Risk factors for Gleason score upgrade from prostate biopsy to radical prostatectomy

Accurate identification of prostate cancer Gleason grade group remains an important component of the initial management of clinically localized disease. However, Gleason score upgrading (GSU) from biopsy to radical prostatectomy can occur in up to a third of patients treated with surgery. Concern for disease undergrading remains a source of diagnostic uncertainty, contributing to both over-treatment of low-risk disease as well as under-treatment of higher-risk prostate cancer. This review examines the published literature concerning risk factors for GSU from time of biopsy to prostatectomy final pathology. Risk factors identified for Gleason upgrading include patient demographic and clinical factors including age, body mass index, race, prostate volume, and biomarker based assays, including prostate-specific antigen (PSA) density, and testosterone values. In addition, prostate magnetic resonance imaging (MRI) findings have also been associated with GSU. Biopsy-specific characteristics associated with GSU include lower number of biopsy cores and lack of targeted methodology, and possibly increasing percent biopsy core positivity. Recognition of risk factors for disease undergrading may prompt confirmatory testing including repeat sampling or imaging. Continued refinements in imaging guided biopsy techniques may also reduce sampling error contributing to undergrading.


Introduction
Prostate cancer (PCa) is the most commonly diagnosed and second-leading cause of cancer-specific mortality among men in the United States [1].Patients are screened for PCa with serum prostate-specific antigen (PSA), which offers value in both diagnosis and risk stratification [2].Recent technological improvements have altered diagnosis and staging strategies for the disease, including multiparametric magnetic resonance imaging (mpMRI), and prostate-specific membrane antigen (PSMA) positron emission tomography (PET) [3,4].However, prostate needle biopsy remains the standard diagnostic test [5,6].Increasing evidence has supported the use MRI-ultrasound-guided fusion prostate biopsy to improve the accuracy of Gleason score (GS) findings and thus PCa risk stratification [7][8][9].Nonetheless, prostate biopsy only samples a small proportion of the prostate and is prone to error.Despite refinements in biopsy technique, patients commonly experience GS discordance and upgrade between biopsy and final pathology on radical prostatectomy (RP).GS upgrading (GSU), defined as an increase in Gleason grade group (GGG) by ≥ 1 or in combined GG by ≥ 1, may occur in as many as 35% of patients who ultimately undergo RP [10].Modifications to the International Society of Urological Pathology standards for PCa grading have also instituted changes to the reporting of proportions of high-grade disease and tertiary patterns and aggregate reporting of MRI-targeted biopsies to better convey disease risk [11].
The accuracy of GGG assessments from biopsy has assumed greater importance in recent years due to the increasing utilization of conservative treatment options, as well as differences in the intensity and modality of treatment.In particular, disease misclassification at biopsy may cause patients to be inappropriately identified as candidates for conservative management.Such patients who are undergraded are more likely to have adverse pathological features, including extra-prostatic extension and biochemical recurrence [12].GSU, even after adjusting for known preoperative variables, is a strong predictor of biochemical recurrence after local treatment [13].Additionally, undergrading of clinically significant PCa could lead to missed escalation of treatment such as combination androgen deprivation therapy-radiation therapy, or omission of pelvic lymph node dissection at the time of RP, further impacting survival.Conversely, overgrading of GSs can lead to overtreatment of less aggressive PCa cases.
Because of its important link to prognosis, researchers have attempted to characterize risk of GSU.Proper identification of patients at risk of GSU could prompt second-level tests such as additional biopsy, prostate MRI, or genomic testing, and may allow for the development of more personalized PCa treatment decisions.The objective of this review is to discuss the risk factors for GSU upon prostatectomy.

Evidence acquisition
A comprehensive literature search for eligible records was conducted from inception until February 2024.All studies were identified and selected from the PubMed database.We employed a broad search strategy with major search terms that included: "Gleason score", "upgrade*," "downgrade*", "risk", "PSA", "MRI", "multivariate analysis", "odds ratio", etc. which were searched in all fields.The literature search was restricted to the English language and did not include any year restrictions.
