Self-healing propagation of a longitudinally varying vector optical beam for distance measurement.

Based on the longitudinal manipulation of polarization, a special vector optical beam (VOB) with customized polarization variation in propagation direction can be generated, whose properties and applications remain to be studied. Here, the self-healing propagation behaviors of the longitudinally varying VOB after an opaque object are investigated, and the localized polarization responses on the object distance are revealed. On this basis, characteristic parameters are defined to measure the distance of object, achieving a minimum relative error of 0.63% in a longitudinal range of 300 mm. Besides, the correlations and uncoupling methods of object distance and size are discussed. Our studies open new ways to use the structural properties of VOB and may be instructive for laser measurement.

Tight focusing of the vector optical field with polarization varying along complex curves of the Poincaré sphere.

Different from the scalar optical field with spatially uniform polarization, the vector optical field exhibits inhomogeneous distribution of polarization on the cross section. Manipulating the variation of polarization in a single optical beam is important to acquire a flexible and controllable focused optical field. Previous studies mainly focused on the vector optical field with its polarization varying along a circular trajectory of the Poincaré sphere. Here, we demonstrate the tight focusing behaviors of the vector optical field with the polarization varying along complex curves of the Poincaré sphere, which is generated by the joint modulation of azimuthal phase and amplitude distributions of orthogonally polarized components. The longitudinal polarization component with a multipolar pattern in rotational symmetry can be achieved with similar distribution of the total focused field. The transverse and longitudinal spin angular momentum distributions in the focal space are discussed. Approximately pure transverse spin angular momentum can be constructed and manipulated in the focal space, which provides the possibility to manipulate the 3D spin flux for the applications of nano and spin photonics.

Deep learning methods for fully automated dental age estimation on orthopantomograms.

OBJECTIVES: This study aimed to use all permanent teeth as the target and establish an automated dental age estimation method across all developmental stages of permanent teeth, accomplishing all the essential steps of tooth determination, tooth development staging, and dental age assessment. METHODS: A three-step framework for automatically estimating dental age was developed for children aged 3 to 15. First, a YOLOv3 network was employed to complete the tasks of tooth localization and numbering on a digital orthopantomogram. Second, a novel network named SOS-Net was established for accurate tooth development staging based on a modified Demirjian method. Finally, the dental age assessment procedure was carried out through a single-group meta-analysis utilizing the statistical data derived from our reference dataset. RESULTS: The performance tests showed that the one-stage YOLOv3 detection network attained an overall mean average precision 50 of 97.50 for tooth determination. The proposed SOS-Net method achieved an average tooth development staging accuracy of 82.97% for a full dentition. The dental age assessment validation test yielded an MAE of 0.72 years with a full dentition (excluding the third molars) as its input. CONCLUSIONS: The proposed automated framework enhances the dental age estimation process in a fast and standard manner, enabling the reference of any accessible population. CLINICAL RELEVANCE: The tooth development staging network can facilitate the precise identification of permanent teeth with abnormal growth, improving the effectiveness and comprehensiveness of dental diagnoses using pediatric orthopantomograms.

Automatic Classification and Segmentation of Multiclass Jaw Lesions in Cone-Beam Computed Tomography using Deep Learning.

OBJECTIVES: To develop and validate a modified deep learning (DL) model based on nnU-net for classifying and segmenting five-class jaw lesions using cone-beam computed tomography (CBCT). METHODS: A total of 368 CBCT scans (37 168 slices) were used to train a multi-class segmentation model. The data underwent manual annotation by two oral and maxillofacial surgeons (OMSs) to serve as ground truth. Sensitivity, specificity, precision, F1-score, and accuracy were used to evaluate the classification ability of the model and doctors, with or without artificial intelligence assistance. The dice similarity coefficient (DSC), average symmetric surface distance (ASSD) and segmentation time were used to evaluate the segmentation effect of the model. RESULTS: The model achieved the dual task of classifying and segmenting jaw lesions in CBCT. For classification, the sensitivity, specificity, precision, and accuracy of the model were 0.871, 0.974, 0.874 and 0.891, respectively, surpassing oral and maxillofacial radiologists (OMFRs) and OMSs, approaching the specialist. With the model's assistance, the classification performance of OMFRs and OMSs improved, particularly for odontogenic keratocyst (OKC) and ameloblastoma (AM), with F1-score improvements ranging from 6.2% to 12.7%. For segmentation, the DSC was 87.2% and the ASSD was 1.359 mm. The model's average segmentation time was 40 ± 9.9 s, contrasting with 25 ± 7.2 min for OMSs. CONCLUSIONS: The proposed DL model accurately and efficiently classified and segmented five classes of jaw lesions using CBCT. In addition, it could assist doctors in improving classification accuracy and segmentation efficiency, particularly in distinguishing confusing lesions (e.g., AM and OKC).

