There is currently a range of reported follow-up biopsy intervals of patients on AS, with no available data supporting the use of one frequency over another and no reports regarding the safety of modifying follow-up intervals based on clinical risk factors.

Prospective data regarding stability of individual patients’ molecular risk scores over time is also lacking; however, important inferences can be made from the clinical validation studies that may help encourage individual tailoring of AS follow-up schedules. In addition to the ability of ProMark and GPS to predict the initial presence of adverse pathology, Decipher biopsy predicts a 10-year metastasis risk after prostatectomy. GPS and CCP score boasts 15-year follow-up data with regard to risk of clinical recurrence (GPS) and metastasis-free survival (CCP) [3, 5]. Applied as prospective selection criteria for an AS cohort, the hypothesis is that patients with the lowest likelihood of adverse pathology would also likely have the most stable molecular risk scores over time, intuitively demonstrating the possibility for decreased frequency of monitoring without sacrificing mortality benefit.

In patients that have molecular risk stratification at the time of initial diagnosis which demonstrates a low likelihood of adverse pathology, a longer interval to re-biopsy or conversion to a watchful waiting protocol could be considered [48]. Conversely, in patients with a higher likelihood of adverse pathology, a shorter interval to re-biopsy or a more confident decision for immediate treatment may be made. While this rationale is based on robust retrospective follow-up data, definitive support for this strategy would come in the form of randomization of patients to more intensive or less intensive follow-up based on their molecular risk profile. Molecular risk profiling adds prognostic precision to known clinical parameters; it does not replace them. Clinicians must still exercise judgment regarding factors of patient life expectancy, comorbidity, and patient desire when using a molecular risk score to help plan the frequency of follow-up in an individualized AS protocol (Fig. 12.1).

Recent results from large active surveillance cohorts suggest that some patients with otherwise classified “insignificant” cancer, and some that fulfill additional stringent clinical selection criteria, still have a risk of metastasis and death [1, 2, 50]. In the Sunnybrook experience, the initial Gleason score of the 15 patients who died from prostate cancer is not specified; however, it does report that 16 of 28 patients (66%) who developed metastatic disease had Gleason ≤6 cancer and 7 of them fulfilled Epstein criteria. The average patient in this cohort was almost 68 years old; thus, one might estimate as this cohort is enriched with younger patients, this risk may increase further [1]. In the Hopkins experience, there were 2 prostate cancer deaths per 100 person-years in the very low-risk patient group (n = 926) and 3 additional low-risk patients with lymph node or distant metastasis per 100 person-years [2]. In the ProtecT Study Group’s comparison of monitoring, surgery, or radiotherapy for localized prostate cancer, metastases developed in men in the active monitoring arm at a rate of 6.3 events per 1000 person-years during 10 years of follow-up. This was substantially higher than patients treated with surgery or radiation. Additionally, there were eight prostate cancer deaths in men with Gleason 6 cancer at initial diagnosis across all three trial arms [50].

Gleason pattern 3 cancers are molecularly more benign than their Gleason pattern 4 and 5 counterparts. This is not unexpected given the strong prognostic value of Gleason score alone and the overwhelmingly positive outcomes of the contemporary active surveillance cohorts previously discussed. There are also, however, several studies that suggest there are some Gleason pattern 3 cancers that have molecular characteristics of more aggressive disease. This is also not unexpected as there are still patients with low-risk disease from these active surveillance cohorts who develop metastasis and/or die of prostate cancer. A closer look at the molecular characteristics of these potential outliers is warranted in order to push the area under the curve for prediction of negative outcomes in patients evaluated for active surveillance as close to 1 as possible. It is likely that the Gleason 6 cancers with adverse molecular features de-differentiate prior to metastasizing, since there are virtually no well documented cases of pure, surgically confirmed Gleason 6 that have metastasized.

Analysis of 340 Gleason 6 prostate cancers using the Decipher genomic classifier revealed high-risk and intermediate-risk scores in 7 and 13% of patients, respectively [51]. Thus, 25 patients had a seven times higher risk of metastasis, and 43 patients had a two times higher risk of metastasis than the remaining 276 patients in this cohort. All tumors were rereviewed by expert genitourinary pathologists using the ISUP 2005 Gleason grading criteria. While there was a significant proportion of pattern 4 disease identified upon rereview, examining the prospective cohort patients specifically, there was no statistically significant difference between the Decipher scores of those upgraded and those that were not. It is still unclear whether aggressive behaviour of the small minority of Gleason 6 cancers with adverse genetic features is generally due to co-existent occult higher grade cancer, or due to de-differentiation of Gleason 6 cancer. Both are possible.

