Efficacy of Treatments for T3bN0M0 Prostate Cancer Patients: A Systematic Review and Meta‐Analysis

ABSTRACT

Background

Patients with prostate cancer (PC) invading into the seminal vesicles (SV) constitute a specific subgroup of high‐risk patients (staged as T3b) with a particularly high risk of disease recurrence, even when lymph node or distant metastases are absent (N0M0).

The aim of the current systematic review and meta‐analysis was to investigate efficacy of available treatments for men with PC invading the SV without metastasis.

Methods

A systematic review was performed according to the Cochrane guidance and reported according to the guidelines for the Preferred Reporting Items for Systematic Reviews and Meta‐analyses. The review protocol was registered prospectively in the PROSPERO database. Studies identified in MEDLINE, EMBASE, and trial registries were selected by two independent reviewers. Treatment modalities included radical prostatectomy with pelvic lymphadenectomy; external beam radiotherapy (EBRT), high‐dose rate or low‐dose rate brachytherapy, androgen deprivation therapy (ADT), or a combination of treatments. Primary outcome measure was 5‐year biochemical recurrence free survival (BRFS).

Results

Twenty‐five studies, comprising 2 881 patients, were included. For EBRT and ADT pooled estimated 5‐year BRFS was 68%, varying from 56% in studies on low‐dose radiotherapy (< 74 Gray [Gy]), up to 78% in studies on high‐dose radiotherapy (≥ 74 Gy). For EBRT with brachytherapy and ADT pooled estimated 5‐year BRFS was 83%. For radical prostatectomy pooled estimated 5‐year BRFS was 32%. When combining surgery with adjuvant treatment (EBRT or ADT) pooled estimated 5‐year BRFS was 54% and 63%, respectively.

Conclusions

These results indicate a role for high‐dose radiotherapy (≥ 74 Gy) and long‐term ADT in T3bN0 PC rather than primary surgical treatment. In addition to dose‐escalated radiotherapy to the entire prostate gland (≥ 74 Gy), selected patients should be offered focal dose‐escalation to the tumor, by high‐dose rate brachytherapy. If proven safe for this patient group, focal dose‐escalation could also be offered with EBRT.

1. Background

Prostate cancer (PC) has emerged as the most common malignancy in men and a major cause of cancer deaths [1]. Approximately 10%–15% of PC patients present with high‐risk features based on T‐stage, Gleason score, or prostate specific antigen (PSA) [2]. Patients with PC invading into the seminal vesicles (SV) constitute a specific subgroup of high‐risk patients (staged as T3b) with a particular high risk of disease recurrence [3].

Traditionally, clinical T‐staging relies on digital rectal examination (DRE), which has been shown to carry a risk of underestimation of non‐palpable T3/4 lesions. In the last decades the multiparametric MRI (mpMRI), which includes T2‐weighted images, diffusion weighted imaging (DWI) and dynamic contrasted‐enhanced (DCE) imaging, has become an important tool in local staging of PC. It has been shown that mpMRI better reflects pathological T‐stage than DRE when diagnosing non‐organ confined disease (T ≥ 3), with a sensitivity between 51% and 59% (MRI) versus 12% and 41% (DRE) and higher cumulative true positive and negative rates (61% vs. 51%) [4, 5, 6].

T3b tumors in itself carry a significant risk of lymph node metastasis with subsequent poor outcome [7]. However, even in patients without lymph node metastases or distant metastases at diagnosis (T3bN0M0) biochemical recurrence rates are as high as 80%, especially after radical prostatectomy [8, 9]. Whether these patients could benefit from adjuvant therapy (radiotherapy and/or androgen deprivation therapy [ADT]) after radical prostatectomy is unclear. Furthermore, the efficacy of primary radiotherapy in T3bN0M0 patients remains to be elucidated. Due to technical advantages, it is increasingly possible to escalate the radiotherapy dose to the tumor in a safe way, both with external beam radiotherapy and brachytherapy (high‐dose rate and low‐dose rate).

In this systematic review we aim to identify the optimal treatment strategy for men with prostate cancer invading the seminal vesicles without lymph node or distant metastases.

2. Material and Methods

The systematic review was performed according to the Cochrane guidance and reported according to the guidelines for the Preferred Reporting Items for Systematic Reviews and Meta‐analyses [10]. The review protocol was registered prospectively at the PROSPERO database (CRD42022327367).

2.1. Search Strategy and Selection Criteria

A literature search was performed in MEDLINE, EMBASE (via Ovid) and trial registries (hand‐search) from January 2000 to September 11, 2024. The following keywords and all known synonyms for these keywords were used: “prostate cancer,” “seminal vesicles,” “T3,” “radiotherapy,” “prostatectomy,” “androgen deprivation therapy.” The search strategy can be found in Supporting Information A. Additional searches were performed by manual cross‐referencing in included studies and systematic reviews. Studies in a language other than English or Dutch were excluded.

Eligible studies investigated the efficacy of one of the following treatment modalities for T3bN0M0 prostate cancer patients:

• –Radical prostatectomy, with pelvic lymphadenectomy, irrespective of surgical technique used (robotic or nonrobotic)
• –External beam radiotherapy (EBRT)
• –Brachytherapy, either low‐dose rate (LDR) or high‐dose rate (HDR)
• –Androgen deprivation therapy (ADT)
• –Multimodality therapy combining two or three of the treatments above

In the selection process, the following deviations from the PROSPERO protocol occurred: Initially, studies were included if the median/mean follow‐up was at least 5 years with at least 50 T3b participants. However, due to a lack of patients identified as T3b in radiotherapy studies, the minimal number of patients was lowered to 30. Second, only studies on photon‐based radiotherapy techniques were included.

