Results of a randomized phase II trial of intense androgen deprivation therapy prior to radical prostatectomy in men with high-risk localized prostate cancer

Abstract

Purpose:

This multicenter randomized phase 2 trial investigates the impact of intense androgen deprivation on radical prostatectomy (RP) pathologic response and radiographic and tissue biomarkers in localized prostate cancer (NCT02903368).

Materials and Methods:

Eligible patients had a Gleason score ≥4+3=7, PSA >20 ng/mL or T3 disease and lymph nodes <20 mm. In Part 1, patients were randomized 1:1 to apalutamide, abiraterone acetate, prednisone, and leuprolide, (AAPL) or abiraterone, prednisone, leuprolide (APL) for 6 cycles (1 cycle=28 days) followed by RP. Surgical specimens underwent central review. The primary endpoint was the rate of pathologic complete response (pCR) or minimum residual disease (MRD, tumor ≤5 mm). Secondary endpoints included PSA response, positive margin rate, and safety. MRI and tissue biomarkers of pathologic outcomes were explored.

Results:

The study enrolled 118 patients at four sites. Median age was 61 years and 94% of patients had high-risk disease. The combined pCR or MRD rate was 22% in the AAPL arm and 20% in the APL arm (difference: 1.5%, one-sided 95% CI: −11%, 14%, one-sided p=0.4). No new safety signals were observed. There was low concordance and correlation between post-therapy MRI-assessed and pathologically-assessed tumor volume. PTEN-loss, ERG positivity and presence of intraductal carcinoma were associated with extensive residual tumor.

Conclusion:

Intense neoadjuvant hormone therapy in high-risk prostate cancer resulted in favorable pathologic responses (tumor ≤5 mm) in 21% of patients. Pathologic responses were similar between treatment arms. Part 2 of this study will investigate the impact of adjuvant hormone therapy on biochemical recurrence.

Introduction:

Fifteen percent of newly-diagnosed localized prostate cancer is high-risk.(1) Despite definitive treatment, many patients develop recurrence; 15-year prostate cancer-specific mortality is 22–38% for patients with prostate specific antigen (PSA) >20 ng/mL, Gleason 8–10, or cT3 disease.(1, 2)

Neoadjuvant systemic therapy is a standard of care for many solid tumor malignancies. It is associated with improved survival, surgical down-staging, and provides the opportunity for in vivo assessment of response.(3–5) Neoadjuvant therapy followed by radical prostatectomy (RP) has been investigated in prostate cancer. Historic trials investigated luteinizing hormone releasing hormone (LHRH) agonists +/− first generation antiandrogens.(6) These trials included a minority of patients with high-risk disease, did not systematically evaluate pathologic responses, and had limited long-term follow-up. The advent of more potent hormonal agents including abiraterone acetate(7–10) (henceforth abiraterone), a CYP17 inhibitor, and apalutamide(11, 12), a next generation androgen receptor (AR) antagonist, provides an opportunity to investigate these agents neoadjuvant to RP.

We conducted a series of phase 2 studies investigating differing durations and combinations of novel hormonal agents that have built our understanding of the neoadjuvant paradigm in prostate cancer(Table 1).(13–16) These studies included systematic central pathology review and embedded intraprostatic pharmacodynamic biomarkers to optimize the appropriate neoadjuvant regimen. These studies demonstrated that: 1) prostate androgens are profoundly lower with abiraterone versus LHRH agonist(16); 2) non-castrating therapy was insufficient to produce pathologic responses(14); and 3) abiraterone, prednisone, enzalutamide and leuprolide resulted in numerically (though not statistically significant) higher pathologic responses compared to enzalutamide and leuprolide(13).

Table 1.

Overview of modern clinical trials of intense neoadjuvant androgen deprivation therapy prior to radical prostatectomy.

