Publications by authors named "Chenyang Shen"

62 Publications

Improving robustness of a deep learning-based lung-nodule classification model of CT images with respect to image noise.

Phys Med Biol 2021 Nov 24. Epub 2021 Nov 24.

Department of Radiation Oncology, UT Southwestern Medical Center, 6363 Forest Park Rd. BL10.202G, MC9315, Dallas, Texas, 75390-9315, UNITED STATES.

Objective: Robustness is an important aspect to consider, when developing methods for medical image analysis. This study investigated robustness properties of deep neural networks (DNNs) for a lung nodule classification problem based on CT images and proposed a solution to improve robustness.

Approach: We firstly constructed a class of four DNNs with different widths, each predicting an output label (benign or malignant) for an input CT image cube containing a lung nodule. These networks were trained to achieve Area Under the Curve of 0.891-0.914 on a testing dataset. We then added to the input CT image cubes noise signals generated randomly using a realistic CT image noise model based on a noise power spectrum at 100 mAs, and monitored the DNN's output change. We defined $SAR_{5} (\%)$ to quantify the robustness of the trained DNN model, indicating that for $5\%$ of CT image cubes, the noise can change the prediction results with a chance of at least $SAR_{5} (\%)$. To understand robustness, we viewed the information processing pipeline by the DNN as a two-step process, with the first step using all but the last layers to extract representations of the input CT image cubes in a latent space, and the second step employing the last fully-connected layer as a linear classifier to determine the position of the sample representations relative to a decision plane. To improve robustness, we proposed to retrain the last layer of the DNN with a Supporting Vector Machine (SVM) hinge loss function to enforce the desired position of the decision plane.

Main Results: $SAR_{5}$ ranged in $47.0\sim 62.0\%$ in different DNNs. The unrobustness behavior may be ascribed to the unfavorable placement of the decision plane in the latent representation space, which made some samples be perturbed to across the decision plane and hence susceptible to noise. The DNN-SVM model improved robustness over the DNN model and reduced $SAR_{5}$ by $8.8\sim 21.0\%$.

Significance: This study provided insights about the potential reason for the unrobustness behavior of DNNs and the proposed DNN-SVM model improved model robustness.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1088/1361-6560/ac3d16DOI Listing
November 2021

Antiviral Therapy with Entecavir following Antituberculosis Therapy Alleviates Liver Injury and Restores Innate Immunity in Tuberculosis Patients Coinfected with Hepatitis B Virus.

Evid Based Complement Alternat Med 2021 2;2021:2884151. Epub 2021 Nov 2.

Infection Department, Hangzhou Yuhang First People's Hospital, Yuhang Branch Second Affiliated Hospital Zhejiang University School of Medicine, No. 369, Yingbin Road, Nanyuan Street, Yuhang District, Hangzhou 311100, Zhejiang, China.

Objective: Coinfection of tuberculosis (TB) and viral hepatitis may increase the risk of antituberculosis treatment-induced hepatotoxicity, which is regarded as a common cause of termination of the first-line antituberculosis drugs. The study aimed at investigating the protective effects of antiviral therapy on the liver and innate immunity in patients with TB-HBV coinfection.

Methods: A total of 100 patients with TB-HBV coinfection were recruited and split into antituberculosis and antiviral groups, 50 per group, according to odd or even date of hospital admission from December 2019 to October 2020. The patients in the anti-TB group received antituberculosis therapy, and those in the antiviral group received antiviral therapy. The clinical effectiveness; HBV-DNA negative conversion rate; liver function assessment involving alanine aminotransferase (ALT), aspartate aminotransferase (AST), and total bilirubin (TBIL); immune function evaluation including CD4, CD8, CD4/CD8, and CD3 T cells; inflammatory cytokines containing tumor necrosis factor- (TNF-), interleukin-6 (IL-6), and interferon- (IFN-); and intestinal microflora including bifidobacterium, lactobacillus, enterobacterium, enterococcus, and clostridium were main outcome measures after treatment.

Results: It was found that the total response rate in the antiviral group was significantly higher than the anti-TB group after treatment (  = 3.157, =0.017). There was a significant difference in HBV-DNA negative conversion rates between the antiviral group and anti-TB group (82% vs. 58%,  = 6.384, =0.001). The ALT, AST, and TBIL in the two groups were all increased after treatment ( < 0.05), but the antiviral group indicated a rise of the above indices compared to the anti-TB group ( < 0.05). The two groups showed a rise on the concentration of CD3, CD4, and CD4/CD8 T cells and a decline on the CD8 T cells after treatment ( < 0.05), but these changes in the antiviral group were more evident to those in the anti-TB group ( < 0.05). There was an increase on the IFN- level and decrease on the TNF- and IL-6 levels in both groups after treatment ( < 0.05), but the antiviral group revealed a higher level of IFN- with lower levels of TNF- and IL-6 compared to the anti-TB group ( < 0.05). After treatment, the number of bifidobacteria and lactobacilli was increased, and the number of enterobacteria, enterococci, and clostridium were decreased in the two groups ( < 0.05), while these changes in the antiviral group were more remarkable compared to the anti-TB group ( < 0.05). There was no significant difference in the incidence of adverse reactions between the two groups (2 = 0.267, =0.731).

Conclusion: Antiviral therapy for tuberculosis-HBV coinfected patients could inhibit HBV replication, providing protection against liver damage, improving innate immunity, and balancing intestinal microflora.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1155/2021/2884151DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC8577919PMC
November 2021

Efficient Zn-Alloyed Low-Toxicity Quasi-Two-Dimensional Pure-Red Perovskite Light-Emitting Diodes.

ACS Appl Mater Interfaces 2021 Nov 10;13(46):55412-55419. Epub 2021 Nov 10.

State Key Laboratory of Luminescent Materials and Devices and Institute of Polymer Optoelectronic Materials and Devices, South China University of Technology, Wushan Road 381, Guangzhou 510640, China.

Metal halide perovskites have attracted extensive attention in next-generation solid-state lighting and displays due to their fascinating optoelectronic properties. However, the toxicity of lead (Pb) impedes their practical application. Herein, we report an efficient Zn-alloyed quasi-two-dimensional (quasi-2D) pure-red perovskite light-emitting device (PeLED) by introducing zinc ions (Zn) into the perovskite lattice and partially substituting Pb. The substitution of Zn is confirmed by X-ray diffraction, X-ray photoelectron spectroscopy, grazing-incidence wide-angle X-ray scattering, and transmission electron microscopy measurements. In addition, the vacancy defect density of Pb and the halogen is reduced by the introduction of Zn in the PEA(CsMA)(ZnPb)I perovskite system, which leads to a more ordered crystal orientation, compact morphology, and increased photoluminescence quantum efficiency. Benefiting from the improved photoelectric properties, a maximum EQE of 9.5% and a luminescence of 453 cd m are achieved for the Zn-alloyed PeLEDs, with a maximum emission peak of 658 nm and stable electroluminescence spectra under various applied biases.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1021/acsami.1c16242DOI Listing
November 2021

Improving dose calculation accuracy in preclinical radiation experiments using multi-energy element resolved cone beam CT.

Phys Med Biol 2021 Nov 9. Epub 2021 Nov 9.

Department of Radiation Oncology, UT Southwestern Medical Center, 6363 Forest Park Rd. BL10.202G, MC9315, Dallas, Texas, 75390-9315, UNITED STATES.

