Evaluation of a cycle-generative adversarial network-based cone-beam CT to synthetic CT conversion algorithm for adaptive radiation therapy

PurposeImage-guided radiation therapy could benefit from implementing adaptive radiation therapy (ART) techniques. A cycle-generative adversarial network (cycle-GAN)-based cone-beam computed tomography (CBCT)-to-synthetic CT (sCT) conversion algorithm was evaluated regarding image quality, image segmentation and dosimetric accuracy for head and neck (H&N), thoracic and pelvic body regions.MethodsUsing a cycle-GAN, three body site-specific models were priorly trained with independent paired CT and CBCT datasets of a kV imaging system (XVI, Elekta). sCT were generated based on first-fraction CBCT for 15 patients of each body region. Mean errors (ME) and mean absolute errors (MAE) were analyzed for the sCT. On the sCT, manually delineated structures were compared to deformed structures from the planning CT (pCT) and evaluated with standard segmentation metrics. Treatment plans were recalculated on sCT. A comparison of clinically relevant dose-volume parameters (D₉₈, D₅₀ and D₂ of the target volume) and 3D-gamma (3%/3mm) analysis were performed.ResultsThe mean ME and MAE were 1.4, 29.6, 5.4 Hounsfield units (HU) and 77.2, 94.2, 41.8 HU for H&N, thoracic and pelvic region, respectively. Dice similarity coefficients varied between 66.7 ± 8.3% (seminal vesicles) and 94.9 ± 2.0% (lungs). Maximum mean surface distances were 6.3 mm (heart), followed by 3.5 mm (brainstem). The mean dosimetric differences of the target volumes did not exceed 1.7%. Mean 3D gamma pass rates greater than 97.8% were achieved in all cases.ConclusionsThe presented method generates sCT images with a quality close to pCT and yielded clinically acceptable dosimetric deviations. Thus, an important prerequisite towards clinical implementation of CBCT-based ART is fulfilled.     

Deep learning methods to generate synthetic CT from MRI in radiotherapy: A literature review

PurposeIn radiotherapy, MRI is used for target volume and organs-at-risk delineation for its superior soft-tissue contrast as compared to CT imaging. However, MRI does not provide the electron density of tissue necessary for dose calculation. Several methods of synthetic-CT (sCT) generation from MRI data have been developed for radiotherapy dose calculation. This work reviewed deep learning (DL) sCT generation methods and their associated image and dose evaluation, in the context of MRI-based dose calculation.MethodsWe searched the PubMed and ScienceDirect electronic databases from January 2010 to March 2021. For each paper, several items were screened and compiled in figures and tables.ResultsThis review included 57 studies. The DL methods were either generator-only based (45% of the reviewed studies), or generative adversarial network (GAN) architecture and its variants (55% of the reviewed studies). The brain and pelvis were the most commonly investigated anatomical localizations (39% and 28% of the reviewed studies, respectively), and more rarely, the head-and-neck (H&N) (15%), abdomen (10%), liver (5%) or breast (3%). All the studies performed an image evaluation of sCTs with a diversity of metrics, with only 36 studies performing dosimetric evaluations of sCT.ConclusionsThe median mean absolute errors were around 76 HU for the brain and H&N sCTs and 40 HU for the pelvis sCTs. For the brain, the mean dose difference between the sCT and the reference CT was <2%. For the H&N and pelvis, the mean dose difference was below 1% in most of the studies. Recent GAN architectures have advantages compared to generator-only, but no superiority was found in term of image or dose sCT uncertainties. Key challenges of DL-based sCT generation methods from MRI in radiotherapy is the management of movement for abdominal and thoracic localizations, the standardization of sCT evaluation, and the investigation of multicenter impacts.     