Clinical Outcomes	NeoAbi(16)		NeoEnza(14)		NeoAbiEnza
	12wA (n=27)	24wA (n=29)	E (n=25)	EDL (n=23)	ELAP (n=50)	EL (n=25)
Median PSA pre-RP visit, ng/mL	0.060	0.040	0.51	0.040	0.030	0.020
≥ypT3	59% n=16	48% n=14	72% n=18	61% n=14	50% n=25	56% n=14
Positive Nodes	11% n=3	24% n=7	4.0% n=1	26% n=6	10% n=5	12% n=3
Positive Margins	19% n=5	10% n=3	16% n=17	22% n=13	18% n=9	12% n=3
pCR (%)	3.7% n=1	10% n =3	0% n=0	4.3% n=1	10% n=5	8.0% n=2
MRD (largest cross-sectional dimension ≤ 5 mm)	0% n=0	14% n=4	-	-	20% n=10	8.0% n=2
MRD (largest cross-sectional dimension ≤ 3 mm)	-	-	0% n=0	13% n=3	-	-
pCR or MRD (largest cross-sectional dimension ≤ 3 or ≤ 5 mm)	3.7% n=1	24% n=7	0% n=0	17% n=4	30% n=15	16% n=4
RCB ≤0.25 cm3	44% n=12	52% n=15	36% n=9	74% n=17	68% n=34	68% n=17

A=Abiraterone; E=Enzalutamide; EDL=Enzalutamide, dutasteride, leuprolide; ELAP=Enzalutamide, leuprolide, abiraterone, prednisone; EL=Enzalutamide, leuprolide; PSA=Prostate specific antigen; RP=Radical prostatectomy, pCR=Pathologic complete response; MRD=Minimum residual disease; RCB=Residual cancer burden.

While pathologic response has been used as a surrogate for long-term survival in other malignancies, the utility in prostate cancer remains under investigation. To assess the impact of pathologic response on recurrence rates, we conducted a pooled analysis of contemporary neoadjuvant clinical trials.(17) In this analysis of 72 patients, no patient with a pathologic complete response (pCR) or minimum residual disease (MRD) experienced PSA recurrence.(18) A subsequent analysis of a larger cohort (n=117) demonstrates that pCR/MRD continues to predict freedom from PSA failure.(19)

We designed this multicenter, open-label, randomized phase 2 trial of apalutamide, abiraterone, prednisone, and leuprolide compared to abiraterone, prednisone, and leuprolide for 6-months followed by RP (NCT02903368). Herein (Part 1), we report the effect of neoadjuvant therapy on pathologic responses, and imaging and tissue biomarkers. Part 2 (not reported) investigates the impact of an additional year of adjuvant therapy on 3-year biochemical progression-free survival.

Methods:

Patients:

Eligible patients had histologically confirmed prostatic adenocarcinoma and were RP candidates based on urologic oncologist assessment. Patients had: 1) Gleason ≥4+3=7, or 2) Gleason score 3+4=7 with PSA >20 ng/mL or T3 disease [by clinical staging or multiparametric prostate magnetic resonance imaging (mpMRI) T2 imaging]. Patients had ≥3 biopsy cores positive for cancer, tumor >1 cm on mpMRI, or T3 on mpMRI. Evidence of metastatic disease on radiologic imaging was not permitted and lymph nodes were required to be ≤20 mm as assessed by local radiology review (Supplementary Data). All patients provided written informed consent.

Design and Treatment:

This was a multicenter, phase 2, randomized trial with two parts: a neoadjuvant component (Part 1) and an adjuvant component post-RP (Part 2). During Part 1, patients were randomly assigned 1:1 to apalutamide (240 mg/day orally), abiraterone (1000 mg/day orally), prednisone (5 mg/twice daily orally), and leuprolide (22.5 mg every 12 weeks intramuscularly) (AAPL) versus abiraterone (1000 mg/day orally), prednisone (5 mg/day orally) and leuprolide (22.5 mg every 12 weeks intramuscularly) (APL) for 24 weeks followed by RP, stratified by disease risk (intermediate versus high, defined by National Comprehensive Cancer Network (NCCN) criteria).(20) Participating institutional review boards approved the study (Supplementary Data).