Cone-beam CT (CBCT) in modern pre-clinical small-animal radiation research platforms provides volumetric images for image guidance and experiment planning purposes. In this work, we implemented multi-energy element-resolved (MEER) CBCT using three scans with different kVps on a SmART platform (Precision X-ray Inc.) We performed comprehensive calibration tasks achieve sufficient accuracy for this quantitative imaging purpose. For geometry calibration, we scanned a ball bearing phantom and used an analytical method together with an optimization approach to derive gantry-angle specific geometry parameters. Intensity calibration and correction included the corrections for detector lag, glare, and beam hardening. The corrected CBCT projection images acquired at 30, 40 and 60 kVp in multiple scans were used to reconstruct CBCT images using the Feldkamp-Davis-Kress reconstruction algorithm. After that, an optimization problem was solved to determine images of relative electron density (rED) and elemental composition (EC) that are needed for Monte Carlo-based radiation dose calculation. We demonstrated effectiveness of our CBCT calibration steps by showing improvements in image quality and successful material decomposition in cases with a small animal CT calibration phantom and a plastinated mouse phantom. It was found that artifacts induced by geometry inaccuracy, detector lag, glare and beam hardening were visually reduced. CT number mean errors were reduced from 19\% to 5\%. In the CT calibration phantom case, median errors in H, O, and Ca fractions for all the inserts were below 1\%, 2\%, and 4\% respectively, and median error in rED was less than 5\%. Compared to standard approach deriving material type and rED via CT number conversion, our approach improved Monte Carlo simulation-based dose calculation accuracy in bone regions. Mean dose error was reduced from 47.5\% to 10.9\%.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1088/1361-6560/ac37fcDOI Listing
November 2021

Design and experimental validation of a unilateral magnet for MRI-guided small animal radiation experiments.

J Magn Reson 2021 11 16;332:107062. Epub 2021 Sep 16.

innovative Technology Of Radiotherapy Computations and Hardware (iTORCH) Laboratory, Department of Radiation Oncology, University of Texas Southwestern Medical Center, Dallas 75390, TX, USA. Electronic address:

Small animal radiation experiments are of paramount importance for the advancement of human radiation therapy. These experiments use a dedicated radiation platform to deliver radiation to small animals, such as mice and rats, similar to how human radiation therapy is performed. By acquiring images immediately before radiation delivery to guide positioning of the animals, image guidance plays a critical role to ensure accuracy of the experiments. Recently, MR-based image guidance has been enabled in human radiation therapy. This paper proposes a new concept using a unilateral magnet-based MRI scanner to realize image guidance for small animal radiation experiments. We reported our design, optimization, construction, and characterization of the magnet. The magnet was designed using eight 2-inch neodymium magnet cubes arranged in a modified Halbach ring configuration. The ring has an opening to allow for animal positioning. We considered a spherical region of interest (ROI) located outside of the ring's plane to allow radiation delivery to the ROI without obstruction of the magnet. An optimization problem was formulated and solved to determine the positions and orientations of the magnet cubes to generate a magnetic field with desired properties in the ROI. The optimization improved the average magnetic flux density from 55 mT to 72 mT and reduced variation from 1.2 T/m to 1.0 T/m. We constructed the magnet using 3D-printed templates to hold the neodymium magnet cubes with the optimized positions and orientations. We measured the spatial distribution of the magnetic flux density. The measurement results and computed results agreed with an average difference of 0.35% through the ROI.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.jmr.2021.107062DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC8546776PMC
November 2021

Human health risk assessment of groundwater nitrate at a two geomorphic units transition zone in northern China.

J Environ Sci (China) 2021 Dec 26;110:38-47. Epub 2021 Mar 26.

Key Laboratory of Drinking Water Science and Technology, Research Center for Eco-Environmental Science Chinese Academy of Sciences, Beijing 100085, China; University of Chinese Academy of Sciences, Beijing 100049, China. Electronic address:

To assess groundwater nitrate contamination and its human health risks, 489 unconfined groundwater samples were collected and analyzed from Zhangjiakou, northern China. The spatial distribution of principle hydrogeochemical results showed that the average concentrations of ions in descend order was HCO, SO, Na, Ca, Cl, NO, Mg and K, among which the NO concentrations were between 0.25 and 536.73 mg/L with an average of 29.72 mg/L. In total, 167 out of 489 samples (~ 34%) exceeded the recommended concentration of 20 mg/L in Quality Standard for Groundwater of China. The high NO concentration groundwater mainly located in the northern part and near the boundary of the two geomorphic units. As revealed by statistical analysis, the groundwater chemistry was more significantly affected by anthropogenic sources than by the geogenic sources. Moreover, human health risks of groundwater nitrate through oral and dermal exposure pathways were assessed by model, the results showed that about 60%, 50%, 32% and 26% of the area exceeded the acceptable level (total health index>1) for infants, children, adult males and females, respectively. The health risks for different groups of people varied significantly, ranked: infants> children> adult males>adult females, suggesting that younger people are more susceptible to nitrate contamination, while females are more resistant to nitrate contamination than males. To ensure the drinking water safety in Zhangjiakou and its downstream areas, proper management and treatment of groundwater will be necessary to avoid the health risks associated with nitrate contamination.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.jes.2021.03.013DOI Listing
December 2021

Perovskite Light-Emitting Diodes with EQE Exceeding 28% through a Synergetic Dual-Additive Strategy for Defect Passivation and Nanostructure Regulation.

Adv Mater 2021 Oct 20;33(43):e2103268. Epub 2021 Sep 20.

State Key Laboratory of Luminescent Materials and Devices and Institute of Polymer Optoelectronic Materials and Devices, South China University of Technology, Guangzhou, 510640, P. R. China.

Quasi-2D perovskites have long been considered to have favorable "energy funnel/cascade" structures and excellent optical properties compared with their 3D counterparts. However, most quasi-2D perovskite light-emitting diodes (PeLEDs) exhibit high external quantum efficiency (EQE) but unsatisfactory operating stability due to Auger recombination induced by high current density. Herein, a synergetic dual-additive strategy is adopted to prepare perovskite films with low defect density and high environmental stability by using 18-crown-6 and poly(ethylene glycol) methyl ether acrylate (MPEG-MAA) as the additives. The dual additives containing COC bonds can not only effectively reduce the perovskite defects but also destroy the self-aggregation of organic ligands, inducing the formation of perovskite nanocrystals with quasi-core/shell structure. After thermal annealing, the MPEG-MAA with its CC bond can be polymerized to obtain a comb-like polymer, further protecting the passivated perovskite nanocrystals against water and oxygen. Finally, state-of-the-art green PeLEDs with a normal EQE of 25.2% and a maximum EQE of 28.1% are achieved, and the operating lifetime (T ) of the device in air environment is over ten times increased, providing a novel and effective strategy to make high efficiency and long operating lifetime PeLEDs.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1002/adma.202103268DOI Listing
October 2021

A hierarchical deep reinforcement learning framework for intelligent automatic treatment planning of prostate cancer intensity modulated radiation therapy.

Phys Med Biol 2021 Jun 23;66(13). Epub 2021 Jun 23.

Medical Artificial Intelligence and Automation (MAIA) Laboratory, Department of Radiation Oncology, University of Texas Southwestern Medical Center, Dallas, TX 75390, United States of America.

