Automated Kidney Stone Composition Analysis with Photon-Counting Detector CT, a Performance Study-A Phantom Study.

Area of Science:

• Medical Imaging
• Urology
• Radiology

Background:

• Urolithiasis stone composition is critical for treatment selection, particularly for uric acid (UA) stones amenable to chemolitholysis.
• Photon-counting detector CT (PCDCT) offers advanced spectral analysis for differentiating UA from non-UA stones.
• Assessing the accuracy of PCDCT for stone composition analysis is essential for clinical application.

Purpose of the Study:

• To evaluate the accuracy of PCDCT spectral data analysis in differentiating uric acid (UA) and non-UA urinary stones.
• To determine the detection rates of urinary stones using PCDCT.
• To compare the predictive accuracy of PCDCT for stone composition against the gold standard (infrared spectroscopy).

Main Methods:

• 148 urinary stones with known compositions (via infrared spectroscopy) were scanned using PCDCT within an abdomen phantom.
• PCDCT stone detection rates were assessed based on stone size and volume.
• The accuracy of PCDCT in predicting UA versus non-UA stone composition was evaluated.

Main Results:

• PCDCT achieved an automated stone detection rate of 86.5%, with higher rates for larger stones (up to 95.4% for >5 mm).
• The analysis demonstrated a negative predictive value (NPV) of 94.5% and a positive predictive value (PPV) of 66.7% for UA stone prediction.
• Optimal diagnostic performance for UA differentiation was observed for stones >30 mm³, yielding 91.7% sensitivity and 92.4% specificity.

Conclusions:

• PCDCT provides robust automated detection of urinary stones, with performance varying by stone size.
• The algorithm shows high accuracy in predicting non-UA stones (NPV 94.5%), aiding in identifying patients unsuitable for chemolitholysis.
• PCDCT-based automated analysis is a promising tool for stone composition assessment and guiding urolithiasis treatment strategies.