Conversion of single-energy computed tomography to parametric maps of dual-energy computed tomography using convolutional neural network.

OBJECTIVES: We propose a deep learning (DL) multi-task learning framework using convolutional neural network (CNN) for a direct conversion of single-energy CT (SECT) to three different parametric maps of dual-energy CT (DECT): Virtual-monochromatic image (VMI), effective atomic number (EAN), and relative electron density (RED). METHODS: We propose VMI-Net for conversion of SECT to 70, 120, and 200 keV VMIs. In addition, EAN-Net and RED-Net were also developed to convert SECT to EAN and RED. We trained and validated our model using 67 patients collected between 2019 and 2020. SECT images with 120 kVp acquired by the DECT (IQon spectral CT, Philips) were used as input, while the VMIs, EAN, and RED acquired by the same device were used as target. The performance of the DL framework was evaluated by absolute difference (AD) and relative difference (RD). RESULTS: The VMI-Net converted 120 kVp SECT to the VMIs with AD of 9.02 Hounsfield Unit, and RD of 0.41% compared to the ground truth VMIs. The ADs of the converted EAN and RED were 0.29 and 0.96, respectively, while the RDs were 1.99% and 0.50% for the converted EAN and RED, respectively. CONCLUSIONS: SECT images were directly converted to the three parametric maps of DECT (ie, VMIs, EAN, and RED). By using this model, one can generate the parametric information from SECT images without DECT device. Our model can help investigate the parametric information from SECT retrospectively. ADVANCES IN KNOWLEDGE: Deep learning framework enables converting SECT to various high-quality parametric maps of DECT.

Novel tongue-positioning device to reduce tongue motions during radiation therapy for head and neck cancer: Geometric and dosimetric evaluation.

This study aimed to assess the performance of a tongue-positioning device in interfractional tongue position reproducibility by cone-beam computed tomography (CBCT). Fifty-two patients treated with radiation therapy (RT) while using a tongue positioning device were included in the study. All patients were treated with 28 or 30 fractions using the volumetric modulated arc therapy technique. CBCT images were acquired at the 1st, 7th, 11th, 15th, 19th, 23th, and 27th fractions. Tongues on planning computed tomography (pCT) and CBCT images were contoured in the treatment planning system. Geometric differences in the tongue between pCT and CBCT were assessed by the Dice similarity coefficient (DSC) and averaged Hausdorff distance (AHD). Two-dimensional in vivo measurements using radiochromic films were performed in 13 patients once a week during sessions. The planned dose distributions were compared with the measured dose distributions using gamma analysis with criteria of 3%/3 mm. In all patients, the mean DSC at the 1st fraction (pCT versus 1st CBCT) was 0.80 while the mean DSC at the 27th fraction (pCT versus 27th CBCT) was 0.77 with statistical significance (p-value = 0.015). There was no statistically significant difference in DSC between the 1st fraction and any other fraction, except for the 27th fraction. There was statistically significant difference in AHD between the 1st fraction and the 19th, 23th, and 27th fractions (p-value < 0.05). In vivo measurements showed an average gamma passing rate of 90.54%. There was no significant difference between measurements at the 1st week and those at other weeks. The tongue geometry during RT was compared between pCT and CBCT. In conclusion, the novel tongue-positioning device was found to minimize interfractional variations in position and shape of the tongue.

3D star shot analysis using MAGAT gel dosimeter for integrated imaging and radiation isocenter verification of MR-Linac system.

PURPOSE: This study aims to investigate a star shot analysis using a three-dimensional (3D) gel dosimeter for the imaging and radiation isocenter verification of a magnetic resonance linear accelerator (MR-Linac). METHODS: A mixture of methacrylic acid, gelatin, and tetrakis (hydroxymethyl) phosphonium chloride, called MAGAT gel, was fabricated. One MAGAT gel for each Linac and MR-Linac was irradiated under six gantry angles. A 6 MV photon beam of Linac and a 6 MV flattening filter free beam of MR-Linac were delivered to two MAGAT gels and EBT3 films. MR images were acquired by MR-Linac with a clinical sequence (i.e., TrueFISP). The 3D star shot analysis for seven consecutive slices of the MR images with TrueFISP was performed. The 2D star shot analysis for the central plane of the gel was compared to the results from the EBT3 films. The radius of isocircle (ICr ) and the distance between the center of the circle and the center marked on the image (ICd ) were evaluated. RESULTS: For MR-Linac with MAGAT gel measurements, ICd at the central plane was 0.46 mm for TrueFISP. Compared to EBT3 film measurements, the differences in ICd and ICr for both Linac and MR-Linac were within 0.11 and 0.13 mm, respectively. For the 3D analysis, seven consecutive slices of TrueFISP images were analyzed and the maximum radii of isocircles (ICr_max ) were 0.18 mm for Linac and 0.73 mm for MR-Linac. The tilting angles of radiation axis were 0.31° for Linac and 0.10° for MR-Linac. CONCLUSION: The accuracy of 3D star shot analysis using MAGAT gel was comparable to that of EBT3 film, having a capability for integrated analysis for imaging isocenter and radiation isocenter. 3D star shot analysis using MAGAT gel can provide 3D information of radiation isocenter, suggesting a quantitative extent of gantry-tilting.

Adaptive Image Rescaling for Weakly Contrast-Enhanced Lesions in Dedicated Breast CT: A Phantom Study.

Purpose: Dedicated breast CT is an emerging volumetric X-ray imaging modality for diagnosis that does not require any painful breast compression. To improve the detection rate of weakly enhanced lesions, an adaptive image rescaling (AIR) technique was proposed. Materials and Methods: Two disks containing five identical holes and five holes of different diameters were scanned using 60/100 kVp to obtain single-energy CT (SECT), dual-energy CT (DECT), and AIR images. A piece of pork was also scanned as a subclinical trial. The image quality was evaluated using image contrast and contrast-to-noise ratio (CNR). The difference of imaging performances was confirmed using student's t test. Results: Total mean image contrast of AIR (0.70) reached 74.5% of that of DECT (0.94) and was higher than that of SECT (0.22) by 318.2%. Total mean CNR of AIR (5.08) was 35.5% of that of SECT (14.30) and was higher than that of DECT (2.28) by 222.8%. A similar trend was observed in the subclinical study. Conclusion: The results demonstrated superior image contrast of AIR over SECT, and its higher overall image quality compared to DECT with half the exposure. Therefore, AIR seems to have the potential to improve the detectability of lesions with dedicated breast CT.