Intensity Modulated Moderately Hypofractionationated Radiotherapy Versus Stereotactic Body Radiotherapy for Prostate Cancer (PACE-C): Early Toxicity Results From a Randomised Open-label Phase III Non-inferiority Trial

Abstract

Background

Moderately hypofractionated radiotherapy (MHRT) is standard for prostate cancer. Stereotactic body radiotherapy (SBRT) is also effective, but randomised data on toxicity for higher risk patients is lacking. We aimed to compare early toxicity of these treatments.

Methods

The phase 3 PACE-C trial recruited men aged ≥18 years with intermediate/high-risk histologically confirmed prostate adenocarcinoma (≤T3a, ≤Gleason 4+4, PSA≤30ng/ml) and performance status 0-2 from 53 centres (UK, Ireland, New Zealand). Participants were centrally randomised (1:1; non-blinded; permuted blocks size four;six), stratified by centre and risk group), to MHRT (60 gray;20 fractions;4 weeks) or SBRT (36.25 gray;5 fractions;1-2 weeks). Six months androgen deprivation therapy was planned. Primary endpoint is freedom from biochemical/clinical failure. Coprimary outcomes for this analysis were worst grade 2 or higher (G2+) Radiation Therapy Oncology Group (RTOG) gastrointestinal or genitourinary events within 12 weeks of radiotherapy; analysis included randomised patients by study treatment received. Late toxicity and efficacy data are awaited as the trial remains in follow-up.

Findings

Between 13Nov2019 and 24Jun2022, 1208 men were randomised (MHRT: 601; SBRT: 607). Ethnicity was reported as 95% White, 2% Black, 1% Asian, 1% Chinese. This analysis includes 608 men receiving MHRT and 584 receiving SBRT.

Within 12 weeks of treatment, G2+ RTOG genitourinary toxicity was reported in 166/608 (27·3%;95% CI 23·8 to 31·1) receiving MHRT and 162/582 (27·8%; 95% CI 24·3 to 31·7) receiving SBRT (absolute difference 0·5% [95% CI -4·7 to 5·7]; p=0·89). For G2+ genitourinary CTCAE, rates were 170/604 (28·1%) after MHRT and 195/581 (33·6%) after SBRT (p=0·050). Grade 3 CTCAE genitourinary toxicity was seen in three (0·5%) patients receiving MHRT and three (0·5%) patients receiving SBRT.

G2+ RTOG gastrointestinal toxicity was reported in 69/608 (11·4%; 95% CI 9·0 to 14·2) receiving MHRT and 74/584 (12·7%; 95% CI 10·2 to 15·8) receiving SBRT (absolute difference 1·4% [95% CI -2·5 to 6·2]; p=0·53) within 12 weeks of treatment. For G2+ gastrointestinal CTCAE, rates were MHRT: 60/604 (9·9%) SBRT: 96/581 (16·5%) (p=0·001). Grade 3 CTCAE gastrointestinal toxicity was seen in three patients (0·5%) receiving MHRT and four patients (0·5%) receiving SBRT. There were no treatment-related deaths.

Interpretation

Despite an accelerated treatment schedule and a larger treated volume than PACE-B, SBRT and MHRT had similar rates of early RTOG toxicity. CTCAE bowel effects were more common after SBRT.

Registration

ClinicalTrials.gov: NCT01584258

Funding

The Royal Marsden Cancer Charity

Introduction

Prostate cancer remains an increasing global health challenge, with the recent Lancet Commission predicting that annual global cases of prostate cancer will increase from 1·4 million to 2·9 million by 2040 ¹. Many countries have inadequate access to radiotherapy, partly due to the need for treatment fractionation over many weeks, requiring more linear accelerators and staff than if treatment could be effectively abbreviated. In the UK, more men are cured of prostate cancer by radiotherapy than any other treatment ² Radiotherapy technology has undergone a revolution over the last two decades, increasing accuracy which has reduced side effects and increased cure rates ^3–6.

Building on previous work showing that moderately hypofractionated radiotherapy (MHRT) (around 3 gray (Gy) per fraction) is non-inferior to longer conventionally fractionated schedules ^3,4,7, the logical next step was to test whether non-inferiority could be demonstrated for ultra-hypofractionated schedules (6Gy per fraction or more). The PACE study comprises multiple cohorts (PACE-A, PACE-B, and PACE-C) each evaluating stereotactic body radiotherapy (SBRT) with independent randomisations. PACE-A compared quality of life following SBRT to that following prostatectomy ⁸. PACE-B compared 4-8 weeks of daily radiotherapy with 5 fraction SBRT and showed that the shorter treatment was non-inferior for biochemical control ⁸. Early RTOG toxicity in PACE-B was similar between arms, although CTCAE genitourinary toxicity was worse after SBRT ⁹. PACE-A and PACE-B recruited men with very early prostate cancer where hormone therapy was not given. PACE-C was designed to test whether SBRT was non-inferior to MHRT for men with higher risk prostate cancer. Here we present the first results of this trial, focussing on toxicity comparison during the first 3 months after radiotherapy.

Methods

Study design and participants

PACE-C is a multicentre, international, phase 3, open label randomised controlled trial aiming to demonstrate non-inferiority of SBRT compared to MHRT with respect to biochemical control. Secondary endpoints are listed in the PACE protocol (in Supplementary materials). The trial was approved by the London (UK) Chelsea Research Ethics Committee (ref 11/LO/1915) and the relevant institutional review boards in Ireland and New Zealand. The trial is sponsored by The Royal Marsden Hospital NHS Foundation Trust and funded by The Royal Marsden Cancer Charity. The study was overseen by a trial steering and an independent data monitoring committee. The study was prospectively registered (clinicaltrials.gov: NCT01584258). The Royal Marsden/Institute of Cancer Research patient and carers panel provided input on the PACE-C trial design and patient documentation and participated in the Trial Management Group.

