Search

Objectives An accurate and rapid diagnosis is crucial for the appropriate treatment of pulmonary tuberculosis (TB). This study aims to develop an artificial intelligence (AI)–based fully automated CT image analysis system for detection, diagnosis, and burden quantification of pulmonary TB. Methods From December 2007 to September 2020, 892 chest CT scans from pathogen-confirmed TB patients were retrospectively included. A deep learning–based cascading framework was connected to create a processing pipeline. For training and validation of the model, 1921 lesions were manually labeled, classified according to six categories of critical imaging features, and visually scored regarding lesion involvement as the ground truth. A “TB score” was calculated based on a network-activation map to quantitively assess the disease burden. Independent testing datasets from two additional hospitals (dataset 2, n = 99; dataset 3, n = 86) and the NIH TB Portals (n = 171) were used to externally validate the performance of the AI model. Results CT scans of 526 participants (mean age, 48.5 ± 16.5 years; 206 women) were analyzed. The lung lesion detection subsystem yielded a mean average precision of the validation cohort of 0.68. The overall classification accuracy of six pulmonary critical imaging findings indicative of TB of the independent datasets was 81.08–91.05%. A moderate to strong correlation was demonstrated between the AI model–quantified TB score and the radiologist-estimated CT score. Conclusions The proposed end-to-end AI system based on chest CT can achieve human-level diagnostic performance for early detection and optimal clinical management of patients with pulmonary TB. Key Points • Deep learning allows automatic detection, diagnosis, and evaluation of pulmonary tuberculosis. • Artificial intelligence helps clinicians to assess patients with tuberculosis. • Pulmonary tuberculosis disease activity and treatment management can be improved. Supplementary Information The online version contains supplementary material available at 10.1007/s00330-021-08365-z.

Accurate diagnosis and precise and effective treatment are currently the two magic weapons for dealing with cancer. However, a single marker is often associated with multiple cellular events, which is not conducive to accurate diagnosis, and overly mild treatment methods often make the treatment effect unsatisfactory. In this paper, we construct a Au/Pd octopus nanoparticle-DNA nanomachine (Au/Pd ONP-DNA nanomachine) as a fully automatic diagnosis and treatment logic system. In this system, multiple DNA components are targeting detection units, Au/Pd ONPs act as carriers, and Au/Pd ONPs with an 808 nm laser is the treatment unit. In order to achieve the purpose of precise treatment, we will detect two secondary markers under the premise of detecting one major tumor marker. When all of the designated targets are detected (the logic system input is (1, 1, 1), and the output is (1, 1)), the 808 nm laser can be programmed to automatically radiate tumors and perform photothermal therapy and photodynamic therapy. In vivo and in vitro experiments show that this logic system not only can accurately identify tumor cells but also has considerable therapeutic effects.

Objective To assess and validate the performance of a new tool developed for segmenting and characterizing lacunas in postoperative MR images of epilepsy patients. Methods A MATLAB‐based pipeline was implemented using SPM12 to produce the 3D mask of the surgical lacuna and estimate its volume. To validate its performance, we compared the manual and automatic lacuna segmentations obtained from 51 MRI scans of epilepsy patients who underwent temporal lobe resections. Results The code is consolidated as a tool named ResectVol, which can be run via a graphical user interface or command line. The automatic and manual segmentation comparison resulted in a median Dice similarity coefficient of 0.77 (interquartile range: 0.71‐0.81). Significance Epilepsy surgery is the treatment of choice for pharmacoresistant focal epilepsies, and despite the extensive literature on the subject, we still cannot predict surgical outcomes accurately. As the volume and location of the resected tissue are fundamentally relevant to this prediction, researchers commonly perform a manual segmentation of the lacuna, which presents human bias and does not provide detailed information about the structures removed. In this study, we introduce ResectVol, a user‐friendly, fully automatic tool to accomplish these tasks. This capability enables more advanced analytical techniques applied to surgical outcomes prediction, such as machine‐learning algorithms, by facilitating coregistration of the resected area and preoperative findings with other imaging modalities such as PET, SPECT, and functional MRI ResectVol is freely available at https://www.lniunicamp.com/resectvol.

Objective This study aimed to validate a deep learning-based fully automatic calcium scoring (coronary artery calcium [CAC]_auto) system using previously published cardiac computed tomography (CT) cohort data with the manually segmented coronary calcium scoring (CAC_hand) system as the reference standard. Materials and methods We developed the CAC_auto system using 100 co-registered, non-enhanced and contrast-enhanced CT scans. For the validation of the CAC_auto system, three previously published CT cohorts (n = 2985) were chosen to represent different clinical scenarios (i.e., 2647 asymptomatic, 220 symptomatic, 118 valve disease) and four CT models. The performance of the CAC_auto system in detecting coronary calcium was determined. The reliability of the system in measuring the Agatston score as compared with CAC_hand was also evaluated per vessel and per patient using intraclass correlation coefficients (ICCs) and Bland-Altman analysis. The agreement between CAC_auto and CAC_hand based on the cardiovascular risk stratification categories (Agatston score: 0, 1-10, 11-100, 101-400, > 400) was evaluated. Results In 2985 patients, 6218 coronary calcium lesions were identified using CAC_hand. The per-lesion sensitivity and false-positive rate of the CAC_auto system in detecting coronary calcium were 93.3% (5800 of 6218) and 0.11 false-positive lesions per patient, respectively. The CAC_auto system, in measuring the Agatston score, yielded ICCs of 0.99 for all the vessels (left main 0.91, left anterior descending 0.99, left circumflex 0.96, right coronary 0.99). The limits of agreement between CAC_auto and CAC_hand were 1.6 ± 52.2. The linearly weighted kappa value for the Agatston score categorization was 0.94. The main causes of false-positive results were image noise (29.1%, 97/333 lesions), aortic wall calcification (25.5%, 85/333 lesions), and pericardial calcification (24.3%, 81/333 lesions). Conclusion The atlas-based CAC_auto empowered by deep learning provided accurate calcium score measurement as compared with manual method and risk category classification, which could potentially streamline CAC imaging workflows.

Purpose: Registration of Laparoscopic Ultrasound (LUS) to a pre-operative scan such as Computed Tomography (CT) using blood vessel information has been proposed as a method to enable image-guidance for laparoscopic liver resection. Currently, there are solutions for this problem that can potentially enable clinical translation by bypassing the need for a manual initialisation and tracking information. However, no reliable framework for the segmentation of vessels in 2D untracked LUS images has been presented. Methods: We propose the use of 2D UNet for the segmentation of liver vessels in 2D LUS images. We integrate these results in a previously developed registration method, and show the feasibility of a fully automatic initialisation to the LUS to CT registration problem without a tracking device. Results: We validate our segmentation using LUS data from 6 patients. We test multiple models by placing patient datasets into different combinations of training, testing and hold-out, and obtain mean Dice scores ranging from 0.543 to 0.706. Using these segmentations, we obtain registration accuracies between 6.3 and 16.6 mm in 50% of cases. Conclusions: We demonstrate the first instance of deep learning (DL) for the segmentation of liver vessels in LUS. Our results show the feasibility of UNet in detecting multiple vessel instances in 2D LUS images, and potentially automating a LUS to CT registration pipeline.