Automatic localization of cephalometric landmarks based on convolutional neural network.

INTRODUCTION: Cephalometry plays an important role in the diagnosis and treatment of orthodontics and orthognathic surgery. This study intends to develop an automatic landmark location system to make cephalometry more convenient. METHODS: In this study, 512 lateral cephalograms were collected, and 37 landmarks were included. The coordinates of all landmarks in the 512 films were obtained to establish a labeled dataset: 312 were used as a training set, 100 as a validation set, and 100 as a testing set. An automatic landmark location system based on the convolutional neural network was developed. This system consisted of a global detection module and a locally modified module. The lateral cephalogram was first fed into the global module to obtain an initial estimate of the landmark's position, which was then adjusted with the locally modified module to improve accuracy. Mean radial error (MRE) and success detection rate (SDR) within the range of 1-4 mm were used to evaluate the method. RESULTS: The MRE of our validation set was 1.127 ± 1.028 mm, and SDR of 1.0, 1.5, 2.0, 2.5, 3.0, and 4.0 mm were respectively 45.95%, 89.19%, 97.30%, 97.30%, and 97.30%. The MRE of our testing set was 1.038 ± 0.893 mm, and SDR of 1.0, 1.5, 2.0, 2.5, 3.0, and 4.0 mm were respectively 54.05%, 91.89%, 97.30%, 100%, 100%, and 100%. CONCLUSIONS: In this study, we proposed a new automatic landmark location system on the basis of the convolutional neural network. The system could detect 37 landmarks with high accuracy. All landmarks are commonly used in clinical practice and could meet the requirements of different cephalometric analysis methods.

DeepLNAnno: a Web-Based Lung Nodules Annotating System for CT Images.

Lung cancer is one of the most common and fatal types of cancer, and pulmonary nodule detection plays a crucial role in the screening and diagnosis of this disease. A well-trained deep neural network model can help doctors to find nodules on computed tomography(CT) images while requiring lots of labeled data. However, currently available annotating systems are not suitable for annotating pulmonary nodules in CT images. We propose a web-based lung nodules annotating system named as DeepLNAnno. DeepLNAnno has a unique three-tier working process and loads of features like semi-automatic annotation, which not only make it much easier for doctors to annotate compared to some other annotating systems but also increase the accuracy of the labels. We invited a medical group from West China Hospital to annotate the CT images using our DeepLNAnno system, and collected a large number of labeled data. The results of our experiments demonstrated that a usable nodule-detection system is developed, and good benchmark scores on our evaluation data are obtained.

Three-Dimensional Planning in Maxillofacial Fracture Surgery: Computer-Aided Design/Computer-Aided Manufacture Surgical Splints by Integrating Cone Beam Computerized Tomography Images Into Multislice Computerized Tomography Images.

This study aimed to evaluate an innovative workflow for maxillofacial fracture surgery planning and surgical splint designing. The maxillofacial multislice computerized tomography (MSCT) data and dental cone beam computerized tomography (CBCT) data both were obtained from 40 normal adults and 58 adults who suffered fractures. The each part of the CBCT dentition image was registered into MSCT image by the use of the iterative closest point algorithm. Volume evaluation of the virtual splints that were designed by the registered MSCT images and MSCT images of the same object was performed. Eighteen patients (group 1) were operated without any splint. Twenty-one (group 2) and 19 patients (group 3) used the splints designed according to the MSCT images and registered MSCT images, respectively. The authors' results showed that the mean errors between the 2 models ranged from 0.53 to 0.92 mm and the RMS errors ranged from 0.38 to 0.69 mm in fracture patients. The mean errors between the 2 models ranged from 0.47 to 0.85 mm and the RMS errors ranged from 0.33 to 0.71 mm in normal adults. 72.22% patients in group 1 recovered occlusion. 85.71% patients in group 2, and 94.73% patients in group 3 reconstructed occlusion. There was a statistically significant difference between the MSCT images based splints' volume and the registered MSCT splints' volume in patients (P <0.05). The MSCT images based splints' volume was statistically significantly distinct from the registered MSCT splints' volume in normal adults (P <0.05). There was a statistically significant difference between the MSCT images based splints' volume and the registered MSCT splints' volume in patients and normal adults (P <0.05). The occlusion recovery rate of group 3 was better than that of group 1 and group 2. The way of integrating CBCT images into MSCT images for splints designing was feasible. The volume of the splints designed by MSCT images tended to be smaller than the splints designed by the integrated MSCT images. The patients operated with splints tended to regain occlusion. The patients who were operated with the splints which were designed according to registered MSCT images tended to get occlusal recovered.

