RESUMO
The prevalence of brain tumor disorders is currently a global issue. In general, radiography, which includes a large number of images, is an efficient method for diagnosing these life-threatening disorders. The biggest issue in this area is that it takes a radiologist a long time and is physically strenuous to look at all the images. As a result, research into developing systems based on machine learning to assist radiologists in diagnosis continues to rise daily. Convolutional neural networks (CNNs), one type of deep learning approach, have been pivotal in achieving state-of-the-art results in several medical imaging applications, including the identification of brain tumors. CNN hyperparameters are typically set manually for segmentation and classification, which might take a while and increase the chance of using suboptimal hyperparameters for both tasks. Bayesian optimization is a useful method for updating the deep CNN's optimal hyperparameters. The CNN network, however, can be considered a "black box" model because of how difficult it is to comprehend the information it stores because of its complexity. Therefore, this problem can be solved by using Explainable Artificial Intelligence (XAI) tools, which provide doctors with a realistic explanation of CNN's assessments. Implementation of deep learning-based systems in real-time diagnosis is still rare. One of the causes could be that these methods don't quantify the Uncertainty in the predictions, which could undermine trust in the AI-based diagnosis of diseases. To be used in real-time medical diagnosis, CNN-based models must be realistic and appealing, and uncertainty needs to be evaluated. So, a novel three-phase strategy is proposed for segmenting and classifying brain tumors. Segmentation of brain tumors using the DeeplabV3+ model is first performed with tuning of hyperparameters using Bayesian optimization. For classification, features from state-of-the-art deep learning models Darknet53 and mobilenetv2 are extracted and fed to SVM for classification, and hyperparameters of SVM are also optimized using a Bayesian approach. The second step is to understand whatever portion of the images CNN uses for feature extraction using XAI algorithms. Using confusion entropy, the Uncertainty of the Bayesian optimized classifier is finally quantified. Based on a Bayesian-optimized deep learning framework, the experimental findings demonstrate that the proposed method outperforms earlier techniques, achieving a 97â¯% classification accuracy and a 0.98 global accuracy.
RESUMO
Vanadium-based phosphate cathode materials (e.g., K3V2(PO4)3) have attracted widespread concentration in cathode materials in potassium-ion batteries owing to their stable structure but suffer from low capacity and poor conductivity. In this work, an element doping strategy is applied to promote its electrochemical performance so that K3.2V2.8Mn0.2(PO4)4/C is prepared via a simple sol-gel method. The heterovalent Mn2+ is introduced to stimulated multiple electron reactions to improve conductivity and capacity, as well as interlayer spacing. Galvanostatic intermittent titration technique (GITT) and in situ X-ray diffraction results further confirm that Mn-doping in the original electrode can obtain superior electrode process kinetics and structural stability. The prepared K3.2V2.8Mn0.2(PO4)4/C exhibits a high-capacity retention of 80.8% after 1 500 cycles at 2 C and an impressive rate capability, with discharge capacities of 87.6 at 0.2 C and 45.4 mA h g-1 at 5 C, which is superior to the majority of reported vanadium-based phosphate cathode materials. When coupled K3.2V2.8Mn0.2(PO4)4/C cathode with commercial porous carbon (PC) anode as the full cell, a prominent energy density of 175 Wh kg-1 is achieved based on the total active mass. Overall, this study provides an effective strategy for meliorating the cycling stability and capacity of the polyanion cathodes for KIB.
