4Kscore: Blurring the Lines Between Subjective and Objective Assessment

In consultation, you tell your perfectly healthy 57-year-old patient that his new 4Kscore 9 months postbiopsy equals 1.2. As you all know, we get decimals now to split hairs between 2 points. First question of many to be considered: Are decimals truly helpful? Not likely, but that nuance is far from our most important 4Kscore dilemma.

His 4Kscore prior to his recent biopsy a year prior was higher and crossed the border to what we technically define as moderate risk. Specifically, this patient’s 4K had slowly increased 4 times over 5 years from 5 to 7 to 8 and then finally to 9. We realize, however, that low risk, moderate risk, and high risk don’t factor in patient age/longevity. I will discuss this further in my commentary.

Of course the patient seems elated about his new 4K of 1.2, but similarly has done his own reading and has questions.

To give you some background, his PSA over 7 years went from 0.8 to 2.93. His percent free fractions, once applicable based on total PSA, were always somewhat concerning as well and always in the low teen range. He also had 3 MRIs dating back over the same 7 years in which there was first a 5×8 mm Prostate Imaging Reporting & Data System (PI-RADS) 2 lesion, one that subsequently remained PI-RADS 2 but had increased to 13×7 mm and finally 15×7 mm. His prostate consistently measured around 32 g with a small median lobe.

All factors considered, he opted for biopsy and underwent a Koelis technology fusion 20-core transperineal biopsy. Pathology revealed 1 core of high-grade (HG) prostatic intraepithelial neoplasia (PIN) in the targeted lesion and no inflammation.

His new 7- to 8-month postbiopsy PSA is 2.83 (stable) and his percent free PSA is similarly 13%. His first question, looking at his new wonderful 4Kscore of 1.2 is “Doctor, doesn’t this new 4Kscore just represent the fact that you can now check off ‘prior biopsy negative’ in the algorithm? I mean everything else is the same.” “Well technically, you’re not wrong,” replies the urologist. In fact, there are many things that we can consider.

Where are we today with the 4Kscore? Is a 4Kscore often only as accurate as the entered parameters in the algorithm? Surely it is, but it does currently represent the best screening blood test we have available. Do we have the means to further enhance its sensitivity, specificity, and positive and negative predictive value?

I believe it is safe to say that the urology community has been less diligent about performing digital rectal examination (DRE) since the institution of more specific PSA parameters, and of course even more recently since the more liberal use of MRI. Perhaps the necessity of including DRE findings as part of the algorithm in a 4Kscore reinvigorated their performance. So how recent does the DRE need to be for us to check off “no nodule” in the algorithm? We can all agree that nodules don’t grow overnight, but is it 6 months, 9 months, 1 year? Not to be cynical or critical of colleagues, but if we review records of a 4Kscore done by a prior urologist, do we always trust that the DRE was actually performed within a reasonable time or even thoroughly?

Let us now consider the parameter of prior biopsy. What exactly constitutes a prior biopsy? How long ago was the procedure? How many cores were sampled and from what volume gland? Are you familiar with the urologist who performed the biopsy and consequently assume that the submitted cores properly represented all regions making it truly systematic? Of course we all understand the typical slow growth kinetics of the majority of prostate cancer. But should a negative biopsy 9 or even 4 years ago apply even if we assume that the sampling was the appropriate number of cores for the gram size of that particular gland?

There also needs to be some discussion about HG PIN or atypical small acinar proliferation. While these are not considered malignant entities, there’s plenty of prior literature to support follow-up saturation biopsies of these areas. This practice obviously has declined considerably in the era of MRIs. Having said that, however, we are also using 4Kscores after biopsies that demonstrated atypical small acinar proliferation or HG PIN and subsequently checking off “prior biopsy negative.” This seems to be a somewhat contradictory philosophy. Shouldn’t there be a spot in the algorithm for these premalignant entities that make the 4Kscore different or perhaps higher than someone who truly has an unequivocally benign biopsy? How can we simultaneously claim that these histological entities might warrant repeat biopsies yet also not recognize this increased risk on a subsequent 4K algorithm?

There is also something to be said about the transrectal US itself. There are a myriad of machines with different imaging quality and of course tremendous variation in certain urologists’ ability to localize what appear to be suspicious-appearing lesions on US. Once again, this is going to truly affect our assessment of prior negative biopsy, especially when lesions noted on MRI are not subsequently seen by certain urologists performing a transrectal US-guided biopsy.