Pathology and Immunohistochemistry:

A pathologist (RL) blinded to patient outcomes centrally reviewed all RP specimens. pCR and MRD (largest cross-sectional dimension of residual tumor measuring 1–5 mm) were tabulated. Residual cancer burden (RCB) was measured as the calculated tumor volume corrected by tumor cellularity(21). American Joint Committee on Cancer 8th edition was used for surgical staging.(22) Presence of intraductal carcinoma was assessed. RP specimens were analyzed for AR, ERG, PTEN, Ki67, and programmed death ligand 1 (PD-L1) immunohistochemistry (IHC) using established assays (Supplementary Data) and were centrally reviewed by a blinded pathologist (HY).(13) Semi-quantitative scores were used to evaluate IHC stains (Supplementary Table 1).

Multiparametric Prostate Magnetic Resonance Imaging:

All patients underwent a mpMRI at baseline and a subset post-neoadjuvant therapy prior to RP. A blinded radiologist (FF) centrally reviewed all mpMRI studies (Supplementary Data).

Statistical Analysis:

The primary endpoint for Part 1 was pCR or MRD rate at RP with planned reporting following Part 1. The sample size of 120 patients provided 81% power to distinguish a pCR or MRD rate of 35% in AAPL treated patients (n=60) from 15% in APL treated patients (n=60) using a Chi-square test for binomial proportion with one-sided type 1 error of 0.05. Being consistent with the design, the primary comparison between arms was based on the chi-square test for binomial proportion, with one-sided p-value of ≤ 0.05 considered statistically significant. The difference in response rate was presented with one-sided 95% confidence interval. Secondary endpoints include PSA response, pathologic staging, positive margins, intra-, peri-, post-operative adverse events, safety, and mpMRI and tissue-based biomarkers associated with pathologic outcomes (Supplementary Data). We conducted a descriptive exploratory analysis to evaluate the predicted pathologic RP outcomes using the Memorial Sloan Kettering Cancer Center (MSKCC) pre-RP nomogram (Supplementary Data).

Results:

Baseline Characteristics:

Overall, 119 patients were enrolled from 11/2016–11/2018 at four institutions. One patient withdrew prior to treatment initiation, leading to analysis of 118 patients (Figure 1). The median age was 61 (range 46–72) years. Patients had NCCN intermediate (n=7, 5.9%) or high-risk (n=111, 94%) disease. Baseline characteristics were balanced between arms (Table 2). Treatment exposure is delineated in Supplementary data.

Figure 1.

Consort diagram. Newly diagnosed unfavorable intermediate and high-risk patients with prostate cancer were randomly assigned 1:1 to abiraterone, apalutamide, leuprolide, and prednisone versus abiraterone acetate, leuprolide, and prednisone for 24 weeks followed by radical prostatectomy. *Three patients discontinued treatment due to toxicity after 1, 2, and 4 cycles of therapy.

Table 2.

Baseline clinicopathologic characteristics.

	AAPL (N=59)		APL (N=59)
	N	%, Median (Range)	N	%, Median (Range)
Age at Study Entry, years	59	62 (47–72)	59	58 (46–72)
Time from Diagnosis to Study Entry, months	59	1.9 (0.82–6.2)	59	2.0 (0–6.7)
ECOG Performance Status
0	58	98%	59	100%
1	1	1.7%	-	-
Race
White	55	93%	48	81%
Black or African American	3	5.1%	6	10%
Asian	-	-	2	3.4%
Other	1	1.7%	3	5.1%
Clinical T Stage at Diagnosis
< T3	27	46%	18	31%
≥ T3	32	54%	41	70%
Number of Biopsies Obtained	58	12 (6–21)	58	12 (1–16)
Number of Positive Biopsy	57	7 (2–14)	58	7 (1–15)
Percent Positive Biopsy Cores	57	67% (17%−100%)	58	68% (17%−100%)
Biopsy Gleason Score
7	19	32%	15	25%
8	13	22%	22	37%
9	27	46%	21	36%
10	-	-	1	1.7%
Biopsy Primary + Secondary Gleason Score
3+4*	6	10%	3	5.1%
3+5	-	-	1	1.7%
4+3	13	22%	12	20%
4+4	13	22%	21	36%
4+5	23	39%	16	27%
5+4	4	6.8%	5	8.5%
5+5	-	-	1	1.7%
PSA at Study Entry, ng/mL	50	7.1 (1.0–75)	25	8.6 (2.1–78)
<10 ng/mL	35	59	27	46
10–20 ng/mL	13	22	15	25
>20 ng/mL	11	19	17	29
NCCN Risk Group **
High	56	95%	55	93%
Intermediate	3	5.1%	4	6.8%

AAPL= Apalutamide, abiraterone, prednisone, leuprolide; APL=Abiraterone, prednisone, leuprolide; ECOG=Eastern Cooperative Oncology Group; PSA=Prostate specific antigen; NCCN=National Comprehensive Cancer Network; RCB=Residual cancer burden

^*These patients had T3 disease or PSA>20 ng/mL.

