Ultrasound Med Biol. 2024 Feb 17:S0301-5629(24)00026-7. doi: 10.1016/j.ultrasmedbio.2024.01.012. Online ahead of print.

ABSTRACT

OBJECTIVE: Although ultrasound is a common tool for breast cancer screening, its accuracy is often operator-dependent. In this study, we proposed a new automated deep-learning framework that extracts video-based ultrasound data for breast cancer screening.

METHODS: Our framework incorporates DenseNet121, MobileNet, and Xception as backbones for both video- and image-based models. We used data from 3907 patients to train and evaluate the models, which were tested using video- and image-based methods, as well as reader studies with human experts.

RESULTS: This study evaluated 3907 female patients aged 22 to 86 years. The results indicated that the MobileNet video model achieved an AUROC of 0.961 in prospective data testing, surpassing the DenseNet121 video model. In real-world data testing, it demonstrated an accuracy of 92.59%, outperforming both the DenseNet121 and Xception video models, and exceeding the 76.00% to 85.60% accuracy range of human experts. Additionally, the MobileNet video model exceeded the performance of image models and other video models across all evaluation metrics, including accuracy, sensitivity, specificity, F1 score, and AUC. Its exceptional performance, particularly suitable for resource-limited clinical settings, demonstrates its potential for clinical application in breast cancer screening.

CONCLUSIONS: The level of expertise reached by the video models was greater than that achieved by image-based models. We have developed an artificial intelligence framework based on videos that may be able to aid breast cancer diagnosis and alleviate the shortage of experienced experts.

PMID:38369431 | DOI:10.1016/j.ultrasmedbio.2024.01.012

J Emerg Med. 2023 Mar 17:S0736-4679(23)00052-5. doi: 10.1016/j.jemermed.2023.02.005. Online ahead of print.

ABSTRACT

BACKGROUND: The adoption of point-of-care ultrasound (POCUS) has greatly improved the ability to rapidly evaluate unstable emergency department (ED) patients at the bedside. One major use of POCUS is to obtain echocardiograms to assess cardiac function.

OBJECTIVES: We developed EchoNet-POCUS, a novel deep learning system, to aid emergency physicians (EPs) in interpreting POCUS echocardiograms and to reduce operator-to-operator variability.

METHODS: We collected a new dataset of POCUS echocardiogram videos obtained in the ED by EPs and annotated the cardiac function and quality of each video. Using this dataset, we train EchoNet-POCUS to evaluate both cardiac function and video quality in POCUS echocardiograms.

RESULTS: EchoNet-POCUS achieves an area under the receiver operating characteristic curve (AUROC) of 0.92 (0.89-0.94) for predicting whether cardiac function is abnormal and an AUROC of 0.81 (0.78-0.85) for predicting video quality.

CONCLUSIONS: EchoNet-POCUS can be applied to bedside echocardiogram videos in real time using commodity hardware, as we demonstrate in a prospective pilot study.

PMID:38369413 | DOI:10.1016/j.jemermed.2023.02.005

Neuroscience. 2024 Feb 16:S0306-4522(24)00061-7. doi: 10.1016/j.neuroscience.2024.02.011. Online ahead of print.

ABSTRACT

Brain Computer Interface (BCI) is a highly promising human-computer interaction method that can utilize brain signals to control external devices. BCI based on functional near-infrared spectroscopy (fNIRS) is considered a relatively new and promising paradigm. fNIRS is a technique of measuring functional changes in cerebral hemodynamics. It detects changes in the hemodynamic activity of the cerebral cortex by measuring oxyhemoglobin and deoxyhemoglobin (HbR) concentrations and inversely predicts the neural activity of the brain. At the present time, Deep learning (DL) methods have not been widely used in fNIRS decoding, and there are fewer studies considering both spatial and temporal dimensions for fNIRS classification. To solve these problems, we proposed an end-to-end hybrid neural network for feature extraction of fNIRS. The method utilizes a spatial-temporal convolutional layer for automatic extraction of temporally valid information and uses a spatial attention mechanism to extract spatially localized information. A temporal convolutional network (TCN) is used to further utilize the temporal information of fNIRS before the fully connected layer. We validated our approach on a publicly available dataset including 29 subjects, including left-hand and right-hand motor imagery (MI), mental arithmetic (MA), and a baseline task. The results show that the method has few training parameters and high accuracy, providing a meaningful reference for BCI development.