PACE-C recruited men planned to have radiotherapy with curative intent. Eligible patients were aged ≥18 years, with World Health Organisation performance status 0-2, life expectancy of ≥5 years and histologically confirmed prostate adenocarcinoma. All patients had NCCN intermediate or high-risk disease ¹⁰. Intermediate risk patients had Gleason 3+4 or Gleason 4+3 disease, a PSA 10-20 ng/mL and T1/T2 stage tumours. High risk patients had one or maximum two of the following risk factors: MRI stage T3a, Gleason 4+4 (maximum of 50% of cores taken permitted to be positive), PSA >20-30 ng/ml. Patients with Gleason 4+3 or any high-risk factors had distant staging with bone scan, CT, PET or whole body diffusion-weighted MRI. A biopsy, meeting the above criteria, was needed within 18 months of randomisation, except for those progressing on active surveillance and now opting for radical treatment by virtue of PSA trajectory or MRI progression. Active surveillance patients were stratified as intermediate risk. For those on 5-alpha reductase inhibitors (5-ARI), PSA thresholds were halved. Treating clinicians had discretion to exclude patients for co-morbid conditions making radiotherapy inadvisable (e.g. inflammatory bowel disease, significant urinary tract symptoms). Detailed inclusion criteria are given in the protocol. All participants provided voluntary, written, informed consent prior to study entry.

Randomisation and masking

Patients were randomised 1:1 to either MHRT or SBRT. Randomisation was undertaken centrally by the Institute of Cancer Research Clinical Trials and Statistics Unit (ICR-CTSU; London, UK), with allocation by computer generated random permuted blocks (size 4 and 6) and stratified by centre and risk group (intermediate or high). Treatment assignment was open label to participants and researchers.

Procedures

The protocol recommended bowel preparation (micro-enemas) immediately prior to radiotherapy planning, and moderate bladder filling. The radiotherapy planning CT scan took place at least 7 days after fiducial placement, if used. A radiotherapy planning MRI scan was strongly recommended to be registered to the CT scan (preferably by fiducial match) for improved prostate anatomical definition. The clinical target volume (CTV) was the prostate and proximal 1cm of seminal vesicles (in Gleason 3+4 intermediate risk disease) and the prostate plus the proximal 2cm of seminal vesicles for Gleason 4+3 or high-risk disease.

Recommended MHRT CTV to planning target volume (PTV) margins were 3mm for the high dose PTV (60 Gy) and 6mm for the low dose PTV (47 Gy), which included the proximal 2cm of SV, where required. Variation in margins in the range 4-8mm for both PTVs were permitted. Recommended SBRT CTV to PTV expansion was 5mm except 3mm posterior for the high dose PTV (36·25 Gy) and 5mm for the low dose PTV (30 Gy) which included the proximal 2cm of SV, where required. Variations in the margins up to 4-5mm isotropic for the high dose PTV and 6mm for the low dose PTV were permitted. Dose constraints were applied to organs at risk and were consistent with PACE-B ⁹.

MHRT PTV dose was 60 Gy in 20 daily fractions over 4 weeks, using IMRT/VMAT and prescribed to the median dose. SBRT PTV dose was 36·25 Gy in 5 fractions over 1-2 weeks (i.e. daily or alternate days, at centre discretion), with an additional mandatory CTV dose target of 40 Gy (no margin) to the prostate and proximal 1cm of seminal vesicles, both prescribed to the isodose covering 95% PTV. Usual SBRT practice in the UK is to allow dose heterogeneity within the target such that the point maximum of the plan may be up to 133% of the prescription dose. Detailed prescription objectives, along with minor variations permitted, are given in the protocol (Supplementary Appendix B).

Radiotherapy was mandated to commence within 12 weeks of randomisation, with ≤8 weeks strongly recommended. Daily image-guided radiotherapy (IGRT) to the prostate (preferably incorporating fiducials) was mandated. Repeat static imaging was required for SBRT for fraction delivery beyond 3 minutes. Intra-fractional motion monitoring was permitted but not required. A radiotherapy quality assurance programme was undertaken for each centre to ensure contouring and planning consistency with trial protocol and quality of radiotherapy treatments: details in protocol and RTQA Guidelines (Supplementary Appendix A & B).

Androgen deprivation therapy (ADT) was mandatory for at least 6 months. LHRH analogues, LHRH antagonists or bicalutamide was permitted. Prolongation of ADT was permitted (up to a maximum total duration of 18 months) if required due to diversion of healthcare resources to the COVID-19 pandemic (protocol amendment approved 25th June 2020).

Ethnicity was reported by the hospital staff either self-declared by the patient or using medical records at trial entry. During the early or “acute” period, participants were assessed alternate weeks during MHRT and on the final fraction for SBRT, and weeks 2, 4, 8 and 12 after the end of radiotherapy. Two clinician reported outcomes were collected: Radiation Therapy Oncology Group (RTOG; gastrointestinal and genitourinary domains) at baseline and every visit (SupplT1) ¹¹; Common Terminology Criteria for Adverse Events (CTCAE v4.03 ¹²) at baseline, after treatment, and at follow-up weeks 2, 4, 8 and 12 weeks (SupplT2). Paper questionnaires collected four patient-reported outcome (PRO) measures: Expanded Prostate Cancer Index Composite Short Form (EPIC-26 ¹³) and the Vaizey Faecal Incontinence Score¹⁴, at baseline, weeks 4 and 12; International Prostate Symptom Score (IPSS ¹⁵), at baseline, weeks 2, 4, 8 and 12; the International Index of Erectile Function 5-question (IIEF-5 ¹⁶) score, at baseline and week 12. Subsequent follow up is ongoing.