Improvement of image quality for pancreatic cancer using deep learning-generated virtual monochromatic images: Comparison with single-energy computed tomography

PurposeTo construct a deep convolutional neural network that generates virtual monochromatic images (VMIs) from single-energy computed tomography (SECT) images for improved pancreatic cancer imaging quality.Materials and methodsFifty patients with pancreatic cancer underwent a dual-energy CT simulation and VMIs at 77 and 60 keV were reconstructed. A 2D deep densely connected convolutional neural network was modeled to learn the relationship between the VMIs at 77 (input) and 60 keV (ground-truth). Subsequently, VMIs were generated for 20 patients from SECT images using the trained deep learning model.ResultsThe contrast-to-noise ratio was significantly improved (p < 0.001) in the generated VMIs (4.1 ± 1.8) compared to the SECT images (2.8 ± 1.1). The mean overall image quality (4.1 ± 0.6) and tumor enhancement (3.6 ± 0.6) in the generated VMIs assessed on a five-point scale were significantly higher (p < 0.001) than that in the SECT images (3.2 ± 0.4 and 2.8 ± 0.4 for overall image quality and tumor enhancement, respectively).ConclusionsThe quality of the SECT image was significantly improved both objectively and subjectively using the proposed deep learning model for pancreatic tumors in radiotherapy.     

Can different Catphan phantoms be used in a multi-centre audit of radiotherapy CT image quality?

PurposeTo determine the variation between Catphan image quality CT phantoms, specifically for use in a future multi-centre image quality audit.Method14 Catphan phantoms (models 503, 504 and 604) were scanned on a Canon Aquilion Prime CT scanner using a single scan protocol. Measurements were made of noise in the uniformity section, visibility of low contrast targets and contrast, x-ray attenuation and CT number for 5 materials in the sensitometry section. Scans were also acquired using one phantom and varying reconstruction field of view, image slice thickness, effective tube-current-time product and iterative reconstruction settings to determine how the degree of inter-phantom variability compared with the magnitude of changes from scan parameter alteration.ResultsAcross all phantoms the mean CT value in the uniformity section was 7.0 (SD 0.9) range: 4.9–8.1 HU. For the different materials the CT numbers were air: −1004 ± 5, Polymethylpentene: −190 ± 2, Polystyrene: −42 ± 2, Delrin: 321 ± 5 and Teflon: 898 ± 8 HU. Consistency of low contrast targets through visual scoring was good. Measured contrast was lower (p < 0.001) with more variability for 504 versus 604 models. All phantoms produced identical tube current settings with x-ray tube current modulation, indicating no x-ray attenuation differences. The degree of change in image quality metrics between phantoms was small compared with results when scan parameters were varied.ConclusionCatphan phantoms model 604 showed minimal differences and will be used for multi-centre inter-comparison work, with the consistency between phantoms appropriate for measuring possible variations in image quality.     

Enhancement of Hounsfield unit distribution in cone-beam CT images for adaptive radiation therapy: Evaluation of a hybrid correction approach

PurposeThis study aims to evaluate the accuracy of a hybrid approach combining the histogram matching (HM) and the multilevel threshold (MLT) to correct the Hounsfield Unit (HU) distribution in cone-beam CT (CBCT) images.Methods and MaterialsCBCT images acquired for ten prostate cancer patients were processed by matching their histograms to those of deformed planning CT (pCT) images obtained after applying a deformable registration (DR) process. Then, HU values corresponding to five tissue types in the pCT were assigned to the obtained CBCT images (CBCT_HM-MLT). Finally, the CBCT_HM-MLT images were compared to the deformed pCT visually and using different statistical metrics.ResultsThe visual assessment and the profiles comparison showed that the high discrepancies in the CBCT images were significantly reduced when using the proposed approach. Furthermore, the correlation values indicated that the CBCT_HM-MLT were in good agreement with the deformed pCT with correlation values ranging from 0.9893 to 0.9962. In addition, the root mean squared error (RMSE) over the entire volume was reduced from 64.15 ± 9.50 to 51.20 ± 6.76 HU. Similarly, the mean absolute error in specific tissue classes was significantly reduced especially in the soft tissue-air interfaces. These results confirmed that applying MLT after HM worked better than using only HM for which the correlation values were ranging from 0.9878 to 0.9955 and the RMSE was 55.95 ± 10.43 HU.ConclusionEvaluation of the proposed approach showed that the HM + MLT correction can improve the HU distribution in the CBCT images and generate corrected images in good agreement with the pCT.     

