Improving the Specificity of Uniparametric Low-Field Magnetic Resonance Imaging in Prostate Cancer Assessment in Resource-Limited Settings

Abstract

Background: Due to infrastructural and financial constraints, public institutions in low- and middle-income countries (LMICs) use low-field magnetic resonance imaging (LFMRI) systems, which limits the possibility for global guideline conformity. Nonetheless, existing systems are being used to broaden the scope of prostate cancer (PCa) imaging assessment and disease characterization. Because LMICs have a dearth of subspecialty radiologists, general radiologists would need more exposure to meet the growing needs of the subspecialties, conform to universal best practice, and improve clinical outcomes. Purpose: The purpose of this study was to evaluate MRI reports for PCa evaluation generated by general radiologists at a tertiary hospital on a LFMRI (<0.5T) system. Methods: An assessment tool (”LFMRI”) was developed, based on PIRADS criteria, to evaluate prostate MRI reports of images generated from a LFMRI. Two independent observers (urologist and radiologist) rated the reports. Another general radiologist used the tool to generate new reports from the same images. These reports were reassessed by the same raters. The average assigned scores were categorized as poor (0–4), fair (5–8), or good (9–12). Results: Six of seven (85.7%) reviewed reports were poor with only one (14.7%) providing useful information for clinical decision. Inter-rater reliability was moderate (48.3%). The revised reports revealed additional clinically useful information in all cases and a mean total score improvement of 7.3 (0.69). Inter-rater reliability improved to 78.7%95% confidence interval, CI (0.5, 1.0), P < 0.0001. Conclusion: The images generated from LFMRI in resource-constrained settings may provide requisite information for PCa evaluation. A simple tool (LFMRI) can guide general radiologists in making useful and enhanced clinical decisions based on LFMRI reports.