Trial Design
TRANSLATE is a parallel group, superiority, multicentre RCT comparing LATP against TRUS biopsy, with respect to detection of clinically significant PCa. The study features two-stages: The first stage is a randomised internal pilot which aims to demonstrate the feasibility of identifying and randomising men with a suspicion of PCa; The second stage comprises the full definitive trial, aiming to assess the effectiveness of LATP versus TRUS biopsy in diagnosing clinically significant PCa, whilst additionally investigating the two biopsy techniques in terms of their risk of significant infection-related and other complications, tolerability, impact on quality of life, and cost-effectiveness.
Randomisation
Eligible men who consent to join TRANSLATE are randomised to one of the two interventions in a 1:1 ratio using minimisation, minimised by recruiting centre and Magnetic Resonance Imaging (MRI) lesion location (i.e. ‘No significant lesion’, ‘Significant lesion, including anterior’, ‘Significant lesion, but not anterior’). The minimisation algorithm includes a probabilistic element, and a small number of participants are initially randomised by simple randomisation, to ensure the unpredictability of biopsy allocation. Minimisation by recruitment centre will help ensure that any centre-effect will be equally distributed in the trial arms, and enable potential practical issues associated with biopsy delivery or site closure to be accounted for. There is some evidence that MRI lesion location can affect PCa detection, therefore it is important for the two prostate biopsy techniques to be balanced across this potentially important prognostic factor. The randomisation is performed via a secure centralised web-based system (“Registration/Randomisation and Management of Product”) via the REDCap database interface at the Oxford Clinical Trials Research Unit (OCTRU), consistent with UK Clinical Research Collaboration-approved standard operating procedures. This system ensures prospective registration and allocation concealment until the point at which the patient enters the trial.
Eligibility
Potential participants are eligible for the TRANSLATE trial if they are biopsy-naïve men aged ≥ 18 years who, during investigation for suspicion of possible PCa, require a prostate biopsy, with a prostate-specific antigen (PSA) value above the age-adjusted upper limit of normal regardless of the pre-biopsy MRI result, or an abnormal pre-biopsy MRI on a 1.5 Tesla or higher MRI scanner, or an abnormal prostate digital rectal examination (DRE) (regardless of serum PSA or MRI result). All participants are considered suitable to tolerate an LATP biopsy procedure by the local clinical team, able to give informed consent, and able to understand written English. The following patients will not be able to join the study: any individuals with a previous prostate biopsy; dysuria on the day of biopsy or untreated urinary tract infection (UTI); immunocompromised (due to history of prior immunocompromising medical condition, or medications e.g. steroids or methotrexate); those who may need enhanced antibiotic prophylaxis (such as those with an indwelling catheter, or who have recurrent UTIs); previous abdomino-perineal resection (i.e. absent rectum); unable to recline adequately in Lloyd-Davis / lithotomy position (e.g. hip surgery, contractures); unable to have a pre-biopsy MRI (e.g. pacemaker, eGFR); PSA > 50ng/ml (> 25ng/ml if patient is on Finasteride) (to exclude patients with locally advanced/metastatic PCa, easily detectable by either biopsy technique).
Blinding
The study participants and clinical team delivering the biopsy cannot be blinded to the allocated procedure due to the nature of the intervention. The Chief Investigators (CIs) will remain blinded to treatment allocation for the overall clinical trial participants (i.e. knowledge of treatment allocation is limited to those participants at the CIs’ own site). In instances where serious adverse events (SAEs) are reported, the CIs will be unblinded to complete a full causality assessment. The DSMC and TSC will be blinded to biopsy allocation when reviewing data. The trial statistician and data entry personnel are not blinded to allocation. The remaining members of the trial management team are blinded to allocation until the completion of data analysis.
Objectives
The primary objective of the TRANSLATE trial is to compare TRUS biopsy versus LATP biopsy in the detection of clinically significant PCa (defined as Gleason Grade Group (GGG) ≥ 2, i.e. any Gleason pattern ≥ 4 disease).
Secondary objectives are to investigate if there are any differences between the two biopsy techniques in the following:
-
Rates of infection
-
Participant reported tolerability of the procedure
-
Participant reported biopsy-related complications
-
Quality of Life
-
Histological parameters, including GGG ≥ 3 PCa
-
Number of subsequent prostate biopsy procedures required and details of the subsequent biopsies
-
Burden and rate of detection of clinically insignificant (GGG1) PCa
-
SAE incidence.