PMID:38369007 | DOI:10.1016/j.neuroscience.2024.02.011

Talanta. 2024 Feb 7;272:125757. doi: 10.1016/j.talanta.2024.125757. Online ahead of print.

ABSTRACT

Currently, it is of great urgency to develop a rapid pre-classification and screening method for suspected drugs as the constantly springing up of new psychoactive substances. In most researches, psychoactive substances classification approaches depended on the similar chemical structures and pharmacological action with known drugs. Such approaches could not face the complicated circumstance of emerging new psychoactive substances. Herein, mass spectrometry imaging and convolutional neural networks (CNN) were used for preliminary screening and pre-classification of suspected psychoactive substances. Mass spectrometry imaging was performed simultaneously on two brain slices as one was from blank group and another one was from psychoactive substance-induced group. Then, fused neurotransmitter variation mass spectrometry images (Nv-MSIs) reflecting the difference of neurotransmitters between two slices were achieved through two homemade programs. A CNN model was developed to classify the Nv-MSIs. Compared with traditional classification methods, CNN achieved better estimation accuracy and required minimal data preprocessing. Also, the specific region on Nv-MSIs and weight of each neurotransmitter that affected the classification most could be unraveled by CNN. Finally, the method was successfully applied to assist the identification of a new psychoactive substance seized recently. This sample was identified as cannabinoids, which greatly promoted the screening process.

PMID:38368831 | DOI:10.1016/j.talanta.2024.125757

Spectrochim Acta A Mol Biomol Spectrosc. 2024 Feb 15;312:124030. doi: 10.1016/j.saa.2024.124030. Online ahead of print.

ABSTRACT

Whole slide imaging (WSI) of Hematoxylin and Eosin-stained biopsy specimens has been used to predict chemoradiotherapy (CRT) response and overall survival (OS) of esophageal squamous cell carcinoma (ESCC) patients. This retrospective study collected 279 specimens in 89 non-surgical ESCC patients through endoscopic biopsy between January 2010 and January 2019. These patients were divided into a CRT response group (CR + PR group) and a CRT non-response group (SD + PD group). The WSIs have segmented approximately 1,206,000 non-overlapping patches. Two experienced pathologists manually delineated the eight types of tissues on 32 WSIs, including esophagus tumor cell (TUM), cancer-associated stroma (CAS), normal epithelium layer (NEL), smooth muscle (MUS), lymphocytes (LYM), Red cells (RED), debris (DEB), uneven areas (UNE). The chemoradiotherapy response prediction models were built using maximum relevance-minimum redundancy (MRMR) feature selection and least absolute shrinkage and selection operator (LASSO) regression. However, pathological features with p < 0.1 were selected and integrated to be further screened using a LASSO Cox regression model to build a multivariate Cox proportional hazards model for predicting the OS. The testing accuracy of the tissue classification model was 91.3 %. The pathological model created using two CAS in-depth features and eight TUM in-depth features performed best for the prediction of treatment response and achieved an AUC of 0.744. For the prediction of OS, the testing AUC of this model at one year and three years were 0.675 and 0.870, respectively. The TUM model showed the highest AUC at one year (0.712). With its high accuracy rate, the deep learning model has the potential to transform from bench to bedside in clinical practice, improve patient's quality of life, and prolong the OS rate.

PMID:38368818 | DOI:10.1016/j.saa.2024.124030

Water Res. 2024 Feb 14;253:121314. doi: 10.1016/j.watres.2024.121314. Online ahead of print.

ABSTRACT

Dam (reservoir)-induced alterations of flow and water temperature regimes can threaten downstream fish habitats and native aquatic ecosystems. Alleviating the negative environmental impacts of dam-reservoir and balancing the multiple purposes of reservoir operation have attracted wide attention. While previous studies have incorporated ecological flow requirements in reservoir operation strategies, a comprehensive analysis of trade-offs among hydropower benefits, ecological flow, and ecological water temperature demands is lacking. Hence, this study develops a multi-objective ecological scheduling model, considering total power generation, ecological flow guarantee index, and ecological water temperature guarantee index simultaneously. The model is based on an integrated multi-objective simulation-optimization (MOSO) framework which is applied to Three Gorges Reservoir. To that end, first, a hybrid long short-term memory and one-dimensional convolutional neural network (LSTM_1DCNN) model is utilized to simulate the dam discharge temperature. Then, an improved epsilon multi-objective ant colony optimization for continuous domain algorithm (ε-MOACOR) is proposed to investigate the trade-offs among the competing objectives. Results show that LSTM _1DCNN outperforms other competing models in predicting dam discharge temperature. The conflicts among economic and ecological objectives are often prominent. The proposed ε-MOACOR has potential in resolving such conflicts and has high efficiency in solving multi-objective benchmark tests as well as reservoir optimization problem. More realistic and pragmatic Pareto-optimal solutions for typical dry, normal and wet years can be generated by the MOSO framework. The ecological water temperature guarantee index objective, which should be considered in reservoir operation, can be improved as inflow discharge increases or the temporal distribution of dam discharge volume becomes more uneven.