PURPOSE Cardiac relaxometry techniques, particularly T1 mapping, have recently gained clinical importance in various cardiac pathologies. Myocardial T1 and extracellular volume are usually calculated from manual identification of left ventricular epicardial and endocardial regions. This is a laborious process, particularly for large volume studies. Here we present a fully automated relaxometry framework (FASTR) for segmental analysis of T1 maps (both native and postcontrast) and partition coefficient (λ). METHODS Patients (N = 11) were imaged postacute myocardial infarction on a 1.5T clinical scanner. The scan protocol involved CINE-SSFP imaging, native, and post-contrast T1 mapping using the Modified Look-Locker Inversion (MOLLI) recovery sequence. FASTR consisted of automatic myocardial segmentation of spatio-temporally coregistered CINE images as an initial guess, followed by refinement of the contours on the T1 maps to derive segmental T1 and λ. T1 and λ were then compared to those obtained from two trained expert observers. RESULTS Robust endocardial and epicardial contours were achieved on T1 maps despite the presence of infarcted tissue. Relative to experts, FASTR resulted in myocardial Dice coefficients (native T1: 0.752 ± 0.041; postcontrast T1: 0.751 ± 0.057) that were comparable to interobserver Dice (native T1: 0.803 ± 0.045; postcontrast T1: 0.799 ± 0.054). There were strong correlations observed for T1 and λ derived from experts and FASTR (native T1: r = 0.83; postcontrast T1: r = 0.87; λ: r = 0.78; P

For three-weekly hemodialysis, a single-pool Kt/V target of at least 1.4 together with a minimal dialysis dose Kt at 45 L for men and 40 L for women per each session is currently recommended. Fully automatic online calculation of Kt and Kt/V from conductivity or UV-absorbance measurements in the dialysate is standardly implemented on some hemodialysis monitors and makes it possible to estimate the dialysis dose without the need for blood or dialysate samples. Monitoring the UV-absorbance of the spent dialysate is the most direct method for estimating Kt/V as it does not require an estimate of V. Calculation of ionic dialysance from conductivity measurements is the most direct method for estimating Kt and BSA-scaled dialysis dose. Both ionic dialysance monitoring and UV-absorbance monitoring may help detect a change in urea clearance occurring during the session, but this change must be interpreted differently depending on the monitoring being considered. An abrupt decrease in urea clearance results in a decrease in ionic dialysance but, paradoxically, a sudden increase in estimated urea clearance provided by dialysate UV-absorbance monitoring. Healthcare teams who monitor both ionic dialysance and UV-absorbance in their hemodialysis units must be clearly informed of this difficulty.

Our objective is the calibration of mathematical tumor growth models from a single multiparametric scan. The target problem is the analysis of preoperative Glioblastoma (GBM) scans. To this end, we present a fully automatic tumor-growth calibration methodology that integrates a single-species reaction-diffusion partial differential equation (PDE) model for tumor progression with multiparametric Magnetic Resonance Imaging (mpMRI) scans to robustly extract patient specific biomarkers i.e., estimates for ( i ) the tumor cell proliferation rate , ( ii ) the tumor cell migration rate , and ( iii ) the original, localized site(s) of tumor initiation . Our method is based on a sparse reconstruction algorithm for the tumor initial location (TIL). This problem is particularly challenging due to nonlinearity, ill-posedeness, and ill conditioning. We propose a coarse-to-fine multi-resolution continuation scheme with parameter decomposition to stabilize the inversion. We demonstrate robustness and practicality of our method by applying the proposed method to clinical data of 206 GBM patients. We analyze the extracted biomarkers and relate tumor origin with patient overall survival by mapping the former into a common atlas space. We present preliminary results that suggest improved accuracy for prediction of patient overall survival when a set of imaging features is augmented with estimated biophysical parameters. All extracted features, tumor initial positions, and biophysical growth parameters are made publicly available for further analysis. To our knowledge, this is the first fully automatic scheme that can handle multifocal tumors and can localize the TIL to a few millimeters.

Since microbial detection is an important aspect for the prevention and control of foodborne diseases, an ideal detection system with high sensitivity, strong specificity, and timeliness is needed. Here, we proposed a fluorescent and opt-electric recording bacterial identification device (FORBID) for fully automatic real-time photoelectric sensing analysis of microbials by integrating the metabolic characteristics of microbial and selective substrate catalysis. It simplifies the testing process (one-step) and decreases the need of professional technicians. Besides, the system exhibits ultrasensitive (1 CFU/mL) and specific detection (99%) in both microbials, Escherichia coli and Pseudomonas aeruginosa. More importantly, the timeliness of this system is even better than that of the traditional culture methods. It is believed that this system can be extended to the detection of other microorganisms and provide a potential alternative for the detection of pathogens.

Background Microalgae are emerging as promising sustainable sources for biofuels, biostimulants in agriculture, soil bioremediation, feed and human nutrients. Nonetheless, the molecular mechanisms underpinning microalgae physiology and the biosynthesis of compounds of biotechnological interest are largely uncharacterized. This hinders the development of microalgae full potential as cell-factories. The recent application of omics technologies into microalgae research aims at unraveling these systems. Nevertheless, the lack of specific tools for analysing omics raw data generated from microalgae to provide biological meaningful information are hampering the impact of these technologies. The purpose of ALGAEFUN with MARACAS consists in providing researchers in microalgae with an enabling tool that will allow them to exploit transcriptomic and cistromic high-throughput sequencing data. Results ALGAEFUN with MARACAS consists of two different tools. First, MARACAS (MicroAlgae RnA-seq and Chip-seq AnalysiS) implements a fully automatic computational pipeline receiving as input RNA-seq (RNA sequencing) or ChIP-seq (chromatin immunoprecipitation sequencing) raw data from microalgae studies. MARACAS generates sets of differentially expressed genes or lists of genomic loci for RNA-seq and ChIP-seq analysis respectively. Second, ALGAEFUN (microALGAE FUNctional enrichment tool) is a web-based application where gene sets generated from RNA-seq analysis as well as lists of genomic loci from ChIP-seq analysis can be used as input. On the one hand, it can be used to perform Gene Ontology and biological pathways enrichment analysis over gene sets. On the other hand, using the results of ChIP-seq data analysis, it identifies a set of potential target genes and analyses the distribution of the loci over gene features. Graphical representation of the results as well as tables with gene annotations are generated and can be downloaded for further analysis. Conclusions ALGAEFUN with MARACAS provides an integrated environment for the microalgae research community that facilitates the process of obtaining relevant biological information from raw RNA-seq and ChIP-seq data. These applications are designed to assist researchers in the interpretation of gene lists and genomic loci based on functional enrichment analysis. ALGAEFUN with MARACAS is publicly available on https://greennetwork.us.es/AlgaeFUN/.

The intima media thickness (IMT) of the common carotid artery (CCA) can be used to predict the risk of atherosclerosis. Many image segmentation techniques have been used for IMT measurement. However, severe noise in the ultrasound image can lead to erroneous segmentation results. To improve the robustness to noise, a fully automatic method, based on an improved Otsu’s method and an adaptive wind-driven optimization technique, is proposed for estimating the IMT (denoted as “improved Otsu-AWDO”). First, an advanced despeckling filter, i.e., “ Nagare’s filter” is used to address the speckle noise in the carotid ultrasound images. Next, an improved fuzzy contrast method (IFC) is used to enhance the region of the intima media complex (IMC) in the blurred filtered images. Then, a new method is used for automatic extraction of the region of interest (ROI). Finally, the lumen intima interface and media adventitia interface are segmented from the IMC using improved Otsu-AWDO. Then, 156 B-mode longitudinal carotid ultrasound images of six different datasets are used to evaluate the performance of the automatic measurements. The results indicate that the absolute error of proposed method is only 10.1 ± 9.6 (mean ± std in μm). Moreover, the proposed method has a correlation coefficient as high as 0.9922, and a bias as low as 0.0007. From comparison with previous methods, we can conclude that the proposed method has strong robustness and can provide accurate IMT estimations.

Speckle tracking echocardiography (STE) enables quantification of myocardial deformation by a generation of spatiotemporal strain curves or time-strain curves (TSCs). Currently, only assessment of peak global longitudinal strain is employed in clinical practice because of the uncertainty in the accuracy of STE. We describe a supervised machine learning, physiologically constrained, fully automatic algorithm, trained with labeled data, for classification of TSCs into physiologic or artifactual classes. The data set of 415 healthy patients, with three cine loops per patient, corresponding to the three standard 2-D longitudinal views, was processed using a previously published, in-house STE software termed K-SAD. We report an accuracy of 86.4% for classifying TSCs as physiologic, artifactual and undetermined curves. The positive predictive value for a physiologic strain curve is 89%. This is as a necessary step for a similar separation of pathologic conditions, to allow full utilization of the temporal information concealed in layer-specific segmental TSCs.