RESUMO
Assessing crop seed phenotypic traits is essential for breeding innovations and germplasm enhancement. However, the tough outer layers of thin-shelled seeds present significant challenges for traditional methods aimed at the rapid assessment of their internal structures and quality attributes. This study explores the potential of combining terahertz (THz) time-domain spectroscopy and imaging with semantic segmentation models for the rapid and non-destructive examination of these traits. A total of 120 watermelon seed samples from three distinct varieties, were curated in this study, facilitating a comprehensive analysis of both their outer layers and inner kernels. Utilizing a transmission imaging modality, THz spectral images were acquired and subsequently reconstructed employing a correlation coefficient method. Deep learning-based SegNet and DeepLab V3+ models were employed for automatic tissue segmentation. Our research revealed that DeepLab V3+ significantly surpassed SegNet in both speed and accuracy. Specifically, DeepLab V3+ achieved a pixel accuracy of 96.69 % and an intersection over the union of 91.3 % for the outer layer, with the inner kernel results closely following. These results underscore the proficiency of DeepLab V3+ in distinguishing between the seed coat and kernel, thereby furnishing precise phenotypic trait analyses for seeds with thin shells. Moreover, this study accentuates the instrumental role of deep learning technologies in advancing agricultural research and practices.
RESUMO
PURPOSE: Clinical next-generation sequencing is an effective approach for identifying pathogenic sequence variants that are medically actionable for participants and families but are not associated with the participant's primary diagnosis. These variants are called secondary findings (SFs). According to the literature, there is no report of the types and frequencies of SFs in a large pediatric cohort which includes substantial African-American participants. We sought to investigate the types (including American College of Medical Genetics and Genomics [ACMG] and non-ACMG recommended gene lists), frequencies, and rates of SFs, as well as the effects of SF disclosure on the participants and families of a large pediatric cohort at the Center for Applied Genomics at The Children's Hospital of Philadelphia (CHOP). METHODS: We systematically identified pathogenic (P) and likely pathogenic (LP) variants in established disease-causing genes, adhering to ACMG v3.2 secondary finding guidelines and beyond. For non-ACMG secondary findings, akin to incidental findings in clinical settings, we utilized a set of criteria focusing on pediatric onset, high penetrance, moderate to severe phenotypes, and the clinical actionability of the variants. This criteria-based approach was applied rather than using a fixed gene list to ensure that the variants identified are likely to impact participant health significantly. To identify and categorize these variants, we employed a clinical-grade variant classification standard per ACMG/AMP recommendations; additionally, we conducted a detailed literature search to ensure a comprehensive exploration of potential secondary findings relevant to pediatric participants. RESULTS: We report a distinctive distribution of 1,464 P/LP SF variants in 16,713 participants. There were 427 unique variants in ACMG genes and 265 in non-ACMG genes. The most frequently mutated genes among the ACMG and non-ACMG gene lists were TTR (41.6%) and CHEK2 (7.16%), respectively. Overall, variants of possible medical importance were found in 8.76% of participants in both ACMG (5.81%) and non-ACMG (2.95%) genes.
RESUMO
Highly complex cognitive works require more brain power. The productivity of a person suffers due to this strain, which is sometimes referred to as a mental burden or psychological load. A person's mental health and safety in high-stress working conditions can be improved with the help of mental workload assessment. A photoplethysmogram (PPG) signal is a non-invasive and easily acquired physiological signal that contains information related to blood volume changes in the micro-vascular bed of tissues and can indicate psychologically relevant information to assess a person's mental workload (MW). An individual under a high MW possesses an increase in sympathetic nervous system activity, which results in morphological changes in the PPG waveform. In this work, a time-frequency analysis framework is developed to capture these distinguishing PPG features for the automatic assessment of MW. In particular, a cross-wavelet coherence (WTC) approach is proposed to extract simultaneous time-frequency information of the PPG during MW relative to the resting PPG. The suggested technique is validated on a publicly available data set of 22 healthy individuals who took part in an N-back task with PPG recording. Under three different fixed window lengths, images are obtained using WTC between PPG records during N-back task activity and rest. The images are used further to obtain PPG classification in two broad classes of low and high MW using a customized pre-trained Inception-V3 model. The best validation and test accuracy of 93.86% and 93.07%, respectively obtained in the window setting of 1200 samples used for WTC image creation.
