (ISUP 5 or advanced) over the rounds of screening (their Table 3), including both screening-detected and interval cancers:

    -- MRI-targeted biopsy group (6575 men):

        -- first screening visit: 3 cases

        -- if PSA <3ng/mL in previous round: 4 cases

        -- if PSA at least 3 ng/mL and MRI-negative, absent, or incomplete in prior round: 0 cases

        -- if PSA at least 3 ng/mL, MRI-positive, absent or negative in prior round: 1 case

    -- systematic biopsy group (6578 men):

        -- first screening visit: 7 cases

        -- if PSA <3ng/mL in previous round: 5 cases

        -- if PSA at least 3 ng/mL and MRI-negative, absent, or incomplete in prior round: 0 cases

        -- if PSA at least 3 ng/mL, MRI-positive, absent or negative in prior round: 2 cases

    -- totals: MRI-targeted biopsy had 8 cases; systematic biopsy had 14 cases (both with small numbers of cases)

 

-- targeted biopsy group: for men who had ISUP grade 1, 63 of 73 men (86.3%) had systematic biopsy on a later occasion, 15 of whom (23.8%) had the cancer upgraded

-- adverse events: 5 men had severe adverse events that led to hospitalization (3 in the systematic biopsy group and 2 in the MRI targeted biopsy group, of which 4 of the 5 events were considered likely to have been related to the biopsy procedure

 

Commentary:

-- prostate cancer is the second most commonly diagnosed cancer globally in men, and the most common one in the US

-- the results of studies on PSA screening finding high PSA and then biopsy, have typically found very high rates of false positive results and mixed results on important long-term outcomes:

    -- the European Randomized Study of Screening for Prostate Cancer (ERPSC) and the GOTEBERG-1 trial, both randomized studies initiated in the 1990s, found a decrease in prostate cancer mortality, though a high risk of prostate cancer overdiagnosis

    -- the Prostate Lung Colorectal and Ovarian trial (PLCO) in the US and the Cluster Randomized Trial of PSA Testing for Prostate Cancer (CAP) in the UK also had PSA testing and systematic biopsy, and neither study showed prostate cancer mortality benefit, but both showed significant overdiagnosis

    -- the Scandinavian Prostate Cancer Group Number 4 study (SPCG-4) had 18 years of follow-up found greater treatment efficacy for radical prostatectomy than observation in patients with early, localized prostate cancer, though this was not shown in the Prostate Cancer Intervention Versus Observation Trial ( PIVOT) with 22 years of follow-up as well as the Prostate Testing for Cancer and Treatment (ProtecT) trial with 10 years of follow-up

    --  so, overall a mixed bag. And part of the issue may be that the large number of overdiagnosed cases could lead to dilution of the potential benefit for cancer interventions

-- there are a few other studies finding that targeted biopsies of suspected lesions on MRI might reduce the risk of prostate cancer overdiagnosis, and this approach has been found to be non-inferior to systematic prostate biopsies

-- the American Urological Assocation in their 2023 guidelines state that "Clinicians may use MRI prior to initial biopsy to increase the detection of GG2+ prostate cancer. (Conditional Recommendation; Evidence Level: Grade B)", where GG2+ means Gleason Grade 2 or more (which is at least 3+4)

 

-- a systematic review and meta-analysis found that the overall sensitivity of PSA for prostate cancer was 0.93 (0.88-0.96) and specificity was a quite low 0.20 (0.12-0.33), and the hierarchical summary receiver operator characteristic curve was 0.72 (0.68-0.76) in a pooled study of 14,489 patients, using the PSA cutpoint of 4 ng/mL, where a 0.7-0.8 score is considered "acceptable" but only a score>0.8 is considered "excellent",   (https://bmcmedicine.biomedcentral.com/articles/10.1186/s12916-021-02230-y ). A cutpoint of a lower PSA number (eg, 3 ng/mL as in Sweden) would find more cancers but lead to more testing and biopsies (though adding MRI as a first step would decrease the number of biopsies, per the above article) . There is argument in the medical literature that a “normal” PSA should incorporate age in its interpretation, such that for someone 40-49yo it should be 0-2.5 ng/mL, 50-59yo 0-3.5 ng/mL and 4.5 for those 60-69yo (age is a risk factor independent of BPH/prostate size)