Patients were able to withdraw from the study at any point, after which no further data would be collected. Medical teams could withdraw the patient if they felt this was in their best interests.

Outcomes

The trial’s primary outcome is freedom from biochemical or clinical failure, the data for which are not yet mature. Secondary oncological endpoints of recommencement of ADT, progression-free, disease-specific and overall survival and late toxicity endpoints are not yet mature. The co-primary endpoints for this pre-planned safety analysis were the percentage of RTOG Grade 2 or higher (G2+) gastrointestinal and genitourinary toxicity at any point during or within 12 weeks of completion of radiotherapy (the early or “acute” period).

Secondary clinician reported outcomes were percentage of RTOG G2+ gastrointestinal and genitourinary at 12 weeks and percentage of CTCAE G2+ gastrointestinal and genitourinary during the early period and at 12 weeks. Individual pre-specified CTCAE G2+ gastrointestinal, and genitourinary, erectile function and other pre-specified CTCAE parameters including hot flushes and fatigue were further secondary outcomes. Secondary PROs were EPIC-26 composite scores (urinary incontinence/irritative, urinary obstructive, bowel and sexual domains) reported as a score and as the percentage of patients experiencing a minimally clinically important difference (MCID) and 2-times MCID for the domain-specific quality of life. The following PROs are also reported: single EPIC urinary, bowel and sexual bother questions, IPSS total and QoL scores, Vaizey total score, and IIEF-5 total score. The EPIC urinary and bowel sub-domain composite scores and MCID at 12 weeks were the main PROs of interest.

Exploratory analyses considered G2+ toxicity in the early period by machine platform (Cyberknife, MR-Linac or standard linac) and the centre ‘learning’ effect of administering SBRT (<10 patients vs > 10 patient previous recruited to SBRT at a centre in PACE B). In addition, a post-hoc analysis was carried out assessing correlation between prior TURP and toxicity.

Statistical analysis

PACE-C is powered for non-inferiority of time to biochemical or clinical failure with a sample size of 1182 patients to exclude a hazard ratio of 1·37. Therefore, for this safety analysis there was sufficient (at least 80%) power to exclude a difference of 6% in RTOG G2+ gastrointestinal toxicity, if the prevalence was 12% in MHRT (as seen in PACE-B¹⁰), and a difference of 8% in RTOG G2+ genitourinary toxicity, if the prevalence was 27% in MHRT (as seen in PACE-B¹⁰) at the 2-sided 0·025 significance level. Analyses are by treatment received, with participants included if they had one or more fractions of MHRT or SBRT regardless of their allocated treatment (safety population). Patients are included in each measure if at least one of the relevant outcome measures is available i.e. patients with missing data were excluded from analysis of that endpoint; completeness of data is provided for each outcome. A statistical analysis plan was written prior to commencing analysis and is available in the Supplementary materials. All analyses presented were pre-specified unless stated otherwise.

Percentage of G2+ events over the early period are presented with 95% confidence intervals (CI) and robustness of results assessed in the presence of baseline toxicity using logistic regression. Confidence intervals for proportions were calculated using Wilson score method with corresponding difference in proportions tested using a continuity corrected chi-square test. Length of SBRT treatment sub-group analysis was pre-planned. SBRT schedule (daily or alternate daily) was largely determined at a centre-level. To account for this confounding effect and support interpretation of any differences between 1- and 2-week SBRT schedules an indicator variable, “centre choice”, was defined by SBRT duration (for SBRT patients) and centre’s use of a 1-week schedule for at least 90% (determined post-hoc) of their SBRT patients (for MHRT patients). This “centre-choice” main effect and a “centre choice” by treatment interaction term were included in a post-hoc logistic regression analysis with the interaction term of primary interest in assessing the effect of 1 vs 2 weeks SBRT on the odds of toxicity. For PRO, descriptive statistics are presented for continuous variables at each time point including baseline. Frequency and percentages are used for categorical data including patients with MCID changes. Statistical comparisons used Mann Whitney test for continuous scores, Chi-square trend test for ordinal and Fishers exact test for binary variables. To reduce the impact of multiple comparisons, a p-value <0·01 was considered statistically significant for secondary endpoints

Baseline assessments needed to take place before the start of radiotherapy. Early assessments taken after the start of radiotherapy and up to 18 weeks post-treatment were included to avoid loss of data. It was expected that most missing data on clinical outcomes would be missing at random as the patients had a good prognosis; an exploration of missingness confirmed minimal impact on results. Similarly, for each PRO measure, a review of baseline characteristics for patients with and without any data during the early period showed no obvious differences.

PRO scores were calculated in accordance with the relevant manuals. EPIC-26 scores were rescaled to a 0-100 point scale, with higher scores representing better quality of life (QoL) ¹⁷. MCID in EPIC-26 subdomain scores were: urinary incontinence 8 points, urinary obstructive 6 points, bowel 5 points, sexual 11 points, hormonal 5 points ¹⁸. IPSS severity categories were assessed as none (0 points), mild (1-7 points), moderate (8-19 points), severe (20-35 points) ¹⁹. The IIEF-5 total score was calculated and ranged from 1 (most severe) to 25 (no erectile dysfunction). The Vaizey total score ranged from 0 (no problems) to 24 (very severe problems with incontinence).