Objectives To simulate clinical deployment, evaluate performance, and establish quality assurance of a deep learning algorithm (U-Net) for detection, localization, and segmentation of clinically significant prostate cancer (sPC), ISUP grade group ≥ 2, using bi-parametric MRI. Methods In 2017, 284 consecutive men in active surveillance, biopsy-naïve or pre-biopsied, received targeted and extended systematic MRI/transrectal US-fusion biopsy, after examination on a single MRI scanner (3 T). A prospective adjustment scheme was evaluated comparing the performance of the Prostate Imaging Reporting and Data System (PI-RADS) and U-Net using sensitivity, specificity, predictive values, and the Dice coefficient. Results In the 259 eligible men (median 64 [IQR 61–72] years), PI-RADS had a sensitivity of 98% [106/108]/84% [91/108] with a specificity of 17% [25/151]/58% [88/151], for thresholds at ≥ 3/≥ 4 respectively. U-Net using dynamic threshold adjustment had a sensitivity of 99% [107/108]/83% [90/108] (p > 0.99/> 0.99) with a specificity of 24% [36/151]/55% [83/151] (p > 0.99/> 0.99) for probability thresholds d3 and d4 emulating PI-RADS ≥ 3 and ≥ 4 decisions respectively, not statistically different from PI-RADS. Co-occurrence of a radiological PI-RADS ≥ 4 examination and U-Net ≥ d3 assessment significantly improved the positive predictive value from 59 to 63% (p = 0.03), on a per-patient basis. Conclusions U-Net has similar performance to PI-RADS in simulated continued clinical use. Regular quality assurance should be implemented to ensure desired performance. Key Points • U-Net maintained similar diagnostic performance compared to radiological assessment of PI-RADS ≥ 4 when applied in a simulated clinical deployment. • Application of our proposed prospective dynamic calibration method successfully adjusted U-Net performance within acceptable limits of the PI-RADS reference over time, while not being limited to PI-RADS as a reference. • Simultaneous detection by U-Net and radiological assessment significantly improved the positive predictive value on a per-patient and per-lesion basis, while the negative predictive value remained unchanged.

Purpose Precise segmentation of brain lesions is essential for neurological research. Specifically, resection volume estimates can aid in the assessment of residual postoperative tissue, e.g. following surgery for glioma. Furthermore, behavioral lesion-symptom mapping in epilepsy relies on accurate delineation of surgical lesions. We sought to determine whether semi- and fully automatic segmentation methods can be applied to resected brain areas and which approach provides the most accurate and cost-efficient results. Methods We compared a semi-automatic (ITK-SNAP) with a fully automatic (lesion_GNB) method for segmentation of resected brain areas in terms of accuracy with manual segmentation serving as reference. Additionally, we evaluated processing times of all three methods. We used T1w, MRI-data of epilepsy patients (n = 27; 11 m; mean age 39 years, range 16–69) who underwent temporal lobe resections (17 left). Results The semi-automatic approach yielded superior accuracy (p p > 0.05). Manual segmentation required more human input (30.41 min/subject) and therefore inferring significantly higher costs than semi- (3.27 min/subject) or fully automatic approaches (labor and cost approaching zero). Conclusion Semi-automatic segmentation offers the most accurate results in resected brain areas with a moderate amount of human input, thus representing a viable alternative compared with manual segmentation, especially for studies with large patient cohorts.

Purpose Rheumatoid arthritis (RA) causes joint space narrowing (JSN) as a form of joint destruction. We developed an automatic system that can detect joint locations and compute the joint space difference index (JSDI), which was defined as the chronological change in JSN between two radiographs. This study aims to evaluate the application of "machine vision" for radiographic image of the finger joints. Materials and methods Fifteen RA patients with long-term sustained clinical low disease activity were recruited. All patients underwent hand radiography and power Doppler ultrasonography (PDUS). The JSN was evaluated using the Genant-modified Sharp scoring (GSS) method and the automatic system. Synovial vascularity (SV) was assessed quantitatively using ultrasonography. Results There were no significant differences in the JSDI between the joints with JSN and those without JSN on GSS (p = 0.052). The JSDI of the joints with SV was significantly higher than those without SV (p = 0.043). The JSDI of the no therapeutic response group was significantly higher than those of the response group (p < 0.001). Conclusion Our software can automatically evaluate temporal changes of JSN, which might free rheumatologists / radiologists from the burden of scoring hand radiography.

The constantly increasing number of computed tomography (CT) examinations poses major challenges for radiologists. In this article, the additional benefits and potential of an artificial intelligence (AI) analysis platform for chest CT examinations in routine clinical practice will be examined. Specific application examples include AI-based, fully automatic lung segmentation with emphysema quantification, aortic measurements, detection of pulmonary nodules, and bone mineral density measurement. This contribution aims to appraise this AI-based application for value-added diagnosis during routine chest CT examinations and explore future development perspectives.

Scientific reports 11(1), 23244 (2021). doi:10.1038/s41598-021-02708-y, Published by Macmillan Publishers Limited, part of Springer Nature, [London]

PURPOSE: To develop an automatic method for selecting heartbeats with consistent respiratory phase to improve accuracy of cardiac function quantification in real-time (RT) cardiac MRI. METHODS: The respiratory signal is extracted by a principal component analysis method from RT cine images. Then, a two-step procedure is used to determine the directionality (sign) of the respiratory signal. With the motion in a manually selected region-of-interest as reference, the quality of the extracted respiratory signal is assessed using multi-slice RT cine data from eleven volunteers and ten patients. In addition, the impact of selecting heartbeats with consistent respiratory phase on the cardiac function quantification is evaluated. RESULTS: The extracted respiratory signal using the proposed method exhibits a high, positive correlation with the reference in all cases and is more robust compared to a recently proposed method. Also, for right ventricular (RV) function quantification, selecting heartbeats at expiratory position improves agreement between RT cine and breath-held reference. CONCLUSIONS: The proposed method enables fully automatic extraction and directionality determinations of respiratory signal from RT cardiac cine images, allowing accurate cardiac function quantification.

International audience; The aim of this study is to develop an automated deep-learning-based whole heart segmentation of ECG-gated computed tomography data. After 21 exclusions, CT acquired before transcatheter aortic valve implantation in 71 patients were reviewed and randomly split in a training (n = 55 patients), validation (n = 8 patients), and a test set (n = 8 patients). A fully automatic deep-learning method combining two convolutional neural networks performed segmentation of 10 cardiovascular structures, which was compared with the manually segmented reference by the Dice index. Correlations and agreement between myocardial volumes and mass were assessed. The algorithm demonstrated high accuracy (Dice score = 0.920; interquartile range: 0.906-0.925) and a low computing time (13.4 s, range 11.9-14.9). Correlations and agreement of volumes and mass were satisfactory for most structures. Six of ten structures were well segmented. Deep-learning-based method allowed automated WHS from ECG-gated CT data with a high accuracy. Challenges remain to improve right-sided structures segmentation and achieve daily clinical application.

Importance A high level of surgical skill is essential to prevent intraoperative problems. One important aspect of surgical education is surgical skill assessment, with pertinent feedback facilitating efficient skill acquisition by novices. Objectives To develop a 3-dimensional (3-D) convolutional neural network (CNN) model for automatic surgical skill assessment and to evaluate the performance of the model in classification tasks by using laparoscopic colorectal surgical videos. Design, setting, and participants This prognostic study used surgical videos acquired prior to 2017. In total, 650 laparoscopic colorectal surgical videos were provided for study purposes by the Japan Society for Endoscopic Surgery, and 74 were randomly extracted. Every video had highly reliable scores based on the Endoscopic Surgical Skill Qualification System (ESSQS, range 1-100, with higher scores indicating greater surgical skill) established by the society. Data were analyzed June to December 2020. Main outcomes and measures From the groups with scores less than the difference between the mean and 2 SDs, within the range spanning the mean and 1 SD, and greater than the sum of the mean and 2 SDs, 17, 26, and 31 videos, respectively, were randomly extracted. In total, 1480 video clips with a length of 40 seconds each were extracted for each surgical step (medial mobilization, lateral mobilization, inferior mesenteric artery transection, and mesorectal transection) and separated into 1184 training sets and 296 test sets. Automatic surgical skill classification was performed based on spatiotemporal video analysis using the fully automated 3-D CNN model, and classification accuracies and screening accuracies for the groups with scores less than the mean minus 2 SDs and greater than the mean plus 2 SDs were calculated. Results The mean (SD) ESSQS score of all 650 intraoperative videos was 66.2 (8.6) points and for the 74 videos used in the study, 67.6 (16.1) points. The proposed 3-D CNN model automatically classified video clips into groups with scores less than the mean minus 2 SDs, within 1 SD of the mean, and greater than the mean plus 2 SDs with a mean (SD) accuracy of 75.0% (6.3%). The highest accuracy was 83.8% for the inferior mesenteric artery transection. The model also screened for the group with scores less than the mean minus 2 SDs with 94.1% sensitivity and 96.5% specificity and for group with greater than the mean plus 2 SDs with 87.1% sensitivity and 86.0% specificity. Conclusions and relevance The results of this prognostic study showed that the proposed 3-D CNN model classified laparoscopic colorectal surgical videos with sufficient accuracy to be used for screening groups with scores greater than the mean plus 2 SDs and less than the mean minus 2 SDs. The proposed approach was fully automatic and easy to use for various types of surgery, and no special annotations or kinetics data extraction were required, indicating that this approach warrants further development for application to automatic surgical skill assessment.