Studies were eligible for inclusion if they were (1) original research with experimental design, (2) in the field of urologic oncology, (3) focused on identifying risk factors for GSU, (4) included uni-or multivariate statiscially analysis.Major exclusion criteria included reviews, studies without corresponding full-text such as conference abstracts, ongoing studies, and studies with sample sizes fewer than 50 patients.The final included studies were evaluated in qualitative synthesis.In total, 48 studies were included in this literature review.

Patient demographics and clinical factors
PCa is a spectrum of heterogenous diseases, some of which are so indolent that they may never impact survival in many patients while others may significantly limit life expectancy.Because of this diversity, there has been a push to detail which demographic factors and other patient clinical factors are valuable in aiding the diagnosis of clinically significant PCa.Among these, increasing age, larger body mass index (BMI), race, and lower prostate volume have often been linked to various clinical end points, including incidence, stage at presentation, biochemical recurrence, and mortality [14,15].These risk factors have also been examined as potential risk factors for Gleason upgrading.Table 1 examines identified studies that found significant associations between demographic and clinical factors and GSU.Age has been well described as a risk factor for GSU.In a retrospective analysis of 3,571 men with GG1 disease on transrectal ultrasound (TRUS)-guided biopsy who underwent prostatectomy, Leeman et al. [16] found older age to be a significant predictor of upgrade on surgical pathology (OR 1.05, 95% CI 1.01-1.08,P = 0.005).This complements the findings of a similar analysis by Gershman et al. [17] which, in a retrospective analysis included 1,836 patients with GG1 disease who had prostatectomy.The study reported a similar risk of upgrade on prostatectomy per increased year of age (OR 1.05, 95% CI 1.03-1.07,P < 0.001) [17].However, on subgroup analysis of only those patients age > 60, this association is considerably strengthened (OR 2.31, 95% CI 1.50-3.54,P < 0.001) [17].In an analysis of 7,643 paired prostate biopsy and RP specimen recorded in the Institutional Urology Prostate Cancer Database, Epstein et al. [18] found that men who were upgraded at prostatectomy were on average 2 years older.However, despite the large evidence of an age-upgrade association, the magnitude of this effect size was small, limiting the finding's clinical utility.The introduction of MRI into the PCa diagnostic pathway and further refinements in biopsy technique may reduce lead-time bias and rates of misdiagnosis at initial biopsy with more targeted biopsies.Mazzone et al. [19] examined a cohort of 424 patients receiving systematic and targeted biopsy and found that age was not a significant risk factor for upgrading when MRI-guided biopsy techniques were utilized.
BMI has also been associated with GSU at prostatectomy.Among a sample of 959 urban-residing males, Vora et al. [20] reported a positive association between BMI and GSU (OR 1.04, P = 0.02).Freedland et al. [21], examining the Shared Equal Access Regional Cancer Hospital database, found that obese men (BMI ≥ 30 kg/m 2 ) had 1.89 times greater odds of GSU when compared to non-obese (BMI < 25 kg/m 2 ) (P = 0.003).More recently, this association was also seen independent of biopsy technique [22].Several explanations for this observed association have been offered.First, high BMI may impair proper patient positioning and needle trajectory, making more likely that biopsy insufficiently samples lesions.As such, some studies have suggested that these patients receive more extended biopsy schemes to overcome sampling issues [23].Additionally, the association between obesity and GSU mirrors the association between obesity and PCa prognosis, likely through similar mechanisms.Obesity is closely related to PCa aggressiveness and PCarelated mortality; several mechanisms have been suggested to underly this including tumor proliferation via the insulin-like growth factor-1 (IGF-1) pathway and oxidative stress-related inflammatory signaling [24,25].