Outcomes
Primary Outcome
The primary outcome is the difference in detection rate of clinically significant PCa upon pathology reporting of biopsy samples between the TRUS and LATP biopsy groups, clinically significant disease being defined as a GGG (ISUP) score of 2 or more, i.e., any Gleason Pattern ≥ 4 disease (3). The GGG score is a pathology-based endpoint and is usually available within 7 days of the initial biopsy, although difficult cases or other pathway delays may result in a longer period.
Secondary Outcomes
The secondary outcome measures assessed in TRANSLATE and associated time points are described in Table 1 and presented in detail below.
Table 1
Time points at which outcomes are assessed.
|
Time point
|
Outcome Measure
|
|
Baseline
|
Participant demographics (Charlson Comorbidity Index, cancer history, DRE, PSA, current medications, body mass index, ethnicity)
IIEF questionnaire (Domain A)
I-PSS questionnaire
EQ-5D-5L questionnaire
|
|
Immediately post-procedure
|
ProBE questionnaire (Perception part)
Histological parameters (ISUP grade group*, cancer core length, core involvement, target biopsy cancer parameters)
Burden and rate of detection of clinically significant (GGG ≥ 2) PCa
Burden and rate of detection of GGG ≥ 3 PCa
Burden and rate of detection of clinically insignificant (GGG 1) PCa
|
|
7 days post-procedure
|
ProBE questionnaire (General Symptoms part)
IIEF questionnaire (Domain A)
I-PSS questionnaire
EQ-5D-5L questionnaire
Infection
Complications and SAEs
|
|
35 days post-procedure
|
IIEF questionnaire (Domain A)
I-PSS questionnaire
EQ-5D-5L questionnaire
Infection
Complications and SAEs
Number of subsequent prostate biopsy procedures
|
|
4 months post-procedure
|
IIEF questionnaire (Domain A)
I-PSS questionnaire
EQ-5D-5L questionnaire
Infection
Complications and SAEs
Number of subsequent prostate biopsy procedures and details of the subsequent biopsies
|
* includes the primary outcome measure, the difference in detection rates of clinically significant PCa, defined as Gleason Grade Group ≥ 2, i.e. any Gleason pattern ≥ 4 disease.
The International Index of Erectile Function (IIEF) Domain A (4) is a participant-reported questionnaire with six questions that examine male erectile function. Each question is marked from 0 to 5, with a total score range between 0 and 30, where lower values represent worse outcomes.
The International Prostate Symptom Score (I-PSS) (5) consists of seven sections concerning urinary symptoms, each scored on a scale from 0 to 5 (0 is “not at all” and 5 is “almost always”). The total I-PSS score ranges from 0 to 35, with higher scores indicative of more severe symptoms. The seven subsections are “incomplete emptying”, “frequency”, “intermittency”, “urgency”, “weak stream”, “straining” and “nocturia”. The I-PSS questionnaire also includes a “quality of life (QoL) due to urinary symptoms” measure with a score from 0 to 6 where 0 means “delighted” and 6 is “terrible”.
The EQ-5D-5L is a generic health-related QoL measure consisting of 5 dimensions each with a 5-level answer possibility and a health thermometer visual analogue scale (EQ VAS) (6, 7). The EQ-5D-5L will be converted into a health utility score where 1 represents perfect health and 0 indicates health states equal to death, to report health-related QoL (8). Negative values on the utility scale are possible. The VAS takes values between 0 and 100, where 0 represents worst imaginable health and 100 best imaginable health.
Participant-reported tolerability of the prostate biopsy procedure will be measured using the Prostate Biopsy Effects (ProBE) questionnaire (Perception part) (9) completed by the participant immediately after the procedure. Participants will report biopsy-related complications (such as bleeding, pain, urinary retention) as part of the Resource Use questionnaire at all follow-up timepoints and using the ProBE questionnaire (General Symptoms part) at 7 days after the procedure (9).
Infection in this trial is defined as any infection resulting in hospitalisation as indicated by data collected from patient notes and reported by the clinicians throughout the trial, as well as data reported by the participants in the participant questionnaire at 7 days, 35 days and 4 months post randomisation. A second broader definition of infection will include any infection resulting in hospitalisation as well as any infection not resulting in hospitalisation measured at 7 days, 35 days and 4 months post randomisation.