Cerebral palsy is a neurological condition that is known to affect muscle growth. Detailed investigations of muscle growth require segmentation of muscles from MRI scans, which is typically done manually. In this study, we evaluated the performance of 2D, 3D, and hybrid deep learning models for automatic segmentation of 11 lower leg muscles and two bones from MRI scans of children with and without cerebral palsy. All six models were trained and evaluated on manually segmented T1 -weighted MRI scans of the lower legs of 20 children, six of whom had cerebral palsy. The segmentation results were assessed using the median Dice similarity coefficient (DSC), average symmetric surface distance (ASSD), and volume error (VError) of all 13 labels of every scan. The best performance was achieved by H-DenseUNet, a hybrid model (DSC 0.90, ASSD 0.5 mm, and VError 2.6 cm3 ). The performance was equivalent to the inter-rater performance of manual segmentation (DSC 0.89, ASSD 0.6 mm, and VError 3.3 cm3 ). Models trained with the Dice loss function outperformed models trained with the cross-entropy loss function. Near-optimal performance could be attained using only 11 scans for training. Segmentation performance was similar for scans of typically developing children (DSC 0.90, ASSD 0.5 mm, and VError 2.8 cm3 ) and children with cerebral palsy (DSC 0.85, ASSD 0.6 mm, and VError 2.4 cm3 ). These findings demonstrate the feasibility of fully automatic segmentation of individual muscles and bones from MRI scans of children with and without cerebral palsy.

Timely, accurate and reliable assessment of fetal brain development is essential to reduce short and long-term risks to fetus and mother. Fetal MRI is increasingly used for fetal brain assessment. Three key biometric linear measurements important for fetal brain evaluation are Cerebral Biparietal Diameter (CBD), Bone Biparietal Diameter (BBD), and Trans-Cerebellum Diameter (TCD), obtained manually by expert radiologists on reference slices, which is time consuming and prone to human error. The aim of this study was to develop a fully automatic method computing the CBD, BBD and TCD measurements from fetal brain MRI. The input is fetal brain MRI volumes which may include the fetal body and the mother's abdomen. The outputs are the measurement values and reference slices on which the measurements were computed. The method, which follows the manual measurements principle, consists of five stages: 1) computation of a Region Of Interest that includes the fetal brain with an anisotropic 3D U-Net classifier; 2) reference slice selection with a Convolutional Neural Network; 3) slice-wise fetal brain structures segmentation with a multiclass U-Net classifier; 4) computation of the fetal brain midsagittal line and fetal brain orientation, and; 5) computation of the measurements. Experimental results on 214 volumes for CBD, BBD and TCD measurements yielded a mean $L_1$ difference of 1.55mm, 1.45mm and 1.23mm respectively, and a Bland-Altman 95% confidence interval ($CI_{95}$) of 3.92mm, 3.98mm and 2.25mm respectively. These results are similar to the manual inter-observer variability. The proposed automatic method for computing biometric linear measurements of the fetal brain from MR imaging achieves human level performance. It has the potential of being a useful method for the assessment of fetal brain biometry in normal and pathological cases, and of improving routine clinical practice., 15 pages, 8 figures, presented in CARS 2020, submitted to IJCARS

Conflict detection in sensory input is central to adaptive human behavior. Perhaps unsurprisingly, past research has shown that conflict may even be detected in the absence of conflict awareness, suggesting that conflict detection is an automatic process that does not require attention. To test the possibility of conflict processing in the absence of attention, we manipulated task relevance and response overlap of potentially conflicting stimulus features across six behavioral tasks. Multivariate analyses on human electroencephalographic data revealed neural signatures of conflict only when at least one feature of a conflicting stimulus was attended, regardless of whether that feature was part of the conflict, or overlaps with the response. In contrast, neural signatures of basic sensory processes were present even when a stimulus was completely unattended. These data reveal an attentional bottleneck at the level of objects, suggesting that object-based attention is a prerequisite for cognitive control operations involved in conflict detection., eLife digest : Focusing your attention on one thing can leave you surprisingly unaware of what goes on around you. A classic experiment known as ‘the invisible gorilla’ highlights this phenomenon. Volunteers were asked to watch a clip featuring basketball players, and count how often those wearing white shirts passed the ball: around half of participants failed to spot that someone wearing a gorilla costume wandered into the game and spent nine seconds on screen. Yet, things that you are not focusing on can sometimes grab your attention anyway. Take for example, the ‘cocktail party effect’, the ability to hear your name among the murmur of a crowded room. So why can we react to our own names, but fail to spot the gorilla? To help answer this question, Nuiten et al. examined how paying attention affects the way the brain processes input. Healthy volunteers were asked to perform various tasks while the words ‘left’ or ‘right’ played through speakers. The content of the word was sometimes consistent with its location (‘left’ being played on the left speaker), and sometimes opposite (‘left’ being played on the right speaker). Processing either the content or the location of the word is relatively simple for the brain; however detecting a discrepancy between these two properties is challenging, requiring the information to be processed in a brain region that monitors conflict in sensory input. To manipulate whether the volunteers needed to pay attention to the words, Nuiten et al. made their content or location either relevant or irrelevant for a task. By analyzing brain activity and task performance, they were able to study the effects of attention on how the word properties were processed. The results showed that the volunteers’ brains were capable of dealing with basic information, such as location or content, even when their attention was directed elsewhere. But discrepancies between content and location could only be detected when the volunteers were focusing on the words, or when their content or location was directly relevant to the task. The findings by Nuiten et al. suggest that while performing a difficult task, our brains continue to react to basic input but often fail to process more complex information. This, in turn, has implications for a range of human activities such as driving. New technology could potentially help to counteract this phenomenon, aiming to direct attention towards complex information that might otherwise be missed.

As sarcopenia research has been gaining emphasis, the need for quantification of abdominal muscle on computed tomography (CT) is increasing. Thus, a fully automated system to select L3 slice and segment muscle in an end-to-end manner is demanded. We aimed to develop a deep learning model (DLM) to select the L3 slice with consideration of anatomic variations and to segment cross-sectional areas (CSAs) of abdominal muscle and fat. Our DLM, named L3SEG-net, was composed of a YOLOv3-based algorithm for selecting the L3 slice and a fully convolutional network (FCN)-based algorithm for segmentation. The YOLOv3-based algorithm was developed via supervised learning using a training dataset (n = 922), and the FCN-based algorithm was transferred from prior work. Our L3SEG-net was validated with internal (n = 496) and external validation (n = 586) datasets. Ground truth L3 level CT slice and anatomic variation were identified by a board-certified radiologist. L3 slice selection accuracy was evaluated by the distance difference between ground truths and DLM-derived results. Technical success for L3 slice selection was defined when the distance difference was 2. A fully automatic system was developed for the selection of an exact axial CT slice at the L3 vertebral level and the segmentation of abdominal muscle areas.

Dual-chamber pacemaker is a fully automatic pacemaker with the function of simulating human physiological pacing. It regulates pacing by programming different refractory periods and various special functions, which are closely related to arrhythmia. After in-depth understanding of these special functions, regular electrocardiogram follow-up analysis is required to provide individualized optimal program control and so is appropriate the administration of the pacemaker’s special functions to better provide optimal clinical guidance for patients with arrhythmia.