PMID:38368733 | DOI:10.1016/j.watres.2024.121314

Comput Med Imaging Graph. 2024 Feb 8;113:102348. doi: 10.1016/j.compmedimag.2024.102348. Online ahead of print.

ABSTRACT

Recurrent inference machines (RIM), a deep learning model that learns an iterative scheme for reconstructing sparsely sampled MRI, has been shown able to perform well on accelerated 2D and 3D MRI scans, learn from small datasets and generalize well to unseen types of data. Here we propose the dynamic recurrent inference machine (DRIM) for reconstructing sparsely sampled 4D MRI by exploiting correlations between respiratory states. The DRIM was applied to a 4D protocol for MR-guided radiotherapy of liver lesions based on repetitive interleaved coronal 2D multi-slice T2-weighted acquisitions. We demonstrate with an ablation study that the DRIM outperforms the RIM, increasing the SSIM score from about 0.89 to 0.95. The DRIM allowed for an approximately 2.7 times faster scan time than the current clinical protocol with only a slight loss in image sharpness. Correlations between slice locations can also be used, but were found to be of less importance, as were a majority of tested variations in network architecture, as long as the respiratory states are processed by the network. Through cross-validation, the DRIM is also shown to be robust in terms of training data. We further demonstrate a good performance across a large range of subsampling factors, and conclude through an evaluation by a radiation oncologist that reconstructed images of the liver contour and inner structures are of a clinically acceptable standard at acceleration factors 10x and 8x, respectively. Finally, we show that binning the data with respect to respiratory states prior to reconstruction comes at a slight cost to reconstruction quality, but at greater speed of the overall protocol.

PMID:38368665 | DOI:10.1016/j.compmedimag.2024.102348

J Appl Clin Med Phys. 2024 Feb 18:e14304. doi: 10.1002/acm2.14304. Online ahead of print.

ABSTRACT

BACKGROUND: Artifacts from implantable cardioverter defibrillators (ICDs) are a challenge to magnetic resonance imaging (MRI)-guided radiotherapy (MRgRT).

PURPOSE: This study tested an unsupervised generative adversarial network to mitigate ICD artifacts in balanced steady-state free precession (bSSFP) cine MRIs and improve image quality and tracking performance for MRgRT.

METHODS: Fourteen healthy volunteers (Group A) were scanned on a 0.35 T MRI-Linac with and without an MR conditional ICD taped to their left pectoral to simulate an implanted ICD. bSSFP MRI data from 12 of the volunteers were used to train a CycleGAN model to reduce ICD artifacts. The data from the remaining two volunteers were used for testing. In addition, the dataset was reorganized three times using a Leave-One-Out scheme. Tracking metrics [Dice similarity coefficient (DSC), target registration error (TRE), and 95 percentile Hausdorff distance (95% HD)] were evaluated for whole-heart contours. Image quality metrics [normalized root mean square error (nRMSE), peak signal-to-noise ratio (PSNR), and multiscale structural similarity (MS-SSIM) scores] were evaluated. The technique was also tested qualitatively on three additional ICD datasets (Group B) including a patient with an implanted ICD.

RESULTS: For the whole-heart contour with CycleGAN reconstruction: 1) Mean DSC rose from 0.910 to 0.935; 2) Mean TRE dropped from 4.488 to 2.877 mm; and 3) Mean 95% HD dropped from 10.236 to 7.700 mm. For the whole-body slice with CycleGAN reconstruction: 1) Mean nRMSE dropped from 0.644 to 0.420; 2) Mean MS-SSIM rose from 0.779 to 0.819; and 3) Mean PSNR rose from 18.744 to 22.368. The three Group B datasets evaluated qualitatively displayed a reduction in ICD artifacts in the heart.