RESUMO
The process of brain tumour segmentation entails locating the tumour precisely in images. Magnetic Resonance Imaging (MRI) is typically used by doctors to find any brain tumours or tissue abnormalities. With the use of region-based Convolutional Neural Network (R-CNN) masks, Grad-CAM and transfer learning, this work offers an effective method for the detection of brain tumours. Helping doctors make extremely accurate diagnoses is the goal. A transfer learning-based model has been suggested that offers high sensitivity and accuracy scores for brain tumour detection when segmentation is done using R-CNN masks. To train the model, the Inception V3, VGG-16, and ResNet-50 architectures were utilised. The Brain MRI Images for Brain Tumour Detection dataset was utilised to develop this method. This work's performance is evaluated and reported in terms of recall, specificity, sensitivity, accuracy, precision, and F1 score. A thorough analysis has been done comparing the proposed model operating with three distinct architectures: VGG-16, Inception V3, and Resnet-50. Comparing the proposed model, which was influenced by the VGG-16, to related works also revealed its performance. Achieving high sensitivity and accuracy percentages was the main goal. Using this approach, an accuracy and sensitivity of around 99% were obtained, which was much greater than current efforts.
Assuntos
Neoplasias Encefálicas , Imageamento por Ressonância Magnética , Redes Neurais de Computação , Humanos , Neoplasias Encefálicas/diagnóstico por imagem , Neoplasias Encefálicas/patologia , Imageamento por Ressonância Magnética/métodos , Processamento de Imagem Assistida por Computador/métodos , Interpretação de Imagem Assistida por Computador/métodos , Algoritmos , Sensibilidade e EspecificidadeRESUMO
Persuasive technologies, in connection with human factor engineering requirements for healthy workplaces, have played a significant role in ensuring a change in human behavior. Healthy workplaces suggest different best practices applicable to body posture, proximity to the computer system, movement, lighting conditions, computer system layout, and other significant psychological and cognitive aspects. Most importantly, body posture suggests how users should sit or stand in workplaces in line with best and healthy practices. In this study, we developed two study phases (pilot and main) using two deep learning models: convolutional neural networks (CNN) and Yolo-V3. To train the two models, we collected posture datasets from creative common license YouTube videos and Kaggle. We classified the dataset into comfortable and uncomfortable postures. Results show that our YOLO-V3 model outperformed CNN model with a mean average precision of 92%. Based on this finding, we recommend that YOLO-V3 model be integrated in the design of persuasive technologies for a healthy workplace. Additionally, we provide future implications for integrating proximity detection taking into consideration the ideal number of centimeters users should maintain in a healthy workplace.
RESUMO
Breast cancer (BC) is the leading cause of female cancer mortality and is a type of cancer that is a major threat to women's health. Deep learning methods have been used extensively in many medical domains recently, especially in detection and classification applications. Studying histological images for the automatic diagnosis of BC is important for patients and their prognosis. Owing to the complication and variety of histology images, manual examination can be difficult and susceptible to errors and thus needs the services of experienced pathologists. Therefore, publicly accessible datasets called BreakHis and invasive ductal carcinoma (IDC) are used in this study to analyze histopathological images of BC. Next, using super-resolution generative adversarial networks (SRGANs), which create high-resolution images from low-quality images, the gathered images from BreakHis and IDC are pre-processed to provide useful results in the prediction stage. The components of conventional generative adversarial network (GAN) loss functions and effective sub-pixel nets were combined to create the concept of SRGAN. Next, the high-quality images are sent to the data augmentation stage, where new data points are created by making small adjustments to the dataset using rotation, random cropping, mirroring, and color-shifting. Next, patch-based feature extraction using Inception V3 and Resnet-50 (PFE-INC-RES) is employed to extract the features from the augmentation. After the features have been extracted, the next step involves processing them and applying transductive long short-term memory (TLSTM) to improve classification accuracy by decreasing the number of false positives. The results of suggested PFE-INC-RES is evaluated using existing methods on the BreakHis dataset, with respect to accuracy (99.84%), specificity (99.71%), sensitivity (99.78%), and F1-score (99.80%), while the suggested PFE-INC-RES performed better in the IDC dataset based on F1-score (99.08%), accuracy (99.79%), specificity (98.97%), and sensitivity (99.17%).
