High prostate cancer mortality in Norway evaluated by automated classification of medical entities

Applying recommended definition of aggressive prostate cancer: a validation study using high-quality data from the Cancer Registry of Norway

BACKGROUND

The Prostate Cancer Cohort Consortium (PC3) Working Group proposed a definition for aggressive prostate cancer (PC) for aetiologic epidemiologic research. We aimed to validate this definition as well as a second approach utilising only information on stage at diagnosis.

METHODS

First primary PCs diagnosed 2004 - 2009 in the population-based Janus Serum Bank (JSB) cohort were identified by linkage to the population-based Cancer Registry of Norway (CRN) (n = 3568). The CRN and Norwegian Prostate Cancer Registry provided clinicopathological data for these cases. Approach 1 classified PC as aggressive if it was clinically T4, or N1, or M1, or had a Gleason score ≥8 at diagnosis (as proposed). Approach 2 classified PC as aggressive if CRN stage at diagnosis was 'regional spread' or 'distant metastases'. Both approaches were validated by calculating the sensitivity and positive predictive value (PPV) against PC-death within 10 years of diagnosis.

RESULTS

Overall, 555 died from PC within 10 years. Approach 1 classified 24.7% of cases as aggressive and 13.6% were unclassified due to missing information. Approach 2 classified 19.6% as aggressive and 29% were unclassified. Sensitivity was highest for Approach 1 (0.76, 95% CI: 0.72 - 0.80 vs 0.69, 95% CI: 0.64 - 0.73), while PPVs were similar for both approaches (0.43, 95% CI: 0.40 - 0.46 and 0.40, 95% CI: 0.36 - 0.44). We observed similarly high sensitivity and higher PPVs than those reported by the PC3 Working Group.

CONCLUSIONS

The proposed definition of aggressive PC was applicable and valid in the JSB cohort. Stage at diagnosis can be useful if data on cTNM or Gleason score is unavailable.