CONCLUSION: CycleGAN-generated reconstructions significantly improved both tracking and image quality metrics when used to mitigate artifacts from ICDs.

PMID:38368615 | DOI:10.1002/acm2.14304

Comput Biol Med. 2024 Feb 13;171:108136. doi: 10.1016/j.compbiomed.2024.108136. Online ahead of print.

ABSTRACT

BACKGROUND: Mesenchymal epithelial transformation (MET) is a key molecular target for diagnosis and treatment of non-small cell lung cancer (NSCLC). The corresponding molecularly targeted therapeutics have been approved by Food and Drug Administration (FDA), achieving promising results. However, current detection of MET dysregulation requires biopsy and gene sequencing, which is invasive, time-consuming and difficult to obtain tumor samples.

METHODS: To address the above problems, we developed a noninvasive and convenient deep learning (DL) model based on Computed tomography (CT) imaging data for prediction of MET dysregulation. We introduced the unsupervised algorithm RK-net for automated image processing and utilized the MedSAM large model to achieve automated tissue segmentation. Based on the processed CT images, we developed a DL model (METnet). The model based on the grouped convolutional block. We evaluated the performance of the model over the internal test dataset using the area under the receiver operating characteristic curve (AUROC) and accuracy. We conducted subgroup analysis on the basis of clinical data of the lung cancer patients and compared the performance of the model in different subgroups.

RESULTS: The model demonstrated a good discriminative ability over the internal test dataset. The accuracy of METnet was 0.746 with an AUC value of 0.793 (95% CI 0.714-0.871). The subgroup analysis revealed that the model exhibited similar performance across different subgroups.

CONCLUSIONS: METnet realizes prediction of MET dysregulation in NSCLC, holding promise for guiding precise tumor diagnosis and treatment at the molecular level.

PMID:38367451 | DOI:10.1016/j.compbiomed.2024.108136

Spectrochim Acta A Mol Biomol Spectrosc. 2024 Feb 12;312:124036. doi: 10.1016/j.saa.2024.124036. Online ahead of print.

ABSTRACT

Raman spectroscopy has become a powerful analytical tool highly demanded in many applications such as microorganism sample analysis, food quality control, environmental science, and pharmaceutical analysis, owing to its non-invasiveness, simplicity, rapidity and ease of use. Among them, quantitative research using Raman spectroscopy is a crucial application field of spectral analysis. However, the entire process of quantitative modeling largely relies on the extraction of effective spectral features, particularly for measurements on complex samples or in environments with poor spectral signal quality. In this paper, we propose a method of utilizing a spectral encoder to extract effective spectral features, which can significantly enhance the reliability and precision of quantitative analysis. We built a latent encoded feature regression model; in the process of utilizing the autoencoder for reconstructing the spectrometer output, the latent feature obtained from the intermediate bottleneck layer is extracted. Then, these latent features are fed into a deep regression model for component concentration prediction. Through detailed ablation and comparative experiments, our proposed model demonstrates superior performance to common methods on single-component and multi-component mixture datasets, remarkably improving regression precision while without needing user-selected parameters and eliminating the interference of irrelevant and redundant information. Furthermore, in-depth analysis reveals that latent encoded feature possesses strong nonlinear feature representation capabilities, low computational costs, wide adaptability, and robustness against noise interference. This highlights its effectiveness in spectral regression tasks and indicates its potential in other application fields. Sufficient experimental results show that our proposed method provides a novel and effective feature extraction approach for spectral analysis, which is simple, suitable for various methods, and can meet the measurement needs of different real-world scenarios.

PMID:38367343 | DOI:10.1016/j.saa.2024.124036

Abdom Radiol (NY). 2024 Feb 17. doi: 10.1007/s00261-023-04172-w. Online ahead of print.

ABSTRACT

INTRODUCTION: Accurate diagnosis and treatment of kidney tumors greatly benefit from automated solutions for detection and classification on MRI. In this study, we explore the application of a deep learning algorithm, YOLOv7, for detecting kidney tumors on contrast-enhanced MRI.

MATERIAL AND METHODS: We assessed the performance of YOLOv7 tumor detection on excretory phase MRIs in a large institutional cohort of patients with RCC. Tumors were segmented on MRI using ITK-SNAP and converted to bounding boxes. The cohort was randomly divided into ten benchmarks for training and testing the YOLOv7 algorithm. The model was evaluated using both 2-dimensional and a novel in-house developed 2.5-dimensional approach. Performance measures included F1, Positive Predictive Value (PPV), Sensitivity, F1 curve, PPV-Sensitivity curve, Intersection over Union (IoU), and mean average PPV (mAP).