Deep learning dose prediction for IMRT of esophageal cancer: The effect of data quality and quantity on model performance

PurposeTo investigate the effect of data quality and quantity on the performance of deep learning (DL) models, for dose prediction of intensity-modulated radiotherapy (IMRT) of esophageal cancer.Material and methodsTwo databases were used: a variable database (VarDB) with 56 clinical cases extracted retrospectively, including user-dependent variability in delineation and planning, different machines and beam configurations; and a homogenized database (HomDB), created to reduce this variability by re-contouring and re-planning all patients with a fixed class-solution protocol.Experiment 1 analysed the user-dependent variability, using 26 patients planned with the same machine and beam setup (E26-VarDB versus E26-HomDB). Experiment 2 increased the training set by groups of 10 patients (E16, E26, E36, E46, and E56) for both databases.Model evaluation metrics were the mean absolute error (MAE) for selected dose-volume metrics and the global MAE for all body voxels.ResultsFor Experiment 1, E26-HomDB reduced the MAE for the considered dose-volume metrics compared to E26-VarDB (e.g. reduction of 0.2 Gy for D95-PTV, 1.2 Gy for Dmean-heart or 3.3% for V5-lungs). For Experiment 2, increasing the database size slightly improved performance for HomDB models (e.g. decrease in global MAE of 0.13 Gy for E56-HomDB versus E26-HomDB), but increased the error for the VarDB models (e.g. increase in global MAE of 0.20 Gy for E56-VarDB versus E26-VarDB).ConclusionA small database may suffice to obtain good DL prediction performance, provided that homogenous training data is used. Data variability reduces the performance of DL models, which is further pronounced when increasing the training set.     

Three critical factors affecting automated image species recognition performance for camera traps

Ecological camera traps are increasingly used by wildlife biologists to unobtrusively monitor an ecosystems animal population. However, manual inspection of the images produced is expensive, laborious, and time‐consuming. The success of deep learning systems using camera trap images has been previously explored in preliminary stages. These studies, however, are lacking in their practicality. They are primarily focused on extremely large datasets, often millions of images, and there is little to no focus on performance when tasked with species identification in new locations not seen during training. Our goal was to test the capabilities of deep learning systems trained on camera trap images using modestly sized training data, compare performance when considering unseen background locations, and quantify the gradient of lower bound performance to provide a guideline of data requirements in correspondence to performance expectations. We use a dataset provided by Parks Canada containing 47,279 images collected from 36 unique geographic locations across multiple environments. Images represent 55 animal species and human activity with high‐class imbalance. We trained, tested, and compared the capabilities of six deep learning computer vision networks using transfer learning and image augmentation: DenseNet201, Inception‐ResNet‐V3, InceptionV3, NASNetMobile, MobileNetV2, and Xception. We compare overall performance on “trained” locations where DenseNet201 performed best with 95.6% top‐1 accuracy showing promise for deep learning methods for smaller scale research efforts. Using trained locations, classifications with <500 images had low and highly variable recall of 0.750 ± 0.329, while classifications with over 1,000 images had a high and stable recall of 0.971 ± 0.0137. Models tasked with classifying species from untrained locations were less accurate, with DenseNet201 performing best with 68.7% top‐1 accuracy. Finally, we provide an open repository where ecologists can insert their image data to train and test custom species detection models for their desired ecological domain.     

Simultaneous achievement of accurate CT number and image quality improvement for myocardial perfusion CT at 320-MDCT volume scanning

PurposeTo investigate differences in image-to-image variations between full- and half-scan reconstruction on myocardial CT perfusion (CTP) study.MethodsUsing a cardiac phantom we performed ECG-gated myocardial CTP on a second-generation 320-multidetector CT volume scanner. The heart rate was set at 60 bpm; once per second for a total of 24 s were performed. CT images were acquired at 80- and 120 kVp and subjected to full- and half-scan reconstruction. On images acquired at the same slice level we then measured image-to-image variations, coefficients of variance (CV), and image noise.ResultsThe image-to-image variations with full- and half-scan reconstruction were 1.3 HU vs. 27.2 HU at 80 kVp (p < 0.001) and 0.70 HU vs. 9.3 HU at 120 kVp (p < 0.001) even though the mean HU value was almost the same for both reconstruction methods. The CV of 80- and 120-kVp images of the left ventricular cavity decreased by 0.16% and 0.17%, respectively, with full-scan reconstruction; with half-scan reconstruction it decreased by 3.34% and 2.30%, respectively. Compared with half-scan reconstruction, the image noise was reduced by 27.2% at 80 kVp and by 28.0% at 120 kVp with full-scan reconstruction.ConclusionMyocardial CTP with full-scan reconstruction substantially decreased image-to-image variations and provided accurate CT attenuation.     

