Inter-reader variability and reproducibility of the PI-QUAL score in a multicentre setting

Area of Science:

• Radiology
• Medical Imaging
• Prostate Cancer Diagnostics

Background:

• Prostate MRI is crucial for diagnosing prostate cancer.
• Standardized quality assessment tools like the Prostate Imaging Quality (PI-QUAL) score are needed.
• Assessing the reproducibility of PI-QUAL is essential for its clinical adoption.

Purpose of the Study:

• To evaluate the inter-reader reproducibility of the Prostate Imaging Quality (PI-QUAL) score.
• To assess PI-QUAL reproducibility across different levels of clinical experience.
• To determine PI-QUAL reproducibility between different healthcare institutions.

Main Methods:

• 60 prostate MRI scans from academic and non-academic hospitals were reviewed.
• Three readers (two radiologists with varied experience, one MRI technician) independently used the PI-QUAL scoring sheet.
• Readers were blinded to the acquisition site.

Main Results:

• Moderate agreement was found for individual PI-QUAL scores.
• Good to very good agreement (Gwet AC1: 0.67-0.836) was achieved when scans were grouped (PI-QUAL 1-3 vs. 4-5).
• Higher PI-QUAL scores were assigned to scans from academic hospitals compared to community hospitals (p < 0.001).

Conclusions:

• The PI-QUAL score demonstrates good to very good reproducibility for classifying MRI scans into broad quality groups.
• Reproducibility for individual PI-QUAL scores was moderate, suggesting a need for clearer criteria.
• Scans from academic teaching hospitals were rated higher in quality than those from community hospitals.