RESULTS: A total of 326 patients with 1034 tumors with 7 different pathologies were analyzed across ten benchmarks. The average 2D evaluation results were as follows: Positive Predictive Value (PPV) of 0.69 ± 0.05, sensitivity of 0.39 ± 0.02, and F1 score of 0.43 ± 0.03. For the 2.5D evaluation, the average results included a PPV of 0.72 ± 0.06, sensitivity of 0.61 ± 0.06, and F1 score of 0.66 ± 0.04. The best model performance demonstrated a 2.5D PPV of 0.75, sensitivity of 0.69, and F1 score of 0.72.

CONCLUSION: Using computer vision for tumor identification is a cutting-edge and rapidly expanding subject. In this work, we showed that YOLOv7 can be utilized in the detection of kidney cancers.

PMID:38368481 | DOI:10.1007/s00261-023-04172-w

Nat Commun. 2024 Feb 17;15(1):1456. doi: 10.1038/s41467-024-45856-1.

ABSTRACT

Capturing fine spatial, spectral, and temporal information of the scene is highly desirable in many applications. However, recording data of such high dimensionality requires significant transmission bandwidth. Current computational imaging methods can partially address this challenge but are still limited in reducing input data throughput. In this paper, we report a video-rate hyperspectral imager based on a single-pixel photodetector which can achieve high-throughput hyperspectral video recording at a low bandwidth. We leverage the insight that 4-dimensional (4D) hyperspectral videos are considerably more compressible than 2D grayscale images. We propose a joint spatial-spectral capturing scheme encoding the scene into highly compressed measurements and obtaining temporal correlation at the same time. Furthermore, we propose a reconstruction method relying on a signal sparsity model in 4D space and a deep learning reconstruction approach greatly accelerating reconstruction. We demonstrate reconstruction of 128 × 128 hyperspectral images with 64 spectral bands at more than 4 frames per second offering a 900× data throughput compared to conventional imaging, which we believe is a first-of-its kind of a single-pixel-based hyperspectral imager.

PMID:38368402 | DOI:10.1038/s41467-024-45856-1

Diagn Interv Imaging. 2024 Feb 16:S2211-5684(24)00036-6. doi: 10.1016/j.diii.2024.02.001. Online ahead of print.

ABSTRACT

PURPOSE: The purpose of this study was to evaluate the ability of ultra-high-resolution computed tomography (UHR-CT) to assess stapes and chorda tympani nerve anatomy using a deep learning (DLR), a model-based, and a hybrid iterative reconstruction algorithm compared to simulated conventional CT.

MATERIALS AND METHODS: CT acquisitions were performed with a Mercury 4.0 phantom. Images were acquired with a 1024 × 1024 matrix and a 0.25 mm slice thickness and reconstructed using DLR, model-based, and hybrid iterative reconstruction algorithms. To simulate conventional CT, images were also reconstructed with a 512 × 512 matrix and a 0.5 mm slice thickness. Spatial resolution, noise power spectrum, and objective high-contrast detectability were compared. Three radiologists evaluated the clinical acceptability of these algorithms by assessing the thickness and image quality of the stapes footplate and superstructure elements, as well as the image quality of the chorda tympani nerve bony and tympanic segments using a 5-point confidence scale on 13 temporal bone CT examinations reconstructed with the four algorithms.

RESULTS: UHR-CT provided higher spatial resolution than simulated conventional CT at the penalty of higher noise. DLR and model-based iterative reconstruction provided better noise reduction than hybrid iterative reconstruction, and DLR had the highest detectability index, regardless of the dose level. All stapedial structure thicknesses were thinner using UHR-CT by comparison with conventional simulated CT (P < 0.009). DLR showed the best visualization scores compared to the other reconstruction algorithms (P < 0.032).

CONCLUSION: UHR-CT with DLR results in less noise than UHR-CT with hybrid iterative reconstruction and significantly improves stapes and tympanic chorda tympani nerve depiction compared to simulated conventional CT and UHR-CT with iterative reconstruction.

PMID:38368178 | DOI:10.1016/j.diii.2024.02.001

Diagn Interv Imaging. 2024 Feb 16:S2211-5684(24)00035-4. doi: 10.1016/j.diii.2024.01.010. Online ahead of print.