We have previously proposed an intelligent automatic treatment planning (IATP) framework that builds a virtual treatment planner network (VTPN) to operate a treatment planning system (TPS) to generate high-quality radiation therapy (RT) treatment plans. While the potential of IATP in automating RT treatment planning has been demonstrated, its poor scalability caused by an almost linear growth of network size with the number of treatment planning parameters (TPPs) is a bottleneck, preventing its application in complicate, but clinically relevant treatment planning problems. The decision-making behavior of the trained network is hard to understand. Motivated by the decision-making process of a human planner, this study proposes a hierarchical IATP framework.The hierarchical VTPN (HieVTPN) consists of three networks, i.e. Structure-Net, Parameter-Net, and Action-Net. When interacting with a TPS, the networks are employed in a sequential order in each step to decide the structure to adjust, the TPP to adjust for the selected structure, and the specific adjustment manner for the parameter, respectively. We developed an end-to-end hierarchical deep reinforcement learning scheme to simultaneously train the three networks. We then evaluated the effectiveness of the proposed framework in the treatment planning problems for prostate cancer intensity modulated RT (IMRT) and stereotactic body RT (SBRT). We benchmarked the performance of our approach by comparing plans made by VTPN of a parallel architecture, and the human plans submitted for competition in the 2016 American Association of Medical Dosimetrist (AAMD)/Radiosurgery Society (RSS) Plan Study. We analyzed scalability of the network size with respect to the number of TPPs. Numerical experiments were also performed to understand the rationale of the decision-making behaviors of the trained HieVTPN.Both HieVTPNs for prostate IMRT and SBRT were trained successfully using 10 training patient cases and 5 validation cases. For IMRT, HieVTPN was able to generate high-quality plans for 59 testing patient cases that were not included in training process, achieving an average plan score of 8.62 (±0.83), with 9 being the maximal score. The score was comparable to that of the VTPN, 8.45 (±0.48). For SBRT planning, HieVTPN achieved an average plan score of 139.07 on five testing patient cases compared to the score of 132.21 averaged over the human plans summited for competition in AAMD/RSS plan study. Different from VTPN with network size linearly scaling with the number of TPPs, the network size of HieVTPN is almost independent of the number of TPPs. It was also observed that the decision-making behaviors of HieVTPN were understandable and generally agreed with the human experience.With the scalability and explainability, the hierarchical IATP framework is more favorable than the previous framework in terms of handling treatment planning problems involving a large number of TPPs.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1088/1361-6560/ac09a2DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC8406431PMC
June 2021

Synthetic CT generation from CBCT images via unsupervised deep learning.

Phys Med Biol 2021 05 31;66(11). Epub 2021 May 31.

Medical Artificial Intelligence and Automation (MAIA) Lab, Department of Radiation Oncology, The University of Texas Southwestern Medical Center, Dallas, TX 75390 United States of America.

Adaptive-radiation-therapy (ART) is applied to account for anatomical variations observed over the treatment course. Daily or weekly cone-beam computed tomography (CBCT) is commonly used in clinic for patient positioning, but CBCT's inaccuracy in Hounsfield units (HU) prevents its application to dose calculation and treatment planning. Adaptive re-planning can be performed by deformably registering planning CT (pCT) to CBCT. However, scattering artifacts and noise in CBCT decrease the accuracy of deformable registration and induce uncertainty in treatment plan. Hence, generating from CBCT a synthetic CT (sCT) that has the same anatomical structure as CBCT but accurate HU values is desirable for ART. We proposed an unsupervised style-transfer-based approach to generate sCT based on CBCT and pCT. Unsupervised learning was desired because exactly matched CBCT and CT are rarely available, even when they are taken a few minutes apart. In the proposed model, CBCT and pCT are two inputs that provide anatomical structure and accurate HU information, respectively. The training objective function is designed to simultaneously minimize (1) contextual loss between sCT and CBCT to maintain the content and structure of CBCT in sCT and (2) style loss between sCT and pCT to achieve pCT-like image quality in sCT. We used CBCT and pCT images of 114 patients to train and validate the designed model, and another 29 independent patient cases to test the model's effectiveness. We quantitatively compared the resulting sCT with the original CBCT using the deformed same-day pCT as reference. Structure-similarity-index, peak-signal-to-noise-ratio, and mean-absolute-error in HU of sCT were 0.9723, 33.68, and 28.52, respectively, while those of CBCT were 0.9182, 29.67, and 49.90, respectively. We have demonstrated the effectiveness of the proposed model in using CBCT and pCT to synthesize CT-quality images. This model may permit using CBCT for advanced applications such as adaptive treatment planning.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1088/1361-6560/ac01b6DOI Listing
May 2021

Isolated True Subclavian Aneurysm without Aberrant Subclavian Artery or Coarctation of Descending Aorta.

Ann Vasc Surg 2021 Aug 2;75:294-300. Epub 2021 Apr 2.

Department of Cardiovascular Surgery, National Center for Cardiovascular Diseases and Fuwai Hospital, Chinese Academy of Medical Sciences, Peking Union Medical College, Beijing, China. Electronic address:

Objective: Isolated true subclavian artery aneurysm (SAA) without aberrant subclavian artery or coarctation of descending aorta is a rare peripheral aneurysm. Herein, the experience of our medical center in the treatment of this disease is presented.

Methods: The Division operative log was queried to identify cases of SAA repair between January 2012 and September 2019 that were not associated with coarctation of the aorta or the presence of an aberrant subclavian artery. A total of 22 cases were identified. The characteristics, treatment and clinical outcomes of these cases were assessed.

Results: The mean age of patients was 53.5 ± 14.3 years and 14 patients were male (63.6%). Half of the cases were attributed to atherosclerotic degeneration. The clinical symptoms of aneurysms were varied, including asymptomatic, pulsatile mass of supraclavicular fossa, local pain, upper limb embolism, Horner's syndrome and hoarseness. Aneurysms were located on the right in 17 cases, on the left in 3 cases and on both sides in 2 cases. Fifteen (68%) patients underwent an intervention, of which 11 (50%) underwent an open surgical repair, and 4 (18%) underwent endovascular repair. The mean diameter of the aneurysms was 39.5 ± 20.7 mm in the open surgery group, and 24.0 ± 4.7 mm in the endovascular group. The follow-up duration ranged from 2 months to 12 years. One patient died of cardiogenic disease in the untreated group. Patients undergoing open operative repair had 100% patency of the reconstruction. In the endovascular group, one patient had stent occlusion 2 years after the operation.

Conclusions: The most common cause of isolated subclavian aneurysm without aberrant subclavian artery or coarctation of descending aorta is atherosclerosis. The clinical symptoms of aneurysms are varied, and the aneurysms tend to occur on the right side. Based on the anatomical conditions of SAAs, open surgery and endovascular repair can be used for treatment.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.avsg.2021.01.108DOI Listing
August 2021

MiR-520b inhibits endothelial activation by targeting NF-κB p65-VCAM1 axis.

Biochem Pharmacol 2021 06 2;188:114540. Epub 2021 Apr 2.

Key Laboratory of Cardiovascular Epidemiology & Department of Epidemiology, State Key Laboratory of Cardiovascular Disease, Fuwai Hospital, National Center for Cardiovascular Diseases, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing 100037, China.

MiR-520b belongs to the miR-373/520 family, is expressed only in human and nonhuman primates. Previous reports indicated that the expression of miR-520b was repressed in human atherosclerotic plaque tissue compared with healthy vessels. However, the role of miR-520b in coronary artery disease still remains to be uncovered. In this study, we demonstrated that endothelial cells (ECs) in human atherosclerotic plaques expressed miR-520b and aimed to elucidate the impact of miR-520b on EC activation and inflammatory response. To determine the potential targets of miR-520b, we performed RNA-seq analysis by transfecting miR-520b mimics in ECs. The quantitative real-time PCR (qPCR) validation suggested that miR-520b over-expression reduced pro-inflammatory gene expression (e.g. ICAM1, VCAM1, SELE) while the inhibition of miR-520b induced their expression. By combining bioinformatics prediction and functional assays, we identified that RELA (Nuclear Factor-κB (NF-κB) Transcription Factor P65) was a direct target of miR-520b. Moreover, miR-520b mimics attenuated monocyte adhesion and monocyte trans-endothelial migration (the initial steps of atherosclerotic formation) in response to lipopolysaccharides (LPS) stimulation. Re-expression of a non-miR-targetable version of p65 could rescue the reduced monocyte cell attachment, suggesting that this process is NF-κB p65 dependent. MiR-520b reduced the abundance of NF-κB p65 in cytoplasmic fractions without corresponding increase in nuclear fractions, indicating that this regulation is independent of p65 translocation process. MiR-520b mimics attenuated the activity of VCAM-1 promoter, whereas miR-520b inhibitor activated its activity. However, miR-520b inhibitor had no effect on promoter activity containing the mutated NF-κB p65 binding sites, strongly demonstrating that the impact of miR-520b on VCAM1 gene is mediated by NF-κB p65. Thus, we concluded that miR-520b suppressed EC inflammation and the cross-talk between monocytes and ECs by down-regulating NF-κB p65-ICAM1/VCAM1 axis and might serve as a potential therapeutic target for EC dysfunction and atherosclerosis.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.bcp.2021.114540DOI Listing
June 2021

Improving efficiency of training a virtual treatment planner network via knowledge-guided deep reinforcement learning for intelligent automatic treatment planning of radiotherapy.

