Related Experiment Video
Updated: Apr 11, 2026
Deep Learning-Based Segmentation of Cryo-Electron Tomograms
Published on: November 11, 2022
Voxel-wise deep learning segmentation of hydroxyapatite and iodine in spectral photon-counting CT: A quantitative
Nadine Francis1, Mohamed L Seghier2, Nabil Maalej1
1Department of Physics, Khalifa University of Science and Technology, Abu Dhabi, United Arab Emirates.
Insights
A new deep learning model, SPFF-UNet, accurately distinguishes hydroxyapatite (HA) from iodine in spectral photon-counting CT scans. This advance improves diagnosis of calcific musculoskeletal disease and vascular calcification.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Biomedical Engineering
Background:
- Accurate non-invasive identification of hydroxyapatite (HA) deposits is crucial for diagnosing calcific musculoskeletal disease and vascular calcification.
- Conventional and dual-energy CT methods face challenges distinguishing HA from iodinated contrast due to artifacts and overlapping attenuation.
- Spectral photon-counting CT (SPCCT) offers enhanced spectral and spatial resolution, but deep learning approaches often focus on segmentation or density regression rather than direct material labeling.
Purpose of the Study:
- To develop a spectral-preserving 3D deep learning model for direct voxel-wise classification of HA and iodine concentrations from SPCCT data.
- To evaluate the model's performance against established 3D architectures without requiring material-decomposition preprocessing.
Main Methods:
- Developed SPFF-UNet, a 3D segmentation model integrating spectral squeeze-excitation, EnergyFiLM, and FourierGate to preserve multi-energy information.
- Trained the model for thirteen-class voxel-wise segmentation using five-bin SPCCT data from a phantom with various HA, iodine, and soft-tissue concentrations.
- Compared SPFF-UNet against five established 3D architectures under matched training conditions.
Main Results:
- SPFF-UNet achieved superior macro-averaged performance (Dice 0.72 ± 0.01) on a held-out phantom scan compared to the best comparator, ResUNet++ (Dice 0.66 ± 0.02).
- Performance gains were most significant for mid/low-contrast HA and low-concentration iodine, demonstrating improved sensitivity and precision.
- The model exhibited reduced slice-wise variability and fewer misclassifications between HA and iodine.
Conclusions:
- Preserving spectral information and employing targeted spectral modulation enhances concentration-aware voxel classification from SPCCT data.
- SPFF-UNet represents a promising phantom-based proof-of-concept for accurate HA and iodine differentiation in medical imaging.
- Further in vivo validation is warranted to translate these findings into clinical practice.
Abstract:
Accurate non-invasive identification of hydroxyapatite (HA) deposits is important for diagnosing calcific musculoskeletal disease and quantifying vascular calcification, but conventional and dual-energy CT often struggle to distinguish HA from iodinated contrast because of overlapping attenuation, noise, and beam-hardening artifacts. Spectral photon-counting CT (SPCCT) offers improved energy resolution and spatial fidelity, yet most deep-learning approaches in spectral CT focus on continuous density regression or anatomical segmentation rather than direct voxel-wise material labeling. We developed SPFF-UNet, a spectral-preserving 3D segmentation model for direct classification of HA and iodine concentrations from five-bin SPCCT volumes without material-decomposition preprocessing. A cylindrical phantom containing twelve materials was scanned at 0.1 mm isotropic resolution, including five HA concentrations, three iodine concentrations, three soft-tissue equivalents, and water. SPFF-UNet integrates spectral squeeze-excitation, EnergyFiLM, and FourierGate to preserve and exploit multi-energy information throughout the network. The model was trained for thirteen-class voxel-wise segmentation and compared with five established 3D architectures under matched training conditions. SPFF-UNet achieved the best macro-averaged performance on a held-out phantom scan (Dice 0.72 ± 0.01, IoU 0.59 ± 0.01, sensitivity 0.73 ± 0.01, precision 0.71 ± 0.01), outperforming the strongest comparator, ResUNet++ (Dice 0.66 ± 0.02, IoU 0.46 ± 0.02, sensitivity 0.67 ± 0.02, precision 0.61 ± 0.03). Performance gains were concentrated in mid/low-contrast HA and low-concentration iodine, with reduced slice-wise variability and fewer HA-iodine misclassifications. These results suggest that preserving spectral information and applying targeted spectral modulation can improve concentration-aware voxel classification from SPCCT. This phantom-based proof-of-concept provides a basis for future in vivo validation.
Related Concept Videos
01:25Imaging Studies for Cardiovascular System VI: Calcium -Scoring CT
01:10Computed Tomography
The technique was invented in the 1970s and is based on the principle that as X-rays pass through the body, they are absorbed or reflected at different levels. In the technique, a patient lies on a motorized platform while a computerized axial tomography (CAT) scanner rotates...