Differentiation between indolent/aggressive
screen-detected prostate cancer

Once prostate cancer has been diagnosed, should you recommend
immediate treatment or active observation?
 
Risk indicator 6 is based on the prediction of an “indolent” prostate cancer which is probably not aggressive. An initial attempt to differentiate indolent and aggressive cancers was first made through a nomogram, produced by Kattan et al (J Urol 2003). His nomogram was validated by replacing clinically detected cancers by screen detected cancers, who underwent radical prostatectomy in Rotterdam as part of the European Randomised Study of Screening for Prostate Cancer. Risk indicator 6 applies the results of this validation (Steyerberg et al, J Urol 2007). The data, which resulted in a score sheet and score diagram presented in this publication, are the basis for this risk indicator.
With this indicator the chance of a given prostate cancer to be classified as indolent can be calculated. A pre-requirement is that the following selection criteria are met: Men who have T1c, T2a-c prostate cancer, a PSA < 20ng/ml, Gleason grades <=3, <= 50% of positive sextant biopsies, < 20mm cancer, >= 40mm benign tissue.
As shown in the table, in a population of screen-detected prostate cancers 46% or 32% of the tumours can be classified as indolent by using probability cut-off values 60 or 70%.

The table shows that the use of a cut-off value for the probability of having indolent cancer of 70% assigns 94% of clinically relevant cancers correctly to immediate treatment. There is a chance of 6% that a non-indolent tumour would be classified as indolent and therefore not treated properly by more aggressive management. In the same situation, using a probability cut-off of 60%, 46% of the potentially indolent tumours would be classified correctly as indolent and advised active surveillance. In this situation, 85% of the clinically relevant cancers would be correctly advised to receive immediate treatment but 15% would be misclassified and selected for observation.
 

Proportional distribution of recommendations for treatment versus active surveillance for clinically important (N = 142) and indolent (N=136) using different cut-off values  prostate cancer (total N=278)  . ERSPC

Treatment (Tx)	Clinically relevant PC - treated N (%)	Indolent PC Tx delayed N (%)
No tx if probability indolent >30%	50/142 (35)	126/136 (93)
No tx if probability indolent > 60%	120/142 (85)	62/136 (46)
No tx if probability indolent > 70%	133/142 (94)	43/136 (32)

Pre-requirements for application of risk indicator 6

• Men age 55 to 74 years old
• All parameters are known or can be reliably estimated
• Lateralised sextant biopsies have been taken
• The cancer has been detected by screening and not on clinical grounds (Micturition complaints can be left out of consideration)

With incomplete data or in presence of clinically diagnosed prostate cancer, modified applications are possible. These are indicated in the original publication by Steyerberg et al, J Urol 2007.
Considering the possibility that prostate cancer is classified as indolent will still turn out to be progressive, a structured follow-up protocol is necessary. This is offered within the prospective PRIAS study protocol (Prostate Cancer Research International Active Surveillance), which is available via www.prias-project.org. If you wish to enter patients on line, you can organise access by getting in touch with the PRIAS office, T. +31-10-4632240 of by email: m.roobol@erasmusmc.nl.

Start risk indicator 6

 

edition 2, 2008