Med Phys 2021 Apr 16;48(4):1909-1920. Epub 2021 Feb 16.

Medical Artificial Intelligence and Automation (MAIA) Laboratory, Department of Radiation Oncology, University of Texas Southwestern Medical Center, Dallas, TX, USA.

Purpose: We previously proposed an intelligent automatic treatment planning framework for radiotherapy, in which a virtual treatment planner network (VTPN) is built using deep reinforcement learning (DRL) to operate a treatment planning system (TPS) by adjusting treatment planning parameters in it to generate high-quality plans. We demonstrated the potential feasibility of this idea in prostate cancer intensity-modulated radiation therapy (IMRT). Despite the success, the process to train a VTPN via the standard DRL approach with an ϵ-greedy algorithm was time-consuming. The required training time was expected to grow with the complexity of the treatment planning problem, preventing the development of VTPN for more complicated but clinically relevant scenarios. In this study, we proposed a novel knowledge-guided DRL (KgDRL) approach that incorporated knowledge from human planners to guide the training process to improve the efficiency of training a VTPN.

Method: Using prostate cancer IMRT as a test bed, we first summarized a number of rules in the actions of adjusting treatment planning parameters of our in-house TPS. During the training process of VTPN, in addition to randomly navigating the large state-action space, as in the standard DRL approach using the ϵ-greedy algorithm, we also sampled actions defined by the rules. The priority of sampling actions from rules decreased over the training process to encourage VTPN to explore new policy on parameter adjustment that were not covered by the rules. To test this idea, we trained a VTPN using KgDRL and compared its performance with another VTPN trained using the standard DRL approach. Both networks were trained using 10 training patient cases and five additional cases for validation, while another 59 cases were employed for the evaluation purpose.

Results: It was found that both VTPNs trained via KgDRL and standard DRL spontaneously learned how to operate the in-house TPS to generate high-quality plans, achieving plan quality scores of 8.82 (±0.29) and 8.43 (±0.48), respectively. Both VTPNs outperformed treatment planning purely based on the rules, which had a plan score of 7.81 (±1.59). VTPN trained with eight episodes using KgDRL was able to perform similar to that trained using DRL with 100 epochs. The training time was reduced from more than a week to ~13 hrs.

Conclusion: The proposed KgDRL framework was effective in accelerating the training process of a VTPN by incorporating human knowledge, which will facilitate the development of VTPN for more complicated treatment planning scenarios.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1002/mp.14712DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC8058266PMC
April 2021

Semi-automatic sigmoid colon segmentation in CT for radiation therapy treatment planning via an iterative 2.5-D deep learning approach.

Med Image Anal 2021 02 16;68:101896. Epub 2020 Dec 16.

innovative Technology of Radiotherapy Computation and Hardware (iTORCH) Laboratory. Department of Radiation Oncology, University of Texas Southwestern Medical Center, Dallas, TX 75390, USA; Department of Radiation Oncology, University of Texas Southwestern Medical Center, Dallas, TX 75390, USA. Electronic address:

Automatic sigmoid colon segmentation in CT for radiotherapy treatment planning is challenging due to complex organ shape, close distances to other organs, and large variations in size, shape, and filling status. The patient bowel is often not evacuated, and CT contrast enhancement is not used, which further increase problem difficulty. Deep learning (DL) has demonstrated its power in many segmentation problems. However, standard 2-D approaches cannot handle the sigmoid segmentation problem due to incomplete geometry information and 3-D approaches often encounters the challenge of a limited training data size. Motivated by human's behavior that segments the sigmoid slice by slice while considering connectivity between adjacent slices, we proposed an iterative 2.5-D DL approach to solve this problem. We constructed a network that took an axial CT slice, the sigmoid mask in this slice, and an adjacent CT slice to segment as input and output the predicted mask on the adjacent slice. We also considered other organ masks as prior information. We trained the iterative network with 50 patient cases using five-fold cross validation. The trained network was repeatedly applied to generate masks slice by slice. The method achieved average Dice similarity coefficients of 0.82 0.06 and 0.88 0.02 in 10 test cases without and with using prior information.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.media.2020.101896DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7847132PMC
February 2021

Experimental and numerical studies on kV scattered x-ray imaging for real-time image guidance in radiation therapy.

Phys Med Biol 2021 02 11;66(4):045022. Epub 2021 Feb 11.

Innovative Technology Of Radiotherapy Computations and Hardware (iTORCH) Laboratory, Department of Radiation Oncology, University of Texas Southwestern Medical Center, Dallas, TX, 75235, United States of America. Department of Radiation Oncology, University of Texas Southwestern Medical Center, Dallas, TX, 75235, United States of America.

Motion management is a critical component of image guided radiotherapy for lung cancer. We previously proposed a scheme using kV scattered x-ray photons for marker-less real-time image guidance in lung cancer radiotherapy. This study reports our recent progress using the photon counting detection technique to demonstrate potential feasibility of this method and using Monte Carlo (MC) simulations and ray-tracing calculations to characterize the performance. In our scheme, a thin slice of x-ray beam was directed to the target and we measured the outgoing scattered photons using a photon counting detector with a parallel-hole collimator to establish the correspondence between detector pixels and scatter positions. Image corrections of geometry, beam attenuation and scattering angle were performed to convert the raw image to the actual image of Compton attenuation coefficient. We set up a MC simulation system using an in-house developed GPU-based MC package modeling the image formation process. We also performed ray-tracing calculations to investigate the impacts of imaging system geometry on resulting image resolution. The experiment demonstrated feasibility of using a photon counting detector to measure scattered x-ray photons and generate the proposed scattered x-ray image. After correction, x-ray scattering image intensity and Compton scattering attenuation coefficient were linearly related, with R greater than 0.9. Contrast to noise ratios of different objects were improved and the values in experimental results and MC simulation results agreed with each other. Ray-tracing calculations revealed the dependence of image resolution on imaging geometry. The image resolution increases with reduced source to object distance and increased collimator height. The study demonstrated potential feasibility of using scattered x-ray imaging as a real-time image guidance method in radiation therapy.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1088/1361-6560/abd66cDOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC8283910PMC
February 2021

Simultaneous needle catheter selection and dwell time optimization for preplanning of high-dose-rate brachytherapy of prostate cancer.

Phys Med Biol 2021 03 1;66(5):055028. Epub 2021 Mar 1.

Innovative Technology Of Radiotherapy Computation and Hardware (iTORCH) Laboratory, Department of Radiation Oncology, University of Texas Southwestern Medical Center, Dallas, TX 75390, United States of America. Department of Radiation Oncology, University of Texas Southwestern Medical Center, Dallas, TX 75287, United States of America.

Purpose: Needle catheter positions critically affect the quality of treatment plans in prostate cancer high-dose-rate (HDR) brachytherapy. The current standard needle positioning approach is based on human intuition, which cannot guarantee a high-quality plan. This study proposed a method to simultaneously select needle catheter positions and determine dwell time for preplanning of HDR brachytherapy of prostate cancer.

Methods: We formulated the needle catheter selection problem and inverse dwell time optimization problem in a unified framework. In addition to the dose objectives of the planning target volume (PTV) and organs at risk (OARs), the objective function incorporated a group-sparsity term with a needle-specific adaptive weighting scheme to generate high-quality plans with the minimal number of needle catheters. The optimization problem was solved by a fast-iterative shrinkage-thresholding algorithm. For validation purposes, we tested the proposed algorithm on 10 patient cases previously treated at our institution and compared the resulting plans with plans generated using needle catheters selected manually.