^**NCCN high risk group defined as Gleason ≥8, or PSA>20 ng/mL, or clinical T3 stage or higher.

PSA Outcomes:

Median pre-RP PSA nadir for patients receiving AAPL was <0.01 ng/mL and 0.02 for APL (Supplementary Data, Supplementary Table 2).

Pathologic Outcomes at RP:

Overall, 114 patients had RP pathologic assessment, excluding the four who withdrew prior to RP. Institutional pathology review reported 15 patients with pCR and 22 with MRD. Central pathology review confirmed 13 patients with pCR and 11 with MRD. The centrally-assessed pCR or MRD rate was similar between arms [22% (n=12/55) in AAPL arm versus 20% (n=12/59) in APL arm, one-sided p=0.4](Table 3). The observed difference in pCR or MRD was 1.5% (one-sided 95% CI: −11%, 14%), which was not clinically meaningful. The positive surgical margin rate was numerically lower in the AAPL versus APL arm (7.3% versus 12%, respectively). All patients with positive surgical margins had ypT3 (11/61, 18%) and no patient with a positive margin or ypT3 disease had MRD.

Table 3.

Pathologic outcomes at radical prostatectomy (based on central review).

	AAPL (n=55)		APL (n=59)
	N	%, Median (Range)	N	%, Median (Range)
ypT Stage
T0	7	13	6	10
T2	21	38	19	32
T2 (no sub class)	2		0
T2a	9		7
T2b	1		1
T2c	9		11
T3	27	49	34	58
T3a	12		18
T3b	15		16
Pathology N Stage
N0	51	93%	49	83%
N1	4	7.3%	10	17%
Positive Surgical Margins
No	51	93%	52	88%
Yes	4	7.3%	7	12%
Extracapsular Extension
No	30	55%	26	44%
Yes	25	45%	33	56%
Positive Seminal Vesicle Invasion
No	40	73%	43	73%
Yes	15	27%	16	27%
Pathologic Response(Largest cross-sectional dimension)
pCR	7	13%	6	10%
MRD (≤5 mm)*	5	9.1%	6	10%
pCR or MRD	12	22%	12	20%
0.6–0.9	5	9.1%	7	12%
1.0–1.9	24	44%	15	25%
>=2	14	26%	25	42%
Total Tumor Volume (cc)	55	0.46(0–26)	59	0.84 (0–14)
Percent Cellularity	55	5% (0–80%)	59	5% (0–50%)
RCB (cm3) **	55	0.023 (0–7.8)	59	0.075 (0–6.8)

AAPL=Apalutamide, abiraterone, prednisone, leuprolide; APL=Abiraterone, prednisone, leuprolide; pCR=Pathologic complete response; MRD=Minimum residual disease; RCB=Residual cancer burden.

^*Minimum residual disease was defined as residual tumor in the radical prostatectomy specimen measuring ≤5 mm.

^**Residual cancer burden calculated as tumor volume (cm³) x percent cellularity.

The majority of patients had significant residual tumor, with ypT3 in 49% and 58% of patients in the AAPL and APL arms, respectively. Rates of lymph node involvement were numerically lower in the AAPL arm compared to the APL arm (7.3% versus 17%, respectively). No patient with lymph node involvement had pCR/MRD. Pathologic responders at RP (pCR or MRD) more likely had baseline <T3 stage (75% versus 29%), lower percent of positive biopsies (median 42% versus 73%), and lower maximum percent tumor involvement in biopsy (median 75% versus 90%) compared to non-responders (Supplementary Table 3).