ABSTRACT

PURPOSE: The purpose of this study was to evaluate the capabilities of photon-counting (PC) CT combined with artificial intelligence-derived coronary computed tomography angiography (PC-CCTA) stenosis quantification and fractional flow reserve prediction (FFRai) for the assessment of coronary artery disease (CAD) in transcatheter aortic valve replacement (TAVR) work-up.

MATERIALS AND METHODS: Consecutive patients with severe symptomatic aortic valve stenosis referred for pre-TAVR work-up between October 2021 and June 2023 were included in this retrospective tertiary single-center study. All patients underwent both PC-CCTA and ICA within three months for reference standard diagnosis. PC-CCTA stenosis quantification (at 50% level) and FFRai (at 0.8 level) were predicted using two deep learning models (CorEx, Spimed-AI). Diagnostic performance for global CAD evaluation (at least one significant stenosis ≥ 50% or FFRai ≤ 0.8) was assessed.

RESULTS: A total of 260 patients (138 men, 122 women) with a mean age of 78.7 ± 8.1 (standard deviation) years (age range: 51-93 years) were evaluated. Significant CAD on ICA was present in 126/260 patients (48.5%). Per-patient sensitivity, specificity, positive predictive value, negative predictive value, and diagnostic accuracy were 96.0% (95% confidence interval [CI]: 91.0-98.7), 68.7% (95% CI: 60.1-76.4), 74.3 % (95% CI: 69.1-78.8), 94.8% (95% CI: 88.5-97.8), and 81.9% (95% CI: 76.7-86.4) for PC-CCTA, and 96.8% (95% CI: 92.1-99.1), 87.3% (95% CI: 80.5-92.4), 87.8% (95% CI: 82.2-91.8), 96.7% (95% CI: 91.7-98.7), and 91.9% (95% CI: 87.9-94.9) for FFRai. Area under the curve of FFRai was 0.92 (95% CI: 0.88-0.95) compared to 0.82 for PC-CCTA (95% CI: 0.77-0.87) (P < 0.001). FFRai-guidance could have prevented the need for ICA in 121 out of 260 patients (46.5%) vs. 97 out of 260 (37.3%) using PC-CCTA alone (P < 0.001).

CONCLUSION: Deep learning-based photon-counting FFRai evaluation improves the accuracy of PC-CCTA ≥ 50% stenosis detection, reduces the need for ICA, and may be incorporated into the clinical TAVR work-up for the assessment of CAD.

PMID:38368176 | DOI:10.1016/j.diii.2024.01.010

Acad Radiol. 2024 Feb 16:S1076-6332(24)00050-3. doi: 10.1016/j.acra.2024.01.029. Online ahead of print.

ABSTRACT

RATIONALE AND OBJECTIVES: Accurate and efficient estimation of patient height and weight is crucial to ensure patient safety and optimize the quality of magnetic resonance imaging (MRI) procedures. Several height and weight estimation methods have been proposed for use in adult patient management, but none is widely established. Estimation by the medical technologists for radiology (MTR) based on personal experience remains to be the most common method. This study aimed to compare a novel deep learning (DL)-based 3-dimensional (3D) camera estimation method to MTR staff in terms of estimation accuracy.

METHODS: A retrospective study was conducted to compare the accuracy of height and weight estimation with a DL-based 3D camera algorithm to the accuracy of height and weight estimation by the MTR. Depth images of the patients were captured during the regular imaging workflow on a low field 0.55 T MRI scanner (MAGNETOM Free.Max, Siemens Healthineers, Erlangen, Germany) and then processed retrospectively. Depth images of a total of 161 patients were used to validate the accuracy of the height and weight estimation algorithm. The accuracy of each estimation method was evaluated by computing the proportions of the estimates within 5% and 15% of actual height (PH05, PH15) and within 10% and 20% of actual weight (PW10, PW20). An acceptable accuracy for height estimation was predetermined to be PH05 = 95% and PH15 = 99% and an acceptable accuracy for weight estimation was predetermined to be PW10 = 70% and PW20 = 95%. The bias in height and weight estimation was measured by the mean absolute percentage error (MAPE).