Race has also been examined as a risk factor for GSU at the time of RP.However, the role of race in GSU is conflicting and with limited reports.Vora et al. [20] used both univariate and multivariate analysis to identify significant predictors of GSU while controlling for clinical parameters.On multivariate analysis, African American men had 1.74 times increased odds of upgrade at the time of prostatectomy.Additionally, a study conducted by Sundi et al. [26] also concluded that African American men diagnosed with very lowrisk PCa were more likely to experience upgrade and had worse pathologic outcomes than Caucasian patients (27.3% versus 14.4%, P < 0.001).More recently, however, two studies examining the National Cancer Database and Surveillance, Epidemiology, and End Results (SEER) database, each containing samples of more than 10,000 patients, did not observe a significant association between Black versus White race and GSU [27,28].These findings were consistent among individuals with GS 3+3 and GS 3+4 cancer at biopsy.As such, it remains unclear what role GSU plays in this relationship.Recent advances in understanding of racial disparities in PCa epidemiology have suggested that much of these differences stem from differential access to care at each level, from screening to diagnosis and treatment [29].It is possible that in some studies, differences in timeliness and quality of care may contribute to upgrade.Regardless, further research is needed to better characterize this association and its potential causes.
Several studies have examined the influence of prostate size on risk of GSU.An early study by Uzzo et al. [30] suggested that larger glands were associated with disease under-sampling, resulting in an increased risk of GSU.The Prostate Cancer Prevention Trial confirmed these findings among 369 patients with PSA < 10 ng/mL who underwent RP at their institution; TRUS volume was associated with GG4 pattern or greater on biopsy but not at prostatectomy [31].However, in more recent years, several studies have suggested an inverse relationship-patients with smaller glands are more likely to be upgraded [17,[32][33][34][35]. Davies et al. [33] examined 1,251 patients with D'Amico low risk disease and found that men with prostate volumes at the 25th percentile (36 cm 3 ) were 50% more likely to experience upgrade than men with prostate volumes at the 75th percentile (58 cm 3 ).A later study conducted by Gershman et al. [17] found increased risk for combined GS7 or greater disease with decreasing prostate weight even after controlling for age and PSA among a sample of 1,836 patients with GS6 on initial prostate biopsy.Additionally, Pierorazio et al. [34] found that while smaller volume was associated with upgrading, larger volume was associated with downgrading at prostatectomy.
Several theories have been offered to explain varied findings between small prostates and GSU.In the last decade, advancement in biopsy technology and technique have made possible more targeted and extended biopsy schemes that can improve sampling accuracy in larger glands.Additionally, there is increasing evidence to suggest that PCa found in smaller glands may harbor distinct, more aggressive biology leading to adverse outcomes.One study has suggested that PCa found in the smaller glands are more likely to be androgen-independent, a marker of advanced disease manifestation [35,36].Additionally, larger prostates could also serve as physical barriers for the growth of PCa, reducing the risk that tumor extends beyond the gland [37].However, some studies have suggested that the association between small prostate gland size and adverse clinical outcomes is due to lead-time bias, as improved PSA-based detection is better at identifying PCa in larger prostates, which generally secrete more PSA [38].However, Freedland et al. [35] point out that this association persists even after exclusion of cT1c biopsy PCa, cases most likely to be subject to lead-time bias.Ultimately, prostate size may affect Gleason upgrading, but more studies are needed to better establish the mechanism underlying this relationship.