RESUMO
BACKGROUND: Revascularization for extracranial vertebral artery dissection or vertebral artery atherosclerotic occlusive lesions caused by vertebrobasilar insufficiency or posterior circulation infarction is relatively rare. When bypassing the cervical external carotid artery (ECA) or common carotid artery (CCA) using a radial artery (RA) or saphenous vein (SV) graft, it is difficult to determine whether the recipient site should be the V2 or V3 portion. OBSERVATIONS: In case 1, cervical ECA-RA-V3 bypass was performed for bilateral extracranial vertebral artery dissection with the onset of ischemia, and cervical CCA-SV-V3 bypass was added 12 days later. Nine years after surgery, the bilateral vertebral artery dissection had improved, and the patient still had a patent bypass. In case 2, cervical ECA-RA-V2 bypass was performed for arteriosclerotic bilateral extracranial vertebral artery occlusion. The bypass was patent 5 years after surgery. The postoperative course was uneventful in both patients. LESSONS: The authors present cases of posterior fossa revascularization using the vertebral artery V3 and V2 portions via skull base surgery and note that it is important to consider each patient's individual characteristics when selecting the V3 or V2 portion.
RESUMO
BACKGROUND: Since most of the commonly known oral diseases are explained in link with balance of microbial community, an accurate bacterial taxonomy profiling for determining bacterial compositional network is essential. However, compared to intestinal microbiome, research data pool related to oral microbiome is small, and general 16S rRNA screening method has a taxonomy misclassification issue in confirming complex bacterial composition at the species level. OBJECTIVE: Present study aimed to explore bacterial compositional networks at the species level within saliva of 39 oral disease patients (Dental Caries group: n = 26 and Periodontitis group: n = 13) through comparison with public Korean-specific healthy oral microbiome data. METHODS: Here, we applied comprehensive molecular diagnostics based on qRT-PCR and Sanger sequencing methods to complement the technical limitations of NGS-based 16S V3-V4 amplicon sequencing technology. RESULTS: As a result of microbiome profiling at the genus level, relative frequencies of many nitrate-reducing bacteria within each oral disease group were found to be significantly low compared to the healthy group. In addition, the molecular diagnostics-based bacterial identification method allowed the determination of the correct taxonomy of screened primary colonizers (Streptococcus and Actinomyces unclassification clusters) for each oral disease. Finally, as with the results of microbiome profiling at the genus level, many core-species classified within the saliva of each oral disease group were also related to nitrate-reduction, and it was estimated that various pathogens associated with each disease formed a bacterial network with the core-species. CONCLUSION: Our study introduced a novel approach that can compensate for the difficulty of identifying an accurate bacterial compositional network at the species level due to unclear taxonomy classification by using the convergent approach of NGS-molecular diagnostics. Ultimately, we suggest that our experimental approach and results could be potential reference materials for researchers who intend to prevent oral disease by determining the correlation between oral health and bacterial compositional network according to the changes in the relative frequency for nitrate-reducing species.
Assuntos
Microbiota , RNA Ribossômico 16S , Saliva , Humanos , Saliva/microbiologia , RNA Ribossômico 16S/genética , Feminino , Masculino , Microbiota/genética , Adulto , Sequenciamento de Nucleotídeos em Larga Escala/métodos , Pessoa de Meia-Idade , Cárie Dentária/microbiologia , Cárie Dentária/diagnóstico , Periodontite/microbiologia , Periodontite/diagnóstico , Bactérias/genética , Bactérias/classificação , Bactérias/isolamento & purificaçãoRESUMO
Reactive sulfur species including sulfides, polysulfides and cysteine hydropersulfide play extensive roles in health and disease, which involve modification of protein functions through the interaction with metals bound to the proteins, cleavage of cysteine disulfide (S-S) bonds and S-persulfidation of cysteine residues. Sulfides over a wide micromolar concentration range enhance the activity of Cav3.2 T-type Ca2+ channels by eliminating Zn2+ bound to the channels, thereby promoting somatic and visceral pain. Cav3.2 is under inhibition by Zn2+ in physiological conditions, so that sulfides function to reboot Cav3.2 from Zn2+ inhibition and increase the excitability of nociceptors. On the other hand, polysulfides generated from sulfides activate TRPA1 channels via cysteine S-persulfidation, thereby facilitating somatic, but not visceral, pain. Thus, Cav3.2 function enhancement by sulfides and TRPA1 activation by polysulfides, synergistically accelerate somatic pain signals. The increased activity of the sulfide/Cav3.2 system, in particular, appears to have a great impact on pathological pain, and may thus serve as a therapeutic target for treatment of neuropathic and inflammatory pain including visceral pain.