Results: Compared to the plan with manually selected needle catheters, when normalizing both plans to the same PTV coverage V  = 95%, the plans generated by the proposed algorithm reduced median V from 65% to 64%, but increased median V from 35% to 38%, and V from 14% to 16%. All planning objectives were met. All clinically important dosimetric parameters of OARs were reduced. D of bladder and rectum were reduced from 8.57 Gy to 8.50 Gy and from 7.24 Gy to 6.80 Gy, respectively. D of urethra was reduced from 15.85 Gy to 15.77 Gy. The median number of selected needle catheters was reduced by two. The computational time for solving the proposed optimization problem was ∼90 s using MATLAB.

Conclusion: The proposed algorithm was able to generate plans for prostate cancer HDR brachytherapy preplanning with increased median conformity index (0.73-0.77) and slightly lower median homogeneity index (0.64-0.62) with the number of selected needles reduced by two compared to the manual needle selection approach.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1088/1361-6560/abd00eDOI Listing
March 2021

On the robustness of deep learning-based lung-nodule classification for CT images with respect to image noise.

Phys Med Biol 2020 12 22;65(24):245037. Epub 2020 Dec 22.

Innovative Technology Of Radiotherapy Computations and Hardware (iTORCH) Laboratory, Department of Radiation Oncology, University of Texas Southwestern Medical Center, Dallas, TX 75235, United States of America. Medical Artificial Intelligence and Automation (MAIA) Laboratory, University of Texas Southwestern Medical Center, Dallas, TX 75235, United States of America.

Robustness is an important aspect when evaluating a method of medical image analysis. In this study, we investigated the robustness of a deep learning (DL)-based lung-nodule classification model for CT images with respect to noise perturbations. A deep neural network (DNN) was established to classify 3D CT images of lung nodules into malignant or benign groups. The established DNN was able to predict malignancy rate of lung nodules based on CT images, achieving the area under the curve of 0.91 for the testing dataset in a tenfold cross validation as compared to radiologists' prediction. We then evaluated its robustness against noise perturbations. We added to the input CT images noise signals generated randomly or via an optimization scheme using a realistic noise model based on a noise power spectrum for a given mAs level, and monitored the DNN's output. The results showed that the CT noise was able to affect the prediction results of the established DNN model. With random noise perturbations at 100 mAs, DNN's predictions for 11.2% of training data and 17.4% of testing data were successfully altered by at least once. The percentage increased to 23.4% and 34.3%, respectively, for optimization-based perturbations. We further evaluated robustness of models with different architectures, parameters, number of output labels, etc, and robustness concern was found in these models to different degrees. To improve model robustness, we empirically proposed an adaptive training scheme. It fine-tuned the DNN model by including perturbations in the training dataset that successfully altered the DNN's perturbations. The adaptive scheme was repeatedly performed to gradually improve DNN's robustness. The numbers of perturbations at 100 mAs affecting DNN's predictions were reduced to 10.8% for training and 21.1% for testing by the adaptive training scheme after two iterations. Our study illustrated that robustness may potentially be a concern for an exemplary DL-based lung-nodule classification model for CT images, indicating the needs for evaluating and ensuring model robustness when developing similar models. The proposed adaptive training scheme may be able to improve model robustness.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1088/1361-6560/abc812DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7870572PMC
December 2020

Perioperative baseline -blockers: An independent protective factor for post-carotid endarterectomy hypertension.

Vascular 2021 Apr 9;29(2):270-279. Epub 2020 Aug 9.

National Center for Cardiovascular Disease, Chinese Academy of Medical Sciences, Fuwai Hospital, Beijing, China.

Objective: Post-carotid endarterectomy hypertension is a well-recognized phenomenon closely related to surgical complications. This study aimed to determine whether different kinds of perioperative antihypertensive drugs had a protective effect on post-carotid endarterectomy hypertension and influence on intraoperative hemodynamics.

Method: We retrospectively investigated 102 carotid stenosis patients who underwent conventional endarterectomy with a perioperative baseline antihypertensive regimen. Post-carotid endarterectomy hypertension was defined as a postoperative peak systolic blood pressure ≥160 mmHg and/or a requirement for any additional antihypertensive therapies. We compared the clinical characteristics and types of baseline perioperative antihypertensive drugs between patients with and without post-carotid endarterectomy hypertension and then determined the significant independent effect of antihypertensive drugs on post-carotid endarterectomy hypertension through multivariate regression and detected their influence on intraoperative hypertension (induction-related systolic blood pressure and vasodilators consumption) and hemodynamic depression (intra-arterial systolic blood pressure ≤100 mmHg and/or heart rate ≤50 beats/min). We also investigated adverse events such as stroke, death, myocardial infarction, and cerebral hyperperfusion syndrome during the postoperative hospitalization.

Results: A total of 52/102 (51.0%) patients were defined as having post-carotid endarterectomy hypertension during the first three days postoperative, including eight patients with a postoperative systolic blood pressure that exceeded 160 mmHg at least once, 31 patients requiring postoperative antihypertensive treatment in addition to their baseline regimen, and 13 patients with both. The incidence of stroke/death/myocardial infarction and cerebral hyperperfusion syndrome after conventional endarterectomy during hospitalization were both 1.9%. A significantly increased risk of composite postoperative complications (including cerebral hyperperfusion syndrome, hyperperfusion-related symptoms, transient ischemic attacks, stroke, death, and cardiac complications) was observed in patients with post-carotid endarterectomy hypertension than without (15.4% versus 2.0%,  = 0.032). Patients free of post-carotid endarterectomy hypertension had a higher incidence of perioperative baseline β-blocker use than patients who suffered from post-carotid endarterectomy hypertension (46.0% versus 21%,  = 0.008). In multivariate analysis, β-blocker use was a significant independent protective factor for post-carotid endarterectomy hypertension (OR = 0.356, 95% CI: 0.146-0.886,  = 0.028). Patients taking β-blockers had a lower postoperative peak systolic blood pressure than the β-blocker-naïve population (137.1 ± 12.1 mmHg versus 145.0 ± 11.2 mmHg,  = 0.008), but the postoperative mean systolic blood pressure showed no intergroup difference. However, the incidence of hemodynamic depression during conventional endarterectomy was higher in patients with perioperative β-blocker use than in those without (44.1% versus 25.0%,  = 0.050). The difference in intraoperative hemodynamic depression became more prominent between the β-blocker and non-β-blocker groups (81.8% versus 33.3%,  = 0.014) for whose preoperative baseline heart rate was equal to or lower than 70 beats/min.

Conclusion: The perioperative use of β-blockers is a protective factor for post-carotid endarterectomy hypertension and contributes to stabilizing the postoperative peak systolic blood pressure three days after conventional endarterectomy. However, β-blockers might also lead to intraoperative hemodynamic depression, especially for patients with a low baseline heart rate.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1177/1708538120946538DOI Listing
April 2021

Chemokine-like factor 1 (CKLF1) aggravates neointimal hyperplasia through activating the NF-κB /VCAM-1 pathway.

FEBS Open Bio 2020 09 14;10(9):1880-1890. Epub 2020 Aug 14.

Vascular Surgery Center, State Key Laboratory of Cardiovascular Disease, National Center for Cardiovascular Diseases, Fuwai Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China.

Neointimal hyperplasia (NIH) is a complicated inflammatory process contributing to vascular restenosis. The present study aimed to explore whether chemokine-like factor 1 (CKLF1) aggravates NIH via the nuclear factor-kappa B (NF-κB)/vascular cell adhesion molecule-1 (VCAM-1) pathway. We found the expression of CKLF1 and VCAM-1 significantly increased in human carotid plaques compared to the control. In vivo, CKLF1 overexpression induced a thicker neointimal formation and VCAM-1 expression was correspondingly upregulated. In vitro, CKLF1 activated NF-κB and induced VCAM-1 upregulation in human aortic smooth muscle cells (HASMCs). Functional experiments demonstrated that CKLF1 promoted monocyte adhesion and HASMC migration via VCAM-1. These results suggest CKLF1 accelerates NIH by promoting monocyte adhesion and HASMC migration via the NF-κB/VCAM-1 pathway. Our findings contribute to a better understanding of the mechanisms underlying the causality of CKLF1 on NIH and could prove beneficial in designing therapeutic modalities with a focus on CKLF1.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1002/2211-5463.12942DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7459414PMC
September 2020

Outcomes of Total Aortoiliac Revascularization for TASC-II C&D Lesion with Kissing Self-Expanding Covered Stents.