RESULTS: The retrospective study included 161 adult patients. For 148/161 patients complying with inclusion criteria, DL-based 3D camera algorithm outperformed the MTR in estimating the patient's height and weight in term of accuracy (3D camera: PH05 =98.6%, PH15 =100%, PW10 =85.1%, PW20 =95.9%; MTR: PH05 =92.5%, PH15 =100%, PW10 =75.0%, PW20 =93.2%). MTR had a slightly higher bias in their estimates compared to the DL-based 3D camera algorithm (3D camera: MAPE height=1.8%, MAPE weight=5.6%, MTR: MAPE height=2.2%, MAPE weight=7.5%) CONCLUSION: This study has demonstrated that the estimation of the patient's height and weight by a DL-based 3D camera algorithm is accurate and robust. It has the potential to complement the regular MRI workflows, by providing further automation during patient registration.

PMID:38368163 | DOI:10.1016/j.acra.2024.01.029

J Dent. 2024 Feb 15:104891. doi: 10.1016/j.jdent.2024.104891. Online ahead of print.

ABSTRACT

OBJECTIVES: To evaluate the diagnostic performance of three versions of a deep-learning convolutional neural network in terms of object detection and segmentation using a multiclass panoramic radiograph dataset.

METHODS: A total of 600 orthopantomographies were randomly selected for this study and manually annotated by a single operator using an image annotation tool (COCO Annotator v.11.0.1) to establish ground truth. The annotation classes included teeth, maxilla, mandible, inferior alveolar nerve, dento- and implant-supported crowns/pontics, endodontic treatment, resin-based restorations, metallic restorations, and implants. The dataset was then divided into training, validation, and testing subsets, which were used to train versions 5, 7, and 8 of You Only Look Once (YOLO) Neural Network. Results were stored, and a posterior performance analysis was carried out by calculating the precision (P), recall (R), F1 Score, Intersection over Union (IoU), and mean average precision (mAP) at 0.5 and 0.5-0.95 thresholds. The confusion matrix and recall precision graphs were also sketched.

RESULTS: YOLOv5s showed an improvement in object detection results with an average R=0.634, P=0.781, mAP0.5=0.631, and mAP0.5-0.95=0.392. YOLOv7m achieved the best object detection results with average R=0.793, P=0.779, mAP0.5=0.740, and mAP0.5-0.95=0,481. For object segmentation, YOLOv8m obtained the best average results (R= 0.589, P=0.755, mAP0.5=0.591, and mAP0.5-0.95= 0.272).

CONCLUSIONS: YOLOv7m was better suited for object detection, while YOLOv8m demonstrated superior performance in object segmentation. The most frequent error in object detection was related to background classification. Conversely, in object segmentation, there is a tendency to misclassify True Positives across different dental treatment categories.

CLINICAL SIGNIFICANCE: General diagnostic and treatment decisions based on panoramic radiographs can be enhanced using new artificial intelligence-based tools. Nevertheless, the reliability of these neural networks should be subjected to training and validation to ensure their generalizability.

PMID:38367827 | DOI:10.1016/j.jdent.2024.104891

Curr Biol. 2024 Feb 9:S0960-9822(24)00103-9. doi: 10.1016/j.cub.2024.01.066. Online ahead of print.

ABSTRACT

To detect and escape looming threats, night-flying insects must rely on other senses than vision alone. Nocturnal mosquitoes can evade looming objects in the dark, but how they achieve this is still unknown. Here, we show how night-active female malaria mosquitoes escape from rapidly looming objects that simulate defensive actions of blood-hosts. First, we quantified the escape performance of flying mosquitoes from an event-triggered mechanical swatter, showing that mosquitoes use swatter-induced airflow to increase their escape success. Secondly, we used high-speed videography and deep-learning-based tracking to analyze escape flights in detail, showing that mosquitoes use banked turns to evade the threat. By combining escape kinematics data with numerical simulations of attacker-induced airflow and a mechanistic movement model, we unraveled how mosquitoes control these banked evasive maneuvers: they actively steer away from the danger, and then passively travel with the bow wave produced by the attacker. Our results demonstrate that night-flying mosquitoes can detect looming objects when visual cues are minimal, suggesting that they use attacker-induced airflow both to detect the danger and as a fluid medium to move with away from the threat. This shows that escape strategies of flying insects are more complex than previous visually induced escape flight studies suggest. As most insects are of similar or smaller sizes than mosquitoes, comparable escape strategies are expected among millions of flying insect species. The here-observed escape maneuvers are distinct from those of mosquitoes escaping from odor-baited traps, thus providing new insights for the development of novel trapping techniques for integrative vector management.