A role for magnetic susceptibility in synthetic computed tomography

PurposeRadiotherapy treatment planning based on magnetic resonance imaging (MRI) benefits from increased soft-tissue contrast and functional imaging. MRI-only planning is attractive but limited by the lack of electron density information required for dose calculation, and the difficulty to differentiate air and bone. MRI can map magnetic susceptibility to separate bone from air. A method is introduced to produce synthetic CT (sCT) through automatic voxel-wise assignment of CT numbers from an MRI dataset processed that includes magnetic susceptibility mapping.MethodsVolumetric multi-echo gradient echo datasets were acquired in the heads of five healthy volunteers and fourteen patients with cancer using a 3 T MRI system. An algorithm for CT synthesis was designed using the volunteer data, based on fuzzy c-means clustering and adaptive thresholding of the MR data (magnitude, fat, water, and magnetic susceptibility). Susceptibility mapping was performed using a modified version of the iterative phase replacement algorithm. On patient data, the algorithm was assessed by direct comparison to X-ray computed tomography (CT) scans.ResultsThe skull, spine, teeth, and major sinuses were clearly distinguished in all sCT, from healthy volunteers and patients. The mean absolute CT number error between X-ray CT and sCT in patients ranged from 78 and 134 HU.ConclusionSusceptibility mapping using MRI can differentiate air and bone for CT synthesis. The proposed method is automated, fast, and based on a commercially available MRI pulse sequence. The method avoids registration errors and does not rely on a priori information, making it suitable for nonstandard anatomy.     

Simple low-cost approaches to semantic segmentation in radiation therapy planning for prostate cancer using deep learning with non-contrast planning CT images

PurposeDeep learning has shown great efficacy for semantic segmentation. However, there are difficulties in the collection, labeling and management of medical imaging data, because of ethical complications and the limited number of imaging studies available at a single facility.This study aimed to find a simple and low-cost method to increase the accuracy of deep learning semantic segmentation for radiation therapy of prostate cancer.MethodsIn total, 556 cases with non-contrast CT images for prostate cancer radiation therapy were examined using a two-dimensional U-Net. Initially, all slices were used for the input data. Then, we removed slices of the cranial portions, which were beyond the margins of the bladder and rectum. Finally, the ground truth labels for the bladder and rectum were added as channels to the input for the prostate training dataset.ResultsThe highest mean dice similarity coefficients (DSCs) for each organ in the test dataset of 56 cases were 0.85 ± 0.05, 0.94 ± 0.04 and 0.85 ± 0.07 for the prostate, bladder and rectum, respectively. Removal of the cranial slices from the original images significantly increased the DSC of the rectum from 0.83 ± 0.09 to 0.85 ± 0.07 (p < 0.05). Adding bladder and rectum information to prostate training without removing the slices significantly increased the DSC of the prostate from 0.79 ± 0.05 to 0.85 ± 0.05 (p < 0.05).ConclusionsThese cost-free approaches may be useful for new applications, which may include updated models and datasets. They may be applicable to other organs at risk (OARs) and clinical targets such as elective nodal irradiation.     