Should we have a 4K parameter for MRI? This is obviously the hot topic question. How can we possibly assume that the low-risk 4Kscore bears as much relevance when concurrent in a patient with a PI-RADS 4 or PI-RADS 5 lesion, or paradoxically the suspicious moderate or high 4Kscore in the presence of a completely unremarkable MRI? Are there many urologists who wouldn’t biopsy a fairly young and healthy biopsy-naïve patient with a high-suspicion MRI lesion just because the 4Kscore was calculated low risk? Alternatively, should a 4Kscore be calculated as lower with completely unremarkable MRIs or conversely as higher with high-suspicion MRIs?

As the same age-specific reference ranges have not yet been applied to 4Kscores, much like age-specific PSA reference ranges were established many years ago, we must bear in mind that numbers used for low-risk, moderate-risk, and high-risk disease must not be written in stone. I am uncertain as to why we still have these designated risk categories as opposed to just a number. Surely, a 4Kscore of 12 (considered moderate risk) in a 54-year-old healthy male is not necessarily going to be considered moderate risk in a 74-year-old with diabetes and hypertension. Perhaps we need to just solely focus on the number and put it in the context of the patient’s expected longevity. The latter older patient with this 4Kscore is probably not really at moderate risk for disease-specific morbidity or mortality, even in the presence of a couple of lower-volume cores of Gleason 3+4 disease.

The 4Kscore has reportedly yielded a 99% sensitivity and around 95% negative predictive value for determination of Gleason grade 7 or greater disease. We have all therefore counseled innumerable patients on the likelihood of the 4Kscore predicting more clinically relevant disease, as I put it to my patients. I further tell my own patients, that “the number doesn’t necessarily correlate with the likelihood of having prostate cancer, but rather with the likelihood that I will find cancer that we should consider treating. In other words, not all prostate cancers we find require treatment.” Like many colleagues with whom I’ve conversed, I find it difficult to use the term clinically relevant disease in healthy men in their 40s and 50s who are quite likely to warrant or quite commonly opt for treatment of, let’s say, higher-volume Gleason 3+3 disease in 4 cores at a PSA of 4.7. And in the ones who don’t, we certainly can expect a fair amount of them to have greater stage migration within a few years subsequently warranting treatment. We can all counsel this particular young group as to how lead time bias will only subject them to deleterious treatment outcomes earlier than might have been necessary. But isn’t that a decision for the patient to make? Should we discourage prostate biopsy in a 52-year-old male with a PSA of 4.1 (increased from 2.8 a year previously) and a percent free fraction of 16 in a palpably 25- to 30-g prostate because his 4Kscore is 6.3, and thereby deemed low risk? Many might inquire as to why we even ordered a 4Kscore in this patient. But I think we can all agree that many patients might have been reluctant in the first place to have a biopsy with those numbers, and so we collectively agreed to further assess his risk. We might have assumed with those parameters that his 4Kscore might result as 12 or 13, and he then might have been less reluctant to proceed with biopsy.

Lastly, perhaps we need to consider family history. The literature has demonstrated a minimum 2 times greater risk of developing prostate cancer in patients with just 1 immediate family history in their father or brother, let alone a questionable higher risk with >1 immediate family member. This can add another, perhaps relevant parameter to calculation of the 4Kscore and should strongly be considered.

At the end of the day, we are still adding a lot of art to the science; at least we should be. The 4Kscore has certainly helped us encourage biopsies in those in whom we truly believe would benefit from prostate cancer therapy while discouraging biopsies in those who are quite unlikely to have prostate cancer represent a source of morbidity and mortality in their lifetime. It is probably in our best interest, however, to further refine the 4Kscore with other parameters. We need prospective data that help us to determine not just prior biopsies, but more information regarding the years passed since and perhaps number of cores taken. Perhaps we can even have a different parameter score for biopsies that were done with MR fusion. We have to make certain that DREs were performed within a reasonable time. We should have parameters for the presence or absence of MRI, and which category lesions constitute no suspicion, low suspicion, or high suspicion based on PI-RADS classification. And as mentioned, family history is likely to add a predictive parameter. Surely it will be a much longer form to fill out and we are all very pressed for time. But improvement and evolution are the hallmark of what we do in the urology community.

Incidentally, I am the 57-year-old patient. Of course I wonder why my PSA has slowly been incrementing at a young age even if my prostate volume hasn’t changed. I wonder why this “low suspicion” lesion on my MRI has progressively grown. I wonder why my percent free fractions have been dropping and are obviously abnormal. So on my newest really low 4Kscore, when they checked DRE “no nodule,” and “prior biopsy negative” (considering it was 20 cores with magnetic resonance fusion) I at least know that is true and relevant. I might also point out that my particular evaluation deviates from the standard AUA guidelines of follow-up. As such, however, it does provide insight into what now is thought-provoking commentary. What we comfortably recommend to our patients might not suffice when it comes to our own suspicious presentation.