Authors (including presenting author) :
Drs Ngai KS (1), Wong SC (1), Chau CM (1), Ng FH (1), Lam SY (1), Wong T (1), Pan NY (1)
Affiliation :
(1) Department of Diagnostic and Interventional Radiology, Princess Margaret Hospital
Keyword 1: :
Prostate cancer
Keyword 2: :
Multiparametric MRI
Keyword 3: :
PI-RADS v2.1
Keyword 4: :
PI-QUAL Version 2 Scoring System
Introduction :
Prostate cancer is the third most common cancer among men in Hong Kong and fourth leading cause of cancer death. The overall five-year relative survival rate was estimated to be 84%, survival varies significantly by stage, exceeding 97% for stages I to III but dropping to 45% for stage IV. The incidence of prostate cancer has risen steadily, increasing by 69% between 2012 and 2022. Multiparametric MRI (mpMRI) has revolutionized the diagnosis of prostate cancer in terms of more accurate local tumour staging, population screening and active surveillance for some indolent disease. Performing MRI before biopsy also minimizes the risk of overdiagnosis and overtreatment of indolent cancers by reducing unnecessary biopsies. The routine use of MRI has increased steadily, supported by its integration into modern prostate cancer management guidelines. High-quality MRI is critical for accurate cancer detection and effective management. Despite the existence of technical standards such as PI-RADS v2.1, variability in image quality remains a challenge. Factors contributing to this variability include poor adherence to PI-RADS requirements, patient-related issues such as motion and inadequate preparation, and scanner-related limitations. To address these challenges, the updated PI-QUAL v2 tool, building on PI-QUAL v1, offers a reliable approach to assessing prostate MRI image quality based on critical technical prerequisites for optimal diagnostic imaging. It also enhances standardization by supporting MRI evaluations without the need for intravenous contrast, simplifying the scoring process, and ensuring reproducibility of image quality assessments across diverse clinical settings. The final PI-QUAL v2 score (1–3) offers a concise, actionable summary of overall prostate MRI quality that directly informs clinical and technical decisions. For PI-QUAL Score 1, it indicates inadequate scans that should not be assigned PI-RADS or Likert scores, particularly PI-RADS/LIKERT score of 3. Imaging teams are required to investigate and address the causes of poor quality by minimizing motion in future scans (if patient-related) or rescanning with a higher-performing MR system (if machine-related). For PI-QUAL Score 2, it reflects acceptable scans that generally do not require rescanning unless the quality is doubtful or insufficient for a confident diagnosis. For PI-QUAL Score 3, it represents optimal scans that are essential for active surveillance to detect or rule out clinically significant disease, and for post-treatment monitoring to evaluate radiological changes or detect residual/recurrent disease.
Objectives :
The aim of this audit is to assess the diagnostic image quality of mpMRI prostate in Princess Margaret Hospital (PMH) based on PI-QUAL Version 2 Scoring System.
Methodology :
MRI Prostate technical requirements should be standardised according to PI-RADS v. 2.1 recommendations outlined in Prostate Imaging Reporting and Data System Version 2.1: 2019 Update of Prostate Imaging Reporting and Data System Version 2. The essential technical prerequisites for prostate MRI sequences are as follows: T2-WI requires a slice thickness of 3 mm. DWI necessitates a slice thickness of ≤ 4 mm, inclusion of a high b-value sequence (≥ 1400 s/mm², either calculated or acquired), and an ADC map derived from at least two b-values up to 1000 s/mm². DCE requires a 3 mm slice thickness, a temporal resolution of ≤ 15 seconds, and fat suppression or post-processing techniques such as subtraction or heat maps with 3D sequences. After verifying the technical parameters, visual scoring for diagnostic quality images is conducted following the recommendations outlined in PI-QUAL version 2: An Update of a Standardised Scoring System for the Assessment of Image Quality of Prostate MRI. The overall quality of all prostate-relevant sequences will be assessed collectively, with individual scores provided for each sequence (respectively T2-WI, DWI +/- DCE). This is to offer detailed insights into specific areas that may require improvement. Based on the individual scores of each domain, a final PI-QUAL score is assigned to each MRI study to assess its overall diagnostic quality. PI-QUAL Score 1 indicates inadequate quality scans where PI-RADS or Likert scores cannot be assigned. PI-QUAL Score 2 reflects acceptable quality scans, with potential areas for improvement identified through individual domain scores. PI-QUAL Score 3 represents optimal quality scans that fully meet technical and visual standards, making them ideal for active surveillance of indolent disease and post-treatment monitoring. 100% of MRI prostate scans should be of acceptable or optimal quality studies (PI-QUAL Score 2 or above). Data were collected retrospectively over 4 month period from 1/7/2024 to 31/10/2024 through the Radiology information System (RIS) in PMH. The keyword search were “PIRADS” OR “PI-RADS” and “reported by all radiologists”. MRI prostate images were co-reviewed on the PACS system by a resident radiologist and a fellow radiologist. T2-WI, DWI and DCE MRI sequences image quality were each accessed and analysed using PI-QUAL v2 scoring sheet. PI-QUAL score was then obtained. Data were then input into Excel for further analysis.
Result & Outcome :
Total of 49 MRI prostate cases were included in phase I during the period from 1 July to 31 October 2024. Of those cases, 48 cases were mpMRI prostate. 1 case was biparametric MRI without contrast medium. Final PI QUAL score was then calculated. Majority of prostate MRI studies (91%) were classified as PI QUAL score 2, indicating acceptable diagnostic image quality. However, a small percentage of cases (9%) were categorized as inadequate quality scans (PI QUAL score 1), which do not meet the critical technical or image quality requirements. To enhance diagnostic image quality, specific areas within each sequence can be optimized. For the T2 WI sequence, motion artefacts were the main limitation, impairing delineation of prostate structures. These arose from bowel, bladder, and pelvic floor motion. Various mitigation strategies exist, but there is currently no convincing evidence that these measures significantly improve diagnostic image quality, particularly for tumour detection or staging. Some centres would recommend use of the anti-peristaltic agent such as Buscopan, unless contraindicated. For DWI, the key issue was low signal-to-noise ratio, especially at lower magnetic field strengths (1.5T) and for high B value DWI images. The current B50/B800-derived B1600 protocol yielded suboptimal images, and a trial of calculated B1600 images with denoising in 1.5T MRI yielded improved image quality. Future work will focus on further DWI optimisation, noise thresholds for calculated sequences. Comparison of 1.5T and 3T scanners showed that although 3T may offer higher SNR, it also introduces more susceptibility artefacts from rectal gas and metal, and current evidence does not clearly favour 3T over 1.5T. Adequate image quality is achievable at 1.5T with optimised protocols, so prostate MRI will continue on both 1.5T and 3T systems. For DCE imaging, most cases were downgraded because fat saturation was absent, contrary to PI RADS v2.1 technical requirements. Fat saturation will therefore be reintroduced to improve protocol compliance and overall image quality. Lastly, at PMH, an additional T2 WI SPACE sequence was routinely acquired for MRI guided prostate biopsy. To optimise scan time and streamline workflow, this sequence will no longer be performed routinely, given its limited clinical utility, resulting in a time saving of 5 minutes per study. If required for specific clinical indications, the sequence will be reinstated for selected cases. By addressing these limitations and implementing optimization strategies, we aim to enhance the diagnostic image quality of prostate MRI.