Outcome measures included at least biochemical recurrence free survival (BRFS). T3b status was confirmed either by MRI, ultrasound, digital rectal examination or pathological assessment (biopsy or prostatectomy). N0 status was confirmed either by pelvic lymph node dissection (PLND) or imaging (computed tomography [CT], MRI or prostate‐specific membrane antigen (PSMA) positron emission tomography CT [PSMA‐PET/CT]). Studies were included when at least 90% of patients underwent PLND/clinical lymph node staging and were staged as cN0/pN0. Randomized controlled trials and cohort studies were included, whereas case reports and conference abstracts were excluded. In case of overlapping patient cohorts, the study with the largest cohort was included. Two investigators (A.P., C.R.) independently reviewed each study for eligibility based on title and abstract using Endnote 21. Subsequently, the full text of presumably eligible studies was evaluated to decide if the study fulfilled the inclusion criteria. In case of discrepancies, consensus was made after discussion. The agreement on the selection of studies between the investigators was calculated by means of a weighted κ. When randomized controlled trials (RCTs) met the inclusion criteria, but did not publish outcome data on this specific patient subgroup separately, authors were contacted and requested to provide the relevant data.

2.2. Assessment of Study Quality

The ‘Risk of Bias In Non‐randomized Studies of Intervention tool (ROBINS‐I) was used for assessment of methodological quality of the non‐randomized studies by two investigators (A.P., C.R.) [11]. Risk of bias was either assessed as “low risk,”, “moderate risk”, “serious risk,”, “critical risk” or “no information” in the following seven domains: bias due to confounding; bias due to selection; bias in classification of interventions; bias due to deviations from intended interventions; bias due to missing data; bias in measurement of outcomes; bias in selection of the reported result; overall judgment. The following domains were identified as possible confounding variables: Gleason score, age and PSA.

We assessed the risk of bias in randomized studies with the Cochrane's “Risk of bias” tool 2.0. Risk of bias was either assessed as “low risk,” “some concerns,” “high risk,” or “no information” in the following domains: arising from the randomization process, due to deviations from intended interventions, due to missing outcome data, measurement of outcome and selection of reported result [12].

2.3. Data Extraction

A data extraction form was designed including information on study design, method of patient inclusion, patient characteristics (age, body mass index), Gleason score, PSA level, evaluation of tumor and nodal status, treatment modality with detailed information (surgical technique, lymphadenectomy, radiotherapy technique, dose and target volume, type and duration of androgen deprivation therapy) and outcome. In case number of events were only reported graphically (e.g., in Kaplan‐Meier curves) the data were extracted by counting the events and censoring from the graph (C.R.).

2.4. Data Synthesis

Primary outcome was 5‐year BRFS, based on PSA, which varied across studies and treatment modalities. Generally, when clinical failure preceded biochemical failure (e.g., with ADT use), this was considered as failure. For the primary analysis, we determined the event rate and, if necessary, recalculated the (hypothetical) cohort size based on the total number of events in 5 years (C.R.). If available, 3‐, 7‐ and 10‐year BRFS were analysed as well. For studies with more than one study arm (e.g., RCTs), the different experimental study arms were analysed separately, as if they were independent studies.

To correlate outcome to the prescribed radiotherapy dose, the equivalent dose in 2 Gy fractions (EQD2) was calculated:

$$\mathit{EQD}2=D\frac{d+\alpha /\beta }{2+\alpha /\beta },$$

With EQD2 being dose in 2 Gray (Gy) fractions biologically equivalent to dose D given in fractions of size d. α/β is a measure for radiosensitivity of the irradiated tissue (assuming α/β = 1.5 for prostate cancer) [13].

We performed an arm‐level meta‐analysis combining data from single‐armed studies and (randomized) controlled trials. A random‐effects meta‐analysis was used to estimate overall effect sizes and test the influence of moderator variables (including treatment modality) on effect size heterogeneity. Meta‐analyses results are displayed separately for different treatment modalities. No formal statistical comparison was performed to compare results between the different treatment modalities. Results are reported as effect size [upper limit; lower limit 95% CI]. I² statistics and forest plots were used to assess for heterogeneity across studies. Heterogeneity was graduated: null for I² = 0%, minimal for 0% < I² ≤ 25%, low for 25% < I² ≤ 50%, moderate for 50% < I² ≤ 75%, and high for I² > 75%. Statistical software R was used for the statistical analysis (version 4.3.3) with the metafor package (1.9‐9). Data extracted from the studies is available in Supporting information B.

3. Results

3.1. Study Selection

A total of 8503 studies were retrieved, of which 8146 remained after removal of duplicates (Figure 1). Based on title and abstract, 420 studies were relevant for full text screening. Nine authors of prospective trials were contacted for subgroup data. From three prospective trials, we were able to receive data on biochemical recurrence status within the T3bN0M0 subgroup [14, 15, 16]. Three studies were excluded because of overlapping study cohorts (studies with the smallest cohorts were excluded) [17, 18, 19]. No additional papers were identified by cross‐referencing. Eventually, 25 studies (n) reporting on 2746 patients (k) were included in the narrative syntheses [8, 14, 15, 16, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40]. A total of 18 studies reporting on 1 885 patients were included in the meta‐analyses [8, 14, 15, 16, 20, 21, 22, 23, 25, 26, 27, 29, 30, 31, 32, 34, 36, 37]. The agreement of the two reviewers regarding eligibility was 99.8% (κ = 0.25 [0.02; 0.48]).

Figure 1.

PRISMA flowchart.