The exogenous anabolic-androgenic steroid (AAS) stanozolol stays one of the most detected substances in professional sports. Its detection is a fundamental part of doping analysis, and the analysis of this steroid has been intensively investigated for a long time. This contribution to the detection of stanozolol doping describes for the first time the unambiguous proof for the existence of 17-epistanozolol-1′N-glucuronide and 17-epistanozolol-2′N-glucuronide in stanozolol-positive human urine samples due to the access to high-quality reference standards. Examination of excretion study samples shows large detection windows for the phase-II metabolites stanozolol-1′N-glucuronide and 17-epistanozolol-1′N-glucuronide up to 12 days and respectively up to almost 28 days. In addition, we present appropriate validation parameters for the analysis of these metabolites using a fully automatic method online solid-phase extraction (SPE) method already published before. Limits of identification (LOIs) as low as 100 pg/ml and other validation parameters like accuracy, precision, sensitivity, robustness, and linearity are given.

Aims: The availability of a true 3D dataset provides an opportunity for automation of left ventricular (LV) and left atrial (LA) measurements. Although manual and automated measurements of 3D volumes are known to correlate, the variance is an important parameter for the individual patient. The reasons for discrepancies remain unexplained. We hence aim to explain the disagreement between automated and manual LV and LA volumes. Methods and Results: A total of 355 patients underwent standard clinical echo, with offline analysis in both fully- (Heart Model, Philips) and semiautomated (3DQ-Adv, Philips) assessment of routine indices of LV and LA function and shape. Each image was classified according to quality using a 4-point scale as well as the American Society for Echocardiography guidelines for appropriate use of contrast. Bland-Altman plots were used to assess agreement, and t tests were used to assess differences in agreement. Predictors of volume discrepancy were sought with linear regression. Measures of LV and LA volumes were greater with automatic than semiautomatic assessment. The difference in left ventricular end-diastolic volume was dependent on the number of regional wall-motion abnormalities (RWMA) (beta = 0.59, P < .04) and image quality (beta = 19.71, P = .02). RWMA predicted the difference in left ventricular end-systolic volume (beta = 0.83, P < .01) and left atrial end-systolic volume (beta = -1.01 P < .01). Conclusion: LV and LA volumes were higher with automatic than semiautomatic assessment. Image quality and RWMA may contribute to this discrepancy. These limitations need to be addressed before fully automatic assessment of 3D echocardiograms can be used in the clinic.

Pulmonary embolism (PE) and other pulmonary vascular diseases, have been found associated with the changes in arterial morphology. To detect arterial changes, we propose a novel, fully automatic method that can extract pulmonary arterial tree in computed tomographic pulmonary angiography (CTPA) images. The approach is based on the fuzzy connectedness framework, combined with 3D vessel enhancement and Harris Corner detection to achieve accurate segmentation. The effectiveness and robustness of the method is validated in clinical datasets consisting of 10 CT angiography scans (6 without PE and 4 with PE). The performance of our method is compared with manual classification and machine learning method based on random forest. Our method achieves a mean accuracy of 92% when compared to manual reference, which is higher than the 89% accuracy achieved by machine learning. This performance of the segmentation for pulmonary arteries may provide a basis for the CAD application of PE.

Knee osteoarthritis (OA) is one major cause of activity limitation and physical disability in older adults. Early detection and intervention can help slow down the OA degeneration. Physicians’ grading based on visual inspection is subjective, varied across interpreters, and highly relied on their experience. In this paper, we successively apply two deep convolutional neural networks (CNN) to automatically measure the knee OA severity, as assessed by the Kellgren-Lawrence (KL) grading system. Firstly, considering the size of knee joints distributed in X-ray images with small variability, we detect knee joints using a customized one-stage YOLOv2 network. Secondly, we fine-tune the most popular CNN models, including variants of ResNet, VGG, and DenseNet as well as InceptionV3, to classify the detected knee joint images with a novel adjustable ordinal loss. To be specific, motivated by the ordinal nature of the knee KL grading task, we assign higher penalty to misclassification with larger distance between the predicted KL grade and the real KL grade. The baseline X-ray images from the Osteoarthritis Initiative (OAI) dataset are used for evaluation. On the knee joint detection, we achieve mean Jaccard index of 0.858 and recall of 92.2% under the Jaccard index threshold of 0.75. On the knee KL grading task, the fine-tuned VGG-19 model with the proposed ordinal loss obtains the best classification accuracy of 69.7% and mean absolute error (MAE) of 0.344. Both knee joint detection and knee KL grading achieve state-of-the-art performance. The code, dataset, and models are released at https://github.com/PingjunChen/KneeAnalysis.

International audience; The purpose of this study is to assess the relevance of computational anatomy for the sex determination in forensic anthropology. A novel groupwise registration algorithm is used, based on keypoint extraction, able to register several hundred full body images in a common space. Experiments were conducted on 83 CT scanners of living individuals from the public VISCERAL database. In our experiments, we first verified that the well-known criteria for sex discrimination on the hip-bone were well preserved in mean images. In a second experiment, we have tested semi-automatic positioning of anatomical landmarks to measure the relevance of groupwise registration for future research. We applied the Probabilistic Sex Diagnosis tool on the predicted landmarks. This resulted in 62% of correct sex determinations, 37% of undetermined cases, and 1% of errors. The main limiting factors are the population sample size and the lack of precision for the initial manual positioning of the landmarks in the mean image. We also give insights on future works for robust and fully automatic sex determination.

Differing forms of self-operating transportation are already among us and some have been in operation now for an extended period of time. From elevators and escalators to airport transit trams, we already use many fully automatic systems. Now such technologies are very publicly and prominently penetrating into the on-road environment of everyday personal vehicle usage. The present article raises and addresses a number of the specific and more general human factors/ergonomic issues associated with such an evolutionary step. One particular concern is that of identified responsibility when such systems fail to perform flawlessly. The ways in which this (r)evolution will impact the social and cultural fabric of affected societies is also considered. Further observations as to the vector of the future characteristics of these vehicular forms and how they and other autonomous systems will affect our world are examined. The very future of the human experience depends upon the ways in which such systems are designed, enacted and integrated into everyday life and these are fundamentally ergonomic endeavours.Practitioner's Summary: The prominence of practitioners working on advanced human-machine systems will increase with public concerns surrounding self-driving vehicles. Driverless cars are not only a technological step but they will also exert widespread effects throughout society. Practitioners should prepare for these broad socio-technical challenges in an evolving, autonomous world.

This study examined the influence of consumer behavior on furan, 2-methylfuran, 3-methylfuran, 2,5-dimethylfuran and 2,3-dimethylfuran exposure in coffee. Coffees brewed using a filter, fully automatic, capsule machine or reconstituted instant coffee were found to have a significant different cup concentrations of furan derivatives. Coffee brewed with the fully automatic machine contained the highest furan and furan derivative concentrations (99.05 µg/L furan, 263.91 µg/L 2-methylfuran, 13.15 µg/L 3-methylfuran and 8.44 µg/L 2,5-dimethylfuran) whereas soluble coffee did not contain detectable levels, thereby contributing least to a consumer’s dietary exposure. Furan and furan derivative concentrations were found to decrease significantly upon cooling, reducing consumer exposure by 8.0–17.2 % on average once the coffee reached drinking temperature 55–60 °C, in ceramic cups. Serving coffee in a ceramic or disposable cup were found to influence the cooling dynamics of the coffee but did not statistically influence the consumers exposure at a given temperature.

Background Segmentation of brain MR images of neonates is a primary step for assessment of brain evolvement. Advanced segmentation techniques used for adult brain MRI are not companionable for neonates, due to extensive dissimilarities in tissue properties and head structure. Existing segmentation methods for neonates utilizes brain atlases or requires manual elucidation, which results into improper and atlas dependent segmentation. Objective The primary objective of this work is to develop fully automatic, atlas free, and robust system to segment and classify brain tissues of newborn infants from magnetic resonance images. Study Design In this study, we propose a fully automatic, atlas-free pipeline based Neural Tree approach for segmentation of newborn brain MRI which utilizes resourceful local resemblance factor such as concerning, connectivity, structure, and relative tissue location. Physical collaboration and uses of an atlas are not required in proposed method and at the same time skirting atlas-associated bias which results in improved segmentation. Proposed technique segments and classify brain tissues both at global and tissue level. Results We examined our results through visual assessment by neonatologists and quantitative comparisons that show first-rate concurrence with proficient manual segmentations. The implementation results of the proposed technique provided a good overall accuracy of 91.82% for the segmentation of brain tissues as compared with other methods. Conclusion The pipelined-based neural tree approach along with local similarity factor segments and classify brain tissues. The proposed automated system have higher dice similarity coefficient as well as computational speed.