A deep learning classifier for digital breast tomosynthesis

PurposeTo develop a computerized detection system for the automatic classification of the presence/absence of mass lesions in digital breast tomosynthesis (DBT) annotated exams, based on a deep convolutional neural network (DCNN).Materials and MethodsThree DCNN architectures working at image-level (DBT slice) were compared: two state-of-the-art pre-trained DCNN architectures (AlexNet and VGG19) customized through transfer learning, and one developed from scratch (DBT-DCNN). To evaluate these DCNN-based architectures we analysed their classification performance on two different datasets provided by two hospital radiology departments. DBT slice images were processed following normalization, background correction and data augmentation procedures. The accuracy, sensitivity, and area-under-the-curve (AUC) values were evaluated on both datasets, using receiver operating characteristic curves. A Grad-CAM technique was also implemented providing an indication of the lesion position in the DBT slice.Results Accuracy, sensitivity and AUC for the investigated DCNN are in-line with the best performance reported in the field. The DBT-DCNN network developed in this work showed an accuracy and a sensitivity of (90% ± 4%) and (96% ± 3%), respectively, with an AUC as good as 0.89 ± 0.04. A k-fold cross validation test (with k = 4) showed an accuracy of 94.0% ± 0.2%, and a F1-score test provided a value as good as 0.93 ± 0.03. Grad-CAM maps show high activation in correspondence of pixels within the tumour regions.Conclusions We developed a deep learning-based framework (DBT-DCNN) to classify DBT images from clinical exams. We investigated also a possible application of the Grad-CAM technique to identify the lesion position.     

A practical methodology to improve the dosimetric accuracy of MR-based radiotherapy simulation for brain tumors

PurposeTo investigate the dosimetric accuracy of synthetic computed tomography (sCT) images generated by a clinically-ready voxel-based MRI simulation package, and to develop a simple and feasible method to improve the accuracy.Methods20 patients with brain tumor were selected to undergo CT and MRI simulation. sCT images were generated by a clinical MRI simulation package. The discrepancy between planning CT and sCT in CT number and body contour were evaluated. To resolve the discrepancies, an sCT specific CT-relative electron density (RED) calibration curve was used, and a layer of pseudo-skin was created on the sCT. The dosimetric impact of these discrepancies, and the improvement brought about by the modifications, were evaluated by a planning study. Volumetric modulated arc therapy (VMAT) treatment plans for each patient were created and optimized on the planning CT, which were then transferred to the original sCT and the modified-sCT for dose re-calculation. Dosimetric comparisons and gamma analysis between the calculated doses in different images were performed.ResultsThe average gamma passing rate with 1%/1 mm criteria was only 70.8% for the comparison of dose distribution between planning CT and original sCT. The mean dose difference between the planning CT and the original sCT were −1.2% for PTV D₉₅ and −1.7% for PTV D_max, while the mean dose difference was within 0.7 Gy for all relevant OARs. After applying the modifications on the sCT, the average gamma passing rate was increased to 92.2%. Mean dose difference in PTV D₉₅ and D_max were reduced to −0.1% and −0.3% respectively. The mean dose difference was within 0.2 Gy for all OAR structures and no statistically significant difference were found.ConclusionsThe modified-sCT demonstrated improved dosimetric agreement with the planning CT. These results indicated the overall dosimetric accuracy and practicality of this improved MR-based treatment planning method.     

Evaluation of the Block Matching deformable registration algorithm in the field of head-and-neck adaptive radiotherapy

Background and purposeTo compare the accuracy of the Block Matching deformable registration (DIR) against rigid image registration (RIR) for head-and-neck multi-modal images CT to cone-beam CT (CBCT) registration.Material and methodsPlanning-CT and weekly CBCT of 10 patients were used for this study. Several volumes, including medullary canal (MC), thyroid cartilage (TC), hyoid bone (HB) and submandibular gland (SMG) were transposed from CT to CBCT images using either DIR or RIR. Transposed volumes were compared with the manual delineation of these volumes on every CBCT. The parameters of similarity used for analysis were: Dice Similarity Index (DSI), 95%-Hausdorff Distance (95%-HD) and difference of volumes (cc).ResultsWith DIR, the major mean difference of volumes was −1.4 cc for MC, revealing limited under-segmentation. DIR limited variability of DSI and 95%-HD. It significantly improved DSI for TC and HB and 95%-HD for all structures but SMG. With DIR, mean 95%-HD (mm) was 3.01 ± 0.80, 5.33 ± 2.51, 4.99 ± 1.69, 3.07 ± 1.31 for MC, TC, HB and SMG, respectively. With RIR, it was 3.92 ± 1.86, 6.94 ± 3.98, 6.44 ± 3.37 and 3.41 ± 2.25, respectively.ConclusionBlock Matching is a valid algorithm for deformable multi-modal CT to CBCT registration. Values of 95%-HD are useful for ongoing development of its application to the cumulative dose calculation.     