Tumor markers
PSA is a widely used biomarker for the detection and monitoring of PCa.Elevated baseline PSA levels are predictive of advanced PCa diagnosis and future cancer mortality [2,39].Despite its diagnostic and prognostic value, the use of PSA as a screening measure has been controversial and has contributed to overtreatment of non-aggressive PCa.Nonetheless, several studies have concluded that increasing PSA before initial biopsy is a significant and independent predictor of GSU [40].Table 2 identifies these studies in addition to those that examine other tumor marker-based assays.Importantly, the relationship between elevated PSA and GSU remains significant after stratifying for low-risk, favorable-intermediate, and unfavorable intermediate disease found at biopsy [40].PSA is also an independent predictor of GSU regardless of the number of biopsy cores and technique used to target biopsy (including MRI-fusion biopsy) [41].Given that this relationship between PSA and GSU persists after controlling for several clinical and demographic factors, PSA continues to serve as a powerful prognostic marker in the contemporary age.While PSA is elevated among patients with PCa, PSA is also closely correlated with prostate gland volume.As such, PSA density (PSAD) has emerged as a useful correction factor that can improve diagnostic specificity for clinically significant prostatic disease [42].PSAD is measured as the ratio of serum PSA concentration to prostate volume measured by TRUS or MRI.A 2020 study of 377 individuals with biopsy GS 3+4 revealed that a PSAD ≥ 0.475 ng/mL 2 was significantly associated with upgrade at prostatectomy independent of PSA values (P < 0.001) [43].Additionally, Maruyama et al. [44], using both univariate an multivariate analysis, identified 403 patients who underwent prostatectomy at their institution, PSAD was significantly associated with upgrade from GG1 to > GG1 disease (OR 1.10).Corcoran et al. [45] also reported that PSAD was the strongest predictor of GSU between initial biopsy and RP specimen analysis.However, they also found that PSAD lost its predictive ability as prostate volume increased [45].While many of the studies described here report PSAD based on TRUS findings, Lai et al. [46] report that the relationship between PSAD and upgrading persists among patients who receive MRI/TRUS fusion biopsy.
PCa is also associated with a reduction of the percent of PSA that is unbound in the serum [47].As a result, percent free PSA (%fPSA) has been clinically incorporated to assess an individual's cancer risk.Visapää et al. [48] report significant GSU among patients with GS5 or GS6 that had low %fPSA.Recently, the prostate health index has emerged as a commercially available test that combines total PSA, %fPSA, and the p2PSA isoform into a single score with superior specificity for clinically significant or aggressive PCa [49].While there is limited research exploring its utility to predict GSU, a recent retrospective study analyzed patients with PSA ≥ 3 ng/mL and found that Prostate Health Index (PHI) values ≥ 55 were significant predictors of GSU in RP specimens (OR 3.64, P = 0.04) [50].
Neutrophil-to-lymphocyte ratio (NLR) is a marker of progression of cancer-related inflammation among various types of cancer [51].Lymphocytes are the dominant mediator of inflammation in early tumorigenesis.With increasing stage, neutrophils systemic involvement of disease is correlated with systemic inflammation driven by neutrophilic infiltration, and as such, higher NLR is associated with more advanced disease.In PCa, NLR has been shown to be associated with more aggressive disease [52].Gokce et al. [53] retrospectively analyzed 210 low-risk PCa patients eligible for AS, who ultimately underwent RP.GSU was significantly more common among patients with an NLR ≥ 2.5 (P = 0.04) but had similar rates of upstaging.Ferro et al. [54] later extended these findings to include platelet and eosinophil-to-lymphocyte ratios as predictors of upgrade.Taken together, these tests may play a role in assessing risks of disease undersampling and pathologic upstaging.
PCa development and progression is androgen dependent.Several retrospective and prospective cohort studies have shown an association between low serum total testosterone (TT) and high-grade disease, extraprostatic extension (EPE), and seminal vesicle invasion [55,56].Pichon et al. [57] prospectively analyzed 937 individuals who underwent prostatectomy at their institution and found that 20.1% of individuals had GSU while 11.6% were upgraded among normal TT patients (P = 0.002).Patients with low TT are subject to more aggressive disease, and more timely definitive treatment is likely warranted in this group to prevent upgrade.Additional studies can help to confirm these early findings and better elucidate the relationship between low TT and GSU among patients without castration-resistant PCa.