Analyses are based on a snapshot of the clinical data taken on March 15, 2024 and PRO data taken on April 5, 2024 and were conducted using Stata version 17, with the exception of Wilson score 95% CIs and corresponding tests which were computed using SAS 9·4. The Independent Data Monitoring Committee gave approval for release of these results, prior to release of the trials’ primary efficacy results.

Role of the funding source

The funder had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data and had final responsibility for the decision to submit for publication.

Results

Between 13/11/2019 and 24/06/2022, 1208 participants were randomised (MHRT: 601; SBRT: 607) from 53 hospitals across the UK, Republic of Ireland and New Zealand (SupplP3). 608 patients received MHRT and 584 received SBRT (Figure 1; analysis population n=1192). Of the 18 patients allocated SBRT who received MHRT, 13 switched treatments due to concerns regarding dose constraints or perceived risk of side effects. No patients crossed over due to toxicity.

Figure 1.

CONSORT diagram showing derivation of the early toxicity analysis population. All patients commencing SBRT completed treatment. Two MHRT patients stopped treatment early for non-toxicity reasons.

Baseline characteristics and radiotherapy details are shown in Table 1 and SupplP7-P8. The groups were well balanced for age and tumour characteristics. Seven percent of patients had a TURP pre-treatment. Fewer patients receiving MHRT had an MRI for radiotherapy planning (248/607; 41%) compared to SBRT (409/584; 70%). Less than half in both groups had fiducial markers placed. No patient received a rectal spacer. Median duration of radiotherapy was 29 days for MHRT (range 21-37 days) and 10 days for SBRT (range 5-21 days). Of those receiving SBRT, 140/584 (24%) received SBRT over 1 week and 444/584 (76%) received SBRT over 2 weeks. All patients received ADT, usually with LHRH analogues (498/608 (82%) for MHRT; 487/584 (83%) SBRT (SupplP9). The majority were planned for 6 months hormone therapy. Hospital-reported ethnicity was reported as 95% White, 2% Black, 1% Asian, 1% Chinese.

Table 1. Baseline demographics, tumour characteristics and treatment details by treatment received.

	MHRT N=608		SBRT N=584
Age at randomisation
Median (IQR)	72·4	(67·1, 75·7)	73·1	(67·9, 76·2)
n (Range)	608	(50·0, 83·6)	584	(50·6, 83·7)
Ethnicity
White	579	(95)	557	(95)
Black/Black British	8	(1)	12	(2)
Asian/Asian British	7	(1)	10	(2)
Mixed	4	(1)	1	(<1)
Chinese or other	5	(1)	2	(<1)
Unknown	5		2
WHO performance status
WHO status 0	474	(78)	445	(76)
WHO status 1	128	(21)	134	(23)
WHO status 2	4	(1)	4	(1)
Missing	2		1
Risk Group at randomisation
Intermediate favourable	6	(1)	11	(2)
Intermediate unfavourable	378	(62)	366	(63)
High	224	(37)	207	(35)
Pre-ADT PSA value
Median (IQR)	8·3	(6·0, 12·4)	8·5	(6·0, 12·4)
n (Range)	608	(0·7, 29·7)	584	(0·2, 28·7)
<10 ng/ml	389	(64)	356	(61)
10-20 ng/ml	191	(31)	194	(33)
>20-30 ng/ml	28	(5)	34	(6)
MRI T staging
<=T1c	25	(4)	23	(4)
T2	409	(67)	398	(68)
T3a	174	(29)	163	(28)
Gleason score
3+3	19	(3)	18	(3)
3+4	361	(59)	332	(57)
4+3	186	(31)	208	(36)
4+4	41	(7)	26	(4)
>4+4	1	(<1)	0	(0)
Radiotherapy details (post randomisation): Planning MRI performed
Yes	248	(41)	409	(70)
No	359	(59)	175	(30)
Unknown	1		0
Fiducial markers placed
Yes	240	(39)	259	(45)
No	368	(61)	323	(55)
Radiotherapy Unit
Conventional Linac	580	(96)	489	(84)
MR Linac	19	(3)	33	(6)
CyberKnife	0		60	(10)
Tomotherapy	2	(<1)	0
Other	6	(1)	2	(<1)

Data are n (%) unless stated otherwise

RTOG and CTCAE data completeness was high (SupplP10, P11), above 95% for most timepoints in both groups. Data completeness for PROs was between 76-85% for EPIC-26 bowel and urinary questions, IPSS and Vaizey (SupplP12, P13, P14, P15). Completion of IIEF-5 scores were lower (SupplP16). Because the visits during treatment were scheduled from the start of treatment and the post-treatment visits were scheduled from the final day of treatment, the week 2 post-treatment visit is scheduled earlier in the SBRT group (SupplP17). For this analysis, data were truncated after the early toxicity reporting period (12 weeks) hence median follow up is not reported.

During the early period, the co-primary endpoint of RTOG G2+ genitourinary toxicity was reported in 166/608 (27·3%; 95% CI 23·8 to 31·1) with MHRT and 162/582 (27·8%; 95% CI 24·3 to 31·7) after SBRT (absolute difference 0·5% [95% CI -4·7 to 5·7]; p=0·89). Grade 3 genitourinary toxicity was seen in eight patients (1·3%) receiving MHRT and nine (1·5%) patients receiving SBRT; Grade 4 genitourinary toxicity was seen in two and three patients respectively (Table 2; Figure 2).