Assuntos
Canais de Cálcio Tipo T , Sulfetos , Canal de Cátion TRPA1 , Sulfetos/farmacologia , Canal de Cátion TRPA1/metabolismo , Humanos , Canais de Cálcio Tipo T/metabolismo , Canais de Cálcio Tipo T/fisiologia , Animais , Zinco/metabolismo , Dor/metabolismo , Dor/tratamento farmacológico , Nociceptores/metabolismo , Nociceptores/efeitos dos fármacosRESUMO
This study aimed to evaluate the blood bacterial microbiota in healthy and febrile cats. High-quality sequencing reads from the 16S rRNA gene variable region V3-V4 were obtained from genomic blood DNA belonging to 145 healthy cats, and 140 febrile cats. Comparisons between the blood microbiota of healthy and febrile cats revealed dominant presence of Actinobacteria, followed by Firmicutes and Proteobacteria, and a lower relative abundance of Bacteroidetes. Upon lower taxonomic levels, the bacterial composition was significantly different between healthy and febrile cats. The families Faecalibacterium and Kineothrix (Firmicutes), and Phyllobacterium (Proteobacteria) experienced increased abundance in febrile samples. Whereas Thioprofundum (Proteobacteria) demonstrated a significant decrease in abundance in febrile. The bacterial composition and beta diversity within febrile cats was different according to the affected body system (Oral/GI, systemic, skin, and respiratory) at both family and genus levels. Sex and age were not significant factors affecting the blood microbiota of febrile cats nor healthy ones. Age was different between young adult and mature adult healthy cats. Alpha diversity was unaffected by any factors. Overall, the findings suggest that age, health status and nature of disease are significant factors affecting blood microbiota diversity and composition in cats, but sex is not.
Assuntos
Microbiota , RNA Ribossômico 16S , Animais , Gatos , RNA Ribossômico 16S/genética , Microbiota/genética , Febre/microbiologia , Febre/sangue , Feminino , Masculino , Bactérias/genética , Bactérias/classificação , Bactérias/isolamento & purificação , Doenças do Gato/microbiologia , Doenças do Gato/sangueRESUMO
Ca2+ influx through Cav3.3 T-type channel plays crucial roles in neuronal excitability and is subject to regulation by various signaling molecules. However, our understanding of the partners of Cav3.3 and the related regulatory pathways remains largely limited. To address this quest, we employed the rat Cav3.3 C-terminus as bait in yeast-two-hybrid screenings of a cDNA library, identifying rat Gß2 as an interaction partner. Subsequent assays revealed that the interaction of Gß2 subunit was specific to the Cav3.3 C-terminus. Through systematic dissection of the C-terminus, we pinpointed a 22 amino acid sequence (amino acids 1789-1810) as the Gß2 interaction site. Coexpression studies of rat Cav3.3 with various Gßγ compositions were conducted in HEK-293 cells. Patch clamp recordings revealed that coexpression of Gß2γ2 reduced Cav3.3 current density and accelerated inactivation kinetics. Interestingly, the effects were not unique to Gß2γ2, but were mimicked by Gß2 alone as well as other Gßγ dimers, with similar potencies. Deletion of the Gß2 interaction site abolished the effects of Gß2γ2. Importantly, these Gß2 effects were reproduced in human Cav3.3. Overall, our findings provide evidence that Gß(γ) complexes inhibit Cav3.3 channel activity and accelerate the inactivation kinetics through the Gß interaction with the Cav3.3 C-terminus.