The number of subsequent prostate biopsy procedures will be collected using the participant questionnaire at 35 days and 4 months post randomisation and reported by clinicians in the 4 months post-biopsy case report form.
Burden and rate of detection of clinically insignificant PCa will be reported by clinicians in the histology report at 7 days post randomisation. Clinically insignificant PCa will be defined as GGG1 disease (3). We will also report on rate of detection of GGG ≥ 3 PCa as an alternative definition of clinically significant disease.
Sample size
The target sample size is 1,042 randomised participants (521 per group). This assumes a detection rate of clinically significant PCa in previously biopsy naïve individuals through TRUS biopsy following a pre-biopsy MRI of 45%, in line with the reported literature and data collected from 792 patients in Oxford over a 12-month period (10). We consider an improvement of 10% (i.e. from 45–55% from TRUS to LATP) in the rate of detection of clinically significant PCa (defined GGG ≥ 2, i.e. any Gleason pattern ≥ 4 disease) through LATP to be a clinically significant primary outcome difference. This sample size assumes 90% power and 5% significance to detect a statistically significant difference between the two biopsy procedures in the primary outcome.
Statistical Analysis
General Analysis Principles
We will carry out all analyses on the intention-to-treat population (ITT) (i.e. participants will be analysed in the group they were randomised to regardless of actual biopsy received). We do not anticipate significant numbers of protocol deviations, but if any occur, we will repeat the primary analysis in the Per Protocol (PP) population. The PP population is a subset of the ITT population and will exclude participants with major protocol deviations (e.g. due to clinician decision, participant preference, or withdrawal), or where the participant did not satisfy the eligibility criteria, or where the biopsy was not conducted in accordance with the protocol as reported by sites (e.g. greater than or fewer than the required range of biopsy core numbers).
Standard descriptive statistics will be used to describe the participant baseline demographics for each of the two biopsy groups, reporting means and standard deviations (SD) or medians and interquartile ranges (IQR) as appropriate for continuous variables and numbers and percentages for binary and categorical variables. Analyses will be conducted using Stata (StataCorp LP, www.stata.com) or other well-validated statistical software. All comparative outcomes will be presented as suitable measures of intervention effect and reported together with 95% confidence intervals, and all tests will be carried out at a 5% two-sided significance level.
Statistical Interim analyses
No formal interim analyses of the primary or key secondary outcomes with stopping guidelines are planned. An independent DSMC will review the accumulating data at regular intervals and may recommend pausing or stopping the trial in the event of safety concerns.
The first stage internal pilot will have a formal ‘stop/go’ review of the randomisation numbers after 6 months of recruitment to the RCT. If the target of at least 140 randomisations has been met, the trial will continue to recruit for a further 9 months. Data from these 140 participants will be included in the final analysis alongside data from the study participants recruited once the pilot period ends.
Statistical Significance and Multiple Testing
The trial is powered for a single pre-specified primary endpoint and pre-specified analysis plan, and therefore no adjustment for multiple testing will be conducted.
Description of Study Patient Throughput
The flow of participants through each stage of the trial, including number of participants screened, numbers eligible, numbers and reasons for ineligibility, numbers giving their consent and randomised, receiving the allocated intervention, and analysed for the primary outcome analysis will be provided using a CONSORT PRO extension flow chart (11) as per Fig. 1.
Withdrawal from Intervention and/or Follow-up
The number and percentage of participants withdrawn from the trial will be summarised by intervention group, with an indication of whether the withdrawal was before or after the biopsy procedure. In addition, details on the average time (in days) to withdrawal from randomisation and the reasons for withdrawal will also be summarised, where known. If any deaths occur during the study, the number and percentage as well as time from biopsy to death (in days) will be summarised.
Missing Data
The number and percentage of participants with available data for each outcome at each time point will be summarised by intervention group. Patterns of missingness will be investigated, and the suitability of missing data assumptions considered. Reasons for data missingness, if known, will be explored and described.
