CT-free kidney single-photon emission computed tomography for glomerular filtration rate

Area of Science:

• Nuclear Medicine
• Artificial Intelligence
• Medical Imaging

Background:

• Quantitative SPECT/CT kidney imaging is crucial for estimating glomerular filtration rate (GFR).
• Conventional methods rely on CT for attenuation correction, increasing radiation exposure and procedure time.
• Developing CT-free quantitative SPECT methods is essential for patient safety and diagnostic efficiency.

Purpose of the Study:

• To develop and validate an AI-based approach for CT-free quantitative SPECT kidney imaging.
• To estimate GFR using SPECT alone, eliminating the need for CT-based attenuation correction.
• To assess the accuracy, safety, and efficiency of the proposed AI method compared to conventional SPECT/CT.

Main Methods:

• A deep learning model (residual U-Net with edge attention) was trained on 1000 SPECT/CT scans.
• The model generated synthetic attenuation maps (µ-maps) from SPECT data for kidney segmentation.
• GFR was calculated using AI-derived kidney segmentations and compared to conventional SPECT/CT measurements.

Main Results:

• The AI model achieved high accuracy in kidney segmentation, with a Dice score of 0.818 ± 0.056.
• AI-based CT-free SPECT yielded GFR values nearly identical to conventional SPECT/CT (109.3 ± 17.3 vs. 109.2 ± 18.4 mL/min).
• The CT-free method reduced radiation exposure by up to 78.8% and decreased segmentation time from 40 minutes to under 1 minute.

Conclusions:

• AI can effectively replace CT in quantitative SPECT kidney imaging for GFR estimation.
• This CT-free approach maintains quantitative accuracy while significantly improving patient safety and procedural efficiency.
• The developed AI method represents a promising advancement for safer and faster kidney function assessment.