3.2. Risk of Bias Assessment

For the nonrandomized studies, all showed at least moderate risk of bias, with 13 studies showing high risk of bias in at least one domain (Figure 2, Supporting Information C). In general, the risk of bias resulting from the patient selection procedure was considered low in most studies, as in most studies all consecutive patients during a certain period were included. Most studies were at moderate or serious risk of bias in at least one of the confounding domains. Classification of interventions showed a moderate risk of bias in most studies, as the type and extent of PLND was often not defined, or EBRT doses/techniques varied across the cohort. Risk of bias due to deviations from intended interventions was classified serious risk in studies in which not all patients underwent PLND. Risk of bias due to missing data was classified as serious in case patients did not have sufficient follow‐up. For the domains “measurements of outcomes” and “selection of the reported results” risk of bias was low in most studies.

Figure 2.

Summary plots for the risk of bias for (A) the ROBINS‐I tool for nonrandomized studies and (B) Risk‐of‐bias tool for randomized controlled trials. [Color figure can be viewed at wileyonlinelibrary.com]

For the five randomized studies, risk of bias for “random sequence generation” was low in four studies and “allocation concealment” was low in one study. Risk of bias for “selective reporting” was considered ‘some concerns’ in two studies because they amended the protocol during the inclusion period to mandate (postoperative) PSA collection.

3.3. Study Characteristics and Descriptive Synthesis

Characteristics of the included studies are shown in Table 1. Of the included 25 studies, ten studies investigated the combination of radiotherapy + ADT [14, 15, 16, 20, 21, 22, 24, 29, 36, 41], including EBRT + ADT (n = 6), EBRT + HDR brachytherapy + ADT (n = 3), EBRT + LDR brachytherapy + ADT (n = 1) and LDR brachytherapy + ADT (n = 1).

Table 1.

Baseline characteristics of the included studies.