Table 2. Maximal genitourinary and gastrointestinal RTOG toxicities and CTCAEs and other CTCAEs recorded during the early toxicity period.

Toxicity/Adverse event* n patients (%)	MHRT N=608						SBRT N=584
Grade	0	1	2	3	4	Missing	0	1	2	3	4	Missing
RTOG genitourinary †	119 (20)	323 (53)	156 (26)	8 (1)	2 (<1)	0	111 (19)	309 (53)	150 (26)	9 (2)	3 (1)	2
CTCAE genitourinary	50 (8)	384 (64)	167 (28)	3 (<1)	0	4	44 (8)	342 (59)	192 (33)	3 (1)	0	3
Frequency	85 (14)	387 (64)	132 (22)	0	0	4	62 (11)	361 (62)	158 (27)	0	0	3
Urgency	192 (32)	363 (60)	48 (8)	0	0	5	166 (29)	356 (61)	59 (10)	0	0	3
Urinary retention	393 (65)	181 (30)	28 (5)	1 (<1)	0	5	366 (63)	182 (31)	32 (6)	2 (<1)	0	3
Dysuria	425 (70)	166 (27)	12 (2)	0	0	5	377 (65)	172 (30)	31 (5)	1 (<1)	0	3
Urinary tract obstruction	500 (83)	87 (14)	16 (3)	0	0	5	473 (81)	82 (14)	26 (4)	0	0	3
Non-infective cystitis/ UTI	478 (79)	105 (17)	17 (3)	3 (<1)	0	5	433 (75)	131 (23)	17 (3)	0	0	3
Incontinence	475 (79)	114 (19)	15 (2)	0	0	4	446 (77)	130 (22)	5 (1)	0	0	3
Haematuria	579 (96)	21 (3)	3 (2)	0	0	5	558 (96)	21 (4)	2(<1)	0	0	3
RTOG gastrointestinal ‡	233 (38)	306 (50)	64 (11)	5 (1)	0	0	207 (36)	301 (52)	72 (12)	2 (<1)	0	2
CTCAE gastrointestinal	254 (42)	290 (48)	57 (9)	3 (<1)	0	4	171 (29)	314 (54)	92 (16)	4(1)	0	3
Diarrhea	417(69)	163 (27)	23 (4)	1 (<1)	0	4	337 (58)	206 (35)	36 (6)	2 (<1)	0	3
Proctitis	423(70)	159 (26)	21 (3)	1 (<1)	0	4	339 (58)	203 (35)	38 (7)	1 (<1)	0	3
Constipation	505(84)	81 (13)	18 (3)	0	0	4	450 (77)	111 (19)	20 (3)	0	0	3
Rectal pain	532(88)	65 (11)	6 (1)	1 (<1)	0	4	479 (82)	89 (15)	13 (2)	0	0	3
Haemorrhage	541(90)	59 (10)	4 (1)	0	0	4	474 (82)	101 (17)	6 (1)	0	0	3
Nausea	587(97)	16 (3)	1(<1)	0	0	4	552 (95)	26 (4)	2(<1)	1 (<1)	0	3
Other pre-specified AEs**
Erectile dysfunction	93 (17)	203 (36)	183 (32)	85 (15)	0	45	90 (16)	201 (37)	180 (33)	78 (14)	0	35
Fatigue	151 (25)	407 (67)	46 (8)	0	0	4	153 (26)	384 (66)	41 (7)	3 (1)	0	3
Hot flashes	143 (24)	381 (63)	75 (12)	5 (1)	0	4	153 (26)	361 (62)	60 (10)	7 (1)	0	3

^*pre-specified individual genitourinary/gastrointestinal CTCAE terms are shown

^**pre-specified non-genitourinary/gastrointestinal CTCAE terms reported in >10% of patients or any G3+ are shown

^†G3 events were MHRT: 8 cystitis,2 cystitis/frequency, 2 urethra stricture post operative, 1 retention, SBRT:8 cystitis/frequency, 7 cystitis, 2 retention, 2 dysuria, 1 haematuria; G4 events were MHRT: 3 urethra stricture post operative, 1 retention SBRT: 5 urethra stricture post operative, 2 retention.

^‡G3 events were MHRT: 4 diarrhoea, 1 pain, SBRT: 2 diarrhoea, 1 incontinence

Figure 2.

Percentage of patients with Grade 1,2 and 3 toxicity for a) RTOG genitourinary, b) RTOG gastrointestinal, c) CTCAE genitourinary and d) CTCAE gastrointestinal

At week 12, RTOG G2+ genitourinary toxicity was 25/575 (4·4%; 95% CI 2·9 to 6·4) for MHRT and 37/558 (6·6%; 95% CI 4·8 to 9·1) for SBRT (p=0·12) (SupplP18, P19, P20, P21, P22).

CTCAE G2+ genitourinary toxicity was seen in 170/604 (28·1%; 95% CI 24·6 to 32·0) after MHRT and 195/581 (33·6%;95% CI 29·8 to 37·6) after SBRT (absolute difference of 5·4%; p=0·050) during the early toxicity period (Figure 2). Grade 3 GU toxicity was seen in three (0·5%) patients after MHRT and three (0·5%) patients receiving SBRT (table 2). No Grade 4 or higher CTCAE toxicity was reported. Individual genitourinary CTCAE parameters are shown in Table 2, SupplP23, P24.

