Image Pattern Research Articles

Predictive reconstruction of missing geological events and patterns in real-life 3D post-stack seismic images: a novel U-Net based deep learning approach

Predictive reconstruction of missing geological events and patterns in real-life 3D post-stack seismic images: a novel U-Net based deep learning approach

Factors associated with the detection of atrial fibrillation in patients with embolic stroke of undetermined source

BackgroundDetection of atrial fibrillation (AF) in patients with embolic stroke of undetermined source (ESUS) is important for the secondary prevention of stroke. We investigated the factors associated with the detection of newly diagnosed AF in ESUS patients during follow-up.MethodsPatients with acute ischemic stroke classified as ESUS were included. All patients underwent transthoracic echocardiography and Holter to detect the source of embolism. Structural, electrophysiological markers of left atrial cardiopathy (i.e., left atrial enlargement [LAE], non-sustained tachycardia [NSAT]) as well as lesion patterns of ischemic stroke were examined. Implantable loop recorder (ILR) was implanted in selective patients. Sensitivity and positive predictive value analysis was used to assess the predictive value for AF detection.ResultsAmong 312 patients with ESUS, AF was detected in 24 (7.7%) patients during follow-up. Patients with AF had a higher prevalence of LAE, NSAT, and the imaging pattern of confluent plus additional lesions in a single vascular territory. Multivariable analysis showed that ILR implantation (hazards ratio 11.497 [95% confidence interval 3.795–34.818]), LAE (3.204 [1.096–9.370]), NSAT (4.070 [1.378–12.018]), and confluent plus additional lesions (4.977 [1.649–15.019]) were independent predictors of AF detection. The sensitivity of detecting AF in those with LAE, NSAT, or confluent plus additional lesions pattern was 91.7%. The positive predictive value of detecting AF in those with LAE, NSAT and confluent plus additional lesions pattern was 40.0%.ConclusionIn conclusion, patients with LAE, NSAT, or confluent plus additional lesions may benefit from ILR monitoring detecting new AF.

“Neutrophil-To-Lymphocyte Ratio and Its Relation with Myocardial Markers and Varying Severity of Myocardial Ischemia on Lexiscan in High Risk Patients”.

Abstract Background: Coronary artery disease (CAD) is mainly caused by atherosclerosis that is considered as an immune-inflammatory condition. The neutrophil to lymphocyte ratio (NLR) is widely available inflammatory biomarker that has been found a very useful marker in acute coronary syndrome. Objectives: The aim of this study was to check the utility of NLR inflammatory biomarker with varying severities of myocardial ischemia and to correlate it with cardiac troponin elevation and abnormal myocardial stress imaging patterns. Methods: All patients with anginal chest pain or angina equivalents who underwent regadenoson SPECT myocardial perfusion imaging, commonly known as Lexiscan, with positive stress results, defined as reversible myocardial ischemia were included. The corresponding NLRs, cardiac troponins, and stress imaging parameters were reviewed. Results: A total 92 patients were studied. There was no statistically significant difference in the elevation of NLR inflammatory biomarker across varying severities of myocardial ischemic categories but NLR correlated positively with the cardiac troponin T elevation. Cardiac troponin T (cTnT) elevation was significantly associated with the severity of myocardial ischemia (p<0.05). The inter-tertile NLR comparison revealed an inversely proportional relationship of stress left ventricular ejection fraction (LVEF) with NLR (p<0.01). Conclusion: NLR is not influenced by the severity of myocardial ischemia, but correlates positively with cardiac troponin T elevation. The high NLR translates into more depressed left ventricular function on stress testing.

A Novel Encoder Decoder Architecture with Vision Transformer for Medical Image Segmentation

Brain tumor image segmentation is one of the most critical tasks in medical imaging for diagnosis, treatment planning, and prognosis. Traditional methods for brain tumor image segmentation are mostly based on Convolution Neural Network (CNN), which have been proved very powerful but still have limitations to effectively capture long-range dependencies and complex spatial hierarchies in MRI images. Variability in the shape, size, and location of tumors may affect the performance and may get stuck into suboptimal outcomes. In these regards, new encoder-decoder architecture with the VisionTranscoder(ViT) is proposed, to enhance brain tumor detection and classification. The proposed VisionTranscoder exploits a transformer's ability in modeling global context through self-attention mechanisms, providing more inclusive interpretation of the intricate patterns in medical images and classification by capturing both local and global features. The proposed VisionTranscoder maintains the Vision Transformer in its encoder for processing images as sequences of patches to capture global dependencies often outside the view of traditional CNNs. Then the segmentation map is rebuilt at a high level of fidelity with the decoder through upsampling and skips connections to maintain detailed spatial information. The risk of overfitting is hugely reduced by design and advanced regularization techniques with extensive data augmentation. The dataset contains 7,023 human brain MRI images, all of which are in four different classes: glioma, meningioma, no tumor, and pituitary. Images from the 'no tumor' class, indicating an MRI scan without any detectable tumor, were taken from the Br35H dataset . The results show the efficiency of VisionTranscoder over a wide set of brain MRI scans, producing an accuracy of 98.5% with a loss of 0.05. This performance underlines the ability of it to accurately segment and classify a brain tumor without overfitting.

Reversible Cerebral Vasoconstriction Syndrome: A Retrospective study from South India.