Assuntos
Canais de Cálcio Tipo T , Subunidades beta da Proteína de Ligação ao GTP , Animais , Humanos , Ratos , Canais de Cálcio Tipo R , Canais de Cálcio Tipo T/metabolismo , Canais de Cálcio Tipo T/genética , Canais de Cálcio Tipo T/química , Proteínas de Transporte de Cátions , Subunidades beta da Proteína de Ligação ao GTP/metabolismo , Subunidades beta da Proteína de Ligação ao GTP/genética , Subunidades beta da Proteína de Ligação ao GTP/química , Subunidades gama da Proteína de Ligação ao GTP/metabolismo , Subunidades gama da Proteína de Ligação ao GTP/genética , Subunidades gama da Proteína de Ligação ao GTP/química , Células HEK293 , Cinética , Técnicas de Patch-Clamp , Ligação ProteicaRESUMO
Early detection of plant leaf diseases accurately and promptly is very crucial for safeguarding agricultural crop productivity and ensuring food security. During their life cycle, plant leaves get diseased because of multiple factors like bacteria, fungi, weather conditions, etc. In this work, the authors propose a model that aids in the early detection of leaf diseases using a novel hierarchical residual vision transformer using improved Vision Transformer and ResNet9 models. The proposed model can extract more meaningful and discriminating details by reducing the number of trainable parameters with a smaller number of computations. The proposed method is evaluated on the Local Crop dataset, Plant Village dataset, and Extended Plant Village Dataset with 13, 38, and 51 different leaf disease classes. The proposed model is trained using the best trail parameters of Improved Vision Transformer and classified the features using ResNet 9. Performance evaluation is carried out on a wide aspects over the aforementioned datasets and results revealed that the proposed model outperforms other models such as InceptionV3, MobileNetV2, and ResNet50.
RESUMO
CONTEXT: Indeterminate thyroid nodules (ITNs) lead to diagnostic surgeries in many countries. Use of molecular testing (MT) is endorsed by several guidelines, but costs are limitative, especially in public healthcare systems like in Canada. OBJECTIVES: Primary objective: evaluate the clinical value of Thyroseq® v3 (TSv3) using benign call rate (BCR) in a real-world practice. Secondary objective: assess cost-effectiveness of MT. DESIGN: This is a multicentric prospective study. SETTING: This study was conducted in 5 academic centers in Quebec, Canada. PATIENTS OR OTHER PARTICIPANTS: 500 consecutive patients with Bethesda III (on 2 consecutive cytopathologies) or IV and TIRADS 3 or 4 nodules measuring 1 to 4 cm were included. INTERVENTION: MT was performed between November 2021 and November 2022. Patients with a positive TSv3 were referred to surgery. Patients with a negative TSv3 were planned for follow-up by ultrasonography for a minimum of 2 years. MAIN OUTCOME MEASURE: The BCR, corresponding to the proportion of ITNs with negative TSv3 results, was assessed. RESULTS: 500 patients underwent TSv3 testing, with a BCR of 72.6% (95% CI: 68.5-76.5; p<0.001). 99.7% of patients with a negative result avoided surgery. The positive predictive value of TSv3 was 68.2% (95% CI: 58.5-76.9). The cost-benefit analysis identified that the implementation of MT would yield cost savings of $6.1 million over the next 10 years. CONCLUSIONS: Use of MT (TSv3) in a well-selected population with ITNs led to a BCR of 72.6%. It is cost-effective and prevents unnecessary surgeries in a public healthcare setting.
