The MSKCC-Malmö four-kallikrein model for prediction of prostate cancer risk


The collaborative research between the teams at Memorial Sloan-Kettering Cancer Center in New York and in Malmö in Sweden continues to  refine a model that may allow us to accurately predict which patients are at clinically significant risk for prostate cancer based solely on simple tests run on a single blood sample.

Initial research by these teams (see Vickers et al., 2008) suggested that a panel of four kallikrein markers (total PSA, free PSA, intact PSA, and kallikrein-related peptidase 2) might be a highly accurate predictor of biopsy outcome. In a paper from May 2010, they had also shown that this model could be used to dramatically reduce biopsy rates in men with a PSA level ≥ 3.0 ng/ml. However, the degree of clinical significance of biopsy-detectable cancers in men classified as low risk by this panel of markers was still not clarified. In their most recent paper, Vickers et al. have used data from the Malmö Diet and Cancer study to investigate this question.

The Malmö Diet and Cancer study is based on a population-based cohort of 11,063 Swedish men living in Malmö and aged between 45 and 73 years all of whom provided a baseline blood sample at study enrollment (between 1991 and 1996). The researchers were able to use the Swedish Cancer Registry to identify a total of 943 men among this study cohort (8.5 percent) who had been diagnosed with prostate cancer by December 31, 2006. The subsequent use of PSA testing among the study cohort was low during the follow-up period from 1991 to 2006. As a consequence, this patient cohort effectively represents a cohort of men exhibiting the natural history of early stage, unscreened prostate cancer over a period of 15 years — from 1991 to 2006.

Vickers et al. were able to use their previously described statistical model to predict and then validate subsequent prostate cancer diagnosis in men in the Malmö cohort who had a total PSA ≥ 3.0 ng/ml at baseline. The results of their analysis show that:

  • 474 men with a total PSA ≥ 3.0 ng/ml at baseline had a clinical diagnosis of prostate cancer by December 31, 2006.
  • Another 469 men had a total PSA < 3.0 ng/ml at baseline and were also diagnosed with prostate cancer by December 31, 2006.
  • Median follow-up for the men not diagnosed with prostate cancer at all was 12 years.
  • Compared to biopsying all men with a PSA ≥ 3 ng/ml at baseline, if only those men with a PSA  ≥ 3 ng/ml at baseline and a ≥ 20 percent risk of a positive biopsy based on the four kallikrein model were to be biopsied …
    • 421/1,000 men (42.1 percent) would be classified as low risk by the panel and would not be recommended for biopsy.
    • 21/421 men (5.0 percent) classified as low risk by the panel and not recommended for biopsy would be clinically diagnosed with prostate cancer within 5 years.
    • Only 2/421 men  (0.5 percent) classified as low risk by the panel and not recommended for biopsy would be clinically diagnosed with advanced prostate cancer (clinical T3/4 or metastases) within 5 years.

Vickers et al. note that the reason they selected clinical diagnosis of prostate cancer within 5 years as the outcome for this study was based on their (very reasonable) assumption that cancers diagnosed after more than 5 years would very likely be identified during a further round of screening today (especially if their PSA was notably higher than 3.0 ng/ml).

So what is the bottom line to this study?

In the first place it appears to demonstrate clearly that men with a total PSA level ≥ 3.0 ng/ml but defined as low risk by the panel of four kallikrein markers are unlikely to develop incurable prostate cancer. (This assumes that the men would have no further PSA or other test within the 5-year follow-up that might still allow early diagnosis and potentially curative therapy.)

Secondly, it shows that the panel of four kallikrein markers can reduce the number of unnecessary prostate biopsies by about 40 percent among men with a total PSA level ≥ 3.0 ng/ml. (This is a substantial reduction. There are estimated to be some 750,000 unnecessary biopsies carried out in the USA each year.)

Third, it shows that, of the men with a risk of < 20 percent of a positive biopsy based on the four kallikrein panel, only 5.0 percent would be clinically diagnosed with prostate cancer within 5 years, and that assumes that they would have no further PSA or other test within that time frame that might still allow early diagnosis and potentially curative therapy.

But what about the 469 men with a PSA < 3.0 ng/ml who went on to get prostate cancer within the Malmö study follow-up period? If one assumes that most of these men were to receive further PSA testing within 3-5 years, it seems likely that the majority of these men would also be diagnosed sufficiently early to receive curative treatment, but clearly this aspect of the model still needs to be carefully tested.

Vickers and his colleagues are currently testing their model in a cohort of US patients with serial PSA and DRE data. Broad clinical application of this model is still not likely to occur in the US in the near future, but it is becoming clear that this model offers a potential method for projecting risk of clinically significant prostate cancer based on a simple blood sample taken every 3-5 years (or annually in men at potentially high risk) without the need for immediate biopsy in a majority of patients.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: