In keeping with current literature, our study reinforces one of the key challenges in the risk stratification of PCa patients: the notable high discordance rates between the GGG assigned at PBx and RP. [9–11] We found a concordance rate of 58% in our population, with 20% of our cases being downgraded and 22% of our cases being upgraded. Several factors may contribute to this discordance. Interobserver variability between reporting pathologists is well recognised and is likely a significant factor. [14] While this is minimised in our department, with a select number of consultant histopathologists reporting PBx, variation cannot be eliminated entirely. Additionally, sampling error due to tumour heterogeneity is likely a contributory factor. Prostate tumours often contain multiple Gleason grades within the same tumour and the specific area biopsied may not reflect the highest Gleason grade present in that tumour. [15] Furthermore, PCa may not only be heterogeneous but may contain multiple foci of PCa, with one study finding two or more adenocarcinomas in 87% of their RP specimens. [16] Although MRI targeted biopsies and extensive prostate sampling aim to reduce this issue, it likely still contributes to discordance.
In 2016 none of the patients who underwent a RP in SVUH had a TP PBx as their diagnostic biopsy method. Our study offers real-world data illustrating the trend shift from TRUS PBx to TP PBx over a seven-year period, rising annually with TP biopsies ultimately accounting for 55% of patients who underwent a RP in our institution. No significant difference was found in the overall concordance rates between TP PBx (55%) and TRUS PBx (59%) in our cohort and this is reflective of the findings presented in the literature by Evans et al and Uleri et al [10, 17]. However, when examining upgrading in isolation, our study diverged from the findings of Evans, Qu and Scott et al, who reported lower rates of upgrading following TP biopsies. [10, 18, 19] In contrast to their findings, our data indicates that while overall concordance rates were similar, there was a significant association between upgrading and TP biopsies (p = < 0.01). This finding is clinically significant, as upgrading at RP is associated with poorer outcomes. [20] A potential explanation for this observation is that those deemed clinically lower-risk patients with normal MRIs are more likely to be selected for TP biopsy over TRUS biopsy, possibly contributing to the higher observed rate of upgrading at RP.
Our study also found a significant association of both EPE (p = 0.001) and SVI (p = 0.001) with GGG upgrading at RP. These finding align with previous research highlighting that EPE and SVI are predictors of worse outcomes and as such are factors that upstage a tumour to pT3a & pT3b in the TNM classification system respectively. [20, 21] Moreover, EPE and a positive surgical margin status have been significantly associated with early biochemical recurrence in patients who have undergone RP [22, 23] and while a positive surgical margin status (103/471) did not reach statistical significant in our study it was observed more frequently in RP specimens that were upgraded. These observations are relevant for both clinicians and histopathologists. For clinicians the presence of these histological features may signal a more aggressive disease and as such help inform post-operative management and decision making around adjuvant therapy. [24] For histopathologists the identification of adverse pathological features should prompt increased vigilance when assigning GGG to RP specimens given such features are associated with higher GGG. [25, 26]
Our study has several strengths. The large sample size of 588 patients provides a robust data set for analysis. Additionally, the trend analysis documenting the shift in biopsy techniques over seven years offers valuable insights into evolving clinical practices. The detailed assessment of multiple pathological outcomes, including GGG concordance and adverse pathological features, provides a comprehensive view of the factors influencing grading discordance. However, there are notable limitations to our work. Firstly, the retrospective nature of the study may introduce selection bias and limit the ability to establish causal relationships. Secondly, the study’s findings are based on data from a single institution, which again adds to the selection bias and lack of variability in work practices. Thirdly, we did not perform multivariable regression analyses to adjust for potential confounders, which may affect the validity of our findings. Lastly, sensitivity analyses were not conducted to assess the robustness of our results. These limitations should be considered when interpreting the results of this study.
In conclusion, our study highlights the ongoing challenge in accurately predicting final pathological outcomes for PCa patients from initial biopsy results. Despite the increasing adoption of TP-guided biopsies, we found no significant difference in overall GGG concordance rates between TP and TRUS biopsies. However, TP biopsies were significantly associated with GGG upgrading, which has important clinical implications. Our findings support the need for continued efforts to improve PCa risk stratification.