The accuracy of abdominal multi-parametric quantification based on dynamic contrast-enhanced ultrasound (DCEUS) is limited by out-of-plane severe distortion induced by respiratory motion. This study developed a fully automatic respiratory gating scheme by using principal component analysis to remove distortions and disturbances in free-breathing DCEUS-based quantification. Taking the known in-vitro perfusions as ground truths, we further evaluated the respiratory gating accuracy from multiple perspectives in a controllable rotary distortion flow model with out-of-plane severe distortion induced by respiratory motion. Compared with those without respiratory gating, the signal-to-clutter ratio and correlation coefficients of perfusion parameters with respiratory gating improved by 3.99 ± 1.71 dB (p 0.01) and 0.39 ± 0.17 (p 0.01), respectively. The corresponding mean square error decreased by 1893.9 ± 763.16 (p 0.001). Results with respiratory gating were significantly consistent with the ground truths without respiratory motion disturbances. Quantitative results demonstrated the effective removal of out-of-plane serious distortion and other disturbances induced by the respiratory kinetics on DCEUS-based quantification. The accuracy and robustness of DCEUS-based quantification were significantly improved by the proposed respiratory gating strategy, which not only benefits physicians in identifying transient hemodynamic characteristics in tumor angiogenesis, but also provides accurate diagnoses for abdominal tumors.

Objectives To evaluate the accuracy of automatic deep learning-based method for fully automatic segmentation of the mandible from CBCTs. Setting and sample population CBCT-derived mandible fully automatic segmentation. Methods 40 CBCT scans from healthy patients (20 females and 20 males, mean age 23,37 ± 3,34) were collected and a manual mandible segmentation were carried out by using Mimics software. Twenty CBCT scans were randomly selected and used for training the artificial intelligence model file. The remaining 20 CBCT segmentation masks were used to test the accuracy of the CNN automatic method by comparing the segmentation volumes of the 3D models obtained with automatic and manual segmentations. The accuracy of the CNN-based method was also assessed by using the DICE Score coefficient (DSC) and by the surface-to-surface matching technique. The Intraclass correlation coefficient (ICC) and Dahlberg's formula were used respectively to test the intra-observer reliability and method error. Independent Student's t-test was used for between-groups volumetric comparison. Results Measurements were highly correlated with an ICC value of 0,937 while the method error was 0,24 mm3. A difference of 0,71 (± 0,49) cm3 was found between the methodologies but it was not statistically significant (p>0.05). The matching percentage detected was 90.35% (± 1.88) (tolerance 0.5 mm) and 96.32% ± 1.97% (tolerance 1.0 mm). The differences, measured as DSC in percentage, between the assessments done with both methods were, respectively, 2,8% and 3,1%. Conclusion The tested deep learning CNN-based technology is accurate and performs as well as an experienced image reader but at much higher speed, which is of significant clinical relevance.

Diabetic retinopathy (DR) is a disease resulting from diabetes complications, causing non-reversible damage to retina blood vessels. DR is a leading cause of blindness if not detected early. The currently available DR treatments are limited to stopping or delaying the deterioration of sight, highlighting the importance of regular scanning using high-efficiency computer-based systems to diagnose cases early. The current work presented fully automatic diagnosis systems that exceed manual techniques to avoid misdiagnosis, reducing time, effort and cost. The proposed system classifies DR images into five stages—no-DR, mild, moderate, severe and proliferative DR—as well as localizing the affected lesions on retain surface. The system comprises two deep learning-based models. The first model (CNN512) used the whole image as an input to the CNN model to classify it into one of the five DR stages. It achieved an accuracy of 88.6% and 84.1% on the DDR and the APTOS Kaggle 2019 public datasets, respectively, compared to the state-of-the-art results. Simultaneously, the second model used an adopted YOLOv3 model to detect and localize the DR lesions, achieving a 0.216 mAP in lesion localization on the DDR dataset, which improves the current state-of-the-art results. Finally, both of the proposed structures, CNN512 and YOLOv3, were fused to classify DR images and localize DR lesions, obtaining an accuracy of 89% with 89% sensitivity, 97.3 specificity and that exceeds the current state-of-the-art results.

This paper proposes a two-dimensional (2D) bidirectional long short-term memory generative adversarial network (GAN) to produce synthetic standard 12-lead ECGs corresponding to four types of signals—left ventricular hypertrophy (LVH), left branch bundle block (LBBB), acute myocardial infarction (ACUTMI), and Normal. It uses a fully automatic end-to-end process to generate and verify the synthetic ECGs that does not require any visual inspection. The proposed model is able to produce synthetic standard 12-lead ECG signals with success rates of 98% for LVH, 93% for LBBB, 79% for ACUTMI, and 59% for Normal. Statistical evaluation of the data confirms that the synthetic ECGs are not biased towards or overfitted to the training ECGs, and span a wide range of morphological features. This study demonstrates that it is feasible to use a 2D GAN to produce standard 12-lead ECGs suitable to augment artificially a diverse database of real ECGs, thus providing a possible solution to the demand for extensive ECG datasets.

OBJECTIVE To determine whether volumetric CT texture analysis (CTTA) using fully automatic tumor segmentation can help predict recurrence-free survival (RFS) in patients with intrahepatic mass-forming cholangiocarcinomas (IMCCs) after surgical resection. MATERIALS AND METHODS This retrospective study analyzed the preoperative CT scans of 89 patients with IMCCs (64 male; 25 female; mean age, 62.1 years; range, 38-78 years) who underwent surgical resection between January 2005 and December 2016. Volumetric CTTA of IMCCs was performed in late arterial phase images using both fully automatic and semi-automatic liver tumor segmentation techniques. The time spent on segmentation and texture analysis was compared, and the first-order and second-order texture parameters and shape features were extracted. The reliability of CTTA parameters between the techniques was evaluated using intraclass correlation coefficients (ICCs). Intra- and interobserver reproducibility of volumetric CTTAs were also obtained using ICCs. Cox proportional hazard regression were used to predict RFS using CTTA parameters and clinicopathological parameters. RESULTS The time spent on fully automatic tumor segmentation and CTTA was significantly shorter than that for semi-automatic segmentation: mean ± standard deviation of 1 minutes 37 seconds ± 50 seconds vs. 10 minutes 48 seconds ± 13 minutes 44 seconds (p < 0.001). ICCs of the texture features between the two techniques ranged from 0.215 to 0.980. ICCs for the intraobserver and interobserver reproducibility using fully automatic segmentation were 0.601-0.997 and 0.177-0.984, respectively. Multivariable analysis identified lower first-order mean (hazard ratio [HR], 0.982; p = 0.010), larger pathologic tumor size (HR, 1.171; p < 0.001), and positive lymph node involvement (HR, 2.193; p = 0.014) as significant parameters for shorter RFS using fully automatic segmentation. CONCLUSION Volumetric CTTA parameters obtained using fully automatic segmentation could be utilized as prognostic markers in patients with IMCC, with comparable reproducibility in significantly less time compared with semi-automatic segmentation.

Intestinal infarction is the fast-evolving endpoint of impaired blood perfusion to an intestinal segment which may have fatal outcome. Early diagnosis and treatment within 6 h reduce mortality. Currently, d-lactate is a promising biomarker, however, not available in the acute clinical setting. The aim of this study is implementation of d-lactate analysis in a routine clinical setting. We used a spectrophotometric method, based on enzymatic oxidation of d-lactate by d-lactate dehydrogenase (D-LDH) coupled to the reduction of nicotinamide-adenine dinucleotide (NAD+). The amount of NADH formed in this reaction is equivalent to d-lactate. The primary concern in this method is interfering NADH formed by oxidation of l-lactate by l-lactate dehydrogenase (L-LDH). A commercially available kit for d-lactate measurement was implemented on our existing automated routine laboratory equipment including pH-inactivation of L-LDH. Our setup fulfilled clinical quality goals. We were able to measure d-lactate with an acceptable performance of the analysis and a short turn-around time. The method can be used to distinguish between the expected cut-off for intestinal ischemia around 0.3 mM and the upper reference limit of 0.05 mM. With a turnaround time of just 9 min, the analysis has potential as a readily available detection of circulating d-lactate for early diagnosis of intestinal ischemia.

