Should all prostate needle biopsy Gleason score 4 + 4 = 8 prostate cancers be high risk? Implications for shared decision-making and patient counselling

Abstract

Objectives:

To estimate the probability of downgrading to Gleason score ≤7 at radical prostatectomy for men with a prostate needle biopsy demonstrating Gleason score 8 (4 + 4).

Methods:

This is a retrospective review of men with Gleason score 8 (4 + 4) prostate cancer on needle biopsy who then underwent a radical prostatectomy at the Karmanos Cancer Institute or the University of Michigan. Men with any pattern 5 on the diagnostic biopsy were excluded. The objective was to estimate the proportion of patients whose tumors were downgraded to Gleason score ≤7 at radical prostatectomy and to identify clinical and biopsy parameters associated with downgrading.

Results:

Median age of our cohort was 63 years (IQR: 59, 67.5) and median follow-up was 15 months (IQR: 7, 37). Of the 105 men that met inclusion criteria, 59% (62/105) were downgraded to Gleason score ≤7 at radical prostatectomy. Having ≤2 cores demonstrating Gleason score 8, ≤50% maximal tumor involvement of any individual core positive for Gleason score 8, or the presence of Gleason pattern 3 (such as 3 + 4, 4 + 3, or 3 + 3) in other biopsy cores were all independently associated with downgrading in our multivariable model. Depending on the absence, presence, or combination of these 3 factors, patients had an estimated 6% to 82% probability of having their tumor downgraded at radical prostatectomy.

Conclusions:

Men with low volume Gleason 8 (4 + 4) and/or the presence Gleason pattern 3 on prostate needle biopsy often have their tumors downgraded at radical prostatectomy. The presence of these preoperative biopsy parameters could affect pretreatment counseling and impact patient management. Published by Elsevier Inc.

1. Introduction

For men with newly diagnosed prostate cancer, risk stratification forms the foundation for patient counseling and treatment recommendations. Patients with newly diagnosed Gleason score (GS) 8 (4 + 4) on needle biopsy are typically considered to have high-risk disease, regardless of other data that may be available from the biopsy cores. Imprecision in AUA, NCCN, and EAU guidelines [1–3] has led to the development of risk stratification models such as the Cancer of the Prostate Risk Assessment (CAPRA) and the Memorial-Sloan Kettering nomograms [4,5]. However, these models are based on the biopsy core demonstrating the highest GS; they do not take into account other biopsy parameters. It remains unknown how biopsy disease volume and the presence of lower grade disease in separate biopsy cores affect pathological downgrading at radical prostatectomy in patients with needle biopsy GS 8 (4 + 4).

Counseling regarding duration of androgen deprivation therapy (ADT) for patients electing radiation or decisions regarding nerve sparing approaches for patients electing radical prostatectomy are made based on needle biopsy pathology. The variability of correlating needle biopsy to radical prostatectomy GS has been well documented [6]. To date, most studies have focused on pathological upgrading at radical prostatectomy; few have examined the possibility of downgrading [6–8].

Certain GS 8 (4 + 4) patients, such as those with markedly abnormal digital rectal exam or clinically enlarged pelvic lymph nodes, are easy to identify as truly high risk and can be counselled accordingly. Identifying which Gleason 8 (4 + 4) patients harbor less aggressive prostate cancer remains a challenge. Different levels of risk among needle biopsy GS 8 (4 + 4) patients should affect counseling and shared decision-making. We hypothesized that clinical and pathologic outcomes may be affected by other biopsy parameters including number of cores demonstrating GS 8 (4 + 4), percent Gleason 8 (4 + 4) involvement of individual cores, and presence of lower grade disease in other cores. In order to aid in counseling of patients with biopsy GS 8 (4 + 4), we aimed to describe clinical and biopsy parameters associated with pathological downgrading at radical prostatectomy and freedom from biochemical recurrence.

2. Materials and methods

This is a retrospective study of patients with prostate cancer treated at the Karmanos Cancer Institute, Detroit, Michigan and the University of Michigan, Ann Arbor, MI. The Karmanos cohort includes patients treated from 2002 to 2015 and the University of Michigan cohort includes patients treated from 2006 to 2014. The primary objective of the study was to estimate the proportion of patients with GS 8 (4 + 4) on prostate needle biopsy whose tumors were downgraded to GS ≤7 upon examination of the prostate after radical prostatectomy.

