Related Experiment Video
Updated: Sep 3, 2025
10:17
Guidelines and Experience Using Imaging Biomarker Explorer IBEX for Radiomics
Published on: January 8, 2018
13.3K
Can Photon-Counting CT Improve Estimation Accuracy of Morphological Radiomics Features? A Simulation Study for
Shobhit Sharma1, Debashish Pal2, Ehsan Abadi3
1Center for Virtual Imaging Trials and Carl E. Ravin Advanced Imaging Laboratories, Durham, NC; Department of Physics, Duke University, Durham, NC.
Academic Radiology
|July 25, 2022
Summary
Deep silicon photon-counting CT (Si-PCCT) offers better spatial resolution for lung lesion imaging. This simulation shows Si-PCCT improves morphological radiomics feature accuracy compared to energy-integrating CT (ECT).
Area of Science:
- Medical Imaging Physics
- Radiomics and Quantitative Imaging
- Detector Technology Development
Background:
- Deep silicon-based photon-counting CT (Si-PCCT) is an emerging technology offering superior spatial resolution due to smaller pixel sizes.
- Accurate estimation of morphological radiomics features in lung lesions is crucial for quantitative imaging and diagnosis.
- Conventional energy-integrating CT (ECT) systems have limitations in spatial resolution that may affect radiomics feature accuracy.
Purpose of the Study:
- To evaluate the impact of Si-PCCT's improved spatial resolution on the estimation accuracy of morphological radiomics features in lung lesions.
- To compare the performance of Si-PCCT against conventional ECT for radiomics feature extraction in simulated lung lesions.
- To assess the influence of imaging parameters (dose, reconstruction kernel, pixel size, distance from isocenter) on feature estimation accuracy for both systems.
Main Methods:
- A dynamic nutrient-access-based stochastic model generated three distinct lung lesion morphologies.
- Lesions were inserted into an anthropomorphic phantom and virtually imaged using DukeSim, simulating both Si-PCCT and ECT systems under various conditions.
- Morphological radiomics features were extracted from segmented lesions using AutoContour and pyradiomics; estimation errors were calculated against ground truth.
Main Results:
- Si-PCCT demonstrated lower mean estimation errors for morphological radiomics features compared to ECT (independent features: 35.9% vs. 54.0%; all features: 54.5% vs. 68.1%).
- Statistically significant error reductions were observed for 8 out of 14 radiomics features with Si-PCCT.
- Estimation accuracy was minimally affected by dose and distance from isocenter, but more strongly influenced by reconstruction kernel and pixel size for both systems.
Conclusions:
- Si-PCCT exhibits superior estimation accuracy for morphological radiomics features in lung lesions compared to conventional ECT.
- The improved spatial resolution of Si-PCCT holds significant potential for enhancing quantitative imaging applications in lung lesion analysis.
- This simulation study underscores the promise of Si-PCCT technology for more precise radiomics-based diagnostics.