Purpose Identification and delineation of the gross tumour and malignant nodal volume (GTV) in medical images are vital in radiotherapy. We assessed the applicability of convolutional neural networks (CNNs) for fully automatic delineation of the GTV from FDG-PET/CT images of patients with head and neck cancer (HNC). CNN models were compared to manual GTV delineations made by experienced specialists. New structure-based performance metrics were introduced to enable in-depth assessment of auto-delineation of multiple malignant structures in individual patients. Methods U-Net CNN models were trained and evaluated on images and manual GTV delineations from 197 HNC patients. The dataset was split into training, validation and test cohorts (n= 142, n = 15 and n = 40, respectively). The Dice score, surface distance metrics and the new structure-based metrics were used for model evaluation. Additionally, auto-delineations were manually assessed by an oncologist for 15 randomly selected patients in the test cohort. Results The mean Dice scores of the auto-delineations were 55%, 69% and 71% for the CT-based, PET-based and PET/CT-based CNN models, respectively. The PET signal was essential for delineating all structures. Models based on PET/CT images identified 86% of the true GTV structures, whereas models built solely on CT images identified only 55% of the true structures. The oncologist reported very high-quality auto-delineations for 14 out of the 15 randomly selected patients. Conclusions CNNs provided high-quality auto-delineations for HNC using multimodality PET/CT. The introduced structure-wise evaluation metrics provided valuable information on CNN model strengths and weaknesses for multi-structure auto-delineation.

Objectives To develop a fully automatic multiview shape constraint framework for comprehensive coronary artery calcium scores (CACS) quantification via deep learning on nonenhanced cardiac CT images. Methods In this retrospective single-centre study, a multi-task deep learning framework was proposed to detect and quantify coronary artery calcification from CT images collected between October 2018 and March 2019. A total of 232 non-contrast cardiac-gated CT scans were retrieved and studied (80 % for model training and 20 % for testing). CACS results of testing datasets (n = 46), including Agatston score, calcium volume score, calcium mass score, were calculated fully automatically and manually at total and vessel-specific levels, respectively. Results No significant differences were found in CACS quantification obtained using automatic or manual methods at total and vessel-specific levels (Agatston score: automatic 535.3 vs. manual 542.0, P = 0.993; calcium volume score: automatic 454.2 vs. manual 460.6, P = 0.990; calcium mass score: automatic 128.9 vs. manual 129.5, P = 0.992). Compared to the ground truth, the number of calcified vessels can be accurate recognized automatically (total: automatic 107 vs. manual 102, P = 0.125; left main artery: automatic 15 vs. manual 14, P = 1.000 ; left ascending artery: automatic 37 vs. manual 37, P = 1.000; left circumflex artery: automatic 22 vs. manual 20, P = 0.625; right coronary artery: automatic 33 vs. manual 31, P = 0.500). At the patient’s level, there was no statistic difference existed in the classification of Agatston scoring (P = 0.317) and the number of calcified vessels (P = 0.102) between the automatic and manual results. Conclusions The proposed framework can achieve reliable and comprehensive quantification for the CACS, including the calcified extent and distribution indicators at both total and vessel-specific levels.

Background We have developed a fully automatic three-dimensional MRI analysis software program for automatic segmentation of knee cartilage using a deep neural network. The purpose of this study was to use this software to clarify the interscan measurement error of the knee cartilage thickness and projected cartilage area ratio at 9 regions and 45 subregions in the knee. Methods Ten healthy volunteers underwent MRI twice in the same day. The software provided cartilage thickness and projected cartilage area ratio (thickness ≥ 1.5 mm) at 9 regions and 45 subregions of the knee without any manual correction. The interscan measurement error was calculated at each region and subregion from the data of nine donors, except for one donor who had body motion during the MRI examination. Results The interscan measurement error of cartilage thickness was less than 0.10 mm at all 9 regions and at 39 subregions among 45 subregions. The measurement errors ranged from 0.03 to 0.21 mm. The intraclass correlation coefficients (ICC) of cartilage thickness were higher than 0.75 at all 9 regions and 41 subregions. The interscan measurement error of the projected cartilage area ratio ranged from 0.01 to 0.03 for all 9 regions. Conclusions This study clarified the interscan measurement error of the knee cartilage thickness and projected cartilage area ratio.

Background An artificial intelligence method could accelerate the clinical implementation of tumour-stroma ratio (TSR), which has prognostic relevance in colorectal cancer (CRC). We, therefore, developed a deep learning model for the fully automated TSR quantification on routine haematoxylin and eosin (HE) stained whole-slide images (WSI) and further investigated its prognostic validity for patient stratification. Methods We trained a convolutional neural network (CNN) model using transfer learning, with its nine-class tissue classification performance evaluated in two independent test sets. Patch-level segmentation on WSI HE slides was performed using the model, with TSR subsequently derived. A discovery (N=499) and validation cohort (N=315) were used to evaluate the prognostic value of TSR for overall survival (OS). Findings The CNN-quantified TSR was a prognostic factor, independently of other clinicopathologic characteristics, with stroma-high associated with reduced OS in the discovery (HR 1.72, 95% CI 1.24-2.37, P=0.001) and validation cohort (2.08, 1.26-3.42, 0.004). Integrating TSR into a Cox model with other risk factors showed improved prognostic capability. Interpretation We developed a deep learning model to quantify TSR based on histologic WSI of CRC and demonstrated its prognostic validity for patient stratification for OS in two independent CRC patient cohorts. This fully automatic approach allows for the objective and standardised application while reducing pathologists' workload. Thus, it can potentially be of significant aid in clinical prognosis prediction and decision-making. Funding National Key Research and Development Program of China, National Science Fund for Distinguished Young Scholar, and National Science Foundation for Young Scientists of China.

Background We aimed to determine the growth-development periods and gender from the cervical vertebrae using the artificial neural network (ANN). Setting and sample population The cephalometric and hand-wrist radiographs obtained from 419 patients aged between 8 and 17 years were included in our study. Materials and methods Our retrospective study consisted of 419 patients' cephalometric and hand-wrist radiographs. The cephalometric radiographs were divided into six cervical vertebrae stages (CVS). Correlations were evaluated between hand-wrist maturation level, CVS, and ages. Twenty-seven vertebral reference points are marked on the cephalometric radiograph, and 32 linear measurements were taken. With the combination of these measurements, 24 different data sets were formed to train ANN. Thus, 24 different ANN models for the determination of the growth-development periods were obtained. According to the results, seven ANN models that have a high success level and clinically applicable were selected. Also, an ANN model was done by all measurements and age for the determination of gender from cervical vertebrae. Results Significantly positive correlations between hand-wrist maturation level, CVS and ages were detected. The ANN-7 model (32 linear measurements and age) accuracy value was found 0.9427. The highest model accuracy, 0.8687, with the least linear measurements, was obtained by drawing 13 linear measurements, using vertical measurements and indents. Gender was determined using ANN (0.8950) on cervical vertebrae data. Conclusion The growth-development periods and gender were determined from the cervical vertebrae by using ANN. The success of the ANN algorithm has been satisfactory. Further studies are needed for a fully automatic decision support system.