Patients were included in the study if the highest GS on any individual core was 8 (4 + 4), no evidence of nodal or distant metastases on preoperative imaging, and radical prostatectomy was performed. Patients were excluded if radical prostatectomy pathology was not available or if Gleason pattern 5 was present in any of the biopsy cores. Biopsy and radical prostatectomy specimens were processed and evaluated according to established institutional protocols. At each institution, the entire prostate was step-sectioned and submitted for microscopic examination. Patients who had neoadjuvant ADT or prior radiotherapy were excluded. Most patients underwent systematic 12-core biopsies. During the time period of this study, prebiopsy prostate MRI was rarely used. Institutional databases and the electronic medical records were reviewed for pertinent clinical and pathological parameters including age, race, prostate-specific antigen (PSA), CAPRA [4] score (3 patients were omitted due to missing cT stage), number of cores demonstrating GS 8 (4 + 4), maximal percent involvement of any individual core with GS 8 (4 + 4), and the presence of lower grade disease (GS ≤7) in separate biopsy cores.

Categorical variables were compared with the chi-squared test. Parametric continuous variables were compared using the student’s t test and nonparametric continuous variables were compared using the Wilcoxon rank-sum test. Variables with an association with downgrading (P < 0.15) in bivariate analysis were entered into a multivariable logistic regression model to obtain odds ratios associated with downgrading at radical prostatectomy. In order to generate a model to predict the probability of downgrading, statistically significant variables (P < 0.05) from the first multivariable model were used to generate a second multivariable model which was then used to calculate the probability of downgrading. As a secondary outcome, time to biochemical recurrence (BCR) was calculated using the Kaplan-Meier method and survival curves were compared using the log rank test. BCR was defined as a postoperative PSA > 0.2ng/ml [9]. All statistical tests were 2-sided with P < 0.05 considered to be statistically significant. This study was approved by both the Wayne State University and the University of Michigan Institution Review Boards.

3. Results

A total of 105 patients met the inclusion criteria for this study (Table 1). Median age of the cohort at the time of prostate biopsy was 63 years (IQR: 59, 67.5). Median follow-up calculated from date of diagnosis to date of BCR or last follow-up was 15 months (IQR: 7, 37). Median prebiopsy PSA was 6.5 ng/ml (IQR: 4.4, 10.3). Median CAPRA score was 5 (IQR: 4, 7).

Table 1.

Baseline demographics and pathological outcomes (n = 105).

Demographic	Value	IQR or %
Age (median)	63	59–67.5
Race
White	38	36%
AA	22	21%
Unknown	45	43%
Prebiopsy PSA (median)	6.5	4.4–10.3
CAPRA (median)	5	4–7
Number of cores	12	8–12
Median follow-up (mo)	15	7–37
Pathological surgical GS
7	62	59%
3 + 4	44	42%
4 + 3	18	17%
8	22	21%
9	21	20%
Pathological T stage (AJCC) [14]
2a	12	11%
2b	38	36%
2c	34	32%
3a	16	15%
3b	5	5%
pN+	5	5%

GS = Gleason score.

Upon examination of the entire prostate after radical prostatectomy, 62/105 (59%) were downgraded to GS ≤ 7 (4 + 3 or 3 + 4), with 42% (44/105) downgraded to Gleason 4 + 3 = 7 and 17% (18/105) downgraded to GS 3 + 4 = 7. There was no difference in the proportion of patients downgraded at each institution (data not shown). Most patients had organ confined disease with 15% (16/105) having extraprostatic extension, 5% (5/105) with seminal vesicle invasion and 5% (5/105) with lymph node metastasis.