[Improving Threshold Segmentation in 3D Reconstruction of Mandible CT Image].

OBJECTIVE: To develop a new threshold segmentation method for mandible image segmentation. METHODS: CT data of 12 volunteers were exported into Mimics 10. 01. An improved method usinga narrowed threshold range (the maximum threshold range that can segment mandible without manual efforts) was developed in 3D reconstruction, and compared with the traditional method. We used dilation operations to make up the information loss of image borders, by which we obtained an approxinate segment result. A precise segment resultwas eventually arrived with the help of logical operations and region growing. We compared mean time consumptions of the two methods, as well as their 3D reconstruction results using Geomagic Studio 11. 0. RESULTS: The new method generated a success rate of 91. 67% (11/12), with a mean time consumption of (319. 7±125. 3) s. The traditional method took much longer time [(1,261. 3±427. 3) s, P<0. 05] than the new method. Compared with the reconstruction results of traditional method, the new method had an outward deviation of (0. 066±0. 011) mm and an inward deviation of (0. 070±0. 008) mm. Such deviations were less than the minimum distance that a naked eye can discern. The lower limit of the widest threshold range which mandible could be isolated was (507. 72± 100. 31) HU, while the upper limit was (1,133. 33±47. 57) HU. CONCLUSION: The new method we proposed can improve the efficiency of threshold segmentation of mandible.

Role of EZH2 protein expression in gastric carcinogenesis among Asians: a meta-analysis.

The present meta-analysis aggregated the results of relevant studies to identify the role of zeste homolog 2 (EZH2) expression in gastric carcinogenesis among Asians. Related articles were found by searching the following electronic databases without language restrictions: PubMed, SpringerLink, Karger Medical and Scientific Publishers, Chinese Biomedical Database (CBM), Chinese National Knowledge Infrastructure (CNKI), and Google Scholar. Meta-analysis was performed using STATA statistical software. Crude odds ratios (ORs) or hazard ratios (HRs) with their corresponding 95 % confidence interval (95 % CI) were calculated. Ten relevant studies, which enrolled a total of 872 gastric cancer patients, were selected for statistical analysis. The most important findings of our meta-analysis was that cancer tissues exhibited higher expression levels of EZH2 protein than normal, adjacent and benign tissues (cancer tissues vs normal tissues: OR = 32.15, 95 % CI 22.58 ~ 45.79, P < 0.001; cancer tissues vs adjacent tissues: OR = 16.10, 95 % CI 11.35 ~ 22.84, P < 0.001; cancer tissues vs benign tissues: OR = 2.66, 95 % CI 1.89 ~ 3.75, P < 0.001; respectively). Furthermore, we observed positive correlations between EZH2 expression and the TNM stage (OR = 2.86, 95 % CI 1.72 ~ 4.75, P < 0.001) as well as lymph node metastasis (OR = 3.02, 95 % CI 2.01 ~ 4.53, P < 0.001) of patients with gastric carcinoma. The correlation between EZH2 expression and gastric cancer prognosis was also evaluated in the meta-analysis. Statistical analysis demonstrated that the overall survival (OS) of EZH2-negative patients was shorter than that of patients with positive expressions of EZH2 (HR = 0.54, 95 % CI = 0.05 ~ 1.03, P = 0.032). Our meta-analysis confirmed the view that EZH2 expression might participate in the development of gastric carcinogenesis. Thus, EZH2 protein may be a valuable biomarker for the diagnosis and prognosis of gastric cancer.