 

-- there is a real concern for low sensitivity of PSA screening: the Prostate Cancer Prevention trial of men randomized to finasteride vs placebo found that those on finasteride had about a 30% decreased risk of developing prostate cancer, entirely for low-grade cancers

    -- BUT, at the end of the above 7-year study, they did a prostate biopsy in 3K men in the placebo group (with informed consent….) who never had PSA >4 or abnormal DRE, with the rather striking finding that 15.2% of the placebo group had prostate cancer (see prostate ca in men with psa under 4 nejm 2004 in dropbox, or  N Engl J Med 2004;350:2239-46 or https://gmodestmedblogs.blogspot.com/2013/10/low-psa-but-prostate-cancer.html ): 

    --for the15.2% who had prostate cancer, breakdown as follows:

        --PSA <0.5 ng/mL, 6.6% had prostate cancer (!!!)

        --PSA 0.6-1 ng/mL,, 10%

        --PSA  1.1-2 ng/mL,, 17%

        --PSA 2.1-3 ng/mL,, 24%

        --PSA 3.1-4 ng/mL,, 27%

            – AND, of those with prostate cancer, 14.9% had high grade cancers (Gleason 7-10):

                -- of those with PSA <0.5 ng/mL, 12.5% had high grade cancer!!!

                -- of those  with PSA 3.1-4, 25% had high-grade lesions!!!

                -- 2 patients with Gleason score of 8 had PSA levels between 0-1ng/mL (one had a PSA of <0.5 ng/mL)

    -- another factor that affects the specificity of PSA is the setting: patients having had ejaculation within the prior 48 hours or so can have a higher PSA (i usually suggest that patients not ejaculate for 4 days just to make sure that the PSA is more reliable, and also repeat an abnormal value which can be pretty significantly different from the initial one). also doing a rectal exam or frequent bike riding can affect PSA results.

 

-- also, the specificity of high PSA is also not so great: BPH, which is very common and increases with age, is the factor most commonly associated with increased PSA levels; prostate inflammation/infection can be associated with very high levels of PSA (up to the 75 ng/mL range) and can remain high for months after treatment of a prostatitis infection

 

-- it is reasonably widely accepted that MRI is a potentially important tool in identifying prostate cancer and that it minimizes overdiagnosis (one of the main obstacles to more systematic PSA screening). But systematic biopsies of men with high PSA levels but negative MRI results have found that 19-28% of these patients have ISUP tumors of at least grade 2. That is why this study was done, since it is really important to know if delaying diagnosis in this substantial portion of patients is safe until the MRI becomes MRI-visible?? 

--This study found several things:

    -- not doing biopsies in men with PSA >3 ng/mL but negative MRIs was associated with a substantial and statistically significant decrease in the detection of clinically insignificant cancers, and there was a very low risk of finding incurable cancers at repeat screening or as interval cancers

        -- ie, there was little evidence that omitting the original pretty intense systematic biopsies (associated with adverse outcomes) and delaying a potential cancer diagnosis and progression was harmful

        -- in fact, if anything, there appeared to be more advanced/high risk cancers (ISUP grade 4 or 5) in the systematic biopsy group, and these are the cancers that often are refractory to any treatment: there were only 5 men in the MRI-targeted group and 7 in the systematic biopsy group who had ISUP 4, out of a total of 26,000 person-years of followup in each group and 5 vs 7 men in the ISUP 5 or advanced group

            -- studies found that systematic biopsies can miss some cancers that are MRI-visible and would be found on MRI-targeted sampling

        -- it seemed that the vast majority of cancers become visible on MRI before they become incurable

    -- interestingly, there were many of the detected cancers in men who had had a PSA <3 ng/ml on the prior screening:

        -- the overall numbers were quite similar for the array of ISUP categories (see their Table 3) between the 2 screening groups, though slightly worse in the systematic biopsy group

        -- does this mean that we should have a lower cutpoint than 3 ng/mL for a concerning PSA value?