Approach to Missing Data in Analyses
The main analyses will be performed on an available case basis. For the primary outcome the effect of missing data will be investigated, and multiple imputation (MI) used if more than ten percent of primary outcome data is missing under a missing at random (MAR) assumption. For outcomes measured at multiple time points and analysed using repeated measures mixed effects models, unavailable observations either due to missed visits or due to a participant leaving the trial early are assumed to be similar to observed outcomes from similar participants at the same time points under a MAR assumption (12). We do not anticipate using MI for missing outcome data in the repeated measures analysis, as MI also assumes a MAR mechanism and is therefore not expected to add value to the analysis model. If the data appear to be missing not at random, then alternative approaches, such as the rctmiss command in Stata (or an equivalent if another statistical package is used) will be utilised to assess the impact of the assumptions on the estimated intervention effect.
Description of Compliance with Intervention
This trial will compare two methods that obtain prostate biopsies to detect clinically significant PCa (i.e. cases of PCa that are likely to require treatment):
A Transrectal biopsy – known as a TRUS – whereby an ultrasound imaging probe is placed in the rectum and through this a biopsy needle is inserted into the prostate gland to take prostate tissue biopsy samples.
A Local Anaesthetic Transperineal biopsy – known as a LATP – whereby an ultrasound probe is placed in the rectum, but instead of the biopsy needle passing up the ultrasound probe and through the wall of the rectum, it passes directly through the skin of the perineum between the anus and the scrotum under local anaesthetic.
Intervention compliance in this trial is defined as the participant receiving the biopsy procedure to which they were randomised. Data on biopsy received will be summarised together with reasons why the allocated biopsy was not administered to the participant where this is relevant. The impact of non-compliance with the allocated intervention will be explored as part of the PP population analysis (as described in the Analysis Methods section below).
Analysis Methods
Analysis of primary outcome
The primary outcome is the difference in detection rate of clinically significant PCa upon pathology reporting of biopsy samples between the TRUS and LATP biopsy groups, clinically significant disease being defined as a GGG (ISUP) score of 2 or more, i.e., any Gleason Pattern ≥ 4 disease (3). Detection rates for the two biopsy intervention groups will be compared using a mixed effects logistic regression model adjusted for the minimisation factors (recruitment centre, and site of prostatic lesion on pre-biopsy MRI as “no significant lesion” versus “significant lesion including anterior” versus “significant lesion, but not anterior”). Recruitment centre will be fitted in the model as a random effect to account for any heterogeneity due to centre, and the radiologist-confirmed MRI pre-biopsy site of prostatic lesion will be fitted as a fixed effect. The adjusted difference in detection rate of clinically significant PCa between the two groups will be presented as an Odds Ratio (OR), along with its associated 95% CI. The primary analysis of the primary outcome will be conducted for the ITT population using available cases and will be repeated in the PP population.
Supporting Analysis
As supporting analyses, we will conduct an unadjusted analysis, and a further analysis adjusting for additional important prognostic factors measured before biopsy. These additional model covariates will include age (years) as a continuous variable, PSA level (ng/ml) as a continuous variable, MRI tumour stage (using three categories: 1c [no lesion]; 2a, 2b, 2c [localised PCa]; 3a, 3b, 4 [locally advanced PCa]) as categorical variables, and cancer risk group (using two categories: breast or PCa family history; no PCa history) as categorical variables. The proportion of participants in each randomised group with positive and negative biopsy results will be tabulated, and the difference between groups reported as ORs and absolute differences together with 95% CI.
Analysis of secondary outcomes
All secondary outcomes will be assessed by intervention group for the ITT population at the follow-up data collection time points. Raw means and SDs will be reported for continuous outcomes, and counts with percentages for binary ones. Continuous outcomes will be analysed using mixed effects linear regression, and logistic regression will be used to analyse binary data. Multilevel models will be used for variables measured at multiple time points (I-PSS, IIEF, EQ-5D-5L) to allow for repeated measures clustered within each participant. All secondary outcome models will be adjusted for baseline measurement where possible and for the following minimisation factors: recruitment site (as a random effect) and radiographer-confirmed MRI lesion location (as a fixed effect). An intervention by time point (as categorical variable) interaction will be included in the model, indicating the follow-up time point to which the outcome refers. The intervention effect at each time point, as well as the overall effect across time points, will be presented as ORs or mean difference (with SDs), depending on outcome. Model assumptions, including approximate normality of the residuals, will be assessed as relevant. All secondary outcome analyses will use two-sided 5% significance and 95% confidence intervals with associated p-values reported throughout.