A Stage-Wise Residual Attention Generation Adversarial Network for Mandibular Defect Repairing and Reconstruction.

Surgical reconstruction of mandibular defects is a clinical routine manner for the rehabilitation of patients with deformities. The mandible plays a crucial role in maintaining the facial contour and ensuring the speech and mastication functions. The repairing and reconstruction of mandible defects is a significant yet challenging task in oral-maxillofacial surgery. Currently, the mainly available methods are traditional digitalized design methods that suffer from substantial artificial operations, limited applicability and high reconstruction error rates. An automated, precise, and individualized method is imperative for maxillofacial surgeons. In this paper, we propose a Stage-wise Residual Attention Generative Adversarial Network (SRA-GAN) for mandibular defect reconstruction. Specifically, we design a stage-wise residual attention mechanism for generator to enhance the extraction capability of mandibular remote spatial information, making it adaptable to various defects. For the discriminator, we propose a multi-field perceptual network, consisting of two parallel discriminators with different perceptual fields, to reduce the cumulative reconstruction errors. Furthermore, we design a self-encoder perceptual loss function to ensure the correctness of mandibular anatomical structures. The experimental results on a novel custom-built mandibular defect dataset demonstrate that our method has a promising prospect in clinical application, achieving the best Dice Similarity Coefficient (DSC) of 94.238% and 95% Hausdorff Distance (HD95) of 4.787.

Deep learning for precise diagnosis and subtype triage of drug-resistant tuberculosis on chest computed tomography.

Deep learning, transforming input data into target prediction through intricate network structures, has inspired novel exploration in automated diagnosis based on medical images. The distinct morphological characteristics of chest abnormalities between drug-resistant tuberculosis (DR-TB) and drug-sensitive tuberculosis (DS-TB) on chest computed tomography (CT) are of potential value in differential diagnosis, which is challenging in the clinic. Hence, based on 1176 chest CT volumes from the equal number of patients with tuberculosis (TB), we presented a Deep learning-based system for TB drug resistance identification and subtype classification (DeepTB), which could automatically diagnose DR-TB and classify crucial subtypes, including rifampicin-resistant tuberculosis, multidrug-resistant tuberculosis, and extensively drug-resistant tuberculosis. Moreover, chest lesions were manually annotated to endow the model with robust power to assist radiologists in image interpretation and the Circos revealed the relationship between chest abnormalities and specific types of DR-TB. Finally, DeepTB achieved an area under the curve (AUC) up to 0.930 for thoracic abnormality detection and 0.943 for DR-TB diagnosis. Notably, the system demonstrated instructive value in DR-TB subtype classification with AUCs ranging from 0.880 to 0.928. Meanwhile, class activation maps were generated to express a human-understandable visual concept. Together, showing a prominent performance, DeepTB would be impactful in clinical decision-making for DR-TB.

LCST-UCST Transition Property of a Novel Retarding Swelling and Thermosensitive Particle Gel.

Super absorbent resin particles used as profile control and water plugging agent remains a deficiency that the particles swells with high speed when absorbing water, resulting in low strength and limited depth of migration. To address this issue, we proposed a thermosensitive particle gel possessing the upper critical solution temperature (UCST), which was synthesized from hydrophobically modified poly(vinyl alcohol)s (PVA) with glutaraldehyde (GA) as a cross-linker. The structure of the hydrogel was characterized by Fourier transform infrared spectrophotometer (FTIR) and nuclear magnetic resonance (NMR). The thermosensitive-transparency measurement and swelling experiment show that the hydrophobic-modified PVA solutions and corresponding hydrogels exhibited thermosensitive phase transition behaviors with lower critical solution temperature (LCST) and UCST. The results indicated that the temperature-induced phase transition behavior of CHPVA hydrogels leads to their retarding swelling property and great potential as an efficient water plugging agent with excellent temperature and salt resistance.

Multi-Label Softmax Networks for Pulmonary Nodule Classification Using Unbalanced and Dependent Categories.