Pre-biopsy imaging
MpMRI has become a central tool in the diagnosis and staging pathway of PCa.Lesions are stratified according to the Prostate Imaging-Reporting and Data System version 2 (PI-RADSv2) scoring system, reflecting probabilities of identifying higher grade PCa.As a result, prostate mpMRI has been evaluated as a staging and prognostic tool that may reduce sampling bias associated with prostate biopsy.Table 3 details studies that examined the relationship of MRI findings and GSU. Park et al. [58] demonstrated that PI-RADSv2-based mpMRI helped identify features associated with aggressive PCa, including GS ≥ 7, larger tumor volume, positive extracapsular extension, and seminal vesicle invasion.Indeed, Brembilla et al. [59] found that mpMRI could identify suitable candidates for extended pelvic node dissection by predicting nodal metastasis.In addition, the role of mpMRI appears to also extend to predicting GSU.Several retrospective analyses spanning multiple institutions have shown that patients with PI-RADS 4 or 5 lesions are significantly more likely to be upgraded on RP [6,43,60].Kamrava et al. [61] show that a PI-RADS 5 lesion is the single most important predictor of GSU (OR 10.56, P < 0.01) and that lack of targeted biopsy is an additional predictor of GSU.Song et al. [62] constructed multivariate models based on the results of 443 patients who underwent RP, and found that predictive accuracy of GSU significantly increased with inclusion of PI-RADSv2 score, suggesting its role in risk classification.

Biopsy-specific factors
Table 4 details studies that identified significant associations between biopsy-based parameters and GSU.
As prostate biopsy techniques have been improved over time, the risks of disease under-sampling and misclassification appear to have declined.Several studies prior to the integration of mpMRI and targeted biopsy modalities into the PCa pathway reported that increasing number of biopsy cores improved the accuracy of initial grading [21,[63][64][65][66]. Indeed, sextant and low-number biopsy protocols were most often associated with upgrading in these studies [21,64].Undersampling may be more pronounced in individuals with larger prostate glands [21].Thus, patients who receive limited sextant biopsies are more likely to be subject to sampling errors.However, this risk has likely substantially decreased in the contemporary era due to the integration of targeted biopsy and 12-core systematic biopsy schemas which are now standard of care [67].Separately, other biopsy metrics such as the percent of cores positive for PCa and percent of biopsy core with tumor involvement have been shown to be associated with biochemical progression in men with organ-confined disease [68,69].As such, it is thought that these measures may also be associated with Gleason upgrading.Fu et al. [69] reported that higher percent tumor involvement was a significant independent predictor of GSU among patients with low-risk PCa.These findings have since been confirmed by numerous additional studies [18,20,27,28,41,63,[70][71][72].The explanation for association between Gleason upgrade and percent tumor involvement on biopsy has been attributed to several factors including tumor diversity, as well as disease sampling.More directly, increased percentage of tumor involvement reflects larger tumor burden within the prostate.Glands which harbor large tumor volume may harbor significant genetic and risk heterogeneity, potentially increasing the chance of initial disease misclassification on biopsy.
In contrast, several studies note a decrease in the risk of GSU with increasing core positivity and tumor involvement.A large retrospective study conducted by Evans et al. [73] showed that when tumor accounted for more than 25% of biopsy volume, the concordance between the biopsy and RP pathological specimens was higher as the percentage increased.Additionally, Zheng et al. [22] evaluated predictors of GSU among patients who underwent cognitive fusion biopsy and found that increasing tumor percentage in biopsy was associated with lower risk of upgrading at prostatectomy.Zhang et al. [74] also found that number of positive cores ≥ 3 was associated with lower risk of upgrading (OR 0.435, P = 0.04).These studies cite the use of targeted or extended biopsy modalities which improve the precision and accuracy of each biopsy to effectively reflect the true pathological state of the prostate.