Radiotherapy and ADT
First author (year)	Study design	Age (median + range/mean + SD)*	Gleason*	PSA (median + range/mean + SD)*	N patients in analysis	Definition BR	Treatment details
Acklin Wehnert 2023	Retrospective	67 (36–80)	47% 6‐7, 53% 8‐10	15.7 (4.4–214.6)	41	Nadir PSA + 2 ng/mL (Phoenix definition)	EBRT: 28×2.5 Gy or 20x3Gy, no elective nodal irradiation. 95% ADT (24 months). EQD2[a/b1.5] 77‐80 Gy
Heemsbergen 2014	RCT	69 (mean) (49–82)	85% 6–7, 15% 8–10*/74% 6–7, 26% 8–10**	39% < 10, 38% 10–20, 24% > 20	38	Nadir PSA + 2 ng/mL (Phoenix definition)	Standard arm: EBRT, 68 Gy, 2 Gy per fraction, no elective nodal irradiation, IMRT/VMAT. EQD2[a/b1.5] 68 Gy. Interventional arm: EBRT, 78 Gy, 2 Gy per fraction, no elective nodal irradiation. EQD2[a/b1.5] 78 Gy. All patients received ADT
Incrocci 2016	RCT	71 (IQR 66–75)	75% 6–7, 25% 8–10	14.6 (IQR 9.2–23.8)	68	Nadir PSA + 2 ng/mL (Phoenix definition)	Standard arm: EBRT, 78 Gy, 2 Gy per fraction, no elective nodal irradiation, IMRT/VMAT. EQD2[a/b1.5] 78 Gy. Interventional arm: EBRT, 64.6 Gy, 3.4 Gy per fraction, 3x/week, no elective nodal irradiation. EQD2[a/b1.5] 90.4 Gy. All patients received ADT
Kerkmeijer 2021	RCT	70 + /−7	67% 6–7, 33% 8–10	15.2 (14.9, standard arm)/16.3 (13.9, boost arm)	82	Nadir PSA + 2 ng/mL (Phoenix definition)	Standard arm: EBRT, 77 Gy, 2.2 Gy per fraction, no elective nodal irradiation, IMRT/VMAT. EQD2[a/b1.5] 81.4 Interventional arm: EBRT, 77 Gy, 2.2 Gy per fraction with integrated boost to visible tumor(s) on MRI up to 95 Gy, 2.7 Gy per fraction, IMRT/VMAT. All patients received ADT. EQD2[a/b1.5] integrated boost up to 113.4 Gy (median dose was 90.5 Gy)
Koutrouvelis 2003	Retrospective	69 (42–90)	91% 6–7, 9% 8–10	10.5 (1.7–143)	39	Nadir PSA + 1.5 ng/mL	LDR‐brachytherapy implantation using 3D stereotactic system and CT‐guidance, Varian BrachyVision, target volume includes 5‐10 mm outside capsule except adjacent to anterior rectal wall, dosage was 120 Gy with 103 Pd or 144 Gy with I‐125. 63% received short‐term neo‐adjuvant ADT
Makino 2021	Retrospective	69 (50–82)	30% 6–7, 70% 8–10	21 (3.7–557.6)	60	Nadir PSA + 2 ng/mL (Phoenix definition)	HDR‐brachytherapy with Ir‐192 afterloading (2×9.5 Gy or 1x13Gy); EBRT 23x2Gy using IMRT (1‐2 weeks after HDR), including internal/external iliac, obturator, anterior sacral nodes; ADT usually 6 months neo‐adjuvant and 2 years adjuvant. Total BED 233 ‐ 246 Gy with a/b ratio 1.5.
Stone 2018	Retrospective	67 (47–80)	55% 6–7, 45% 8–10	47% < 10, 21% 10–20, 32% > 20	53	Nadir PSA + 2 ng/mL (Phoenix definition)	ADT: 3 months neo‐adjuvant, adjuvant 6 months (total 9 months). LDR‐brachytherapy: 103 Pd implantation to the prostate and seminal vesicles. EBRT: 45 Gy/1.8 Gy PLND was performed in patients with seminal vesicle invasion.
Tsumura 2017	Retrospective	72 (53–86)**	54% 6–7, 46% 8–10	46.6 (4.9–481)**	80	Nadir PSA + 2 ng/mL (Phoenix definition)	HDR‐brachytherapy (5×7.5 Gy); EBRT (6 days after HDR) 10x3Gy with dynamic‐arc conformal technique to prostate gland and seminal vesicles; ADT 6 months neo‐adjuvant and 36 months adjuvant (total BED 315 Gy)
Yamazaki 2021	Retrospective	72 (60–89)	45% 6–7, 54% 8–10	42.1 (3.4–398) (EBRT)/31.5 (4.7–486) (EBRT + HDR boost)	249	Nadir PSA + 2 ng/mL (Phoenix definition)	Treatment 1: EBRT 35‐36x2Gy (prostate) and 20‐23x2Gy (whole pelvis), 3D‐CRT; ADT 7.5 months neo‐adjuvant and 20months adjuvant. EQD2[a/b1.5] 70‐72 Gy Treatment 2: EBRT 74 ‐ 80 Gy (prostate) and 46 Gy (pelvic), 3D‐CRT; ADT 11 months neo‐adjuvant and 36months adjuvant. EQD2[a/b1.5] 74‐80 Gy. Treatment 3: HDR‐boost (2x9Gy, 2x10Gy, 3x7Gy, 5×6.3 Gy); EBRT (up to 30 ‐ 51 Gy); ADT 11 months neo‐adjuvant and 36 months adjuvant
Zelefsky 2008	Retrospective	68 (45–84)	70% 6–7, 30% 8–10	34% < 10, 26% 10–20, 40% > 20	166	Nadir PSA + 2 ng/mL (Phoenix definition)	EBRT, 3D‐CRT or IMRT, 66 ‐ 86.4 Gy to prostate, no elective nodal irradiation; ADT 3 months neo‐adjuvant (64%), no adjuvant therapy. EQD2[a/b1.5] 63‐82 Gy.
Radical prostatectomy
Badani 2018	Retrospective	64 (IQR 59–68)	51% 6–7, 49% 8–10	9.1 (IQR 7.2–17.1)	172	PSA ≥ 0.2 ng/mL	RP + PLND (90%), varying techniques, mean number of nodes 7
Bastide 2011	Retrospective	64 (48–75)	72% 6–7, 27% 8–10, 1% unknown	12.5 (mean)	199	PSA ≥ 0.2 ng/mL 2× consecutively	Treatment 1: RP + PLND (95%), technique not specified, mean number of nodes 7 Treatment 2: RP + EBRT. Median dose was 60 Gy (range 50– 70 Gy) delivered to the prostatic fossa and periprostatic tissues, start < 6 months after RP Treatment 3: RP + ADT. Median (range) duration for ADT was 15 (6–48) months. Androgen ablation was initiated < 3 months after RP. LHRH analog was preferred (22 of 26 patients) Treatment 4: RP + EBRT + ADT. Androgen ablation was initiated on the first day of RT. The median (range) duration of androgen ablation was 12 (6–36) months. Adjuvant RT was initiated < 6 months after RP and the median (range) dose was 60 (48–66) Gy. LHRH analog was used.
Bloom 2004	Retrospective	Unknown	58% 6–7, 42% 8–10	38% < 10, 33% 10–20, 29% > 20	103	PSA failure, not specified	RP + PLND (between external iliac vein and obturator nerve)
Freedland 2004	Retrospective	63 + /− 6.6	70% 6–7, 31% 8–10	18.2 + /−18.2	135	PSA ≥ 0.2 ng/mL or 2 × 0.2 ng/mL	RP + LND, type of LND not specified
Galosi 2023	Retrospective	67 + /− 7.2	62% 6–7, 38% 8–10	9.4 + /−8.9	61	PSA ≥ 0.2 ng/mL	RP + PLND, technique not specified
Johnson 2004	Retrospective	63 + /− 7.2	57% 6–7, 43% 8–10	11.3 + /− 9.2	70	PSA ≥ 0.4 ng/mL	RP + PLND, technique not specified
Kneebone 2020	RCT	63 + /− 6.2	85% 6–7, 15 8–10%	7.4 (0.6–137)	33	PSA ≥ 0.2 ng/mL	Treatment 1: RP, type of LND not specified Treatment 2: RP + EBRT. 3DRT/IMRT radiotherapy, target volumes according to consensus guidelines Faculty of Radiation Oncology Genito‐Urinary Group, dose: 32x2Gy, start < 6 months after RP
Lee 2016	Retrospective	69 (IQR 65–72)	56% 6–7, 44% 8–10	19 (IQR 10–40)	68	PSA ≥ 0.2 ng/mL or 2× 0.2 ng/mL	RP (52% laparoscopic, 32% open and 12% robot‐assisted) + PLND (81%)*** (including external, internal and obturator lymph nodes)
Masterson 2005	Retrospective	65 (mean)	7.3 (mean)	14 (mean)	63	PSA ≥ 0.2 ng/mL 2× consecutively	RP, type of LND not specified
Pagano 2016	Retrospective	64 (IQR 59–68)	50% 6–7, 50% 8–10	9.1 (IQR 6.3–17.1)	158	PSA ≥ 0.05 ng/mL 2× consecutively	RP + PLND, technique not specified, mean number of nodes 8.6
Ramírez‐Backhaus 2016	Retrospective	63 + /− 6.2	86% 6–7, 13% 8–10	35% < 10, 29% 10–20, 36% > 20	54	PSA ≥ 0.4 ng/mL	RP + PLND (79%)*** (including obturator +/‐ internal iliac lymph nodes)
Salomon 2003	Retrospective	64 + /− 6.7	77% 6‐7, 23% 8‐10	17.1 + /−19.2	137	PSA ≥ 0.2 ng/mL	RP, type of LND not specified
Siddiqui 2011	Retrospective	65 (48–79) **	75% 6–7, 25% 8–10	3.6 (0.7–16.6)	382	PSA ≥ 0.4 ng/mL	Treatment 1: RP, type of LND not specified Treatment 2: RP + ADT. Administered < 90 days after RP, and included either LHRH agonizts, antiandrogens or orchiectomy
Swanson 2008	RCT	65 (44–79)	81% 6–7, 19% 8–10	45% < 10, 55% > 10	139	PSA ≥ 0.4 ng/mL	Treatment 1: RP + PLND (97%), technique not specified Treatment 2: RP + EBRT. Radiotherapy was directed at the prostate fossa only. A 4‐field or arch technique was allowed. On the 4‐field approach this was defined as a 9 × 9 cm or 10 × 10 cm anteroposterior and posteroanterior portal. The prescribed dose was 60–64 Gy
Vargas 2004	Retrospective	65	5.8 (mean, RP)/6.4 (mean, RP + EBRT)	10.7 (RP)/9.5 (RP + EBRT)**	65	PSA ≥ 0.1 ng/ml or PSA ≥ 0.3 ng/mL	Treatment 1: RP + PLND (94%), technique not specified Treatment 2: RP + EBRT. 6–18 MV photons, median dose 59.4 Gy (range 50.4 – 66.6 Gy) in 1.8–2 Gy fractions using four‐field or bilateral arc technique. Two patients underwent pelvic lymph node radiotherapy.