We examined the data for associations between downgrading at radical prostatectomy and preoperative clinical parameters and biopsy characteristics. A higher proportion of prostatectomy specimens were downgraded to GS ≤ 7 at radical prostatectomy when Gleason pattern 3 was present in any of the other biopsy cores (4 + 3, 3 + 4, and 3 + 3) as opposed to only GS 4 + 4 = 8, 66% (n = 51/77) vs. 39% (n = 11/28), P = 0.013 (Table 2A). There was an inverse association between volume of GS 8 (4 + 4) on biopsy and downgrading at radical prostatectomy. Among patients with ≤50% maximal tumor involvement of any individual cores demonstrating GS 8 (4 + 4), 71% (n = 46/65) of prostatectomy specimens were downgraded at radical prostatectomy compared to 40% (n = 16/40) of those with >50% maximal core involvement of Gleason 8 (4 + 4) (P = 0.002). Additionally, 68% (n = 59/87) of prostatectomy specimens were downgraded when GS 8 (4 + 4) was demonstrated on ≤2 biopsy cores compared with 17% (n = 3/18) of prostatectomy specimens with >2 cores of GS 8 (4 + 4) on biopsy (P < 0.001). PSA was weakly associated with downgrading (P = 0.123); race and age were not associated with downgrading. Mean CAPRA score was similar between patients downgraded (5.2) and not downgraded (5.5) at radical prostatectomy.

Table 2A.

Patient characteristics associated with downgrading to RP ≤Gleason score 7.

	RP ≤ GS7 (n)	RP ≥ GS8 (n)	P
All patients	62	43
Race			0.602
White	21 (34%)	17 (40%)
AA	15 (24%)	7 (16%)
Unknown	26 (42%)	19 (44%)
Mean age (SD)	62.7 (6.49)	63.7 (7.48)	0.445
Mean number of total biopsy cores (SD)	11 (3.9)	11.1 (4.8)	0.604
Mean CAPRA (SD)	5.2 (1.1)	5.5 (1.3)	0.394
Presence of pattern 3			0.013
Pattern 3 and 4	51 (82%)	26 (60%)
Pattern 4 only	11 (18%)	17 (40%)
PSA			0.123
<10 ng/ml	50 (81%)	29 (67%)
≥10 ng/ml	12 (19%)	14 (33%)
Maximal % cancer involvement of any GS8 Core			0.002
≤50%	46 (74%)	19 (44%)
>50%	16 (26%)	24 (56%)
# Cores positive for GS8			<0.001
≤2	59 (95%)	28 (65%)
>2	3 (5%)	15 (35%)

Bold values are statistically significant (P<0.05).

In the first multivariable model, the number of cores demonstrating GS 8 (4 + 4), ≤50% maximal tumor involvement of any individual core demonstrating GS 8 (4 + 4), and presence of Gleason pattern 3 remained independently associated with downgrading (Table 2B). To generate a prediction model to estimate the probability of downgrading at radical prostatectomy, a second multivariable model was generated with the 3 significant predictors (Supplemental Table 1). Depending on the absence, presence, or combination of these 3 significant preoperative predictive factors, the probability of downgrading ranged from 6% to 82% (Fig. 1).

Table 2B.

Multivariable logistic regression analysis of factors associated with downgrading to ≤7 at radical prostatectomy.

	OR	95% CI	P
# Cores positive for GS 8
≤2	5.17	1.25–21.5	0.024
Presence of pattern 3
Pattern 3 and 4	4.20	1.47–12.1	0.008
Maximal % cancer involvement of any GS8 Core
≤50%	3.18	1.18–8.56	0.022
PSA
<10	1.9	0.66–5.44	0.234

Bold values are statistically significant (P<0.05).

GS = Gleason score.

Fig. 1.

Estimated probability of downgrading to Gleason score ≤7 at radical prostatectomy. Error bars display 95% confidence interval. GS = Gleason score.

There were a total of 19 BCR events. Not surprisingly, biopsy GS 8 (4 + 4) patients whose tumors were downgraded (GS ≤ 7) at radical prostatectomy had improved BCR-free survival compared with patients whose tumors remained high-risk (GS ≥ 8) (P < 0.001, Supplemental Fig. 1). More importantly, for the purpose of pretreatment counseling and shared decision-making, we wished to determine how preoperative factors affect BCR-free survival. Kaplan-Meier curves stratified by preoperative biopsy disease volume, PSA, and presence of pattern 3 in separate biopsy cores are shown in Fig. 2. There were statistically significant differences in time to BCR in patients with ≤2 biopsy cores of GS 8 (4 + 4) compared with ≥2 biopsy cores of GS 8 (4 + 4) (P < 0.001) as well as in patients with PSA < 10 compared with PSA ≥ 10 (P < 0.001).

Fig. 2.