RESUMO
INTRODUCTION AND IMPORTANCE: Direct vertebrovertebral fistulas (VVFs) involving the V3 segment of the vertebral artery (VA) are rare. Endovascular treatment (EVT) can be used to obliterate these VVFs. CASE PRESENTATION: Case 1 was a 30-year-old male with limb weakness. He had grade V muscle strength in his limbs. Angiography confirmed a low-flow direct VVF of the V3 segment. The right VA was well developed. Coiling of the VVF and its parent VA obliterated the VVF. The vertebrobasilar arteries had sufficient blood from the right VA. Postoperatively, the patient recovered well. Case 2 was a 51-year-old male with headache and weakness of the limbs. He had grade IV muscle strength in his limbs. Angiography revealed a high-flow direct VVF in the V3 segment. The left VA was well developed. Coiling of the VVF and its parent VA obliterated the VVF. The vertebrobasilar arteries had sufficient blood from the left VA. Postoperatively, the patient recovered well. CLINICAL DISCUSSION: Direct VVF of the V3 segment is difficult to treat, and EVT, including reconstructive and deconstructive approaches, can be an effective treatment option. According to this case report and literature review, reconstructive EVT is ideal; however, it is difficult to perform. Currently, deconstructive EVT may be the mainstream option for treating direct VVF of the V3 segment. CONCLUSION: In certain cases where the contralateral VA is well developed, coiling the fistula and the parent VA is still an effective treatment.
RESUMO
A bridge disease identification approach based on an enhanced YOLO v3 algorithm is suggested to increase the accuracy of apparent disease detection of concrete bridges under complex backgrounds. First, the YOLO v3 network structure is enhanced to better accommodate the dense distribution and large variation of disease scale characteristics, and the detection layer incorporates the squeeze and excitation (SE) networks attention mechanism module and spatial pyramid pooling module to strengthen the semantic feature extraction ability. Secondly, CIoU with better localization ability is selected as the loss function for training. Finally, the K-means algorithm is used for anchor frame clustering on the bridge surface disease defects dataset. 1363 datasets containing exposed reinforcement, spalling, and water erosion damage of bridges are produced, and network training is done after manual labelling and data improvement in order to test the efficacy of the algorithm described in this paper. According to the trial results, the YOLO v3 model has enhanced more than the original model in terms of precision rate, recall rate, Average Precision (AP), and other indicators. Its overall mean Average Precision (mAP) value has also grown by 5.5%. With the RTX2080Ti graphics card, the detection frame rate increases to 84 Frames Per Second, enabling more precise and real-time bridge illness detection.
RESUMO
BACKGROUND: The atypia of undetermined significance (AUS) category is heterogeneous, leading to variations in its use. To prevent excessive usage, the AUS rate should be ≤10%. Although this recommendation aims to maintain diagnostic quality, it lacks supporting data. The AUS:Malignant (AUS:M) ratio has been proposed as a metric tool to evaluate AUS use. Furthermore, integrating ThyroSeq v3 (TSV3) positive call rate (PCR) and the molecular-derived risk of malignancy (MDROM) have been put forward as performance improvement tools. The authors reviewed their AUS:M ratios, TSV3 PCR, MDROM, and ROM. METHODS: Thyroid aspirates evaluated in the laboratory (from August 2022 to September 2023) by seven cytopathologists (CPs) were identified. AUS:M ratio, MDROM, ROM, and TSV3 PCR results for the laboratory and each CP were recorded and analyzed. RESULTS: A total of 2248 aspirates were identified (462 AUS and 80 malignant). The AUS:M ratio for the laboratory was 5.8 (CPs range, 2.8 to 7.3). The TSV3 PCR for the laboratory was 23% (CPs range, 11% to 41%). The MDROM for the laboratory was 19% (CPs range, 9% to 31%), whereas the ROM was 36% (CPs range, 29% to 50%). Linear regression analysis of AUS:M ratio versus TSV3 PCR and MDROM demonstrated a moderate positive correlation but a weak negative correlation to the ROM. Deviations from established targets were attributed to multiple factors. CONCLUSION: The findings of this study underscore the importance of using a combination of metrics to evaluate diagnostic practices. By dissecting the practice patterns of each CP, the authors can measure different aspects of their performance and provide individualized feedback.