Abbreviations: 3D‐CRT, 3D conformal radiotherapy; ADT, androgen deprivation therapy; BED, biologically effective dose; BR, biochemical recurrence; EBRT, external beam radiotherapy; EQD2, equivalent dose in 2 Gy fractions; Gy, gray; IMRT, intensity modulated radiotherapy; IQR, interquartile range; LDR, low‐dose rate; LND, lymph node dissection; PSA, prostate specific antigen; RCT, randomized controlled trial; RP, radical prostatectomy; SD, standard deviation; VMAT, volumetric modulated arc technique.

^*Values based on entire patient cohort.

^**Values based on analyzed patients only.

^***Not all patients underwent PLND, however subgroups analyses were performed including only pN0 patients.

Fifteen studies investigated the efficacy of RP + PLND [8, 23, 25, 26, 27, 28, 30, 31, 32, 33, 34, 35, 37, 38, 39]. Of those, three compared RP + PLND versus RP + PLND + EBRT; one compared RP + PLND versus RP + PLND + /−EBRT + /−ADT; and one compared RP + PLND versus RP + PLND + ADT.

Median patient age varied between 63 and 72 years and median PSA level varied between 3.6 and 46.6 ng/ml (Table 1). Gleason 6–7 histology was found in 30%–91% of patients. The definition of biochemical recurrence used varied between studies: studies on radiotherapy generally used the Phoenix definition (nadir PSA + 2 ng/mL), while most studies on surgery used a postoperative PSA threshold value of ≥ 0.2 ng/ml. Alternatively, ≥ 0.05 ng/ml, ≥ 0.3 ng/mL and ≥ 0.4 ng/mL were used as thresholds. As can be appreciated from Supporting information D, Gleason score and PSA levels were comparable between studies on radiotherapy versus studies on surgery. Of note, median patient age for surgery patients was lower compared to radiotherapy patients.

3.4. Radiotherapy

Ten studies (k = 891) reported on the efficacy of radiotherapy (EBRT and/or brachytherapy) + ADT [14, 15, 16, 20, 21, 22, 24, 29, 36, 41], including four studies with two or three treatment arms [14, 15, 16, 21].

Six studies reported on EBRT + ADT [14, 15, 16, 20, 21, 36]. Data from three RCTs was included (randomizing between two dose levels), and one retrospective study compared low‐dose EBRT (EQD2 70–72 Gy) with high‐dose EBRT (EQD2 74–80 Gy) [14, 15, 16, 21]. No elective pelvic irradiation was performed in any study, except for Yamazaki et al [21]. EBRT techniques were heterogenous and included 3D conformal radiotherapy/4‐field techniques, intensity modulated radiotherapy (IMRT) and volumetric modulated arc therapy (VMAT).

Equivalent doses in 2 Gy fractions (EQD2) to the tumor ranged between 62.8 and 113.4 Gy. 5‐year BRFS varied between 49% and 87% (Table 2, Figure 2).

Table 2.

Outcome data for the included studies.