Impact and future directions
In the modern PSA era, the number of patients diagnosed with low-risk PCa (clinical T1c or T2a or GS ≤ 6), has risen substantially [75].To limit overtreatment in patients with clinically indolent PCa, less invasive management strategies including active surveillance have been increasingly used to delay definitive management [76].However, as was most notably discovered by D'Amico et al. [77], roughly 40% of men with GS ≤ 6 at biopsy may have occult, higher-grade disease at the time of prostatectomy.The management and prognosis of PCa is heavily driven by GS.Despite improvement in biopsy approaches, difference between biopsy and postoperative pathology remains an enduring source of uncertainty in the management of localized PCa [78].Gleason upgrading can lead to unnecessary delays in treatment while downgrading can lead to overtreatment, impacting a patient's quality of life and healthcare expenditure [69].As such, understanding factors that underly the discrepancy between biopsy and prostatectomy GS is an area for continued refinement.
Appreciating the possibility of Gleason undegrading contributes to the personalized discussion of PCa decision making.For example, patients with initial "low risk" PCa but with significant risk factors for upgrading including high PSA values and high suspicion MRI scans should be counseled about their risks for disease undersampling.As such, close evaluation including confirmatory testing and repeat prostate biopsy in the setting of risk factors may be warranted.
Improvements in image-guided biopsy modalities can also reduce sampling error that miss higher Gleason grade PCa.Prostate biopsy has traditionally been guided with conventional ultrasound.The use of mpMRI and MRI-TRUS fusion biopsy has led to improvements in the diagnostic assessment for clinically significant PCa, but possibilities of undersampling remain as noted by the presence of GSU on final pathology, especially for PI-RADS 4-5 lesions [79,80].Conversely, the addition of mpMRI to the PCa diagnostic pathway has caused risk inflation, often doubling the proportion of National Comprehensive Cancer Network (NCCN) high-risk patients identified at biopsy [81].In fact, the use of MRI-targeted biopsy has been associated with increased proportions of downgrading on final pathology, possibly due to sampling bias in heterogenous lesions [82].
Because of these limitations, advancements in imaging have been offered and are being evaluated to improve the accuracy of clinical disease sampling.For example, micro-ultrasound (microUS) technology represents a novel technology that may further improve diagnostic accuracy of prostate biopsy.Compared to TRUS with 6-9 Mhz, microUS operates at 29 Mhz and provides a threefold improvement in spatial resolution [83].Indeed, several studies have shown that the use of microUS alone is non-inferior to MRI-TRUS fusion biopsy for clinically significant Pca (csPCa) detection [84].Additionally, microUS may also display improved sensitivity while retaining similar specificity when compared to MRI-TRUS [85][86][87][88].Large randomized controlled trials such as the optimization of prostate biopsy-micro-ultrasound versus MRI (OPTIMUM) trial are ongoing to assess the effectiveness of MRI/microUS fusion biopsy [89].To our knowledge, no study has yet examined rates of GSU in patients biopsied via microUS when compared to other imaging guided techniques.
In addition to advances in imaging, genomic classifiers, such as the Decipher classifier (GenomeDx Biosciences, Vancouver, BC, Canada), have recently been developed as a tissue-based platform to evaluate the expression of 22 genes reflecting pathways involved in cellular proliferation, differentiation, immune modulation, and androgen-receptor signaling [90].The Decipher classifier has been shown to provide robust predictions of disease outcome among patients with high-risk PCa [90].However, little is known of Decipher's ability to estimate the trajectory of untreated low or favorable-risk PCa.One study conducted by Press et al. [91] has shown that elevated Decipher score is associated with short-term biopsy upgrading among patients on active surveillance.Another study by Sheng et al. [92] report that Decipher can independently predict upgrading at prostatectomy.As such, the Decipher genomic classifier may represent a novel approach to predicting GSU and prognostic value in PCa, but larger retrospective studies are necessary to confirm these findings.

Table 1 .
Included studies: patient demographic and clinical factors