Determination of radiation dose and low-dose protocol for digital chest tomosynthesis using radiophotoluminescent (RPL) glass dosimeters

PurposeThis study aimed to determine a low-dose protocol for digital chest tomosynthesis (DTS).MethodsFive simulated nodules with a CT number of approximately 100 HU with size diameter of 3, 5, 8, 10, and 12 mm were inserted into an anthropomorphic chest phantom (N1 Lungman model), and then scanned by DTS system (Definium 8000) with varying tube voltage, copper filter thickness, and dose ratio. Three radiophotoluminescent (RPL) glass dosimeters, type GD-352 M with a dimension of 1.5 × 12 mm, were used to measure the entrance surface air kerma (ESAK) in each protocol. The effective dose (ED) was calculated using the recorded total dose-area-product (DAP). The signal-to-noise ratio (SNR) was determined for qualitative image quality evaluation. The image criteria and nodule detection capability were scored by two experienced radiologists. The selected low-dose protocol was further applied in a clinical study with 30 pulmonary nodule follow-up patients.ResultsThe average ESAK obtained from the standard default protocol was 1.68 ± 0.15 mGy, while an ESAK of 0.47 ± 0.02 mGy was found for a low-dose protocol. The EDs for the default and low-dose protocols were 313.98 ± 0.72 µSv and 100.55 ± 0.28 µSv, respectively. There were small non-significant differences in the image criteria and nodule detection scoring between the low-dose and default protocols interpreted by two radiologists. The effective dose of 98.87 ± 0.08 µSv was obtained in clinical study after applying the low-dose protocol.ConclusionsThe low-dose protocol obtained in this study can substantially reduce radiation dose while preserving an acceptable image quality compared to the standard protocol.     

Assessment of daily dose accumulation for robustly optimized intensity modulated proton therapy treatment of prostate cancer

PurposeTo implement a daily CBCT based dose accumulation technique in order to assess ideal robust optimization (RO) parameters for IMPT treatment of prostate cancer.MethodsTen prostate cancer patients previously treated with VMAT and having daily CBCT were included. First, RO-IMPT plans were created with ± 3 mm and ± 5 mm patient setup and ± 3% proton range uncertainties, respectively. Second, the planning CT (pCT) was deformably registered to the CBCT to create a synthetic CT (sCT). Both daily and weekly sampling strategies were employed to determine optimal dose accumulation frequency. Doses were recalculated on sCTs for both ± 3 mm/±3% and ± 5 mm/±3% uncertainties and were accumulated back to the pCT. Accumulated doses generated from ± 3 mm/±3% and ± 5 mm/±3% RO-IMPT plans were evaluated using the clinical dose volume constraints for CTV, bladder, and rectum.ResultsDaily accumulated dose based on both ± 3mm/±3% and ±5 mm/±3% uncertainties for RO-IMPT plans resulted in satisfactory CTV coverage (RO-IMPT_3mm/3% CTV_V95 = 99.01 ± 0.87% vs. RO-IMPT_5mm/3% CTV_V95 = 99.81 ± 0.2%, P = 0.002). However, the accumulated dose based on ± 3 mm/3% RO-IMPT plans consistently provided greater OAR sparing than ±5 mm/±3% RO-IMPT plans (RO-IMPT_3mm/3% rectum_V65Gy = 2.93 ± 2.39% vs. RO-IMPT_5mm/3% rectum_V65Gy = 4.38 ± 3%, P < 0.01; RO-IMPT_3mm/3% bladder_V65Gy = 5.2 ± 7.12% vs. RO-IMPT_5mm/3% bladder_V65Gy = 7.12 ± 9.59%, P < 0.01). The gamma analysis showed high dosimetric agreement between weekly and daily accumulated dose distributions.ConclusionsThis study demonstrated that for RO-IMPT optimization, ±3mm/±3% uncertainty is sufficient to create plans that meet desired CTV coverage while achieving superior sparing to OARs when compared with ± 5 mm/±3% uncertainty. Furthermore, weekly dose accumulation can accurately estimate the overall dose delivered to prostate cancer patients.     