Kaplan-Meier curves of biochemical recurrence-free survival based on preoperative predictive factors. Number at risk is displayed beneath each time point below the KM curves. Curves are compared with the log rank test. GS = Gleason score. (a) Patients with biopsy pattern 4 only vs. pattern 4 and 3 (P = 0.653). (b) Patients with ≤2 cores of GS 8 vs. ≥3 cores of GS 8 (P < 0.001). (c) Patients with PSA <10 ng/dL or ≥10ng/dL at diagnosis (P < 0.001). (d) Patients with ≤50% maximal tumor involvement of any core demonstrating GS 8 vs. >50% (P = 0.351).

4. Discussion

Biopsy GS is an important parameter to help risk stratify patients with newly diagnosed prostate cancers, and pretreatment counseling depends heavily on the biopsy report. Our data demonstrate that a substantial proportion (59%) of men with biopsy Gleason 8 (4 + 4) cancer are downgraded to GS ≤ 7 at radical prostatectomy. We identified 3 preoperative predictive factors associated with downgrading including ≤2 biopsy cores of GS 8 (4 + 4), ≤50% maximal tumor involvement of the cores demonstrating GS 8 (4 + 4), and the presence of Gleason pattern 3 in separate biopsy cores. The probability of downgrading increased when combinations of these preoperative predictive factors were present.

These data suggest a dichotomy between how patients are placed into risk groups based on prostate biopsy specimens compared with radical prostatectomy specimens. Using biopsy data, each biopsy core is assessed individually and overall risk is assigned based on the core with the highest GS; the other cores are essentially ignored. This contrasts with the postoperative practice where the entirety of the tumor is judged to determine GS and risk stratification.

Pretreatment counseling is based on the biopsy parameters as well as other clinical variables. Our data add more detail to previously published studies documenting the relationship between needle biopsy grade and radical prostatectomy grade [6,8,10,11]. For patients with biopsy GS 8 in previously published series, downgrading is in the range of 47% to 60%. Similar to the current study, Qi et al. found that downgrading was associated with lower PSA, lower percent positive of individual cores, lower number of cores demonstrating GS 8, and the presence of Gleason pattern 3 [12]. Our study adds to these data by focusing only on GS 4 + 4 = 8 (excluding 3 + 5 and 5 + 3), inclusion of a multivariable models to predict the probability of downgrading, and inclusion of the clinical outcome of BCR.

This analysis was designed to help clinicians with pretreatment counseling. A practical implication of our finding is for patients electing radiation therapy. Currently, many high-risk patients electing radiation therapy are advised to undergo 2 to 3 years of ADT [13]. Our data suggest that over half of these patients would be reclassified to intermediate risk if they had a prostatectomy, and almost 20% would be reclassified to favorable intermediate risk. Current guidelines suggest patients with intermediate-risk prostate cancer should receive only 6 months of ADT or potentially radiation without ADT [3,13]. Another clinical implication of these data could be in patients considering a radical prostatectomy. Patients more likely to be downgraded to GS 7 are likely better candidates for an aggressive nerve sparing approach while patients more likely to have final GS 8 prostate cancer may be better suited for a wider excision. Additionally, surgical patients with increased probability of being downgraded may be counselled and reassured that they are less likely to need secondary treatments.

There are several limitations worth noting in our study. This is a retrospective study which allows for the potential for unaccounted confounding. Similar protocols for pathologic review of the biopsy and radical prostatectomy specimens were done at each treating institution without centralized pathologic review. The criteria for pathologic grading evolved during the time frame of this study. In addition, having only 16 BCR events prohibits the inclusion of Cox regression modeling of time to BCR. Due to length of follow-up, we were unable to assess for differences in the development of metastasis, cancer-specific mortality, or overall mortality. Most patients were biopsied prior to wide use of multiparametric MRI and targeted biopsies. It is conceivable that these technologies will reduce grading disparities between needle biopsy and radical prostatectomy.

5. Conclusions

In summary, a predictable subset of patients with GS 8 (4 + 4) on needle biopsy will have less aggressive pathological and clinical outcomes. Presence of certain preoperative biopsy parameters in these patients could affect pretreatment counseling and impact patient management.

Abbreviations:

ADT

Androgen Deprivation Therapy

AUA

American Urological Association

BCR

Biochemical recurrence

CAPRA

Cancer of the Prostate Risk Assessment

EAU

European Association of Urology

GS

Gleason score

IQR

Interquartile range

NCCN

National Comprehensive Cancer Network

PSA

Prostate-Specific Antigen