Ann Vasc Surg 2020 Oct 16;68:434-441. Epub 2020 May 16.

Chinese Academy of Medical Sciences, Peking Union Medical College, National Center for Cardiovascular Disease, National Key Laboratory for Cardiovascular Disease, Fuwai Hospital, Vascular Surgery Center, Beijing, China.

Background: The endovascular approach has been widely used for aortoiliac occlusive disease (AIOD), especially for aortic bifurcation and iliac artery Trans-Atlantic Inter-Society Consensus II (TASC-II) A and B lesions. However, the outcomes of self-expanding covered stents (SECSs) for extensive aortoiliac lesion remain unclear. This study aimed to assess the short-term patency of kissing covered stents for the revascularization of aortoiliac TASC-II C and D diseases.

Methods: Thirty-three patients with TASC-II C and D lesions of AIOD were treated with kissing covered stents. All patients were reviewed under a standard institutional review board protocol. Demographic variables, lesion location and characteristics, stenting configuration, and patency were analyzed.

Results: Thirty-one male and 2 female patients with a mean age of 65.1 ± 10.7 years underwent aortoiliac bifurcation reconstruction with kissing SECSs. Eight patients had TASC-II C lesions, and 25 patients had TASC-II D lesions. Among them, 8 patients had total infrarenal aortoiliac occlusion, of which 5 had juxtarenal aortoiliac lesions. The mean lesion length was 11.6 ± 2.1 cm. Mean diameters of aorta and common iliac artery were 18.3 ± 2.1 and 10.7 ± 1.5 mm, respectively. Among them, the abutting stent configuration was used in 11 patients with short or focal ostial lesions, whereas the crossing stent configuration was used in 22 patients with longer lesions extending into the distal aorta. The mean follow-up was 24.5 ± 7.8 months, the follow-up rate was 93.9% (31 of 33), and 29 patients had follow-up longer than 12 months. Primary patency rate at 12 months was 96.5%, and secondary patency rate was 100%.

Conclusions: The use of kissing SECSs for the revascularization of extensive AIOD is safe and effective. The short-term primary patency rates of endovascular treatment of TASC-II C and D lesions were favorable.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.avsg.2020.04.055DOI Listing
October 2020

Discriminative Transfer Learning for Driving Pattern Recognition in Unlabeled Scenes.

IEEE Trans Cybern 2020 May 12;PP. Epub 2020 May 12.

Driving pattern recognition based on features, such as GPS, gear, and speed information, is essential to develop intelligent transportation systems. However, it is usually expensive and labor intensive to collect a large amount of labeled driving data from real-world driving scenes. The lack of a labeled data problem in a driving scene substantially hinders the driving pattern recognition accuracy. To handle the scarcity of labeled data, we have developed a novel discriminative transfer learning method for driving pattern recognition to leverage knowledge from related scenes with labeled data to improve recognition performance in unlabeled scenes. Note that data from different scenes may have different distributions, which is a major bottleneck limiting the performance of transfer learning. To address this issue, the proposed method adopts a discriminative distribution matching scheme with the aid of pseudolabels in unlabeled scenes. It is able to reduce the intraclass distribution disagreement for the same driving pattern among labeled and unlabeled scenes while increasing the interclass distance among different patterns. Pseudolabels in unlabeled scenes are updated iteratively via an ensemble strategy that preserves the data structure while enhancing the model robustness. To evaluate the performance of the proposed method, we conducted comprehensive experiments on real-world parking lot datasets. The results show that the proposed method can substantially outperform state-of-the-art methods in driving pattern recognition.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1109/TCYB.2020.2987632DOI Listing
May 2020

Spatial distribution of pollution characteristics and human health risk assessment of exposure to heavy elements in road dust from different functional areas of Zhengzhou, China.

Environ Sci Pollut Res Int 2020 Jul 6;27(21):26650-26667. Epub 2020 May 6.

College of Water Conservancy Engineering, Zhengzhou University, Zhengzhou, Henan, 450001, People's Republic of China.

Road dust from different sources directly contacts the human body and has potential effects on public health. In this study, a total number of 87 road dust samples were collected at 29 sampling sites from five different functional areas (commercial area (CA), residential area (RA), educational area (EA), industrial area (IA), and park area (PA)) in Zhengzhou to study the contamination status, distribution, source identification, ecological risk assessment, and spatial distribution of human health risks due to eight heavy elements. The geo-accumulation index (I) and pollution index (PI) revealed that there was very high contamination with Cd and Hg caused by atmospheric deposition, which should be paid special attention. Additionally, the source identification indicated that Cr, Ni, Cu, Zn, Cd, and Pb originate from anthropogenic activities related to traffic, and Hg can originate from medical equipment and agricultural chemicals, while the extremely low level of pollution with As could be explained by geographic sources. Moreover, the calculated ecological risk index values were increased in the order of CA > RA > EA > IA > PA in different functional areas. According to the human health risks of the whole city, children exposed to Pb have the highest health risk, especially for CA and IA, as calculated by the noncarcinogenic hazard index (HI). For adults and children, health risks caused by Cu, Zn, and Pb were higher in the CA, RA, and PA of the downtown area, whereas Cr and Ni had the highest noncarcinogenic exposure risk in northwestern Zhengzhou due to point source pollution. Calculations of the carcinogenic risk (CR) values for Cr, Ni, As, and Cd indicate that the value of Cr is highest (1.17 × 10), especially inside the industrial area (8.55 × 10), which is close to the lower limit of the threshold values (10 to 10). These results can provide a theoretical basis and data support for air treatment, pollution control, and the implementation of public prevention in different functional areas of Zhengzhou.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1007/s11356-020-08942-7DOI Listing
July 2020

Development of a real-time indoor location system using bluetooth low energy technology and deep learning to facilitate clinical applications.

Med Phys 2020 Aug 11;47(8):3277-3285. Epub 2020 May 11.

Medical Artificial Intelligence and Automation (MAIA) Laboratory, Department of Radiation Oncology, University of Texas Southwestern Medical Center, Dallas, TX, 75235, USA.

Purpose: An indoor, real-time location system (RTLS) can benefit both hospitals and patients by improving clinical efficiency through data-driven optimization of procedures. Bluetooth-based RTLS systems are cost-effective but lack accuracy because Bluetooth signal is subject to significant fluctuation. We aim to improve the accuracy of RTLS using the deep learning technique.

Methods: We installed a Bluetooth sensor network in a three-floor clinic building to track patients, staff, and devices. The Bluetooth sensors measured the strength of the signal broadcasted from Bluetooth tags, which was fed into a deep neural network to calculate the location of the tags. The proposed deep neural network consists of a long short-term memory (LSTM) network and a deep classifier for tracking moving objects. Additionally, a spatial-temporal constraint algorithm was implemented to further increase the accuracy and stability of the results. To train the neural network, we divided the building into 115 zones and collected training data in each zone. We further augmented the training data to generate cross-zone trajectories, mimicking the real-world scenarios. We tuned the parameters for the proposed neural network to achieve relatively good accuracy.

Results: The proposed deep neural network achieved an overall accuracy of about 97% for tracking objects in each individual zone in the whole three-floor building, 1.5% higher than the baseline neural network that was proposed in an earlier paper, when using 10 s of signals. The accuracy increased with the density of Bluetooth sensors. For tracking moving objects, the proposed neural network achieved stable and accurate results. When latency is less of a concern, we eliminated the effect of latency from the accuracy and gained an accuracy of 100% for our testing trajectories, significantly improved from the baseline method.