MSKCC Pre-RP Nomogram Prediction:

The predicted rates of organ confined disease, extracapsular extension, seminal vesicle invasion, and lymph node involvement for matched patients undergoing RP alone were 18%, 80%, 40%, and 42%, respectively (Supplementary Table 4). For reference, observed rates of organ confined disease, extracapsular extension, seminal vesicle invasion, and lymph node involvement from our cohort were 47%, 51%, 27%, and 12%, respectively.

Adverse Events:

Treatment was well-tolerated and there were no grade 4–5 adverse events. Treatment-related adverse events were comparable between the arms with the exception of increased any grade and grade 3 maculopapular rash which was more common in the AAPL arm compared to the APL arm (19% versus 0%) (Supplementary Table 5 and 6). Thirteen patients (11%) experienced a grade 3 treatment-related adverse event (n=8 in AAPL, n=5 in APL). Three patients discontinued treatment prior to RP given toxicity (Supplementary Table 7).

Intraoperative complications (n=2) (Supplementary Table 8) and in-hospital complications (n=2) (Supplementary Table 9) were minimal. Post-operative complications were low and similar between arms (Supplementary Table 10 and 11).

Association of mpMRI with Pathologic Responses:

Paired baseline and post-neoadjuvant therapy mpMRIs were available on 71 patients. All patients experience a decline in mpMRI-assessed tumor volume from baseline to post-therapy [median percent decline 91% (range decline 17%−100%)]. Post-therapy mpMRI-assessed median tumor and prostate volume measurements were systematically lower than pathologically-assessed volumes (Supplementary Table 12). There was low concordance and correlation between mpMRI-assessed and pathologically-assessed tumor volume (Supplementary Figure 1A). While concordance between mpMRI-assessed and pathologically-assessed prostate volumes was weak, mpMRI-assessed prostate volume correlated with pathologically-assessed prostate volume (Supplementary Figure 1B). Additionally, mpMRI had moderate specificity for determining the absence of a pCR (80%) and low sensitivity for determining presence of a pCR [one of nine (11%) of pCR identified on mpMRI] (Supplementary Table 13). There was weak correlation with mpMRI-assessed tumor volume and pathologically determined RCB and percent tumor cellularity (Supplementary Figure 2A and Figure 2B). The percent change in tumor volume, from baseline to post-therapy mpMRI, was not associated with RP pathologic response (MRD/pCR) or pathologic T stage (Figure 2A, Figure 2B, Supplementary Figure 3).

Figure 2.

A: Representative images of multiparametric prostate MRI pre- and post-neoadjuvant therapy on patient 20. The patient was a 67 year old male at baseline with intermediate-risk disease (Gleason 4+3, PSA 8.3 ng/mL, cT2). Prior to neoadjuvant therapy, on T2 weighted imaging (T2WI) the tumor is seen as a large grey area with corresponding black area detected on Apparent Diffusion Coefficient (ADC) and white signal on Diffusion Weighted Imaging (DWI). There is enhancement post-contrast on subtraction Dynamic Contrast Enhanced (DCE) imaging. Following neoadjuvant therapy with abiraterone, prednisone, and leuprolide the tumor is no longer visible on T2WI and there is less apparent ADC, DWI, and subtraction DCE, though not totally resolved. B: Representative images from a RP specimen show invasive carcinoma (IC) and intraductal carcinoma (IDC) on hematoxylin and eosin (HE) staining. Immunohistochemistry staining of IC and IDC components is weakly positive for androgen receptor (AR), negative for PTEN, and positive for ERG. This patient at RP had ypT3aN0 disease with negative margins. The residual tumor measured 1.8 × 1.0 cm with total tumor volume of 1.73 cm³ and 10% cellularity.

Association of IHC and Histology in RP Specimens with Pathologic Response:

IHC was performed on 97 RP specimens. IHC profiles were similar between treatment arms (Supplementary Table 13). Notably, 26% of residual tumors (n=25/97) demonstrated negative or weak AR nuclear staining in ≥50% tumor cells.