A quantitative evaluation of deformable image registration based on MV cone beam CT images: Impact of deformation magnitudes and image modalities

Background and purposeTo evaluate the impact of deformation magnitude and image modality on deformable-image-registration (DIR) accuracy using Halcyon megavoltage cone beam CT images (MVCBCT).Materials and methodsPlanning CT images of an anthropomorphic Head phantom were aligned rigidly with MVCBCT and re-sampled to achieve the same resolution, denoted as pCT. MVCBCT was warped with twenty simulated pre-known virtual deformation fields (T_i, i = 1–20) with increasing deformation magnitudes, yielding warped CBCT (wCBCT). The pCT and MVCBCT were registered to wCBCT respectively (Multi-modality and Uni-modality DIR), generating deformation vector fields V_i and V_i′ (i = 1–20). V_i and V_i′ were compared with T_i respectively to assess the DIR accuracy geometrically. In addition, V_i, T_i, and V_i′ were applied to pCT, generating deformed CT (dCT_i), ground-truth CT (G_i) and deformed CT′ (dCT_i′) respectively. The Hounsfield Unit (HU) on these virtual CT images were also compared.ResultsThe mean errors of vector displacement increased with the deformation magnitude. For deformation magnitudes between 2.82 mm and 7.71 mm, the errors of uni-modality DIR were 1.16 mm ~ 1.73 mm smaller than that of multi-modality (p = 0.0001, Wilcoxon signed rank test). DIR could reduce the maximum signed and absolute HU deviations from 70.8 HU to 11.4 HU and 208 HU to 46.2 HU respectively.ConclusionsAs deformation magnitude increases, DIR accuracy continues to deteriorate and uni-modality DIR consistently outperformed multi-modality DIR. DIR-based adaptive radiotherapy utilizing the noisy MVCBCT images is only conditionally applicable with caution.     

能谱CT优化胃肿瘤扫描辐射剂量对肾上腺嗜铬细胞瘤的诊断价值

目的:探讨能谱CT优化胃肿瘤扫描辐射剂量对肾上腺嗜铬细胞瘤的诊断价值。方法:采用回顾性、抽样、随机研究方法选择2012年9月到2017年2月在我院诊治的肾上腺嗜铬细胞瘤患者59例作为研究对象,所有患者都给予常规CT扫描与能谱CT优化胃肿瘤扫描,记录和比较辐射剂量与图像质量。结果:所有病例包膜均完整,边缘清楚,肿瘤内见单发或多发低密度区,肿瘤实质区呈不均匀显著强化。常规CT与能谱CT的图像质量主观评分分别为3.89±0.45分和4.54±0.34分;常规CT与能谱CT图像的胃肿瘤CT值分别为31.94±6.39HU和35.29±5.19HU,对比都有显著差异(P0.05)。能谱CT图像的膀胱和皮下脂肪图像噪声值都显著低于常规CT图像,对比差异都有统计学意义(P0.05);能谱CT扫描的CTDIvol和DLP分别为12.39±3.48mGy和624.10±39.19mGy.cm,都显著低于常规CT扫描的14.09±4.13mGy和653.92±56.29mGy.cm(P0.05)。结论:能谱CT优化胃肿瘤扫描在肾上腺嗜铬细胞瘤诊断中的应用能有效减少辐射剂量与图像噪声,提高图像CT值与主观质量,临床应用价值更高。     

Dual energy computed tomography thermometry during hepatic microwave ablation in an ex-vivo porcine model

