Our study shows a trend towards the surgical treatment of more advanced PCa, in particular, due to higher grade disease, more lymph node metastases, and rising rates of PSM and pT3 across all NCCN risk groups. Radical prostatectomy has been performed in slightly older men with more comorbidities. A higher representation of African American patients was observed, likely as a result of selection bias due to their higher rates of aggressive disease; yet other competing factors (e.g. disparities in access to health care) cannot be ruled out. We also observed a notable and steady rise in pathologic upstaging across all risk groups after radical prostatectomy between 2006 and 2020.

The elevation in pT3 rates is likely multifactorial: in addition to the growing utilization of RP for treating high-risk PCa, more low and favorable intermediate-risk cases have been managed with active surveillance or focal therapy, rather than immediate surgery [13, 14]. Since 2021, AS has become the preferred treatment modality according to the NCCN guidelines for very low and low-risk PCa. Moreover, the shift towards less intense screening strategies might delay diagnosis and result in more advanced disease stages at the time of radical prostatectomy [8, 15]. The increase in pT3 rates within intermediate and especially low-risk categories suggests a shift towards higher-risk PCa among men undergoing surgery, both across and within risk groups. In fact, contemporary low-risk patients undergoing RP likely have more adverse features than in the past. These factors might also have contributed to the more than two-fold increase in pathologic node-positive disease over the last 15 years. As a result, fewer patients might be cured with surgery alone, necessitating more multimodal therapy, which adversely impacts functional outcomes.

The notable increase in tumor pathologic upstaging across all NCCN groups is primarily a consequence of the growing discrepancy between stable rates of clinical and rising pathologic T3 stages. This trend persisted despite greater availability and refined expertise with prostate MRI [16]. This observation may reflect MRI’s limited sensitivity in detecting extraprostatic extension [17]. Additionally, persistent use of DRE alone for clinical staging may understage cases with limited extraprostatic extension. Still, such observations merit further investigation as they raise concern for unanticipated upstaging in this population, which is associated with lower rates of Pentafecta (functional, oncologic, and perioperative success) [18].

Gleason upgrading exhibited differing trends across NCCN risk groups that reflect the evolution of PCa management in the last years. The significant elevation in upgrading among low-risk cases suggests a more selective indication of RP for those with unfavorable features within this category, such as high-volume Gleason 6 or elevated PSA density. Also, Gleason 6 (Grade Group 1) tumors are less visible in multiparametric MRI. In other words, trends in NCCN low-risk group indicate a more selective application of surgery to cases in the highest spectrum of risk. Conversely, the minimal reduction in upgrading rates in intermediate and high-risk patients is relevant as they would be expected to rise amidst the major shift to higher-risk disease observed in RPs. This suggests an improved detection of the highest-grade disease within the prostate before surgery, likely due to the expansion of imaging and targeted biopsies. This hypothesis is supported by evidence demonstrating that multiparametric MRI not only improves the detection of clinically significant PCa [19]but also reduces pathologic upstaging in the RP specimen [20].

Rising PSM rates have been previously reported and largely mirror the significant increase in pT3 rates across all risk groups [8, 10]. PSM increased despite the growing adoption and accumulated experience with the robotic platform over the years. Combined with enhanced precision and magnified visualization, one would expect stable or even declining PSM rates, particularly for organ-confined (pT2) disease. Thus, the observed 20% increase in PSM rates in pT2 cases is concerning for surgical decision-making because more conservative approaches, such as fascia and nerve-sparing techniques, might increase the incidence and extension of PSM, potentially risking RP oncological effectiveness. As surgeons may choose to perform less fascia and nerve-sparing procedures, our findings carry implications on functional outcomes as well. Indeed, this was highlighted by a single-center study of 10,000 RPs which showed a decrease in nerve-sparing procedures as pT3 and PSM rates went up [8]. Consequently, the current rates of continence and potency may no longer be comparable to the early robot-assisted RP series [9, 21, 22]. Although still frequently cited as reference standards for functional outcomes, these series contain higher proportions of low-risk and favorable cases. Therefore, our findings suggest that surgical protocols and patient counseling may need to be revised to more accurately reflect these recent changes.

The strengths of our study lie in its detailed information on demographics, cancer features, and postoperative outcomes coupled with a large hospital-based cohort. This allows for a broader understanding of how the RP population has changed over time. While not a population-based registry, NCDB captures more than 70% of cancers diagnosed and treated annually in the United States across more than 1,500 commission on cancer-accredited facilities [12, 23]. This makes it particularly valuable for assessing real-world practices, in contrast to previous studies addressing changes in RP patient profiles from academic centers, known for achieving superior outcomes [24].

The main limitations of our study are linked to the retrospective analysis of large datasets, which inherently includes risks of bias. However, the NCDB is a substantial national dataset that is well-suited for analyzing trends. Another limitation is the 2005 ISUP major review of the Gleason grading system, which experienced a delay in adoption. Consequently, Gleason scores, particularly from the initial years following this review, may not be entirely comparable even after limiting analysis to the period post-modifications [25]. Lastly, NCDB does not capture MRI utilization and whether clinical tumor stage was based on digital rectal examination (DRE) or MRI findings. As MRI has superior local staging accuracy [26], persistent utilization of DRE can lead to rising pathologic upstaging rates particularly when the population is shifting to higher-risk disease.