Conclusions: The proposed deep neural network composed of a LSTM, a deep classifier and a posterior constraint algorithm significantly improved the accuracy and stability of RTLS for tracking moving objects.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1002/mp.14198DOI Listing
August 2020

Operating a treatment planning system using a deep-reinforcement learning-based virtual treatment planner for prostate cancer intensity-modulated radiation therapy treatment planning.

Med Phys 2020 Jun 28;47(6):2329-2336. Epub 2020 Mar 28.

Medical Artificial Intelligence and Automation (MAIA) Laboratory, Department of Radiation Oncology, University of Texas Southwestern Medical Center, Dallas, TX, 75390, USA.

Purpose: In the treatment planning process of intensity-modulated radiation therapy (IMRT), a human planner operates the treatment planning system (TPS) to adjust treatment planning parameters, for example, dose volume histogram (DVH) constraints' locations and weights, to achieve a satisfactory plan for each patient. This process is usually time-consuming, and the plan quality depends on planer's experience and available planning time. In this study, we proposed to model the behaviors of human planners in treatment planning by a deep reinforcement learning (DRL)-based virtual treatment planner network (VTPN), such that it can operate the TPS in a human-like manner for treatment planning.

Methods And Materials: Using prostate cancer IMRT as an example, we established the VTPN using a deep neural network developed. We considered an in-house optimization engine with a weighted quadratic objective function. Virtual treatment planner network was designed to observe an intermediate plan DVHs and decide the action to improve the plan by changing weights and threshold dose in the objective function. We trained the VTPN in an end-to-end DRL process in 10 patient cases. A plan score was used to measure plan quality. We demonstrated the feasibility and effectiveness of the trained VTPN in another 64 patient cases.

Results: Virtual treatment planner network was trained to spontaneously learn how to adjust treatment planning parameters to generate high-quality treatment plans. In the 64 testing cases, with initialized parameters, quality score was 4.97 (±2.02), with 9.0 being the highest possible score. Using VTPN to perform treatment planning improved quality score to 8.44 (±0.48).

Conclusions: To our knowledge, this was the first time that intelligent treatment planning behaviors of human planner in external beam IMRT are autonomously encoded in an artificial intelligence system. The trained VTPN is capable of behaving in a human-like way to produce high-quality plans.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1002/mp.14114DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7903320PMC
June 2020

An introduction to deep learning in medical physics: advantages, potential, and challenges.

Phys Med Biol 2020 03 3;65(5):05TR01. Epub 2020 Mar 3.

Medical Artificial Intelligence and Automation (MAIA) Laboratory, Department of Radiation Oncology, University of Texas Southwestern Medical Center, Dallas, TX, United States of America. Innovative Technology Of Radiotherapy Computation and Hardware (iTORCH) Laboratory, Department of Radiation Oncology, University of Texas Southwestern Medical Center, Dallas, TX, United States of America.

As one of the most popular approaches in artificial intelligence, deep learning (DL) has attracted a lot of attention in the medical physics field over the past few years. The goals of this topical review article are twofold. First, we will provide an overview of the method to medical physics researchers interested in DL to help them start the endeavor. Second, we will give in-depth discussions on the DL technology to make researchers aware of its potential challenges and possible solutions. As such, we divide the article into two major parts. The first part introduces general concepts and principles of DL and summarizes major research resources, such as computational tools and databases. The second part discusses challenges faced by DL, present available methods to mitigate some of these challenges, as well as our recommendations.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1088/1361-6560/ab6f51DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7101509PMC
March 2020

Synthetic CT generation from CBCT images via deep learning.

Med Phys 2020 Mar 13;47(3):1115-1125. Epub 2020 Jan 13.

Medical Artificial Intelligence and Automation (MAIA) Lab, Department of Radiation Oncology, University of Texas Southwestern Medical Center, Dallas, TX, 75390, USA.

Purpose: Cone-beam computed tomography (CBCT) scanning is used daily or weekly (i.e., on-treatment CBCT) for accurate patient setup in image-guided radiotherapy. However, inaccuracy of CT numbers prevents CBCT from performing advanced tasks such as dose calculation and treatment planning. Motivated by the promising performance of deep learning in medical imaging, we propose a deep U-net-based approach that synthesizes CT-like images with accurate numbers from planning CT, while keeping the same anatomical structure as on-treatment CBCT.

Methods: We formulated the CT synthesis problem under a deep learning framework, where a deep U-net architecture was used to take advantage of the anatomical structure of on-treatment CBCT and image intensity information of planning CT. U-net was chosen because it exploits both global and local features in the image spatial domain, matching our task to suppress global scattering artifacts and local artifacts such as noise in CBCT. To train the synthetic CT generation U-net (sCTU-net), we include on-treatment CBCT and initial planning CT of 37 patients (30 for training, seven for validation) as the input. Additional replanning CT images acquired on the same day as CBCT after deformable registration are utilized as the corresponding reference. To demonstrate the effectiveness of the proposed sCTU-net, we use another seven independent patient cases (560 slices) for testing.

Results: We quantitatively compared the resulting synthetic CT (sCT) with the original CBCT image using deformed same-day pCT images as reference. The averaged accuracy measured by mean absolute error (MAE) between sCT and reference CT (rCT) on testing data is 18.98 HU, while MAE between CBCT and rCT is 44.38 HU.

Conclusions: The proposed sCTU-net can synthesize CT-quality images with accurate CT numbers from on-treatment CBCT and planning CT. This potentially enables advanced CBCT applications for adaptive treatment planning.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1002/mp.13978DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7067667PMC
March 2020

Incorporating human and learned domain knowledge into training deep neural networks: A differentiable dose-volume histogram and adversarial inspired framework for generating Pareto optimal dose distributions in radiation therapy.

Med Phys 2020 Mar 29;47(3):837-849. Epub 2019 Dec 29.

Medical Artificial Intelligence and Automation (MAIA) Laboratory, Department of Radiation Oncology, UT Southwestern Medical Center, Dallas, TX, 75390, USA.

Purpose: We propose a novel domain-specific loss, which is a differentiable loss function based on the dose-volume histogram (DVH), and combine it with an adversarial loss for the training of deep neural networks. In this study, we trained a neural network for generating Pareto optimal dose distributions, and evaluate the effects of the domain-specific loss on the model performance.

Methods: In this study, three loss functions - mean squared error (MSE) loss, DVH loss, and adversarial (ADV) loss - were used to train and compare four instances of the neural network model: (a) MSE, (b) MSE + ADV, (c) MSE + DVH, and (d) MSE + DVH+ADV. The data for 70 prostate patients, including the planning target volume (PTV), and the organs at risk (OAR) were acquired as 96 × 96 × 24 dimension arrays at 5 mm voxel size. The dose influence arrays were calculated for 70 prostate patients, using a 7 equidistant coplanar beam setup. Using a scalarized multicriteria optimization for intensity-modulated radiation therapy, 1200 Pareto surface plans per patient were generated by pseudo-randomizing the PTV and OAR tradeoff weights. With 70 patients, the total number of plans generated was 84 000 plans. We divided the data into 54 training, 6 validation, and 10 testing patients. Each model was trained for a total of 100,000 iterations, with a batch size of 2. All models used the Adam optimizer, with a learning rate of 1 × 10 .

Results: Training for 100 000 iterations took 1.5 days (MSE), 3.5 days (MSE+ADV), 2.3 days (MSE+DVH), and 3.8 days (MSE+DVH+ADV). After training, the prediction time of each model is 0.052 s. Quantitatively, the MSE+DVH+ADV model had the lowest prediction error of 0.038 (conformation), 0.026 (homogeneity), 0.298 (R50), 1.65% (D95), 2.14% (D98), and 2.43% (D99). The MSE model had the worst prediction error of 0.134 (conformation), 0.041 (homogeneity), 0.520 (R50), 3.91% (D95), 4.33% (D98), and 4.60% (D99). For both the mean dose PTV error and the max dose PTV, Body, Bladder and rectum error, the MSE+DVH+ADV outperformed all other models. Regardless of model, all predictions have an average mean and max dose error <2.8% and 4.2%, respectively.

