The Cohort.
Detailed collection methods have been described previously(16, 17) but are repeated here in brief. Men with PCa were identified from three cancer registries in the UK. Collaborating hospitals in these areas were found, and cases from these hospitals were reviewed. Men were included if they were aged <76 years at the date of diagnosis and had clinically localised PCa diagnosed by needle biopsy in 1990–2003 inclusive. The median date of diagnosis was May 2002. Patients treated with radical prostatectomy or radiation therapy within 6 months of diagnosis were excluded. Only initial hormone therapy was permitted. Those with objective evidence of metastatic disease (by bone scan, X-ray, radiograph, computed tomography scan, magnetic resonance imaging, bone biopsy, lymph node biopsy, or pelvic lymph node dissection) or clinical indications of metastatic disease (including pathological fracture, soft tissue metastases, spinal compression, or bone pain), or a prostate-specific antigen (PSA) measurement of >100 ng/ml at or within 6 months of diagnosis, were also excluded. Men who had received hormone therapy prior to the diagnostic biopsy were also excluded, because of the influence of hormone treatment on Gleason score and Grade Group. Men who died within 6 months of diagnosis or had <6 months of follow-up were also excluded.
Original histological specimens from the diagnostic biopsies were requested. Follow-up was conducted by use of the cancer registries, and the cut-off date was 31 December 2012. Deaths were divided into those from prostate cancer and those from other causes, according to WHO standardised criteria (WHO, 2010).
Baseline PSA level was defined as the last prediagnostic PSA measurement within 6 months before diagnosis. If no such PSA value was available, we took the first post diagnostic PSA level within 6 months; failing that, the prediagnostic PSA level measured closest to the date of diagnosis was used. All PSA values after treatment with hormones or orchiectomy or within 3 weeks after a surgical procedure on the prostate were excluded.
National ethics approval was obtained from the Northern Multicentre Research Ethics Committee, and local ethics committee approval was obtained at each of the collaborating hospitals.
Pathological review.
A panel of 3 urological pathologists (DMB, LB, GS) confirmed the diagnosis of adenocarcinoma and reassigned Grade Groups (GG) by using a contemporary and consistent interpretation of the Gleason scoring system. For every core in each case, the presence of cancer was recorded. Also, the length of each core (mm), length of cancer was measured. In some cases there were ‘stromal gaps’ where two malignant foci were separated by benign tissue. The length of any stromal gap was recorded, though stromal gaps were ignored if less than 2mm.
These data allowed calculation of the percentages of each tumour involved core, percentages of cores positive for cancer, and calculation of this data was performed, with and without stromal gaps. Abbreviations for these different measurements was as follows.
Measures of cancer volume on prostatic cores
NPC= Number of positive cancer cores
%+cores = Percentage of positive cancer cores over total cancer cores
MCL= Maximum Cancer length in a single core in a prostatic biopsy series. (mm)
TCL= Total Cancer length: addition of all cancer core lengths in a biopsy series. (mm)
SG=Stromal Gap (mm)
The panel met and discussed all controversial cases and a selection of others to audit the dataset. In keeping with the ISUP 2014 recommendations and the grading in World Health Organization (WHO) 2016, cribriform and glomeruloid glands were all assigned Gleason pattern 4. Analysis of this cohort with regards to pattern and grade of disease has been published previously(17).
Statistical analysis
Survival analysis was performed with a Cox proportional hazards model. The primary endpoint was time to death from prostate cancer. All events were used for estimation of hazard ratios (HRs), and observations were censored on the date of last follow-up or death from other causes. All events were used for estimation of HRs (maximum follow-up 232 months), but follow-up was censored at 10 years for prediction of 10-year risks. HRs were calculated for the interquartile range (IQR), as this is a better method of comparison when the units of the different variables are very different.
Extent of disease was measured in each core by the methods described above. Covariates included in the statistical analysis were Gleason scores by overall grade. This had been shown to be comparable to ‘worst’ grade in a previous paper(18).
The PSA level was modelled as the natural logarithm of [1 + PSA (ng/ml)]. Patients with values of >100 ng/ml were excluded as likely to have metastatic disease. Missing PSA values were imputed by use of a median regression with GG, age and extent of disease as predictors, and PSA as outcome. Missing T stage values were imputed using the median clinical T stage among all patients.
A univariate model was applied to Gleason score, baseline PSA level, T stage, and the multiple different methods of tumour volume measurement. A multivariate Cox proportional hazard model applied performed Gleason score, baseline PSA level, extent of disease, T stage, The primary event of interest was time to death from prostate cancer. A stepwise model selection was performed.
Spearman’s rank correlation was estimated between all variables. All applied tests were two-sided, and P-values of <0.05 were accepted as statistically significant. No P-value adjustment was performed for multiple comparisons. Statistical analyses were performed with r (R Core Team (2018). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. http://www.R-project.org/).