Radiotherapy and ADT
First author (year)	N patients in analysis	Treatment	3‐year BRFS	5‐year BRFS	7‐year BRFS	10‐year BRFS	Other outcome	EQD2[ab1.5], Gy
Acklin Wehnert 2023	41	EBRT + ADT		83%*				77–80
Heemsbergen 2014	19	EBRT + ADT		47%				68
	19	EBRT + ADT		78%				78
Incrocci 2016	31	EBRT + ADT		65%				78
	37	EBRT + ADT		66%				90.4
Kerkmeijer 2021	38	EBRT + ADT		83%				81.4
	44	EBRT + boost + ADT		87%				113.4***
Koutrouvelis 2003	39	LDR BT + ADT		90%				**
Makino 2021	60	EBRT + HDR BT + ADT		94%*		83%*		**
Stone 2018	53	EBRT + LDR BT + ADT		75%		61%	15‐year BRFS 61%; 10‐year DMFS 78%; 15‐year DMFS 78%; 15‐year CSS 70%	**
Tsumura 2017	80	EBRT + HDR BT + ADT		82%		64%
Yamazaki 2021	86	EBRT + ADT		70%		38%		70–72
	39	EBRT + ADT		87%				74–80
	124	EBRT + HDR BT + ADT		80%		68%
Zelefsky 2008	166	EBRT + ADT		49%		32%		63–82
Radical prostatectomy
Badani 2018	172	RP	60%				Median BRFS 66.5 months
Bastide 2011	82	RP		33%	26%
	41	RP + EBRT		44%	28%
	26	RP + ADT		48%	32%
	50	RP + EBRT + ADT		83%	62%
Bloom 2004	103	RP	28%
Freedland 2004	135	RP		36%
Galosi 2023	61	RP					Biochemical progression in 28/61
Johnson 2004	70	RP		47%
Kneebone 2020	33	RP					Biochemical progression in 8/33
	31	RP + EBRT					Biochemical progression in 12/31
Lee 2016	68	RP	35%	18%			1‐year BRFS 64%
Masterson 2005	63	RP		72% (SM‐) 37% (SM + )
Pagano 2016	158	RP	52%	44%
Ramírez‐Backhaus 2016	54	RP		17%			Biochemical progression in 44/54
Salomon 2003	137	RP		34%		10%	10‐year LRFS 78%, 10‐year CSS 87%
Siddiqui 2011	191	RP		32%		16%
	191	RP + ADT		70%		60%
Swanson 2008	68	RP		36%		12%	5‐year LRFS 56%; 10‐year LRFS 34%; 5‐year DMFS 72%; 10‐year DMFS 47%; 5‐year OS 87%; 10‐year OS 51%
	71	RP + EBRT		62%		36%
Vargas 2004	55	RP		18%
	10	RP + EBRT		60%

Abbreviations: ADT, androgen deprivation therapy; BRFS, biochemical recurrence free survival; BT, brachytherapy; CSS, cancer specific survival; DMFS, distant metastasis free survival; Gy, gray; HDR, high‐dose rate; KM, Kaplan‐Meier; LDR, low‐dose rate; LRFS, locoregional failure free survival; OS, overall survival; RP, radical prostatectomy.

^*Measured from Kaplan Meier curve.

^**For studies on brachytherapy, no EQD2 was calculated because of the high fraction doses.

^***Prescribed isotoxic EQD2 was 113.4 Gy. The actual delivered median EQD2 was 90.5 Gy.

Analysing only RCT data, 5‐year BRFS was 47% in one study arm with EQD2 < 74 Gy [14], between 65% and 78% for EQD2 74–80 Gy [14, 15], between 66% and 80% for EQD2 80‐100 Gy, and 87% in one study arm with EQD2 ≥ 100Gy [15, 16].

Analysing all studies, 5‐year BRFS varied between 47% and 70% in studies with EQD2 < 74 Gy [20, 21], between 65% and 87% for EQD2 74 ‐ 80 Gy [21, 36], and between 66% and 87% for EQD2 ≥ 80Gy (Table 2) [15, 16].

Four studies reported on EBRT + brachytherapy + ADT (including three studies on HDR brachytherapy and one study on LDR brachytherapy) [21, 22, 24, 29]. In three studies on EBRT + HDR brachytherapy + ADT, 5‐year BRFS varied between 80% and 94%. 5‐year BRFS in the study on EBRT + LDR brachytherapy + ADT was 75%.

One study reported on LDR brachytherapy + ADT only, with 5‐year BRFS of 90% [41].

3.5. Radical Prostatectomy

Of the 15 studies investigating the efficacy of RP + PLND, 13 studies were retrospective and two studies were randomized controlled trials. One study provided detailed specifications on surgical techniques (Table 1) [30]. Nine studies (k = 773) reported on 5‐year BRFS, which varied between 18% and 47% [8, 23, 26, 27, 28, 30, 31, 32, 34]. One study reported on 5‐year BRFS in case of negative surgical margins (72%) and in case of positive surgical margins (37%) [28].

Of the four studies investigating the efficacy of RP + EBRT, three studies (k = 122) reported on 5‐year BRFS, which varied between 44% and 62% [8, 23, 34]. The median prescribed doses were 60–64 Gy within 6 months after RP. EBRT techniques were heterogenous and included 3D conformal radiotherapy/4‐field techniques or IMRT to the prostatic fossa and periprostatic tissues. No lymph node irradiation was performed.

Of the two studies investigating the efficacy of RP + PLND + ADT, one study reported on 5‐year BRFS, which was 48% [25]. Bastide et al. reported a varying length of ADT use with a median duration of 15 months (range 6–48 months). ADT was started within 3 months, mostly LHRH analogs.

One study (k = 50) reported on the efficacy of RP + EBRT + ADT, with a 5‐year BRFS of 83% and a 7‐year BRFS of 62% [34]. Adjuvant EBRT was initiated within 6 months after RP with a median total dose of 60 Gy (range 48–66 Gy). Adjuvant ADT (LHRH analog) was initiated at the first day of EBRT, with a median duration of 12 months (range 6–36 months).

3.6. Meta‐Analysis

We performed separate meta‐analyses for each treatment regime. For studies focusing on EBRT, pooled estimated 5‐year BRFS was 68% [57; 79] with low heterogeneity (I ² 46%, Figure 3). Subgroup analysis revealed that EQD2 was a significant effect modifier (p = 0.0216, Supporting information E). When analysing studies with high‐dose EBRT ( ≥ 74 Gy), pooled estimated 5‐year BRFS was 78% [66; 89] (I ²  = 0). Analysing studies with low‐dose EBRT ( < 74 Gy), pooled estimated 5‐year BRFS was 56% [41; 71] (I ²  = 51%)

Figure 3.