Conclusion: The MSE+DVH+ADV model performed the best in these categories, illustrating the importance of both human and learned domain knowledge. Expert human domain-specific knowledge can be the largest driver in the performance improvement, and adversarial learning can be used to further capture nuanced attributes in the data. The real-time prediction capabilities allow for a physician to quickly navigate the tradeoff space for a patient, and produce a dose distribution as a tangible endpoint for the dosimetrist to use for planning. This is expected to considerably reduce the treatment planning time, allowing for clinicians to focus their efforts on the difficult and demanding cases.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1002/mp.13955DOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC7819274PMC
March 2020

A manifold learning regularization approach to enhance 3D CT image-based lung nodule classification.

Int J Comput Assist Radiol Surg 2020 Feb 25;15(2):287-295. Epub 2019 Nov 25.

Innovative Technology Of Radiotherapy Computations and Hardware (iTORCH) Laboratory, University of Texas Southwestern Medical Center, Dallas, TX, 75235, USA.

Purpose: Diagnosis of lung cancer requires radiologists to review every lung nodule in CT images. Such a process can be very time-consuming, and the accuracy is affected by many factors, such as experience of radiologists and available diagnosis time. To address this problem, we proposed to develop a deep learning-based system to automatically classify benign and malignant lung nodules.

Methods: The proposed method automatically determines benignity or malignancy given the 3D CT image patch of a lung nodule to assist diagnosis process. Motivated by the fact that real structure among data is often embedded on a low-dimensional manifold, we developed a novel manifold regularized classification deep neural network (MRC-DNN) to perform classification directly based on the manifold representation of lung nodule images. The concise manifold representation revealing important data structure is expected to benefit the classification, while the manifold regularization enforces strong, but natural constraints on network training, preventing over-fitting.

Results: The proposed method achieves accurate manifold learning with reconstruction error of ~ 30 HU on real lung nodule CT image data. In addition, the classification accuracy on testing data is 0.90 with sensitivity of 0.81 and specificity of 0.95, which outperforms state-of-the-art deep learning methods.

Conclusion: The proposed MRC-DNN facilitates an accurate manifold learning approach for lung nodule classification based on 3D CT images. More importantly, MRC-DNN suggests a new and effective idea of enforcing regularization for network training, possessing the potential impact to a board range of applications.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1007/s11548-019-02097-8DOI Listing
February 2020

Adaptive Sample-level Graph Combination for Partial Multiview Clustering.

IEEE Trans Image Process 2019 Nov 15. Epub 2019 Nov 15.

Multiview clustering explores complementary information among distinct views to enhance clustering performance under the assumption that all samples have complete information in all available views. However, this assumption does not hold in many real applications, where the information of some samples in one or more views may be missing, leading to partial multiview clustering problems. In this case, significant performance degeneration is usually observed. A collection of partial multiview clustering algorithms has been proposed to address this issue and most treat all different views equally during clustering. In fact, because different views provide features collected from different angles/feature spaces, they might play different roles in the clustering process. With the diversity of different views considered, in this study, a novel adaptive method is proposed for partial multiview clustering by automatically adjusting the contributions of different views. The samples are divided into complete and incomplete sets, while a joint learning mechanism is established to facilitate the connection between them and thereby improve clustering performance. More specifically, the method is characterized by a joint optimization model comprising two terms. The first term mines the underlying cluster structure from both complete and incomplete samples by adaptively updating their importance in all available views. The second term is designed to group all data with the aid of the cluster structure modeled in the first term. These two terms seamlessly integrate the complementary information among multiple views and enhance the performance of partial multiview clustering. Experimental results on real-world datasets illustrate the effectiveness and efficiency of our proposed method.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1109/TIP.2019.2952696DOI Listing
November 2019

A method to reconstruct intra-fractional liver motion in rotational radiotherapy using linear fiducial markers.

Phys Med Biol 2019 11 21;64(22):225013. Epub 2019 Nov 21.

Department of Physics, University of Texas Arlington, Arlington, TX 76019, United States of America.

Complex intra-fractional motion and deformation of the liver significantly impacts the accuracy of delivered dose in radiotherapy. It limits margin reduction, dose escalation and normal tissue sparing. A critical component of motion management is to accurately reconstruct tumor motion. In this study, we developed a six degrees of freedom projection marker matching method (6-DoF PM) to reconstruct translational and rotational liver tumor motion in a rotational treatment delivery, such as volumetric modulated arc therapy (VMAT). Specifically, we modeled the use of two gold markers implanted in a linear form. The four endpoints of the two gold linear markers were used as tracking surrogates. During delivery, kV x-ray projection images were acquired. A method was developed to automatically identify the 2D marker-endpoints on the projection images. 3D marker positions were determined by solving an optimization problem with the objective function penalizing the distance from the reconstructed 3D position of each fiducial marker endpoint to the corresponding straight line defined by the kV x-ray projection of the endpoints. We performed a series of tests to evaluate different components of the method. For 2D marker endpoints identification, 99.9% of the marker endpoints were identified with an error [Formula: see text] (1 pixel) along both u and v directions. For 3D reconstruction of motion in simulation studies, error of rotational angle was [Formula: see text]° without considering the 2D marker identification error. The rotational angle error was relatively sensitive to the accuracy of 2D marker identification. When the 2D error raised from 0.22 mm to 0.776 mm, the error of 3D rotational angle increased from 0.5° to 2.5°. In the experimental end-to-end tests, the mean root-mean square error of the 3D reconstructed marker positions was 0.75 mm and the mean error of rotational angle was within 1.7°. Our method can accurately determine intra-fractional liver tumor motion in rotational radiotherapy using kV projections of only two linear fiducial markers.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1088/1361-6560/ab4c0dDOI Listing
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC6986893PMC
November 2019

Deep-learning assisted automatic digitization of interstitial needles in 3D CT image based high dose-rate brachytherapy of gynecological cancer.

Phys Med Biol 2019 10 23;64(21):215003. Epub 2019 Oct 23.

Medical Artificial Intelligence and Automation (MAIA) Lab, University of Texas Southwestern Medical Center, Dallas, TX 75235, United States of America. Innovative Technology Of Radiotherapy Computation and Hardware (iTORCH) Lab, University of Texas Southwestern Medical Center, Dallas, TX 75235, United States of America. Department of Radiation Oncology, University of Texas Southwestern Medical Center, Dallas, TX 75235, United States of America.

Digitization of interstitial needles is a complicated and tedious process for the treatment planning of 3D CT image based interstitial high dose-rate brachytherapy (HDRBT) of gynecological cancer. We developed a deep-learning assisted auto-digitization method for interstitial needles. The digitization method consisted of two steps. The first step used a deep neural network with a U-net structure to segment all needles from CT images. The second step simultaneously clustered the segmented voxels into different needle groups and generated the needle central trajectories by solving an optimization problem. We evaluated the effectiveness of the developed method in ten interstitial HDRBT patient cases that were not used in the training of the U-net. Average number of needles per case was 20.7. For the segmentation step, average Dice similarity coefficient between automatic and manual segmentation was 0.93. For the digitization step, Hausdorff distance between needle trajectories determined by our method and manually by qualified medical physicists was ~0.71 mm on average and mean difference of tip positions was ~0.63 mm, which were considered acceptable for HDRBT treatment planning. It took ~5 min to complete the digitization process of an interstitial HDRBT case. The achieved accuracy and efficiency made our method clinically attractive.
View Article and Find Full Text PDF

Download full-text PDF

Source
http://dx.doi.org/10.1088/1361-6560/ab3fcbDOI Listing
October 2019
-->