Radiographic attributes of lung nodules remedy the shortcomings of lung cancer computer-assisted diagnosis systems, which provides interpretable diagnostic reference for doctors. However, current studies fail to dedicate multi-label classification of lung nodules using convolutional neural networks (CNNs) and are inferior in exploiting statistical dependency between the labels. In addition, data imbalance is an indispensable problem to be reckoned with when employing CNNs to perform lung nodule classification. It introduces greater challenges especially in the multi-label classification. In this paper, we propose a method called MLSL-Net to discriminate lung nodule characteristics and simultaneously address the challenges. Particularly, the proposal employs multi-label softmax loss (MLSL) as the performance index, aiming to reduce the ranking errors between the labels and within the labels during training, thereby optimizing ranking loss and AUC directly. Such criterions can better evaluate the classifier's performance on the multi-label imbalanced dataset. Furthermore, a scale factor is introduced based on the investigation of the max surrogate function. Different from preceding usages, the small factor is used so that to narrow the discrepancy of gradients produced by different labels. More interestingly, this factor also facilitates the exploit of label dependency. Experimental results on the LIDC-IDRI dataset as well as another akin dataset demonstrate that MLSL-Net can effectively perform multi-label classification despite the imbalance issue. Meanwhile, the results confirm the responsibility of the factor for capturing label correlations, accordingly leading to more accurate predictions.

Reinforcement of Nanocomposite Hydrogel with Dialdehyde Cellulose Nanofibrils via Physical and Double Network Crosslinking Synergies.

Preparation of tough and high-strength hydrogels for water plugging in oil fields with an easy-scalable method is still considered to be a challenge. In this study, dialdehyde cellulose nanofibril (DA-CNF) prepared by sodium periodate oxidation, polyamine, 2-acrylamido-2-methylpropane sulfonic acid (AMPS) with sulfonate groups and Acrylamide (AM) as raw materials, CNF reinforced nanocomposite hydrogels were prepared in one step by in-situ polymerization. The tensile strength, and texture stability of the obtained nanocomposite hydrogel were determined. The results showed that the tensile strength and toughness of the obtained nanocomposite hydrogel increased four times compared with control sample due to physical and chemical double crosslinking synergies. Moreover, the texture intensity of DA-CNFs reinforced hydrogel still maintains high stability and strength performance under high salinity conditions. Therefore, DA-CNF reinforced hydrogel has potential application value in both normal and high-salinity environments in oil recovery.

Preparation and Performance Evaluation of Amphiphilic Polymers for Enhanced Heavy Oil Recovery.

The continuous growth in global energy and chemical raw material demand has drawn significant attention to the development of heavy oil resources. A primary challenge in heavy oil extraction lies in reducing crude oil viscosity. Alkali-surfactant-polymer (ASP) flooding technology has emerged as an effective method for enhancing heavy oil recovery. However, the chromatographic separation of chemical agents presents a formidable obstacle in heavy oil extraction. To address this challenge, we utilized a free radical polymerization method, employing acrylamide, 2-acrylamido-2-methylpropane sulfonic acid, lauryl acrylate, and benzyl acrylate as raw materials. This approach led to the synthesis of a multifunctional amphiphilic polymer known as PAALB, which we applied to the extraction of heavy oil. The structure of PAALB was meticulously characterized using techniques such as infrared spectroscopy and Nuclear Magnetic Resonance Spectroscopy. To assess the effectiveness of PAALB in reducing heavy oil viscosity and enhancing oil recovery, we conducted a series of tests, including contact angle measurements, interfacial tension assessments, self-emulsification experiments, critical association concentration tests, and sand-packed tube flooding experiments. The research findings indicate that PAALB can reduce oil-water displacement, reduce heavy oil viscosity, and improve swept volume upon injection into the formation. A solution of 5000 mg/L PAALB reduced the contact angle of water droplets on the core surface from 106.55° to 34.95°, shifting the core surface from oil-wet to water-wet, thereby enabling oil-water displacement. Moreover, A solution of 10,000 mg/L PAALB reduced the oil-water interfacial tension to 3.32 × 10-4 mN/m, reaching an ultra-low interfacial tension level, thereby inducing spontaneous emulsification of heavy oil within the formation. Under the condition of an oil-water ratio of 7:3, a solution of 10,000 mg/L PAALB can reduce the viscosity of heavy oil from 14,315 mPa·s to 201 mPa·s via the glass bottle inversion method, with a viscosity reduction rate of 98.60%. In sand-packed tube flooding experiments, under the injection volume of 1.5 PV, PAALB increased the recovery rate by 25.63% compared to traditional hydrolyzed polyacrylamide (HPAM) polymer. The insights derived from this research on amphiphilic polymers hold significant reference value for the development and optimization of chemical flooding strategies aimed at enhancing heavy oil recovery.