Background We developed a fully automatic three-dimensional knee MRI analysis software that can quantify meniscus extrusion and cartilage measurements, including the projected cartilage area ratio (PCAR), which represents the ratio of the subject’s actual cartilage area to their ideal cartilage area. We also collected 3D MRI knee data from 561 volunteers (aged 30–79 years) from the “Kanagawa Knee Study.” Our purposes were to verify the accuracy of the software for automatic cartilage and meniscus segmentation using knee MRI and to examine the relationship between medial meniscus extrusion measurements and cartilage measurements from Kanagawa Knee Study data. Methods We constructed a neural network for the software by randomly choosing 10 healthy volunteers and 103 patients with knee pain. We validated the algorithm by randomly selecting 108 of these 113 subjects for training, and determined Dice similarity coefficients from five other subjects. We constructed a neural network using all data (113 subjects) for training. Cartilage thickness, cartilage volume, and PCAR in the medial femoral, lateral femoral, medial tibial, and lateral tibial regions were quantified by using the trained software on Kanagawa Knee Study data and their relationship with subject height was investigated. We also quantified the medial meniscus coverage ratio (MMCR), defined as the ratio of the overlapping area between the medial meniscus area and the medial tibial cartilage area to the medial tibial cartilage area. Finally, we examined the relationship between MMCR and PCAR at middle central medial tibial (mcMT) subregion located in the center of nine subregions in the medial tibial cartilage. Results Dice similarity coefficients for cartilage and meniscus were both approximately 0.9. The femoral and tibial cartilage thickness and volume at each region correlated with height, but PCAR did not correlate with height in most settings. PCAR at the mcMT was significantly correlated with MMCR. Conclusions Our software showed high segmentation accuracy for the knee cartilage and meniscus. PCAR was more useful than cartilage thickness or volume since it was less affected by height. Relations ips were observed between the medial tibial cartilage measurements and the medial meniscus extrusion measurements in our cross-sectional study. Trial registration UMIN, UMIN000032826; 1 September 2018

In rheumatoid arthritis (RA), the radiographic progression of joint space narrowing (JSN) is evaluated using visual assessments. However, those methods are complicated and time-consuming. We developed an automatic system that can detect joint locations and compute the joint space difference index (JSDI), which was defined as the chronological change in JSN between two radiographs. The purpose of this study was to establish the validity of the software that automatically evaluates the temporal change of JSN. This study consisted of 39 patients with RA. All patients were treated with tocilizumab and underwent hand radiography (left and right hand separately) at 0, 6, and 12 months. The JSN was evaluated using mTSS (modified Total Sharp Score) by one musculoskeletal radiologist as well as our automatic system. Software measurement showed that JSDI between 0 and 12 months was significantly higher than that between 0 and 6 months (p < 0.01). While, there was no significant difference in mTSS between 0, 6, and 12 months. The group with higher disease activity at 0 months had significantly higher JSDI between 0 and 6 months than that with lower disease activity (p = 0.02). The automatic software can evaluate JSN progression of RA patients in the finger joint on X-ray. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (10.1007/s10278-020-00390-6) contains supplementary material, which is available to authorized users.

Objective. Recent advancements in electrode designs and micro-fabrication technology has allowed existence of microelectrode arrays with hundreds of channels for single-cell recordings. In such electrophysiological recordings, each implanted micro-electrode can record the activities of more than one neuron in its vicinity. Recording the activities of multiple neurons may also be referred to as multiple unit activity. However, for any further analysis, the main goal is to isolate the activity of each recorded neuron and thus called single-unit activity. This process may also be referred to as spike sorting or spike classification. Recent approaches to extract SUA are time consuming, mainly due to the requirement of human intervention at various stages of spike sorting pipeline. Lack of standardization is another drawback of the current available approaches. Therefore, in this study we proposed a standard spike sorter: SpikeDeep-Classifier, a fully automatic spike sorting algorithm. Approach. We proposed a novel spike sorting pipeline, based on a set of supervised and unsupervised learning algorithms. We used supervised, deep learning-based algorithms for extracting meaningful channels and removing background activities (noise) from the extracted channels. We also showed that the process of clustering becomes straight-forward, once the noise/artifact is completely removed from the data. Therefore, in the next stage, we applied a simple clustering algorithm (K-mean) with predefined maximum number of clusters. Lastly, we used a similarity-based criterion to keep distinct clusters and merge similar-looking clusters. Main results. We evaluated our algorithm on a dataset collected from two different species (humans and non-human primates (NHPs)) without any retraining. We also validated our algorithm on two publicly available labeled datasets., 33 Pages, 14 Figures, 10 Tables

Various MRI sequences have shown their potential to discriminate parotid gland tumors, including but not limited to T 2‐weighted, postcontrast T 1‐weighted, and diffusion‐weighted images. In this study, we present a fully automatic system for the diagnosis of parotid gland tumors by using deep learning methods trained on multimodal MRI images. We used a two‐dimensional convolution neural network, U‐Net, to segment and classify parotid gland tumors. The U‐Net model was trained with transfer learning, and a specific design of the batch distribution optimized the model accuracy. We also selected five combinations of MRI contrasts as the input data of the neural network and compared the classification accuracy of parotid gland tumors. The results indicated that the deep learning model with diffusion‐related parameters performed better than those with structural MR images. The performance results (n = 85) of the diffusion‐based model were as follows: accuracy of 0.81, 0.76, and 0.71, sensitivity of 0.83, 0.63, and 0.33, and specificity of 0.80, 0.84, and 0.87 for Warthin tumors, pleomorphic adenomas, and malignant tumors, respectively. Combining diffusion‐weighted and contrast‐enhanced T 1‐weighted images did not improve the prediction accuracy. In summary, the proposed deep learning model could classify Warthin tumor and pleomorphic adenoma tumor but not malignant tumor., We used a two‐dimensional convolution neural network to conduct the segmentation and classifications of parotid gland tumors. The results indicated that the deep learning model with diffusion‐related parameters performed better than those with structural MR images. In summary, the deep learning model presented the potential to classify Warthin tumor and pleomorphic adenoma tumor, but not malignant tumor.

BackgroundPositron Emission Tomography (PET) is increasingly utilized in radiomics studies for treatment evaluation purposes. Nevertheless, lesion volume identification in PET images is a critical and still challenging step in the process of radiomics, due to the low spatial resolution and high noise level of PET images. Currently, the biological target volume (BTV) is manually contoured by nuclear physicians, with a time expensive and operator-dependent procedure.This study aims to obtain BTVs from cerebral metastases in patients who underwent L-[11C]methionine (11C-MET) PET, using a fully automatic procedure and to use these BTVs to extract radiomics features to stratify between patients who respond to treatment or not. For these purposes, 31 brain metastases, for predictive evaluation, and 25 ones, for follow-up evaluation after treatment, were delineated using the proposed method. Successively, 11C-MET PET studies and related volumetric segmentations were used to extract 108 features to investigate the potential application of radiomics analysis in patients with brain metastases. A novel statistical system has been implemented for feature reduction and selection, while discriminant analysis was used as a method for feature classification.ResultsFor predictive evaluation, 3 features (asphericity, low-intensity run emphasis, and complexity) were able to discriminate between responder and non-responder patients, after feature reduction and selection. Best performance in patient discrimination was obtained using the combination of the three selected features (sensitivity 81.23%, specificity 73.97%, and accuracy 78.27%) compared to the use of all features. Secondly, for follow-up evaluation, 8 features (SUVmean, SULpeak, SUVmin, SULpeakprod-surface-area, SUVmeanprod-sphericity, surface mean SUV 3, SULpeakprod-sphericity, and second angular moment) were selected with optimal performance in discriminant analysis classification (sensitivity 86.28%, specificity 87.75%, and accuracy 86.57%) outperforming the use of all features.ConclusionsThe proposed system is able i) to extract 108 features for each automatically segmented lesion and ii) to select a sub-panel of 11C-MET PET features (3 and 8 in the case of predictive and follow-up evaluation), with valuable association with patient outcome. We believe that our model can be useful to improve treatment response and prognosis evaluation, potentially allowing the personalization of cancer treatment plans.

3D facial landmarks are known to be diagnostically relevant biometrics for many genetic syndromes. The objective of this study was to extend a state-of-the-art image-based 2D facial landmarking algorithm for the challenging task of 3D landmark identification on subjects with genetic syndromes, who often have moderate to severe facial dysmorphia. The automatic 3D facial landmarking algorithm presented here uses 2D image-based facial detection and landmarking models to identify 12 landmarks on 3D facial surface scans. The landmarking algorithm was evaluated using a test set of 444 facial scans with ground truth landmarks identified by two different human observers. Three hundred and sixty nine of the subjects in the test set had a genetic syndrome that is associated with facial dysmorphology. For comparison purposes, the manual landmarks were also used to initialize a non-linear surface-based registration of a non-syndromic atlas to each subject scan. Compared to the average intra- and inter-observer landmark distances of 1.1 mm and 1.5 mm respectively, the average distance between the manual landmark positions and those produced by the automatic image-based landmarking algorithm was 2.5 mm. The average error of the registration-based approach was 3.1 mm. Comparing the distributions of Procrustes distances from the mean for each landmarking approach showed that the surface registration algorithm produces a systemic bias towards the atlas shape. In summary, the image-based automatic landmarking approach performed well on this challenging test set, outperforming a semi-automatic surface registration approach, and producing landmark errors that are comparable to state-of-the-art 3D geometry-based facial landmarking algorithms evaluated on non-syndromic subjects.