The late-breaking abstract for the results of a US-based study of the accuracy of the new 4Kscore Test™ has just been made available on line. (Thank you to one of our correspondents for bringing this to our early attention.)
According to the abstract of this presentation by Lin et al. (to be given at the upcoming meeting of the American Urological Association in Florida), their study enrolled an initial 300 patients from 26 academic and community urology centers between October 2013 and January 2014. (A further 900 patients were subsequently enrolled, but the initial data analysis is based exclusively on the first 300 patients.)
The data are based on patients who were already scheduled for a prostate biopsy based on prior PSA and other clinical data. For each patient, on the day they were scheduled for biopsy, there was an initial blood sample taken (by phlebotomy) followed by a systematic TRUS-guided biopsy of ≥ 10 cores.
All blood samples were processed at the OPKO laboratory in Nashville, Tennessee. Histopathological examination of biopsy specimens was carried out according to the established practices of each clinical site. All data were evaluated by independent biostatisticians.
Very limited data are provided in the actual abstract, as shown in the following table and comments:
The 4Kscore test does seem to offer a significant degree of differential discrimination of risk compared to a standard clinical algorithm based on patients’ PSA levels and and clinical data.
Furthermore, according to Lin et al.:
- The 4Kscore test would have reduced the number of biopsies required by 41 percent, with a negative predictive value (NPV) of 97 percent.
- Decision analysis demonstrates that clinical decision-making would be improved by the use of the 4Kscore test.
The authors conclude that:
The 4Kscore test, using fresh samples evaluated in a commercial laboratory, yields superior performance compared with a clinical model. The 4Kscore test reports a probability of finding high-grade cancer that can inform clinically relevant, shared decision-making between urologist and patient.
It appears that the 4Kscore test may well have significant potential (compared to PSA and clinical data alone) from a patient perspective. If the cost equation is also correct (i.e., the cost of giving a 4Kscore test to all patients who might need a biopsy is less than the cost of biopsying men who appear not to need a biopsy based on that test), then the 4Kscore test may well have potential from a payer perspective too. However, The “New” Prostate Cancer InfoLink will be watching for data from the full 1,200 men initially enrolled into this prospective study, and specifically for data from that study addressing the original PSA values, clinical stages, and ages of the patients enrolled and a breakout of the results of the 4Kscore data correlated to age, PSA levels, and clinical stage.
What is going to be very important to the uptake of a test like this is defining criteria beyond which the accuracy of the 4Kscore test may be less accurate (e.g., upper and lower levels of PSA, age, etc.). And then there is going to be the question of the accuracy of the test in specific subgroups by ethnicity.
If this test is to be used to tell a patient that he does not need an immediate biopsy, then the test is going to need a high level of accuracy to make a clinician comfortable about giving that recommendation. If that recommendation can not be given with great accuracy, any potential cost-benefit will vanish rapidly because the clinician’s advice to the patient is likely to be, “You may have no cancer at all, or perhaps only a tiny amount of low-risk cancer that we won’t need to treat, … but the only way to tell will be if I give you a biopsy.” And even that recommendation isn’t strictly accurate, of course, because a standard 12-core, systematic, TRUS-guided biopsy is still very likely to “miss” finding small amounts of low-risk tumor (or even small amounts of higher-risk tumor).