A novel approach to integrated prostate cancer diagnostics: insights from MRI, prostate biopsy, and pathology reports in a pilot study

In the diagnostic pathway for early detection of prostate cancer, MRI, biopsies, and pathology are included. The shift towards ‘MRI first’ avoids unnecessary biopsies and reduces overdiagnosis [1] but demands high standards for Prostate Imaging-Reporting and Data System (PI-RADS) [2] grading and accuracy in targeted biopsies. Formulating a comprehensive assessment for optimal patient management is challenging for today's urologists as the volume and complexity of data grow. Quality registers are crucial for evaluating, comparing, and improving patient care. The Swedish National Prostate Cancer Register (NPCR) encompasses over 240 000 patients and is routinely used to monitor and evaluate care quality, aiming to optimise patient outcomes through transparent comparisons [3].

Traditionally, reporting to quality registries relies on manual post-registration, which has disadvantages including time consumption, registration errors, and incomplete data. A multi-professional group within the Swedish NPCR developed a nationally endorsed novel model for integrated prostate cancer diagnostics, tested in routine care in autumn 2023. The model (Figs 1, S1 and S2) is based on ‘single entry’ per specialist and comprises six key components: complete data registration, real-time data registration, structured MRI and pathology review and reporting, structured biopsy sampling and referral, individualised automated feedback from biopsy to radiologist, and summarised outcomes from MRI, biopsy, and pathology in an interactive format.

Fig. 1

Open in figure viewerPowerPoint

An overview of the model with the workflow from MRI to pathology, feedback loop (top) and registry data (bottom). MDT, multidisciplinary team meeting.

The workgroup included urologists, radiologists, pathologist, urology nurse, patient representative, product owner from the Swedish Information Network for Cancer (INCA), and a project manager. The model was evaluated by a national reference group that included key opinion leaders holding leadership positions for the National Clinical Care guidelines for prostate cancer and the organised prostate cancer testing from urology, pathology, and radiology. Data management occurs within the Swedish INCA, with designated personnel ensuring adaptation to a national model with modular data collection for quality registries and structured frameworks within collaborative groups. The system ensures vendor neutrality and compatibility with various data sources.

The model serves as a tool for structured MRI prostate review and reporting based on PI-RADS version 2.1. Radiologists can register directly in the MRI template on the Swedish INCA database or follow their standard clinical workflow in a Picture Archiving and Communication System (PACS) mirrored with the Swedish INCA (Sectra PACS was used in the pilot study, Sectra AB, Linköping, Sweden). A structured report is automatically generated as data are entered into the MRI template. Prostate volume calculation to obtain PSA density is mandatory. For focal lesions, zone, sector, size, and PI-RADS scoring are mandatory fields. The correct zone and sector are automatically generated when the radiologist outlines the lesion in a biopsy template, serving as foundational information for the urologist in subsequent biopsy procedures. For PI-RADS ≥4, grading of extraprostatic extension and seminal vesicle invasion is mandatory, as well as evaluation of potential metastatic lesions in lymph nodes or bone marrow. Data registered in the PACS system are transferred and registered in the MRI template on the Swedish INCA when approved by the radiologist. Feedback from biopsy outcomes can be accessed through the Swedish INCA or automated feedback in a dynamic worklist in the local PACS system.

The urologist records the localisation and number of biopsies in the Swedish INCA template pre-filled with MR lesions, ensuring targeted biopsies are registered correctly. Information regarding transrectal or transperineal biopsy approach and whether biopsies are cognitive or software-assisted is noted. For transperineal biopsies, multiple sectors in the vertical dimension can be registered on the same biopsy. A structured referral to the pathologist is generated in real-time as registrations are made.

The pathologist records data in the Swedish INCA template pre-filled with MRI and biopsy information, resulting in a structured report. For each biopsy, sector(s) and biopsy length are registered. For adenocarcinoma, Gleason grade is selected from a drop-down menu, and if Gleason 7 is recorded, percentage Grade 4 and potential presence of a cribriform pattern are mandatory. A suggestion for the global Gleason score and an International Society of Urological Pathology (ISUP) grade for each lesion is automatically generated. Total cancer length and biopsy length are calculated for systemic biopsies. Invasion (i.e., perineural, vascular) is registered when present. The structured report is auto-populated and can be duplicated to the pathologist's local Laboratory Information System.

After these steps, a synoptic summary of the diagnostic pathway is available with complex information presented clearly and interactively. Lesion localisation and PI-RADS score, number and type of biopsies, and outcomes are summarised, functioning as referral response to the urologist and feedback to the radiologist. The overview can be used during patient meetings or in multidisciplinary conferences to quickly summarise complex information.

This novel model was tested at the urology clinic in Helsingborg over 3 months in autumn 2023. In all, 98 patients underwent biopsies, with all pathway steps verified and monitored. User experiences were collected through surveys, and data transfer between systems and the Swedish INCA was monitored. The pilot study showed effective data flow between pathway components, from radiologist's registration in the local PACS to biopsy referrals, pathology referral responses, and automated feedback to radiologists regarding biopsy outcomes.

Surveys indicated high satisfaction among users. The global rating for using the templates was 8.5/10 among 14 urologists, 9.5/10 among eight radiologists, and 9/10 among five pathologists. Feedback of biopsy outcomes to radiologists was highly appreciated (10/10). Documentation in templates was considered time-saving by radiologists (5 min shorter per case) and pathologists, though urologists experienced a 3-min increase per case. Urologists were pleased with the biopsy template functioning as a pathology referral (8/10).

Using the model requires only one registration per specialist, ensuring complete and accurate data for the quality register. Evaluation can be done nationally, regionally, and at the clinic and individual levels. The pilot study found that the model is highly appreciated by radiologists, urologists, and pathologists.

The model addresses three challenges in MRI prostate diagnostics: the importance of structured review and reporting, and radiologists learning from biopsy feedback [4]. The solution with automated feedback is unique, based on integration between quality registers and PACS. ‘MRI first’ relies on adequate PI-RADS scoring, and challenges with inter-reader variation are well known [5]; hence, the feedback loop in our model is expected to reduce false-positive MRIs.

For urologists, the model's greatest benefit is a concise summary of all relevant information from the integrated pathway. Additionally, biopsy templates provide an automatically generated structured referral to pathology. Narrative free-text reports are still common in pathology and radiology, making complex information communication challenging. The model provides pathologists with a structured review roadmap and results in a structured report, used as feedback to urologists and radiologists. The workflow is time-saving for radiologists and pathologists, and synoptic templates are advantageous [6]. Consolidated information was well received as feedback to radiologists and had added value in multidisciplinary conferences.

A positive outcome is the model's use in organised prostate cancer testing in several large health regions, offering unique quality monitoring and evaluation opportunities [7]. After the pilot study's positive outcomes, the model is being implemented in a health region with 1.4 million inhabitants in southern Sweden. A sustainable, long-term plan for resource allocation, communication, training, and support functions is in place to ensure successful implementation. Once implemented, the model will provide opportunities for validating a wide range of diagnostic and therapeutic metrics. Ongoing efforts include continuous refinement, such as incorporating prostate-specific membrane antigen positron emission tomography/CT data and output from radical prostatectomies.

In summary, we present a novel, nationally accepted model, integrated into clinical practice, to enhance the precision and quality assessment of prostate cancer diagnostics. The model's structure and design can be generalised to other diagnoses where quality assessments of integrated diagnostic pathways are of interest.