             -- the above mentioned Prostate Cancer Prevention trial found advanced cancers in some men with very low PSA levels

             -- but that would mean there would be more clinically insignificant cancers detected, along with the increased testing, biopsies, psychological trauma of a “cancer diagnosis”, and more men who elect to have aggressive therapy (perhaps unnecessarily)

             -- but would the MRI-targeted biopsy approach that decreases the pickup of insignificant cancers compensate for that?

             -- but would a more discriminatory approach to PSA screening help, such as assessing PSA density, which controls for prostate volume in predicting clinically significant prostate cancer (https://www.nature.com/articles/s41598-020-76786-9 ) and has a higher sensitivity (70%) and specificity (79%) than just the PSA by itself, help compensate for using a lower PSA threshold?? Though this test does require transrectal ultrasound assessment of prostate size, taking screening into a more intense, non-primary care setting. And, by the way, one advantage of MRI is that one can also determine prostate volume and thereby calculate teh PSA density

            --for the men with a the most advanced ISUP score and most likely to have incurable cancers:

                -- at first screening, in the MRI-targeted biopsy group, there were 5 such cancers, and 4 of them occurred in men with PSA <3 ng/mL; and 1 occurred in a man with a high PSA but a false negative MRI-guided biopsy (2 were found with PSA>3 and either positive MRI-targeted biopsy)

                -- at first screening, in the systematic biopsy group, there were 7 such cancers, and 5 of them occurred in men with PSA <3 ng/mL; and 2 occurred in a man with a high PSA but a false negative MRI- guided biopsy (2 were found with PSA>3 but none had an MRI-targeted biopsy)

 

    -- there were also striking differences over the cancers found in subsequent screenings: 

        -- the MRI-targeted group: there were many fewer clinically insignificant cancers in subsequent screening rounds

        -- the systematic biopsy group: there were pretty similar numbers of insignificant cancers picked up in subsequent screenings:

            -- perhaps this is related to the slow growth of these ISUP grade 1 cancers, and more of these non-significant cancers being picked-up in the systematic biopsy group

            -- but it should be reinforced that these ISUP grade 1 cancers (the insignificant ones) can progress to bad cancers, and those that are visible on MRI more often need active treatment than those not visible. But these cancers do in general progress slowly (as in this study) and the cumulative incidence of clinically significant cancers detected in each group were not statistically different

 

    -- adherence to biopsy was higher in men in the MRI-targeted group. Unclear why. Perhaps these men were more concerned that the MRI picked up a suspicious lesion (and it was not just a blood test that led to lots of biopsies)??

 

Limitations:

 -- having routine MRIs on patients with high PSAs (which is especially true with a cutpioint of 3 ng/mL as in this study, or even lower ones as might be considered), requires adequate numbers of expensive MRI machines, adequate numbers of well-trained MRI technicians, adequate numbers of well-trained and experienced prostate cancer pathologists (and more of all machines and people if PSA screening thresholds are lowered, or if more men with the current thresholds get PSA screening (a 2018 assessment found that about only 32% of men were screened: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8758274/ ; the National Cncer Institute found decreasing numbers comparing 2005 to 2021, with 37% in 2021: https://progressreport.cancer.gov/detection/prostate_cancer . And more health care dollars spent on this

    -- though, as per this study, there were large benefits of avoiding biopsies with MRIs, with about one-half the adverse events, as well as radiation/surgeries in some men who want any possibility of cancer resolved (and their very common life-long complications of incontinence, impotence, etc), psychological strife from the “cancer” moniker, etc