Polson and colleagues reported that the TMPRSS2/ERG gene fusion (a critical and likely early factor in prostate cancer pathogenesis) is present and expressed at the transcriptional and translational levels in the stem cell compartment from primary human prostate cancers, including 3 of 5 analyzed Gleason 3 + 3 = 6 cancers [52]. This is significant given the possibility that the cancer stem cell population potentially drives the process of metastasis as well as therapy resistance, leading to recurrence and relapse. Additionally, 10–20% of Gleason 6 cancers demonstrate loss of important tumor suppressor genes such as PTEN [53, 54]. The significance of PTEN loss is clearly demonstrated in an animal model in which the combination of PTEN loss and MYC activation is sufficient to lead to genomic instability and lethal metastatic disease [55].

Evidence for a biological field effect in prostate cancer is demonstrated by the predictive value of the presence of PTEN loss, MYC/8q gain, and LPL/8p loss on Gleason pattern 3 biopsies for the presence of un-sampled Gleason pattern 4 cancer. In other words, the presence of these markers in a Gleason pattern 3 core makes it much more likely to have come from a prostate that harbors Gleason pattern 4, suggesting that molecular changes of tumor aggressiveness are present before histologic changes occur. This demonstrates yet again that biologically important markers for adverse pathology are found in Gleason pattern 3 cancer [56]. Testing for PTEN deletion and TMPRSS2/ERG fusion on biopsy tissue is commercially available through Metamark Genetics, Waltham, Massachusetts.

An analysis of biopsy-based genomic and microenvironmental indices to predict 5-year risk of biochemical recurrence (BCR) after local therapy revealed that several individual Gleason 6 tumors had a higher percentage of genome alteration as measured by copy number variation than some Gleason ≥8 tumors with significant overlap in the genomic instability of Gleason 3 + 3 = 6 and higher-grade tumors [57]. Percentage of genome alteration carries strong prognostic value independent from clinical covariates. Every 1% increase imparts a 5–8% decrease in 5-year post-local therapy biochemical recurrence-free survival [57]. Interestingly, among the low- and intermediate-risk prostatectomy cohort, the risk signature was more strongly predictive of biochemical relapse than clinical variables. 89% (95% CI 85–96) of good prognosis (PGA ≤7.49) patients were free from BCR at 5 years compared to 58% (35–96) of poor prognosis (PGA >7.49) patients.

There is data suggesting that the multiple tumor foci detected in patients with prostate cancer have independent origins. However, whole-genome sequencing of multiple metastatic sites from several patients’ primary tumors demonstrated a common clonal origin containing 40–90% of the total mutations and the majority of driver mutations. The implication is that widespread metastases originate commonly from only one of many tumor foci – a foci potentially small enough to be easily missed under normal pathological examination, regardless of its grade [58].

An in-depth analysis of the genetic phylogeny of multifocal prostate cancer in three separate cases identified multiple independent clonal expansions of cells in both neoplastic and morphologically normal prostate tissue [59]. This is important since the mutations defining the clonal expansion of morphological normal tissue were the same as those in cancer. There were large numbers of mutations shared among foci of Gleason 6 and 7 cancer, as well as a smaller but significant number of shared mutations in the adjacent normal tissue.

In describing the genomic correlates to the new grade 1–5 prognostic groups for prostate cancer, Rubin et al. demonstrate that Gleason 6 cancers harbor similar mutations to higher-grade cancers, albeit at a lower frequency [60]. The molecular profile of some low-grade tumors looks remarkably similar to those more frequently found in higher-grade cancer (Fig. 12.2).

Finally, the overexpression of long noncoding RNA SChLAP1, an independent predictor of lethal prostate cancer, was found in 4 of 165 Gleason ≤6 cancers and present in 11 of 334 Gleason 3 + 4 cancers [61].

The plausibility of genetic alterations characteristic of aggressive and even lethal prostate cancer lurking in tissue that lacks the morphologic changes necessary to garner a histologic classification equaling the seriousness of its underlying molecular perturbations has been thoroughly demonstrated (Table 12.2). The implications of these findings are such that clinicians should not fall into the temptation of a one-size-fits-all approach with regard to active surveillance. While current tissue-based molecular risk assessment is an improvement on clinical risk assessment, it is an incremental improvement that is not a panacea. Continued development and integration of molecular tumor analysis with pathological grading/staging is important to achieve the most accurate prognostic information possible. Further progress may come by combining molecular tests into algorithmic formulas in order to reliably detect patients who may currently meet all active surveillance eligibility criteria but may in fact still harbor disease that warrants treatment.