ProBE – Perception and ProBE – General Symptoms
The ProBE – Perception data collected immediately after the biopsy procedure will assess the between group difference in the proportion of participants reporting a moderate or major problem in one or more items. The ProBE – General Symptoms part collected at 7 days post biopsy will assess the between group difference in the proportion of participants reporting a moderate or major problem in one or more complications (9). Models to analyse this data will use mixed effects logistic regression and will be adjusted for the minimisation factors (recruitment centre and lesion location) as described above.
Infection
The primary definition of infection in this study is any symptoms or signs of infection resulting in hospitalisation. A secondary, broader definition of infection will additionally include any infections not requiring hospitalisation as reported through participant post-procedure questionnaire, ProBE-General Symptoms questionnaire or SAE forms.
Complications and subsequent biopsy procedures
The number and percentage of participants experiencing each type of biopsy-related complication, and those experiencing a biopsy-related complication at least once (such as urinary retention, blood in bowel movements, urology admission due to urinary bleeding, and urology admission due to pain), will be presented by intervention group and compared. Biopsy-related complications such as reduced sexual or urinary function will be reported as part of the secondary outcome IIEF (4). The number of participants undergoing a subsequent biopsy procedure over the 4-month trial follow-up period and details of the subsequent biopsies will be summarised by intervention group.
Histology parameters
Histological parameters (ISUP grade group, cancer core length, core involvement, target biopsy cancer parameters) will be summarised descriptively by intervention group and compared. An additional assessment of clinically significant PCa will be conducted based on a definition of GGG ≥ 3 (equivalent to Gleason Score 7 (4 + 3), Gleason Score 8 or Gleason Score 9–10) and compared between the two biopsy types. Participant burden rate will be described and evaluated based on the rate of clinically insignificant PCa detection (i.e. GGG score of 1).
Safety
It is anticipated that the rate of SAEs will be low in this trial. The total number of SAEs will be summarised by intervention group alongside the number and percentage of participants reporting at least one SAE. Details of the events, including expectedness and relatedness of the SAEs, will be presented, together with information on the timing of the events. The proportion of participants with at least one SAE will be compared.
Pre-specified Subgroup Analysis
Subgroup analyses for the primary outcome are planned. We will assess whether the intervention effect differs between patient groups with differing MRI lesion status (recorded as ‘significant lesion including anterior’, ‘significant lesion not anterior’, or ‘no significant lesion’) and prostate volume (normal < 50 cc, large 50–79 cc, extra-large ≥ 80 cc). We hypothesise that the detection rate of clinically significant PCa will be higher in participants with an anterior lesion, and in those with a larger prostate, on MRI receiving an LATP versus TRUS biopsy (13). Where numbers within subgroups are too small, we will consider combining levels within subgroups (e.g. combine prostate volume 50–79 cc with ≥ 80 cc). Intervention allocation by subgroup interaction terms will be used to assess point estimates and confidence intervals and will be displayed using forest plots.
Sensitivity Analyses
Sensitivity analyses will be conducted to assess the robustness of the main trial results. Sensitivity analyses to explore the impact of missing primary outcome PCa detection are outlined in the Approach to Missing Data in the Analyses section.
Supplementary Analyses and Outcomes
Health Economics
Details of the Health Economics analyses are described in the study protocol(2) and will be undertaken separately by a Health Economist.
Number of biopsy cores
The number of recommended biopsy cores to be taken for each type of intervention is stipulated in the protocol such that it is expected to be similar across the two biopsy groups. The number and proportion of participants with too many (more than 18 systematic cores) or too few (less than 6 systematic cores) will be presented for each randomised group.
Bayesian re-design methodological add-on
A methodological Bayesian re-design of the trial data was pre-specified in the study protocol and is being conducted alongside the main trial (2). This re-design intends to assess whether analysing TRANSLATE using novel Bayesian adaptive methods may provide efficiency gains over the more typical trial design outlined in this SAP. The final reporting of trial results will be presented in line with the aspects described above and will be considered independent of the methodological outcomes. Details of the Bayesian methodology employed are described in the Additional file Appendix 1.
Statistical packages
All analysis will be carried out using appropriate validated statistical software such as Stata, SAS, SPLUS or R. The relevant package(s) and version number(s) will be recorded in the Statistical Report.