BackgroundMicrowave thermoablation (MTA) is a treatment method used to destroy hepatic tumors.ObjectiveTo investigate temperature changes during MTA of ex-vivo porcine liver using dual-energy computed tomography (DECT) imaging.MethodsThree MTA experiments were performed using ex-vivo porcine liver (15 × 15 × 15 cm³) and DECT imaging with 80/Sn140 kVp spectral and 0.5-weighted reconstructions. Images were acquired from basic organ temperature to 100 °C with 10 °C difference during microwave heating and cooling phases. Three fluoro-optic thermometers were used for temperature measurements; two were placed at 1 cm and third one positioned at 2 cm distance from the applicator. Tissue temperature, ablation-region-conspicuity (ARC), ablation-region dimensions and image quality were determined. Regression analysis was performed determining thermal sensitivity during heating and cooling phases.ResultsDetermined thermal sensitivities during heating phase were: −0.59 Hounsfield Unit/°C (80 kVp), −0.60HU/°C (0.5-weighted) and −0.59HU/°C (140 kVp); furthermore, during cooling: −0.56HU/°C (80 kVp), −0.55HU/°C (0.5-weighted) and −0.55HU/°C (140 kVp). ARC showed significantly higher (all P < 0.05) values for thermometer positions −1 and −2 compared to −3; however, comparison between positions −1 and −2 was insignificant (P > 0.05). Signal-to-noise ratios were higher for 0.5-weighted but ARC values were higher for 80 kVp images.ConclusionMicrowave thermal sensitivity on tissue was inversely linear with DECT image datasets. Heating phase showed higher influence of temperature on HU compared to cooling; ARC and ablation-region were increased with increase in temperature.     

Deep template-based protein structure prediction

MotivationProtein structure prediction has been greatly improved by deep learning, but most efforts are devoted to template-free modeling. But very few deep learning methods are developed for TBM (template-based modeling), a popular technique for protein structure prediction. TBM has been studied extensively in the past, but its accuracy is not satisfactory when highly similar templates are not available.ResultsThis paper presents a new method NDThreader (New Deep-learning Threader) to address the challenges of TBM. NDThreader first employs DRNF (deep convolutional residual neural fields), which is an integration of deep ResNet (convolutional residue neural networks) and CRF (conditional random fields), to align a query protein to templates without using any distance information. Then NDThreader uses ADMM (alternating direction method of multipliers) and DRNF to further improve sequence-template alignments by making use of predicted distance potential. Finally, NDThreader builds 3D models from a sequence-template alignment by feeding it and sequence coevolution information into a deep ResNet to predict inter-atom distance distribution, which is then fed into PyRosetta for 3D model construction. Our experimental results show that NDThreader greatly outperforms existing methods such as CNFpred, HHpred, DeepThreader and CEthreader. NDThreader was blindly tested in CASP14 as a part of RaptorX server, which obtained the best average GDT score among all CASP14 servers on the 58 TBM targets.     

A quantitative comparison of data evaluation methods to derive diagnostic reference levels for CT from a dosimetric survey: Correlation analysis compared to simple evaluation strategies

ObjectivesTo compare simple and sophisticated evaluation strategies for CT dosimetry surveys with focus on DRLs.MethodsBased on data from a nationwide Austrian CT dose survey, different evaluation strategies are compared. These were pooled data analysis, weight banding excluding data from patients with weights outside ±20 kg of the standard weights (70 and 75.6 kg representing the actual average weight), and a regression method estimating DLP probability distributions for the standard patient for each scanner before calculating quartiles.ResultsIn the abdomen and chest region, weight restriction (?9% and ?4% around 70 and 75.6 kg, respectively, compared to pooled data analysis) and statistically weighting each scanner equally (?9%) have the largest effect on DRLs derived. However, the difference in 3rd quartiles calculated using weight restriction alone compared to regression analysis is relatively small (<1% for 70 ± 20 and ?6% for 75.6 ± 20 kg, respectively, trunk region). In the head/neck region the effect of weight restriction is less than in for scans of the trunk (?1.3% and ?0.2%, respectively); the most prominent changes resulted from excluding scanners with less than 10 patient cases (?5%), and equally weighting scanners rather than cases (?3%).ConclusionFor adult CT examinations (different to a paediatric survey), quite simple evaluation strategies yield results very comparable to those from sophisticated strategies.