Forest plots displaying pooled estimates for 5‐year biochemical recurrence free survival (BRFS) with corresponding 95%‐confidence intervals, for (A) EBRT + ADT, (B) EBRT + HDR brachytherapy + ADT, (C) RP, (D) RP + EBRT, (E) RP + ADT. Studies on EBRT, LDR brachytherapy + ADT; LDR brachytherapy + ADT; prostatectomy + EBRT + ADT are not displayed as only one study was identified and no pooled estimates could be calculated. Studies are sorted on type of study (prospective or retrospective) and subsequently ordered based on treatment outcome. ADT, androgen deprivation therapy; BRFS, biochemical recurrence free survival; EBRT, external beam radiotherapy; HDR, high‐dose rate brachytherapy; RP, radical prostatectomy.

When analysing RCTs only, the pooled estimated 5‐year BRFS was 71% [58; 83] with no heterogeneity. Analysing retrospective studies only, the pooled estimated 5‐year BRFS was 68% [49; 87] with moderate heterogeneity (I ²  = 70%).

Pooled estimated 5‐year BRFS was 83% [71; 95] in studies with dose escalation by means of EBRT + HDR brachytherapy + ADT.

Heterogeneity was moderate when analysing studies on prostatectomy only (I ²  = 63%), with a pooled estimated 5‐year BRFS of 32% [25; 38]. For studies focusing on prostatectomy with adjuvant treatment, pooled estimated 5‐year BRFS varied between 54% [41; 67], prostatectomy with EBRT) and 63% [45; 82], prostatectomy with ADT) with no heterogeneity. For EBRT + LDR brachytherapy + ADT; LDR brachytherapy + ADT; RP + EBRT + ADT, only one study was identified and therefore no pooled estimates could be calculated.

4. Discussion

In this systematic review we have investigated the efficacy of different treatment modalities for men with PC invading the seminal vesicles (cT3b) without lymph node or distant metastasis (N0M0).

All studies were at moderate to high risk of bias, mainly because of concerns due to confounding, classifications of interventions and deviations from interventions. This could impact the findings in this study and therefore findings should be interpreted with caution.

We identified more studies (including more patients) on surgery with or without adjuvant treatment, compared to studies on radiotherapeutic management. For studies on radical prostatectomy without adjuvant treatment 5‐year BRFS was low, whereas studies investigating multimodality treatment showed improved outcomes. These results do not support the use of surgery as the sole treatment for these patients, which is in line with the current European guidelines that advise IMRT/VMAT after radical prostatectomy in case of pT3 disease and ISUP 4‐5 histology [42]. Furthermore, for patients with locally advanced PC and cN0 disease, offering primary radiotherapeutic treatment with IMRT/VMAT in combination with long‐term ADT is advised. For patients with good urinary function and locally advanced PC it is advised to offer IMRT/VMAT with HDR‐brachytherapy boost and long‐term ADT which is in line with a recent systematic review confirming the usefulness of HDR‐brachytherapy boost in locally advanced patients, at the expense of more GU toxicity [43]. Although patient numbers in our systematic review are limited, these recommendations are supported by our finding that BRFS is high in studies using high‐dose radiotherapy (either by EBRT, or in combination with brachytherapy). The ongoing SPCG‐15 trial randomizes patients with locally advanced PC (including T3b disease) between surgery and radiotherapy and will in the future provide level I evidence on the best treatment within this patient group [44].

Five‐year BRFS for radiotherapy varied strongly depending on the technique used as well as the EQD2. These results provide a strong rationale for dose escalation to these tumors that are shown to be at particular high risk for biochemical failure [45]. As dose escalation > 80 Gy to the entire prostate may be associated with increased toxicity, and local recurrence usually occurs at the same site as the dominant primary tumor at baseline, supplementary focal boosting to the primary lesion is proposed as a strategy to improve biochemical control without increasing toxicity [46, 47]. This strategy was prospectively investigated in the phase III multicentre FLAME trial on intermediate‐ and high‐risk PC patients, in which patients either received 77 Gy to the entire prostate, or an additional isotoxic boost up to 95 Gy to the intraprostatic lesion [16]. This trial also included 92 patients with cT3bN0 PC and showed an improved 5‐year BRFS. Secondary analyses have shown that focal boosting is especially beneficial for patients with high‐risk features [45]. As focal doses up to 95 Gy may further improve oncological outcomes in high‐risk patients, more complex imaging and delivery techniques are required to ensure safe dose delivery without increasing toxicity. This is especially relevant for patients with T3b tumors due to the anatomical locations proximate to the rectum and the mobility of the seminal vesicles. A possible solution for this problem could be magnetic resonance guided radiotherapy (MRgRT), with hybrid systems combining linear accelerators with MRI‐scanners. In this way, it is possible to visualize pelvic anatomy in great detail and subsequently perform replanning of the radiation dose distribution before each radiotherapy fraction. This technique potentially increases the therapeutic window of EBRT by improved normal tissue sparing due to margin reduction, correction for deformation of organs and target, and more accurate target dose delivery. Given these evolving radiotherapy techniques and the intrinsic susceptibility of PC to hypofractionation, focal boosting in an ultrahypofractionated way was tested in the phase II hypoFLAME trial [48]. Toxicity was shown to be acceptable and efficacy of this five‐fraction focal boost strategy is currently tested in a multicentre randomized hypoFLAME 3.0 (NCT05705921). Whether this five‐fraction regime is also associated with acceptable toxicity in patients with T3bN0 PC is currently tested in the AFFIRM trial (NCT05373316).