PTEN-loss and ERG+ were associated with a lower rate of pCR/MRD (Table 4). Additionally, PTEN-loss and ERG+ were associated with larger residual tumors and higher RCB. Only 2/34 patients with either PTEN-loss or ERG+ and 1/35 with both PTEN-loss and ERG+ experienced a pCR/MRD. Ki-67 proliferation index, tumor PD-L1 expression, and AR nuclear expression were not associated with pCR/MRD, pathologic T stage, residual tumor size, tumor cellularity or RCB (Supplementary Table 14 and 15). There was no association between nuclear AR expression and ERG/PTEN status (Supplementary Table 16).

Table 4.

Association of immunohistochemistry with pathologic outcomes at radical prostatectomy.

	PTEN		ERG		ERG/PTEN
Pathologic Outcomes	Intact/Reduced (N=38)	Loss (N=59)	Negative (N=52)	Positive (N=45)	PTEN Intact/ Reduced + ERG Negative (N=28)	PTEN Intact/Reduced + ERG Positive OR PTEN Loss + ERG Negative (N=34)	PTEN Loss + ERG Positive (N=35)
	N (%) or Median (Range)	N (%) or Median (Range)	N (%) or Median (Range)	N (%) or Median (Range)	N (%) or Median (Range)	N (%) or Median (Range)	N (%) or Median (Range)
Pathologic Response
Yes	8 (21)	1 (1.7)	6 (12)	3 (6.7)	6 (21)	2 (5.9)	1 (2.9)
No	30 (79)	58 (98)	46 (88)	42 (93)	22 (79)	32 (94)	34 (97)
Pathologic T Stage
T2	20 (53)	16 (27)	21 (40)	15 (33)	15 (54)	11 (32)	10 (29)
T3a	8 (21)	22 (37)	12 (23)	18 (40)	5 (18)	10 (29)	15 (43)
T3b	10 (26)	21 (36)	19 (37)	12 (27)	8 (29)	13 (38)	10 (29)
Largest cross-sectional dimension (cm)	1.2 (0.10–6.0)	1.9 (0.40–3.9)	1.5 (0.10–6.0)	2.0 (0.40–3.9)	1.1 (0.10–6.0)	1.6 (0.40–3.5)	2.1 (0.40–3.9)
Percent cellularity	5 (1–80)	10 (1–40)	5 (1–80)	10 (1–50)	5 (1–80)	7 (1–50)	10 (1–40)
RCB (cm3)	0.034 (0–7.8)	0.14 (<0.01–4.7)	0.038 (0–7.8)	0.13 (<0.01–6.7)	0.034 (0–7.8)	0.069 (<0.01–6.7)	0.17 (<0.01–4.7)

RCB=Residual cancer burden.

Bold values are statistically significant.

The presence of intraductal carcinoma was associated with higher pathologic T stage, larger residual tumors, increased tumor cellularity, and higher RCB (Supplementary Table 17). There was a correlation between biopsy Gleason score and ERG status (higher Gleason correlated with ERG+), otherwise baseline T stage, Gleason score, and PSA did not correlate with PTEN, ERG, Ki-67, AR, or PD-L1 status (Supplementary Table 18).

Discussion:

We demonstrate that intense neoadjuvant therapy prior to RP results in exceptional pathologic responses in a subset of patients, and that, the addition of apalutamide to APL does not improve pathologic outcomes. We demonstrate weak correlation between mpMRI-assessed and pathologically-assessed tumor volumes and other pathologic parameters. Additionally, there was a negative association between pathologic outcomes and the presence of intraductal carcinoma, PTEN-loss and ERG expression in residual RP tumors.

While in castration-resistant prostate cancer (CRPC), the combination of abiraterone and enzalutamide has not proven beneficial(23), we evaluated this approach in hormone sensitive disease as the biologic composition of these disease states differ greatly. Our prior study investigating neoadjuvant enzalutamide and leuprolide +/− abiraterone demonstrated a numerically higher pCR and MRD rate of 30% with combination therapy compared to 16% with enzalutamide and leuprolide.(13) As demonstrated in CRPC, it does not appear that more intense hormone therapy combining CYP17 inhibition and AR antagonism was superior to single agent therapy. We did demonstrate that 21% of patients achieved an exceptional pathologic response. Furthermore, we demonstrate that neoadjuvant therapy prior to RP is feasible with limited toxicity and operative and peri-operative complications.