A Nano-Cleaning Fluid for Downhole Casing Cleaning.

In drilling and completion projects, sludge is formed as a byproduct when barite and oil are mixed, and later sticks to the casing. This phenomenon has caused a delay in drilling progress, and increased exploration and development costs. Since nano-emulsions have low interfacial surface tension, wetting, and reversal capabilities, this study used nano-emulsions with a particle size of about 14 nm to prepare a cleaning fluid system. This system enhances stability through the network structure in the fiber-reinforced system, and prepares a set of nano-cleaning fluids with adjustable density for ultra-deep wells. The effective viscosity of the nano-cleaning fluid reaches 11 mPa·s, and the system is stable for up to 8 h. In addition, this research independently developed an indoor evaluation instrument. Based on on-site parameters, the performance of the nano-cleaning fluid was evaluated from multiple angles by heating to 150 °C and pressurizing to 3.0 Mpa to simulate downhole temperature and pressure. The evaluation results show that the viscosity and shear value of the nano-cleaning fluid system is greatly affected by the fiber content, and the cleaning efficiency is greatly affected by the concentration of the nano-emulsion. Curve fitting shows that the average processing efficiency could reach 60-85% within 25 min and the cleaning efficiency has a linear relationship with time. The cleaning efficiency has a linear relationship with time, where R2 = 0.98335. The nano-cleaning fluid enables the deconstruction and carrying of the sludge attached to the well wall, which accomplishes the purpose of downhole cleaning.

The Gap in the Thickness: Estimating Effectiveness of Pulmonary Nodule Detection in Thick- and Thin-Section CT Images with 3D Deep Neural Networks.

BACKGROUND AND OBJECTIVES: There is a noticeable gap in diagnostic evidence strength between the thick and thin scans of Low-Dose CT (LDCT) for pulmonary nodule detection. When the thin scans are needed is unknown, especially when aided with an artificial intelligence nodule detection system. METHODS: A case study is conducted with a set of 1,000 pulmonary nodule screening LDCT scans with both thick (5.0mm), and thin (1.0mm) section scans available. Pulmonary nodule detection is performed by human and artificial intelligence models for nodule detection developed using 3D convolutional neural networks (CNNs). The intra-sample consistency is evaluated with thick and thin scans, for both clinical doctor and NN (neural network) models. Free receiver operating characteristic (FROC) is used to measure the accuracy of humans and NNs. RESULTS: Trained NNs outperform humans with small nodules < 6.0mm, which is a good complement to human ability. For nodules > 6.0mm, human and NNs perform similarly while human takes a fractional advantage. By allowing a few more FPs, a significant sensitivity improvement can be achieved with NNs. CONCLUSIONS: There is a performance gap between the thick and thin scans for pulmonary nodule detection regarding both false negatives and false positives. NNs can help reduce false negatives when the nodules are small and trade off the false negatives for sensitivity. A combination of human and trained NNs is a promising way to achieve a fast and accurate diagnosis.

DeepLN: A Multi-Task AI Tool to Predict the Imaging Characteristics, Malignancy and Pathological Subtypes in CT-Detected Pulmonary Nodules.