Maximum CTCAE toxicity by week of follow up is shown in Figure 2 (and SupplP25). At week 12, CTCAE G2+ genitourinary toxicity was reported in 42/577 (7·3%; 95% CI 5·4 to 9·8) patients receiving MHRT and 59/564 (10·5%; 95% CI 8·1 to 13·4) receiving SBRT (p=0·074; SupplP26).

The odds of G2+ RTOG / CTCAE genitourinary toxicity with SBRT compared to MHRT was not materially altered when baseline RTOG scores / pre-treatment CTCAE genitourinary symptoms were accounted for. However, in both treatment groups, early RTOG / CTCAE G2+ genitourinary toxicity was significantly increased in those with baseline symptoms compared to those without (odds ratio for G2+ RTOG: 21·1 95% CI 9·2 to 48·5; odds ratio for G2+ CTCAE: 8·7 (95% CI 5·0 to 15·0; post-hoc analysis SupplP27, P28).

Figure 3 (SupplP29, P30, P31) shows proportion of patients meeting the threshold for MCID for each of the five main EPIC-26 domains, with SupplP31 showing the equivalent figures for twice MCID. No statistically significant differences were seen between groups in the urinary incontinence score. For the urinary irritative/obstructive score, there was no difference in the percentage of patients experiencing a MCID drop (66% MHRT vs 70% SBRT) however more SBRT patients experienced a 2*MCID drop at 4 weeks 46% vs 56% (p=0·0051) which persisted at 12 weeks (16% vs 26%; p=0·0002).

Figure 3.

Percentage of patients reaching MCID for each EPIC-26 composite score a) Urinary incontinence b) Urinary irritative/obstructive c) Bowel d) Sexual function e) Hormonal bother

Urinary irritative/obstructive scores showed no difference at 4 weeks but were worse after SBRT at 12 weeks only (SupplP32, P33). For the single question about overall “bother” from symptoms, urinary scores were similar at week 4 (SupplP34, P35).

Total IPSS score increased after treatment from a median baseline of 7 (both groups) to a peak of 14 (week 2, MHRT) and 15 (week 2, SBRT) which was statistically significant (p=0·005) (SupplP36, P37, P38). There were no differences in quality of life as scored by IPSS (SupplP38, P39).

During the early period, the co-primary endpoint of RTOG G2+ gastrointestinal toxicity was seen in 69/608 (11·4%; 95% CI 9·0 to 14·2) receiving MHRT and 74/584 (12·7%; 95% CI 10·2 to 15·8) receiving SBRT (absolute difference 1·4% [-2·5 to 5·2]; p=0·53). Grade 3 toxicity was seen in five patients (0·7%) receiving MHRT and two patients (0·3%) receiving SBRT (Table 2; Figure 2).

The profile of worst RTOG toxicity grade by week of follow-up is shown in Figure 2 (SupplP40, P41, P42). Gastrointestinal toxicity peaks at 4 weeks after the start of radiotherapy in both groups. For MHRT this is during the last week of radiotherapy and for SBRT this is two weeks after completion of radiotherapy.

At week 12, RTOG G2+ gastrointestinal toxicity was 1% MHRT vs 1% SBRT (SupplP40, P22).

CTCAE G2+ gastrointestinal toxicity was seen in 60/604 (9·9%) receiving MHRT and 96/581 (16·5%) receiving SBRT (p=0·0011) during the early period (Figure 2). Grade 3 toxicity was seen in three patients (0·5%) receiving MHRT and four patients (0·5%) receiving SBRT (table2, SupplP20).

At week 12, CTCAE G2+ gastrointestinal toxicity was 2% MHRT vs 2% SBRT (SupplP26, P43). Individual gastrointestinal CTCAE parameters are shown in Table 2, SupplP44, P45.

The odds of G2+ RTOG / CTCAE gastrointestinal toxicity with SBRT compared to MHRT was not materially altered when baseline RTOG scores / CTCAE pre-treatment symptoms were accounted for (SupplP27, P28). However, in both treatment groups, early CTCAE G2+ gastrointestinal toxicity was significantly increased (odds ratio 8·0 95% CI 3·0 to 21·5; post-hoc analysis) if baseline G2+ CTCAE gastrointestinal symptoms were recorded (SupplP27, P28). This effect was not significant for RTOG gastrointestinal toxicity.

More men experienced bowel bother after SBRT at 4 weeks (MCID seen in 204/466 (44%) MHRT versus 244/441 (55%) SBRT; p=0·0006), but no differences persisted at 12 weeks (Figure 3 (SupplP46). Median bowel composite EPIC-26 score is worse at 4 weeks after SBRT (SupplP32, P33), compared to MHRT, but recovers by 12 weeks. Bowel bother scores showed 46/512 men (9%; MHRT) versus 88/499 men (18% SBRT) experienced moderate or big problems with bowel bother at 4 weeks post-treatment (SupplP46, P47). At 12 weeks this was 19/484 (4%) vs 35/491 (7%).

There were no statistically significant differences in Vaizey scores between groups (SupplP47, P48).

No difference was seen in clinician-reported erectile function between MHRT and SBRT treated groups (SupplP48). Quality of life was significantly impacted in the sexual domain where 171/413 (41%) (MHRT) and 156/413 (38%) (SBRT) experienced a drop in quality of life at week 12, with no difference seen between MHRT and SBRT (Figure 3; SupplP31, P49).

Median sexual composite EPIC-26 score shows no difference between arms (SupplP32, P33). For the single question about overall “bother” from symptoms, sexual bother scores were similar between arms (SupplP50, P51).

