Discussion

Our findings demonstrate that Decipher risk groups are significantly associated with critical histopathologic parameters in diagnostic prostate biopsies, including grade group, positive core ratio, and maximum tumor length ratio.

Historically, prostate cancer management has relied upon clinical risk stratification to guide treatment and surveillance decisions. The National Comprehensive Cancer Network (NCCN) guidelines represent the most widely utilized framework for this purpose, categorizing patients into low, favorable intermediate, unfavorable intermediate, high, and very-high risk groups. These classifications are based on established clinicopathologic features, such as clinical T-stage, grade group, serum PSA levels, and a positive core ratio cutoff of 50% [11].

In recent years, the oncology landscape has shifted toward personalized medicine through the development of multi-gene genomic classifiers. In the context of prostate cancer, these assays have shown substantial promise in refining prognosis and optimizing management. While these tools were initially viewed as adjuncts, they are increasingly being integrated into standard clinical practice guidelines [12]. Notably, within the current NCCN framework, the Decipher classifier (GenomeDx Biosciences, Vancouver, BC, Canada) is the only genomic classifier recommended under the advanced diagnostic tools for risk stratification and biomarker assessment [11]. Our study reinforces this integration by showing that genomic risk groups align with histomorphological aggressiveness of tumors.

The Decipher classifier is a tissue-based platform that evaluates 22 differential genes involved in critical biological pathways, including cellular proliferation, differentiation, immune modulation, and androgen-receptor signaling, derived from whole-transcriptome analysis. It assigns each case a score from 0 to 1 and classifies them into low-risk (LR) (<0.45), intermediate-risk (IR) (≥0.45-<0.6), or high-risk (HR) (≥ 0.6) categories. The assay has been extensively validated as both a prognostic and predictive biomarker across multiple clinical endpoints, such as adverse pathology at radical prostatectomy, biochemical recurrence, metastasis, and prostate cancer-specific mortality following treatment [13,14]. Research has established that Decipher independently predicts oncological outcomes; for instance, patients with low genomic classifier (GC) scores exhibit significantly lower 10-year metastasis rates compared to those with high GC scores (4% versus 16%) following radiotherapy alone. Such data suggest a potential for treatment de-escalation, particularly by omitting short-term androgen deprivation therapy (ADT) in select NCCN intermediate-risk patients [11,15-17]. However, while its utility in surgical specimens is well-documented, there remains a comparative paucity of data regarding its application and association with histopathology in the diagnostic or follow-up biopsy settings [13].

Within our institution’s genitourinary pathology practice, we have observed a marked increase in the clinical utilization of the Decipher genomic assay in the biopsy setting. The primary clinical driver for this trend is the need to distinguish between biologically indolent disease and high mortality phenotypes at an early diagnostic stage, a distinction that conventional prognostic markers, including Gleason scoring, may not sufficiently provide [18].

While our study established that grade groups (GG) correlate significantly with Decipher risk groups, this association was primarily driven by the differences between HR and LR cohorts. Notably, GG alone failed to predict a categorical difference between the IR and HR groups. In alignment with the validation cohort by Kim et al., the majority of biopsies in our study featured lower-grade groups and were subsequently classified into the LR genomic group [19]. However, we identified exceptional cases, such as a GG1 case in the HR group and a GG5 case in the LR group; based on findings by Shee et al., such discordant cases are expected to exhibit clinical behavior dictated more by their genomic profile than their histopathology over the course of clinical care [20].

Our data also demonstrate that positive core ratios correlate with Decipher risk groups. Specifically, the HR group exhibited the highest positive core ratios, a finding that aligns with Vince et al., who reported an increase in percent positive cores within genomic HR groups [16]. Although percent positive cores is a standard variable in traditional NCCN risk stratification, our findings emphasize that these histopathologic differences are most pronounced when comparing the polar ends of the spectrum (HR versus LR) rather than correlating as a seamless three-tier predictive system together with the genomic classifier.

Additionally, our analysis revealed a significant association between the maximum tumor length ratio and Decipher risk groups, which is a relationship that, to our knowledge, has not been previously detailed in the literature. Much like our other findings, this association was predominantly driven by the divergence between the HR and LR groups.

Finally, although not yet fully incorporated into established clinical predictive models, two aggressive histologic features, extraprostatic extension and intraductal carcinoma, were observed exclusively in the high-risk (HR) group. Additionally, other adverse features, including cribriform architecture and perineural invasion, were more prevalent in the HR cohort. The lack of statistical significance for these differences is likely attributable to the limited sample size. As previously established, the presence of these histomorphologic features is strongly associated with adverse clinical outcomes and treatment resistance [21-24]. Our findings suggest that their presence on biopsy may be a potential morphological surrogate for high genomic risk, warranting evaluation in larger cohorts.

Overall, our study underscores the clinical relevance of histopathologic findings and their significant association with genomic classifier risk groups. Given the well-established biological significance of these histopathologic features, these findings support the biological plausibility of the Decipher genomic classifier in the biopsy setting. At the same time, they also reaffirm the enduring validity and critical importance of histomorphology in prostate cancer management; traditional features remain highly predictive of biological behavior, particularly when evaluated within a two-tier genomic risk framework.

Limitations

Despite the clinical relevance of our findings, this study is not without limitations. First, the cohort included only patients who underwent Decipher genomic classifier (GC) testing, introducing potential selection bias, as Decipher is typically ordered for selected patients in whom additional risk stratification is clinically indicated. Consequently, this cohort may not be fully representative of all patients with biopsy-proven prostatic adenocarcinoma, potentially limiting the generalizability of our findings. Second, the retrospective design and relatively modest cohort size (N=73) may have limited the statistical power to detect associations involving less common histopathologic features, such as neuroendocrine differentiation or tumor necrosis. Third, because the study focused on recently diagnosed cases, the follow-up duration was insufficient, precluding evaluation of clinically meaningful long-term outcomes such as biochemical recurrence, metastasis-free survival, prostate cancer-specific survival, and overall survival. Finally, as this was a single-institution study, further multicenter prospective investigations with larger datasets are warranted to validate these associations and facilitate the seamless integration of the Decipher genomic classifier into routine pathologic practice.