 -- fewer than half of the men invited to participate in the study were willing to actually participate, potentially leading to a bias, since those men who decided to participate may have been in fundamental ways different from those who did not (perhaps they were more health conscious, perhaps had lifestyle issues or demographics that were different or perhaps they had comorbidities that were different, etc., and this could skew the results and affect their generalizability)

-- this study used a cutpoint for “high” PSA of 3.0 ng/mL.  This is not the standard in the US or many other countries. And this cutpoint will pick up more cancers, but also more false positives. It would have been useful if they calculated their results int his study for the PSA 4-10 ng/mL group

-- the median age of men in this study was 56yo

    -- this young age and short follow-up is problematic in terms of the quality and generalizability:

        -- prostate cancer incidence increases with age (so confining the group to those ages 50-60 with median age of 56 misses lots of men; as does the likelihood of adverse events from the biopsies

        -- and, longer term mortality is somewhat higher in older men: a SEER study of 116,796 men diagnosed with prostate cancer between 1992-1997 and followed 20 years found that the risk of men dying from prostate cancer increased progressively with age at the time of diagnosis: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9454626/

        -- the short-term length of this study (only 3.9 years of follow-up) is also problematic in a cancer that usually is relatives slow-growing over many years

    -- this study was also a single-center Swedish study, and we know of several factors that may increase the risk of prostate cancer, including age, but also diet (high animal fat, low vegetables, foods with low phytoestrogens, high omega-3 fatty acids, high alcohol and low coffee), smoking, obesity, lack of physical activity. Pretty much all of these have some but not totally consistent relationships with prostate cancer risk. But to the extent that any of these do, their prevalence varies dramatically in different areas of the world, so having a study in one area of Sweden may not reflect the results that would occur more globally

    -- this study relied on traditional systematic prostate biopsies. Newer techniques (eg transperineal biopsy, or image-guided fusion) might have produced different results. Would be good to have studies using the newer biopsy methods; they also used multiparametric MRI, though biometric MRI without contrast has some evidence suggesting that might reduce the incidence of false positive results. And transperineal biopsies do seem to decrease the risk of infection associated with the transrectal approach and does not seem to require antibiotics prior to the biopsy (and these are typically pretty aggressive antibiotics, such as fluoroquinolones, with the potential adverse effects of  changing the gut microbiome (and perhaps leading to more C difficile infections (https://journals.asm.org/doi/10.1128/iai.05496-11), antibiotic resistance, and the litany of other adverse effects

 

So, this study found that doing multiparametric MRI in men with PSA values 3-10 ng/mL would reduce the number of prostate cancer biopsies considerably and likely reduce the number of men who have aggressive treatments and the significant risk of permanent adverse effects. In particular, this study raises the possibility that:

-- MRI screening of men with a high PSA, then followed by targeted biopsies, might reduce the incidence of false positive results (ie overdiagnosis) vs proceeding to systematic biopsies

    -- and those who do have significant cancers have been found do have lower Gleason grade cancers and typically could be followed expectantly very closely. And their outcomes in this short study was the same in terms of later identifying significant cancers

--it would be great to have further studies on different populations of men in different countries, using the currently largely accepted PSA cutpoint of 4ng/mL, longer follow-up and perhaps with some of the newer biopsy techniques…

 

geoff

-----------------------------------

If you would like to be on the regular email list for upcoming blogs, please contact me at gmodest@bidmc.harvard.edu

  

to get access to all of the blogs:  go to http://gmodestmedblogs.blogspot.com/ to see the blogs in reverse chronological order

  -- click on 3 parallel lines top left, if you want to see blogs by category, then click on "labels" and choose a category​

  -- or you can just click on the magnifying glass on top right, then type in a name in the search box and get all the blogs with that name in them

 

if you would the article, please email me.

 

please feel free to circulate this to others. also, if you send me their emails (gmodest@bidmc.harvard.edu), i can add them to the list