There were no statistically significant differences in IIEF-5 scores between groups (SupplP51, P52). A high proportion of men in both groups recorded erectile dysfunction in the severe range, both at baseline (194/464 (42%) MHRT vs 197/445 (44%) SBRT) and at 12 weeks (272/363 (75%) vs 267/361 (74%)), with no differences between groups.

Grade 1 hot flushes and Grade 1 fatigue were common in both groups (SupplP53). The most common QOL impact was in the hormonal domain: 312/466 (67%) (MHRT) and 311/455 (68%) (SBRT) experienced a drop in quality of life due to hormonal symptoms (Figure 3; SupplP54). This persisted at 12 weeks. No effect of radiotherapy fractionation was seen.

There was evidence that centres that chose to use the 2-week SBRT schedule reported significantly higher rates of RTOG and CTCAE G2+ genitourinary (but not gastrointestinal) toxicity than centres that chose to use the 1-week schedule. This was evident in both the SBRT and MHRT groups and we suggest that this indicates a “centre effect” perhaps due to differential reporting practices for genitourinary toxicity (SupplP27, P28; P55, P57; P56, P58, P59, P60). For RTOG genitourinary toxicity, the interaction between centre’s choice of SBRT schedule and treatment received (SBRT or MHRT) did not reach statistical significance although for CTCAE genitourinary toxicity it did, suggesting that 2-week SBRT could, after accounting for the centre effect, be associated with less CTCAE G2+ toxicity than 1-week SBRT.

Toxicity events by radiotherapy machine type (CK vs non-CK; MRL vs non-MRL) are presented in SupplP61, P62. Small numbers preclude formal comparison. Of all centres recruiting, 34 (74%) of sites recruited <10 patients to PACE A/B. When assessing differences by centre experience, G2+ genitourinary toxicity was similar but early gastrointestinal toxicity was higher for SBRT at sites who had previously treated <10 patients with SBRT (RTOG 46/276 (17%) vs 28/306 (9%); CTCAE 60/275 (22%) vs 36/306 (12%) (SupplP63, P64). There was no evidence of higher genitourinary or gastrointestinal toxicity (RTOG or CTCAE) for patients who had a prior TURP (SupplP65; post-hoc analysis), although small numbers preclude formal comparison.

Discussion

To the best of our knowledge, PACE-C is the first phase III randomised trial testing SBRT against standard fractionation in a cohort including patients with high-risk prostate cancer. We have shown no significant difference in the co-primary endpoints of RTOG genitourinary and gastrointestinal toxicity with rates of toxicity remarkably similar to PACE-B.

As the target volume was larger than PACE-B (due to more seminal vesicle inclusion) and the standard arm dose was lower than PACE-B (60 Gy vs 62 Gy) it was hypothesized that toxicity could be comparatively higher than that seen in PACE-B for the SBRT arm. We saw in PACE-B that CTCAE is a more sensitive marker of toxicity, recording higher levels of symptoms than RTOG, which was the preferred instrument at the time of trial design (2011). In PACE-C we have shown that after SBRT a higher proportion of patients experience Grade ≥2 early CTCAE gastrointestinal toxicity but no statistically significant difference in CTCAE early genitourinary toxicity was seen. There was no difference in Grade 3 or higher toxicity, which was low in both groups, and no difference in either domain using the RTOG symptom scale.

These findings were mirrored in the PROs where bowel quality of life dropped more frequently in the SBRT group (55%) compared to MHRT (44%; figures quoted for MCID).

A higher rate of gastrointestinal early toxicity may be concerning if it translated into late toxicity. However, in the PACE-B trial we reported very similar differences in CTCAE early gastrointestinal toxicity (7·5% standard arm vs 14·9% SBRT arm) and yet cumulative gastrointestinal toxicity in the late toxicity period (to 2 years) were 12·3% for the standard arm and 12·5% for the SBRT arm ²⁰. This is consistent with the PATRIOT trial comparing SBRT delivered every-other-day versus weekly which showed worse early bowel and bladder toxicity for the shorter schedule, but no difference in late toxicity ^21,22. Nevertheless, even short-term gastrointestinal toxicity may be troublesome for some men and should form part of the informed consent process. Technology continues to iterate, to reduce toxicity through smaller margins, daily adaptation or intrafraction motion monitoring ²³.

In contrast, PACE-B showed a higher early G2+ CTCAE genitourinary toxicity rate for SBRT (standard arm 22·3%, SBRT 29·2%; p=0·010) which did translate into a difference in cumulative CTCAE Grade 2+ late toxicity 19·8% vs 32·3% (p=0·0001). This difference in cumulative incidence is driven by a flare of toxicity which occurs around 12 months after SBRT, and then settles such that the risk of genitourinary toxicity between 2 years and 5 years is no different ⁸.

We have shown that the risk of late toxicity is impacted by both the development of early toxicity and the presence of baseline symptoms ²⁴. This provides a potential way to stratify men for SBRT, as those with bladder issues at baseline are those most likely to experience toxicity after radiotherapy. These men may be better served by MHRT.

The PACE-C results to date mirror the results seen in the HYPO-RT-PC trial, which randomised men to either 78 Gy in 39 fractions (standard arm) or 42·7 Gy in 7 fractions, using non-SBRT techniques. This trial showed a higher chance of early genitourinary and gastrointestinal toxicity in the 7-fraction arm, yet aside from a flare of genitourinary symptoms at 12 months, no other differences in long-term sequelae of ultra-hypofractionation compared to conventional fractionation were seen ²⁵.

