Synthetic Low-Energy Monochromatic Image Generation in Single-Energy Computed Tomography System Using a Transformer-Based Deep Learning Model.

While dual-energy computed tomography (DECT) technology introduces energy-specific information in clinical practice, single-energy CT (SECT) is predominantly used, limiting the number of people who can benefit from DECT. This study proposed a novel method to generate synthetic low-energy virtual monochromatic images at 50 keV (sVMI50keV) from SECT images using a transformer-based deep learning model, SwinUNETR. Data were obtained from 85 patients who underwent head and neck radiotherapy. Among these, the model was built using data from 70 patients for whom only DECT images were available. The remaining 15 patients, for whom both DECT and SECT images were available, were used to predict from the actual SECT images. We used the SwinUNETR model to generate sVMI50keV. The image quality was evaluated, and the results were compared with those of the convolutional neural network-based model, Unet. The mean absolute errors from the true VMI50keV were 36.5 ± 4.9 and 33.0 ± 4.4 Hounsfield units for Unet and SwinUNETR, respectively. SwinUNETR yielded smaller errors in tissue attenuation values compared with those of Unet. The contrast changes in sVMI50keV generated by SwinUNETR from SECT were closer to those of DECT-derived VMI50keV than the contrast changes in Unet-generated sVMI50keV. This study demonstrated the potential of transformer-based models for generating synthetic low-energy VMIs from SECT images, thereby improving the image quality of head and neck cancer imaging. It provides a practical and feasible solution to obtain low-energy VMIs from SECT data that can benefit a large number of facilities and patients without access to DECT technology.

The effects of distance between the imaging isocenter and brain center on the image quality of cone-beam computed tomography for brain stereotactic irradiation.

In linear accelerator-based stereotactic irradiation (STI) for brain metastasis, cone-beam computed tomography (CBCT) image quality is essential for ensuring precise patient setup and tumor localization. However, CBCT images may be degraded by the deviation of the CBCT isocenter from the brain center. This study aims to investigate the effects of the distance from the brain center to the CBCT isocenter (DBI) on the image quality in STI. An anthropomorphic phantom was scanned with varying DBI in right, anterior, superior, and inferior directions. Thirty patients undergoing STI were prospectively recruited. Objective metrics, utilizing regions of interest included contrast-to-noise ratio (CNR) at the centrum semiovale, lateral ventricle, and basal ganglia levels, gray and white matter noise at the basal ganglia level, artifact index (AI), and nonuniformity (NU). Two radiation oncologists assessed subjective metrics. In this phantom study, objective measures indicated a degradation in image quality for non-zero DBI. In this patient study, there were significant correlations between the CNR at the centrum semiovale and lateral ventricle levels (rs = - 0.79 and - 0.77, respectively), gray matter noise (rs = 0.52), AI (rs = 0.72), and NU (rs = 0.91) and DBI. However, no significant correlations were observed between the CNR at the basal ganglia level, white matter noise, and subjective metrics and DBI (rs < ± 0.3). Our results demonstrate the effects of DBI on contrast, noise, artifacts in the posterior fossa, and uniformity of CBCT images in STI. Aligning the CBCT isocenter with the brain center can aid in improving image quality.

Dual-energy computed tomography-based iodine concentration as a predictor of histopathological response to preoperative chemoradiotherapy for pancreatic cancer.

To explore predictors of the histopathological response to preoperative chemoradiotherapy (CRT) in patients with pancreatic cancer (PC) using dual-energy computed tomography-reconstructed images. This retrospective study divided 40 patients who had undergone preoperative CRT (50-60 Gy in 25 fractions) followed by surgical resection into two groups: the response group (Grades II, III and IV, evaluated from surgical specimens) and the nonresponse group (Grades Ia and Ib). The computed tomography number [in Hounsfield units (HUs)] and iodine concentration (IC) were measured at the locations of the aorta, PC and pancreatic parenchyma (PP) in the contrast-enhanced 4D dual-energy computed tomography images. Logistic regression analysis was performed to identify predictors of histopathological response. Univariate analysis did not reveal a significant relation between any parameter and patient characteristics or dosimetric parameters of the treatment plan. The HU and IC values in PP and the differences in HU and IC between the PP and PC (ΔHU and ΔIC, respectively) were significant predictors for distinguishing the response (n = 24) and nonresponse (n = 16) groups (P < 0.05). The IC in PP and ΔIC had a higher area under curve values [0.797 (95% confidence interval, 0.659-0.935) and 0.789 (0.650-0.928), respectively] than HU in PP and ΔHU [0.734 (0.580-0.889) and 0.721 (0.562-0.881), respectively]. The IC value could potentially be used for predicting the histopathological response in patients who have undergone preoperative CRT.

Clinical target volume segmentation based on gross tumor volume using deep learning for head and neck cancer treatment.

Accurate clinical target volume (CTV) delineation is important for head and neck intensity-modulated radiation therapy. However, delineation is time-consuming and susceptible to interobserver variability (IOV). Based on a manual contouring process commonly used in clinical practice, we developed a deep learning (DL)-based method to delineate a low-risk CTV with computed tomography (CT) and gross tumor volume (GTV) input and compared it with a CT-only input. A total of 310 patients with oropharynx cancer were randomly divided into the training set (250) and test set (60). The low-risk CTV and primary GTV contours were used to generate label data for the input and ground truth. A 3D U-Net with a two-channel input of CT and GTV (U-NetGTV) was proposed and its performance was compared with a U-Net with only CT input (U-NetCT). The Dice similarity coefficient (DSC) and average Hausdorff distance (AHD) were evaluated. The time required to predict the CTV was 0.86 s per patient. U-NetGTV showed a significantly higher mean DSC value than U-NetCT (0.80 ± 0.03 and 0.76 ± 0.05) and a significantly lower mean AHD value (3.0 ± 0.5 mm vs 3.5 ± 0.7 mm). Compared to the existing DL method with only CT input, the proposed GTV-based segmentation using DL showed a more precise low-risk CTV segmentation for head and neck cancer. Our findings suggest that the proposed method could reduce the contouring time of a low-risk CTV, allowing the standardization of target delineations for head and neck cancer.