Although we were not able to compare radiotherapy with ADT to radiotherapy without ADT, a recent individual patient data meta‐analysis including 10 853 patients provided clear evidence for the benefit of ADT treatment intensification with radiotherapy for men with localized PC [49]. For high‐risk patients, it was previously shown that long‐term ADT (3 years) provided superior overall survival compared to short‐term ADT (6 months) [50]. Most of the included studies investigating EBRT with ADT prescribed long‐term ADT (at least 2 years), supporting the importance for these patients, in line with the European guidelines [42].

To the best of our knowledge, this is the first systematic review providing data on the efficacy of different treatment modalities and combinations, in patients with T3bN0M0 PC. However, some limitations need to be addressed. We performed subgroup analyses on treatment modality with small numbers of studies, possibly explained by the fact that only 10‐15% of PC patients present with high‐risk features, and even fewer constitute this specific stage [2]. Second, we were not able to retrieve data from all authors contacted for subgroup data. Third, some studies could not be included in the meta‐analysis because they did not report on 5‐year BRFS, but 3‐, 7‐ or 10‐year outcomes. Fourth, our primary outcome is BCR, which is a surrogate outcome for prostate cancer specific mortality. Ideally, clinical outcome measures such as metastasis‐free survival of prostate cancer‐specific mortality should be included in such studies, which were available in only a limited amount of studies.

Importantly, because most studies were performed and published ten to twenty years ago, varying techniques were used to stage and treat patients. In the last decades the multiparametric MRI (mpMRI), including T2‐weighted images, diffusion weighted imaging (DWI) and contrasted‐enhanced (DCE) imaging, has become an important tool in local staging of PC. Approximately one out of three patients will be upstaged with mpMRI compared to DRE [6]. Patients included in surgical studies are staged pathologically (RP specimen) and therefore reflect true T3b disease, whereas clinically staged patients (by MRI or DRE) in radiotherapy studies are potentially being under‐ of overstaged. As clinical staging methods have low sensitivity for SVI, the clinically staged patients will likely constitute a more unfavorable patient group, as patients staged by RP would partly have been staged as cT1‐2 rather than cT3b. Therefore, results need to be interpreted with caution when comparing different treatment modalities, as study populations will not be completely comparable. To prevent patients with cT1‐2 and pT3b disease from being undertreated when undergoing radiotherapy, risk of SVI is estimated based on a validated nomogram and the seminal vesicles are included into the target volume based on the predicted risk [51].

Furthermore, current guidelines recommend nodal staging by means of PSMA‐PET/CT, as CT and MRI lack sensitivity for direct detection of positive lymph nodes. Sensitivity of CT however, which was the main method of nodal staging in the included studies, is only 40% for the detection of LNM [52]. This directly impacts applicability of our findings, as the patient cohorts within our included studies will harbor a risk of nodal “understaging”, compared to modern staging methods. In a recent meta‐analysis on 68Ga‐PSMA PET, sensitivity and specificity of 77% and 97% respectively, were found for the prediction of LNM [53]. This implies that even after staging with PSMA‐PET/CT, some clinically‐staged patients will actually have pN+ disease and harbor more extensive disease than actually detected.

Finally, we did not identify literature on new second‐generation hormone treatment including abiraterone or enzalutamide. In 2022, the STAMPEDE platform has shown that addition of abiraterone improved metastasis‐free survival in non‐metastatic high‐risk PC patients (including N1, or T3‐4N0 with Gleason score 8‐10 or PSA ≥ 40 ng/mL, or high‐risk relapsing disease) [54]. Even though no data could be extracted for our subgroup of interest, it is likely that these patients will benefit from addition of second‐generation ADT.

5. Conclusions

Although the included studies harbored a significant risk of bias, this systematic review provides rationale for dose‐escalated ( ≥ 74 Gy) radiotherapy with IMRT/VMAT technique combined with long‐term ADT in T3bN0 PC patients, which can increasingly be delivered with acceptable toxicity, due to evolving radiotherapy and imaging techniques. Studies on further (focal) dose escalation with modern staging methods and combinations with second‐generation ADT are warranted to further improve biochemical disease control.

Author Contributions

Conception and design: Casper Reijnen and Robert Jan Smeenk. Acquisition of data: Casper Reijnen, Wilma D. Heemsbergen, Aishly Panneflek, Floris J. Pos, J.P. Michiel Sedelaar, Uulke A. van der Heide, Luca Incrocci and Robert Jan Smeenk. Analysis and interpretation of data: Casper Reijnen, Wilma D. Heemsbergen, Kimberley E. Wever, Aishly Panneflek, Floris J. Pos, J.P. Michiel Sedelaar, Uulke A. van der Heide, Luca Incrocci, and Robert Jan Smeenk. Drafting of the manuscript: Casper Reijnen and Robert Jan Smeenk. Critical revision of the manuscript: Casper Reijnen, Wilma D. Heemsbergen, Kimberley E. Wever, Aishly Panneflek, Floris J. Pos, J.P. Michiel Sedelaar, Uulke A. van der Heide, Luca Incrocci, and Robert Jan Smeenk. Statistical analysis: Casper Reijnen, Kimberley E. Wever, and Robert Jan Smeenk.

Ethics Statement

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supplementary information A: Search strategy.

Supplementary information B: Study characteristics.

Supplementary information C: Bias assessment.

Supplementary information D: Characteristics of included studies (age, Gleason, PSA).

Supplementary information E: Effect size plot.

supmat.

Data Availability Statement

Search strategies and data extraction form can be found in the Supporting information.