Our data support the design of a current phase 3 trial investigating perioperative leuprolide +/− apalutamide in high-risk localized prostate cancer (NCT03767244). This landmark study will define pathologic responses to neoadjuvant therapy, validate pathologic response as a surrogate for metastasis-free survival,(24) and investigate biomarkers of response and resistance.

While a subset of patients had favorable pathologic responses to neoadjuvant therapy, >50% have ypT3 disease at RP. Correlation of pathologic outcomes with biochemical recurrence and other long term outcomes will be critical to validation of this endpoint for future studies. In a pooled analysis of three neoadjuvant trials, we demonstrate a BCR rate of 8% in patients who achieve an exceptional pathologic response and 51% in all other patients.(19) Part 2 of the study will investigate the impact of therapy on 3-year biochemical progression-free survival.

A key component of this work was to investigate biomarkers of resistance in residual RP tumors. As with our prior neoadjuvant study(13), we demonstrate that PTEN-loss and ERG expression were associated with extensive residual tumors at RP. This is consistent with studies in advanced prostate cancer that show the negative prognostic association of these biomarkers.(25) We posit whether patients with PTEN-loss would be more likely to benefit from either a docetaxel-based or targeted approach such as AKT or PI3K inhibitor neoadjuvant strategy.(26) Post-RP therapy escalation strategies may be warranted for patients with significant residual disease or negative biomarkers. Additionally, work is currently ongoing evaluating predictive biomarkers or response and resistance in baseline prostate biopsy specimens.

We demonstrate that intraductal carcinoma is associated with resistance to neoadjuvant hormone therapy. Prior studies have demonstrated that intraductal carcinoma is associated with increased genomic instability(27) and potentially higher prevalence of germline DNA repair alterations(28). This raises the hypothesis that such tumors may be more responsive to strategies that modulate DNA repair efficiency such as platinum chemotherapy or PARP inhibitors.

A unique aspect of our study was integration of paired mpMRI pre- and post-neoadjuvant therapy. We demonstrate that pre-RP mpMRI systematically underestimated tumor and prostate gland volume.(29) However, while we have previously shown that mpMRI-based tumor volume correlates strongly with whole mount pathology assessment of tumor volume in the treatment naïve prostate, this was not the case post-neoadjuvant therapy where we found limited correlation between mpMRI-assessed and pathologically-assessed MRD/pCR.

Despite the randomized design, several limitations exist which are inherent to phase 2 studies. Given that we did not include a control arm of RP alone, we conducted an exploratory analysis to understand the predicted pathologic outcomes of patients matched to several baseline parameters of the individuals in this analysis using the MSKCC pre-RP nomogram model. While direct comparisons cannot be made, encouraging pathologic outcomes were observed. The primary endpoint for Part 1 was pathologic response, which has not been validated as an endpoint associated with improved survival in prostate cancer.

In summary, given suboptimal outcomes of patients with high-risk prostate cancer, alternative approaches that integrate multidisciplinary strategies are warranted to improve survival for patients with aggressive prostate cancer. Neoadjuvant intense androgen deprivation therapy followed by RP is a promising strategy for a subset of patients with high-risk prostate cancer. Tissue and imaging-based biomarkers of responses and resistance will be critical for selecting patients for therapy. Refining selection criteria, beyond currently utilized clinicopathologic parameters, will allow for more optimized personalization and potentially efficacy of therapy.

Abbreviations:

PSA

Prostate specific antigen

RP

Radical prostatectomy

LHRH

Luteinizing hormone-releasing hormone

AR

Androgen receptor

pCR

Pathologic complete response

MRD

Minimum residual disease

mpMRI

Multiparametric prostate magnetic resonance imaging

AAPL

Apalutamide, abiraterone, prednisone, leuprolide

APL

Abiraterone, prednisone, leuprolide

NCCN

National Comprehensive Cancer Network

RCB

Residual Cancer Burden

PD-L1

Program death ligand 1

IHC

Immunohistochemistry

CI

Confidence interval

MSKCC

Memorial Sloan Kettering Cancer Center

CRPC

Castration resistant prostate cancer