PMID:38367617 | DOI:10.1016/j.cub.2024.01.066

J Infect Public Health. 2024 Jan 22;17(4):559-572. doi: 10.1016/j.jiph.2024.01.013. Online ahead of print.

ABSTRACT

Internet of Medical Things (IoMT) is an emerging subset of Internet of Things (IoT), often called as IoT in healthcare, refers to medical devices and applications with internet connectivity, is exponentially gaining researchers' attention due to its wide-ranging applicability in biomedical systems for Smart Healthcare systems. IoMT facilitates remote health biomedical system and plays a crucial role within the healthcare industry to enhance precision, reliability, consistency and productivity of electronic devices used for various healthcare purposes. It comprises a conceptualized architecture for providing information retrieval strategies to extract the data from patient records using sensors for biomedical analysis and diagnostics against manifold diseases to provide cost-effective medical solutions, quick hospital treatments, and personalized healthcare. This article provides a comprehensive overview of IoMT with special emphasis on its current and future trends used in biomedical systems, such as deep learning, machine learning, blockchains, artificial intelligence, radio frequency identification, and industry 5.0.

PMID:38367570 | DOI:10.1016/j.jiph.2024.01.013

Biosens Bioelectron. 2024 Feb 12;251:116128. doi: 10.1016/j.bios.2024.116128. Online ahead of print.

ABSTRACT

Early diagnosis of Alzheimer's disease is crucial to stall the deterioration of brain function, but conventional diagnostic methods require complicated analytical procedures or inflict acute pain on the patient. Then, label-free Surface-enhanced Raman spectroscopy (SERS) analysis of blood-based biomarkers is a convenient alternative to rapidly obtain spectral information from biofluids. However, despite the rapid acquisition of spectral information from biofluids, it is challenging to distinguish spectral features of biomarkers due to interference from biofluidic components. Here, we introduce a deep learning-assisted, SERS-based platform for separate analysis of blood-based amyloid β (1-42) and metabolites, enabling the diagnosis of Alzheimer's disease. SERS substrates consisting of Au nanowire arrays are fabricated and functionalized in two characteristic ways to compare the validity of different Alzheimer's disease biomarkers measured on our SERS system. The 6E10 antibody is immobilized for the capture of amyloid β (1-42) and analysis of its oligomerization process, while various self-assembled monolayers are attached for different dipole interactions with blood-based metabolites. Ultimately, SERS spectra of blood plasma of Alzheimer's disease patients and human controls are measured on the substrates and classified via advanced deep learning techniques that automatically extract informative features to learn generalizable representations. Accuracies up to 99.5% are achieved for metabolite-based analyses, which are verified with an explainable artificial intelligence technique that identifies key spectral features used for classification and for deducing significant biomarkers.

PMID:38367567 | DOI:10.1016/j.bios.2024.116128

J Environ Manage. 2024 Feb 16;354:120313. doi: 10.1016/j.jenvman.2024.120313. Online ahead of print.

ABSTRACT

This paper addresses the critical environmental issue of effectively managing construction and demolition waste (CDW), which has seen a global surge due to rapid urbanization. With the advent of deep learning-based computer vision, this study focuses on improving intelligent identification of valuable recyclables from cluttered and heterogeneous CDW streams in material recovery facilities (MRFs) by optimally leveraging both visual and spatial features (depth). A high-quality CDW RGB-D dataset was curated to capture MRF stream complexities often overlooked in prior studies, and comprises over 3500 images for each modality and more than 160,000 dense object instances of diverse CDW materials with high resource value. In contrast to former studies which directly concatenate RGB and depth features, this study introduces a new depth fusion strategy that utilizes computationally efficient convolutional operations at the end of the conventional waste segmentation architecture to effectively fuse colour and depth information. This avoids cross-modal interference and maximizes the use of distinct information present in the two different modalities. Despite the high clutter and diversity of waste objects, the proposed RGB-DL architecture achieves a 13% increase in segmentation accuracy and a 36% reduction in inference time when compared to the direct concatenation of features. The findings of this study emphasize the benefit of effectively incorporating geometrical features to complement visual cues. This approach helps to deal with the cluttered and varied nature of CDW streams, enhancing automated waste recognition accuracy to improve resource recovery in MRFs. This, in turn, promotes intelligent solid waste management for efficiently managing environmental concerns.

PMID:38367501 | DOI:10.1016/j.jenvman.2024.120313