Related Experiment Video
Updated: May 16, 2026
06:33
Construction of a Preclinical Multimodality Phantom Using Tissue-mimicking Materials for Quality Assurance in Tumor Size Measurement
Published on: July 29, 2013
Accuracy and reliability assessment of CT and MR perfusion analysis software using a digital phantom
Kohsuke Kudo1, Soren Christensen, Makoto Sasaki
1Division of Ultra-High Field MRI and Department of Radiology, Iwate Medical University, 19-1 Uchimaru, Morioka 020-8505, Japan. kokudo@iwate-med.ac.jp).
Radiology
|December 11, 2012
Summary
A new digital phantom accurately assesses CT and MR perfusion software. Commercial software showed errors, while academic MR programs demonstrated reliable results for cerebral blood flow and volume.
Area of Science:
- Medical Imaging
- Radiology
- Quantitative Perfusion Imaging
Background:
- Accurate quantitative analysis of computed tomography (CT) and magnetic resonance (MR) perfusion imaging is crucial for diagnosing and managing cerebrovascular diseases.
- Existing postprocessing software for perfusion imaging varies in accuracy and reliability, necessitating standardized evaluation methods.
Purpose of the Study:
- To develop a digital phantom dataset for CT and MR perfusion imaging based on established tracer kinetic theory.
- To evaluate the accuracy and reliability of various postprocessing algorithms used in CT and MR perfusion analysis.
Main Methods:
- A digital phantom was created with known ground truth values for cerebral blood flow (CBF), cerebral blood volume (CBV), mean transit time (MTT), and tracer arrival delay.
- Concentration-time curves were generated and embedded in human brain images.
- Thirteen CT and thirteen MR algorithms, including commercial and academic software, were used to analyze the phantom data.
- Accuracy was assessed using Pearson correlation coefficients (r), and delay-dependent, CBV-dependent, and MTT-dependent errors were evaluated.
Main Results:
- CT perfusion analysis showed good reproduction of CBV (r > 0.9 in 12 algorithms) but less consistent results for CBF and MTT.
- MR perfusion analysis revealed good correlations (r > 0.9) in half of the commercial programs, while all academic algorithms demonstrated high accuracy for all parameters.
- Most algorithms exhibited delay-induced errors, particularly commercial software.
- CBV and MTT dependency errors were observed in CBF/MTT and CBV calculations, respectively.
Conclusions:
- The developed digital phantom effectively evaluated the performance of CT and MR perfusion analysis software.
- Commercial perfusion analysis software demonstrated significant delay-induced errors and/or insufficient correlation with true values.
- Academic MR perfusion software exhibited superior accuracy and reliability compared to commercial alternatives.