Assuntos
Neoplasias da Glândula Tireoide , Nódulo da Glândula Tireoide , Humanos , Neoplasias da Glândula Tireoide/patologia , Neoplasias da Glândula Tireoide/genética , Neoplasias da Glândula Tireoide/diagnóstico , Biópsia por Agulha Fina , Nódulo da Glândula Tireoide/patologia , Nódulo da Glândula Tireoide/genética , Nódulo da Glândula Tireoide/diagnóstico , Glândula Tireoide/patologia , Citodiagnóstico/métodos , Feminino , Estudos Retrospectivos , CitologiaRESUMO
Retinal segmentation is a crucial step in the early warning of human health conditions. However, retinal blood vessels possess complex curvature, irregular distribution, and contain multi-scale fine structures, which make the limited receptive field of regular convolution challenging to process their vascular details efficiently. Additionally, the encoder-decoder based network leads to irreversible spatial information loss because of multiple downsampling, resulting in over-segmentation and missed segmentation of the vessels. For this reason, we develop a high-resolution network based on Deformable Convolution v3, called HRD-Net. By constructing a high-resolution representation, the network allows special attention to be paid to the details of tiny blood vessels. The proposed feature enhancement cascade module based on Deformable Convolution v3 can flexibly adapt and capture the ever-changing morphology and intricate connections of retinal blood vessels, ensuring the continuity of vessel segmentation. In the output phase of the network, the proposed global aggregation module integrates full-resolution feature maps while suppressing redundant features, achieving an effective fusion of high-level semantic information and spatial detail information. In addition, we have re-examined the selection criteria for activation and normalization methods, and also refine the network architectures from a spatial domain perspective to release redundant computational loads. Testing on the DRIVE, STARE, and CHASE_DB1 datasets indicates that HRD-Net, with fewer parameters, outperforms existing segmentation methods on several evaluation metrics such as F1, ACC, SE, SP, AUC, and IOU.
Assuntos
Aprendizagem , Vasos Retinianos , Humanos , Vasos Retinianos/diagnóstico por imagem , Benchmarking , Retina/diagnóstico por imagem , Salários e Benefícios , Processamento de Imagem Assistida por Computador , AlgoritmosRESUMO
The envelope glycoprotein (Env) trimer on the surface of human immunodeficiency virus type I (HIV-1) mediates viral entry into host CD4+ T cells and is the sole target of neutralizing antibodies. Broadly neutralizing antibodies (bnAbs) that target gp120 V3-glycan of HIV-1 Env trimer are potent and block the entry of diverse HIV-1 strains. Most V3-glycan bnAbs interact, to a different extent, with a glycan attached to N332, but Asn at this position is not absolutely conserved or required for HIV-1 entry based on the prevalence of N332 in different circulating HIV-1 strains from diverse clades. Here, we studied the effects of amino acid changes at position 332 of HIV-1AD8 Envs on HIV-1 sensitivity to antibodies, cold exposure, and soluble CD4. We further investigated how these changes affect Env function and HIV-1 infectivity in vitro. Our results suggest robust tolerability of HIV-1AD8 Env N332 to changes, with specific changes that resulted in extended exposure of gp120 V3 loop, which is typically concealed in most primary HIV-1 isolates. Viral evolution leading to Asn at position 332 of HIVAD8 Envs is supported by the selection advantage of high levels of cell-cell fusion, transmission, and infectivity with high levels of cell surface expression and slightly higher gp120 shedding than most N332 variants. Thus, tolerance of HIV-1AD8 Envs to different amino acids at position 332 provides increased flexibility to respond to changing conditions/environments and evade the immune system. Modeling studies of the distance between N332 glycan and specific bnAbs were in agreement with N332 glycan dependency on bnAb neutralization. Overall, our studies provide insights into the contribution of specific amino acids at position 332 to Env antigenicity, stability on ice, and conformational states. IMPORTANCE: Glycan attached to amino acid asparagine at position 332 of HIV-1 envelope glycoproteins is a main target of a subset of broadly neutralizing antibodies that block HIV-1 infection. Here, we defined the contribution of different amino acids at this position to Env antigenicity, stability on ice, and conformational states.