With PACE-B 5-year biochemical control rates after SBRT of 96%, in the absence of androgen deprivation therapy, reducing toxicity should be a focus of ongoing studies. Whilst rectal spacers have not been shown to reduce early toxicity in all studies ²⁶, they do reduce rectal dose ²⁷ which might affect gastrointestinal toxicity. Alternatively, margin reduction in combination with MRI-guidance has been shown to reduce both gastrointestinal and genitourinary toxicity in a single centre randomised trial ²³. Ongoing efforts seek to identify the structure(s) responsible for genitourinary toxicity after SBRT ²⁸. Improving technique and healthy tissue constraints may then provide a way to minimise toxicity whilst maximising tumour control and patient convenience.

The strengths of this multi-centre, multi-national trial are its size (n=1208), its high level of data completeness and comprehensive radiotherapy quality assurance programme. Whilst there remains some variability in radiation delivery, the dose, fractionation and organ-at-risk constraints were standardised.

Limitations include the timing of post-treatment toxicity assessment, confounded by the markedly different treatment length which introduces bias to the analysis, as toxicity appears to peak earlier for MHRT compared to SBRT. Analyses of SBRT duration are likely confounded by differences in toxicity reporting between centres choosing to use the 1-week SBRT schedule and those choosing the 2-week schedule. The suggestion that 2-weeks SBRT may have less genitourinary toxicity than 1-week requires further investigation. Some analysis presented uses non-randomised data, such as the analysis by platform, and therefore is inherently biased. The margins, although similar and standard at the time of trial conception, were not identical for both arms, which may have introduced a bias in the toxicity comparison. Another limitation is that unfortunately the ethnic diversity of men recruited to the trial is not representative of the populations of the recruiting countries. Efforts to ensure clinical trial populations are representative of the wider community are ongoing in our group.

Here we only report early toxicity, which can be bothersome but transient and hence we await the longer-term toxicity and oncological outcomes before considering SBRT the standard of care for this group. Many countries lack sufficient linacs or workforce to cope with the doubling of demand forecast within 15 years. SBRT goes some way to mitigating this looming crisis and also reduces the carbon footprint of radiotherapy ²⁹. SBRT is often preferred by patients, who may prioritise faster completion of treatment over temporary side effects. It is important to ensure men understand the pros and cons of shorter fractionations, aided by the data presented here.

The PACE-C trial shows that 5-fraction SBRT is associated with a higher chance of temporary bowel side effects, which settle by 12 weeks after treatment. There is no significant difference in clinician-reported bladder side effects. Long term toxicity and oncological outcomes are needed before SBRT can be recommended for this group of patients as a standard of care.

Research in context panel.

Evidence before this study

We searched Pubmed for “randomised trials” AND “stereotactic body radiotherapy” AND “prostate cancer” in the title, published from 01/01/2000 to 31/12/2018 (prior to inception of this study). No publications were found. Prior to starting PACE-C, non-randomised data had suggested acceptable toxicity rates after five-fraction stereotactic body radiotherapy (SBRT) for localised prostate cancer. In 2019 data from the randomised phase III PACE-B study suggested early RTOG toxicity was similar between patients receiving the standard arm (conventional or moderately hypofractionated radiotherapy) versus SBRT. However gastrointestinal (GI) toxicity, as measured on CTCAE scale, showed a higher rate of early side effects for the SBRT schedule. Longer term follow-up shows no difference in longer term gastrointestinal toxicity, more genitourinary toxicity at 2 years, but not thereafter. Subsequently PACE-B has demonstrated non-inferior biochemical control for the 5 fraction SBRT schedule compared to the 4-8 week standard, in the absence of androgen deprivation therapy.

Added value of this study

PACE-C recruited men with higher risk prostate cancer than PACE-B where a larger irradiated volume may be required, and radiotherapy is delivered with androgen deprivation therapy. It is the first randomised trial to compare early toxicity between SBRT and standard fractionation in the unfavourable intermediate and high-risk prostate cancer group. We have shown that short term clinician-reported gastrointestinal toxicity is higher in the SBRT group. However, by 12 weeks after radiotherapy, there is no statistically significant difference between groups.

Implications of all the available evidence

SBRT is associated with a higher rate of temporary bowel side effects compared to moderately hypofractionated radiotherapy, but by 12 weeks after therapy, no difference in side effects is evident. Longer follow-up of the oncological outcomes of PACE-C is required before a change in standard practice can be recommended.

Data sharing

The ICR-CTSU supports the wider dissemination of information from the research it conducts, and increased cooperation between investigators. Trial data is collected, managed, stored, shared and archived according to ICR-CTSU Standard Operating Procedures in order to ensure the enduring quality, integrity and utility of the data. Formal requests for data sharing are considered in line with ICR-CTSU procedures with due regard given to funder and sponsor guidelines. Requests are via a standard proforma describing the nature of the proposed research and extent of data requirements. Data recipients are required to enter a formal data sharing agreement which describes the conditions for release and requirements for data transfer, storage, archiving, publication and Intellectual Property. Requests are reviewed by the Trial Management Group (TMG) in terms of scientific merit and ethical considerations including patient consent. Data sharing is undertaken if proposed projects have a sound scientific or patient benefit rationale as agreed by the TMG and approved by the Independent Data Monitoring and Steering Committee as required.

Restrictions relating to patient confidentiality and consent will be limited by aggregating and anonymising identifiable patient data. Additionally, all indirect identifiers that may lead to deductive disclosures will be removed in line with Cancer Research UK Data Sharing Guidelines.