Objectives: Distinction of malignant pulmonary nodules from the benign ones based on computed tomography (CT) images can be time-consuming but significant in routine clinical management. The advent of artificial intelligence (AI) has provided an opportunity to improve the accuracy of cancer risk prediction. Methods: A total of 8950 detected pulmonary nodules with complete pathological results were retrospectively enrolled. The different radiological manifestations were identified mainly as various nodules densities and morphological features. Then, these nodules were classified into benign and malignant groups, both of which were subdivided into finer specific pathological types. Here, we proposed a deep convolutional neural network for the assessment of lung nodules named DeepLN to identify the radiological features and predict the pathologic subtypes of pulmonary nodules. Results: In terms of density, the area under the receiver operating characteristic curves (AUCs) of DeepLN were 0.9707 (95% confidence interval, CI: 0.9645-0.9765), 0.7789 (95%CI: 0.7569-0.7995), and 0.8950 (95%CI: 0.8822-0.9088) for the pure-ground glass opacity (pGGO), mixed-ground glass opacity (mGGO) and solid nodules. As for the morphological features, the AUCs were 0.8347 (95%CI: 0.8193-0.8499) and 0.9074 (95%CI: 0.8834-0.9314) for spiculation and lung cavity respectively. For the identification of malignant nodules, our DeepLN algorithm achieved an AUC of 0.8503 (95%CI: 0.8319-0.8681) in the test set. Pertaining to predicting the pathological subtypes in the test set, the multi-task AUCs were 0.8841 (95%CI: 0.8567-0.9083) for benign tumors, 0.8265 (95%CI: 0.8004-0.8499) for inflammation, and 0.8022 (95%CI: 0.7616-0.8445) for other benign ones, while AUCs were 0.8675 (95%CI: 0.8525-0.8813) for lung adenocarcinoma (LUAD), 0.8792 (95%CI: 0.8640-0.8950) for squamous cell carcinoma (LUSC), 0.7404 (95%CI: 0.7031-0.7782) for other malignant ones respectively in the malignant group. Conclusions: The DeepLN based on deep learning algorithm represented a competitive performance to predict the imaging characteristics, malignancy and pathologic subtypes on the basis of non-invasive CT images, and thus had great possibility to be utilized in the routine clinical workflow.

Virtual reconstruction of midfacial bone defect based on generative adversarial network.

BACKGROUND: The study aims to evaluate the accuracy of the generative adversarial networks (GAN) for reconstructing bony midfacial defects. METHODS: According to anatomy, the bony midface was divided into five subunit structural regions and artificial defects are manually created on the corresponding CT images. GAN is trained to reconstruct artificial defects to their previous normal shape and tested. The clinical defects are reconstructed by the trained GAN, where the midspan defects were used for qualitative evaluation and the unilateral defects were used for quantitative evaluation. The cosine similarity and the mean error are used to evaluate the accuracy of reconstruction. The Mann-Whitney U test is used to detect whether reconstruction errors were consistent in artificial and unilateral clinical defects. RESULTS: This study included 518 normal CT data, with 415 in training set and 103 in testing set, and 17 real patient data, with 2 midspan defects and 15 unilateral defects. Reconstruction of midspan clinical defects assessed by experts is acceptable. The cosine similarity in the reconstruction of artificial defects and unilateral clinical defects is 0.97 ± 0.01 and 0.96 ± 0.01, P = 0.695. The mean error in the reconstruction of artificial defects and unilateral clinical defects is 0.59 ± 0.31 mm and 0.48 ± 0.08 mm, P = 0.09. CONCLUSION: GAN-based virtual reconstruction technology has reached a high accuracy in testing set, and statistical tests suggest that it can achieve similar results in real patient data. This study has preliminarily solved the problem of bony midfacial defect without reference.

Deep Learning Empowers Lung Cancer Screening Based on Mobile Low-Dose Computed Tomography in Resource-Constrained Sites.

BACKGROUND: Existing challenges of lung cancer screening included non-accessibility of computed tomography (CT) scanners and inter-reader variability, especially in resource-limited areas. The combination of mobile CT and deep learning technique has inspired innovations in the routine clinical practice. METHODS: This study recruited participants prospectively in two rural sites of western China. A deep learning system was developed to assist clinicians to identify the nodules and evaluate the malignancy with state-of-the-art performance assessed by recall, free-response receiver operating characteristic curve (FROC), accuracy (ACC), area under the receiver operating characteristic curve (AUC). RESULTS: This study enrolled 12,360 participants scanned by mobile CT vehicle, and detected 9511 (76.95%) patients with pulmonary nodules. Majority of participants were female (8169, 66.09%), and never-smokers (9784, 79.16%). After 1-year follow-up, 86 patients were diagnosed with lung cancer, with 80 (93.03%) of adenocarcinoma, and 73 (84.88%) at stage I. This deep learning system was developed to detect nodules (recall of 0.9507; FROC of 0.6470) and stratify the risk (ACC of 0.8696; macro-AUC of 0.8516) automatically. CONCLUSIONS: A novel model for lung cancer screening, the integration mobile CT with deep learning, was proposed. It enabled specialists to increase the accuracy and consistency of workflow and has potential to assist clinicians in detecting early-stage lung cancer effectively.