Advanced Diffusion Magnetic reSonance Prostate Imaging for Reducing Extraneous Biopsies (ASPIRE) - Calibration Component (Multi-centre Scanner Calibration) and Validation Component (Multi-centre Validation) (ASPIRE MS)

Overview: ASPIRE constitutes three components:

• Scanner Set-up Component (optional), at all sites (except University College London Hospitals NHS Foundation Trust (UCLH)),
• Multi-centre Calibration Component and
• Multi-centre Diagnostic Accuracy Validation Component Scanner Set-up Component (optional): A multi-centre optional optimisation study conducted across multiple National Health Service (NHS) hospital sites, apart from UCLH, once the advanced MRI sequences (VERDICT, fast VERDICT, and RSI) have been installed or programmed on each scanner by a dedicated research engineering team. Sequence parameters (such as b-value distributions, diffusion times, and echo times) will be standardised as much as possible across platforms, recognising that some vendor-specific adjustments may be required. This component includes a very small, optional cohort of local volunteer patients who are already undergoing a planned, routine MRI scan. Scans in this component are planning to be used as a quality assurance measure once these advanced sequences are installed ahead of the calibration cohort.

This Optional Scanner Set Up Component is not powered to test a clinical hypothesis; instead, its endpoint is the successful implementation of protocols and generation of reliable imaging at each site. As a process measure, success will be declared if the advanced sequences are operational on all scanners with adequate image quality. Men will be referred to this component based on a clinical suspicion of prostate cancer. Patients can expect 1 standard MRI scan plus all additional study sequences. The standard MRI scan can be either a bi-parametric or multiparametric MRI scan depending on the availability of the site where the scan is being performed. Patients will complete this MRI and exit the study with no follow up planned for this component.

Multi-centre Calibration Component: A technical feasibility assessment study involving NHS hospital sites equipped with 3 Tesla MRI scanners from each major MRI vendor (Philips, Siemens, GE) with potential for additional sites to represent duplicates from vendors, covering platform variations. The participating site selection captures the vendor diversity likely to be encountered in a larger trial. At each site, the VERDICT, fast VERDICT, and RSI MRI sequences will be installed and tested. This component involves a small travelling volunteer cohort design of patients who will undergo the advanced diffusion research MRI scan on two different scanners within a short timeframe. By scanning the same individuals at different sites, within a small timeframe, we are enabling direct comparison of measurements and direct calibration of the quantitative parameters can be performed. This component seeks primarily to develop methods (proof-of-concept) and is not designed to evaluate clinical outcomes.

The primary data collected in the Calibration Component are the imaging metrics (e.g., VERDICT Fractional intracellular volume (fIC) values) from each scanner for each subject. These will be analysed to derive transformation formulas that map values from one scanner to an equivalent value on another, effectively creating a calibration curve for each scanner/vendor combination. This component will also record practical aspects such as the success rate of sequence deployment, any scanner-specific issues, and qualitative image quality assessments at each site.

The Calibration Component is not powered to test a clinical hypothesis; instead, its endpoint is the successful implementation of protocols and generation of calibration equations. As a process measure, success will be declared if the advanced sequences are operational on all scanners with adequate image quality, and if the analysis of the travelling volunteer data shows a consistent relationship (correlation) between scanners. For example, preliminary data already suggest that VERDICT fIC values on a Siemens scanner are systematically lower than on a Philips scanner, but with a linear mapping that could adjust one to the other. This component will formally validate this across more subjects and across all vendor pairs.

Men will be identified based on whether they are in current active surveillance for previously diagnosed low-risk prostate cancer or awaiting definitive treatment for clinical suspicion of prostate cancer. Patients will be informed that travel as a part of this component is mandatory and should not provide full informed consent if travel is not possible for them. Participants in this cohort will have 2 scans at 2 different major vendors, one will occur at UCLH and the other at a secondary site within 2 months of the first scan. As soon as the secondary scan is completed, their participation ceases and will resume standard of care treatment.

Multi-centre Diagnostic Accuracy Validation Component: A multi-site diagnostic accuracy study with embedded technical comparison of patients undergoing initial evaluation for suspected prostate cancer at four external NHS sites to assess and compare the diagnostic performance of VERDICT, fast VERDICT, and RSI (additional research sequences) when added to standard MRI as per normal clinical protocol) in the same session. Following SoC imaging, all patients will proceed to the normal diagnostic pathway: those with MRI findings warranting biopsy (typically Prostate Imaging - Reporting and Data System (PI-RADS) or Likert score ≥3) will undergo a targeted biopsy of the identified lesion(s) and any appropriate systematic sampling as per local standard. Patients with negative MRI (score 1-2, indicating low suspicion) generally do not undergo immediate biopsy as part of standard care; these patients will be managed according to standard local practice (which may involve clinical observation and repeat Prostate-Specific Antigen (PSA) or MRI at a later date). Importantly, the results of the advanced diffusion MRI sequences will not be used to direct patient management in the Validation Component.

For the patients who undergo a biopsy, as directed by the clinical findings of their MRI scan, while the results of the biopsy will be blinded, we will know the lesions that were identified on MRI that underwent the biopsy. Using this location information, we will draw the lesion onto the VERDICT, fast-VERDICT, RSI maps to determine the diagnostic accuracy of the MRI with the additional advanced diffusion imaging (each of the three techniques independently), allowing comparison of diagnostic accuracy within the same individuals, which is referred to as a paired within subject design. Each patients MRI will be interpreted both with and without the retrospective advanced diffusion metrics. This design allows evaluation of the new tests against the reference standard without altering the standard-of-care management. It is ethically appropriate because it avoids withholding recommended biopsies or performing unnecessary ones purely for research purposes. There is no randomisation; every participant receives all imaging tests under evaluation. The interpretation of the advanced diffusion MRI results, however, will be retrospective and not influence immediate clinical management, to avoid introducing any experimental risk.

All components are prospective in that participants are enrolled going forward with data collected in real-time. The protocol will be implemented concurrently at multiple centres, leveraging each site's clinical workflow for prostate MRI and biopsy. The Calibration Component will begin concurrently to the Validation Component, and the calibration results will inform the analysis of the validation data.

After all imaging and biopsy procedures, an independent evaluation of the MRI data will be conducted at the sites as per standard-of-care. The MRI for each patient will be interpreted in two ways: (a) using standard MRI alone (using PI-RADS v2.1 or local Likert scoring for suspicion), and (b) using MRI plus each advanced method (e.g., MRI+VERDICT). For the combined interpretation, predefined criteria will be used such as: "Consider an MRI-suspicious lesion as benign (no biopsy needed) if VERDICT fIC is below threshold X" - where X may be a site-calibrated threshold corresponding to ~90% sensitivity for cancer detection. By comparing these interpretations to the biopsy-confirmed outcomes, diagnostic performance characteristics for each approach can be calculated. The primary analysis will focus on specificity (proportion of patients without significant cancer who are correctly identified as such and spared biopsy) and sensitivity (proportion of patients with significant cancer correctly identified) of the new pathway versus standard care.

All Validation Component sites will follow a common master protocol for imaging and data collection to ensure consistency. The study is not randomized; every participant effectively serves as their own control for comparing tests. The Validation Component is designed in accordance with STARD (Standards for Reporting Diagnostic Accuracy Studies) guidelines, including clearly defined index tests and reference standards, blinded interpretation, and paired analysis.