Reversible cerebral vasoconstriction syndrome (RCVS) is a rare cause of stroke characterized by headache, seizures, focal deficits, or encephalopathy. Very little is known about this rare condition from the Indian subcontinent. Here, we present the clinical and imaging characteristics and short-term outcomes of RCVS patients from South India. A single-center retrospective study of all consecutive subjects with a clinical-radiological diagnosis of RCVS from January 2014 to December 2023 with a 3-month completed follow-up was conducted. The clinical features, vascular imaging patterns, and outcomes of patients with ischemic and hemorrhagic forms of RCVS were compared. Of the 22 patients who fulfilled Calabrese et al.'s criteria for RCVS, the majority were women with a mean age of 47.59 (±13.55) years. While headache was the most common presenting symptom in our cohort (18/22, 81.81%), 14/22 (63.6%) developed focal neurologic deficits during the course of illness. Four of 22 patients (18%) did not report headaches during the course of illness. The most common imaging finding at presentation was cortical subarachnoid hemorrhage (SAH) in 9/22 (40.9%), followed by infarcts in 6/22, (27.2%), while 12/22 (54.5%) patients developed new ischemic lesions on repeat imaging. Ischemic and hemorrhagic presentations of RCVS did not differ in terms of clinical presentation or outcome. All patients with ischemic lesions showed diffuse vasospasm on imaging, while those with SAH had both diffuse and focal vascular abnormalities. We present the largest single series of RCVS from India, with a favorable short-term outcome. Although the most common vascular abnormality is diffuse vasospasm, it can remain focal in a quarter of patients.

Analyzing temporal imaging patterns in acute ischemic stroke via DICOM-timestamps

Acute stroke management is time-sensitive, making time data crucial for both research and quality management. However, these time data are often not reliably captured in routine clinical practice. In this proof-of-concept study we analysed image-based time data automatically captured in the DICOM format. We enrolled data from two separate stroke centers (n = 3136 and n = 2089). Data from the first center was additionally separated into groups with large-vessel-occlusion (LVO, n = 1.092), medium-vessel-occlusions (MVO, n = 416), and no occlusion (NVO, n = 1630). The DICOM-tag StudyTime was used to analyze the distribution of scan times throughout the day. Additionally, manually documented onset- and admission were extracted from the patients’ records in a subset of cases (n = 347). Timestamps were compared across centers and occlusion groups, and a probabilistic model was developed to illustrate and compare stroke occurrence patterns throughout the day. The temporal distribution of the scan times at both centers was exceptionally consistent with a peak around noon and a nighttime low. The groups with vessel occlusions showed an earlier peak compared to those without (p < 0.04). The median interval between admission and scan time was 23 min, while the median onset-to-imaging time was 1 h:54 min. This proof-of-concept study indicates that DICOM-timestamps can reveal insights into the temporal patterns of stroke imaging and may be a promising tool for quality control and stroke research in general since they are always automatically captured by imaging devices as opposed to manual data collection in routine clinical practice.

Leptomeningeal Masses or Masquerades: A Spectrum of Diseases with Leptomeningeal Enhancement and their Mimics.

Leptomeningeal enhancement, visible on MRI, can indicate a variety of diseases, both neoplastic and non-neoplastic. This comprehensive pictorial review aims to equip radiologists and trainees with a thorough understanding of the diverse imaging presentations of leptomeningeal disease. Drawing from a retrospective analysis of MRI scans conducted between 1 January 2008 and 30 September 2022, at two tertiary teaching hospitals in Singapore, this review covers a wide range of conditions. Case Collection: The main neoplastic conditions discussed include leptomeningeal carcinomatosis, myelomatosis, schwannoma, CNS lymphoma, and pineal region tumors. Additionally, the review addresses non-neoplastic enhancements such as meningoencephalitis, intracranial hypotension, cerebral ischemia/infarction, epidural lipomatosis, syringomyelia, Sturge-Weber syndrome, and subarachnoid hemorrhage. By highlighting the imaging features and patterns associated with these conditions, the review underscores the critical role of accurate diagnosis and timely management in improving patient outcomes. Enhanced understanding of these conditions can significantly improve patient outcomes through timely and effective therapeutic interventions.

Deep learning-driven UAV vision for automated road crack detection and classification

ABSTRACT Unmanned aerial vehicles are increasingly utilised for monitoring and inspecting critical infrastructure such as power generation grids, oil and gas pipelines and roads. One key task in road maintenance is the detection of cracks, which is crucial for ensuring road safety. Manual detection of cracks is time-consuming and prone to errors, highlighting the need for automated solutions. This study presents an automated method for road crack detection and classification using a hybrid deep learning technique. The proposed approach integrates a pyramid vision transformer and ConvMixer models for feature extraction, enabling the system to learn complex patterns in crack images. Image pre-processing is first performed using a median filtering technique to enhance image quality. The detection and classification of cracks are then carried out using an Elman neural network model, with its hyperparameters optimised through an improved black widow optimisation algorithm. Extensive simulations demonstrate that the proposed method outperforms other deep learning models in terms of performance, providing a reliable and efficient solution for automated crack detection.

Customised Convolutional Neural Network With Transfer Learning for Multi‐Nutrient Deficiency Identification With Pattern and Deep Features in Paddy Image

ABSTRACTMulti‐nutrient deficiency in crops, involving a shortage of essential nutrients such as nitrogen, phosphorus and potassium, impacts plant growing and yield. Accurate recognition is vital for effective nutrient management and maximising productivity. Identification techniques include extractive methods that analyse symptoms and abstractive methods that generate insights from data, with hybrid approaches aiming to improve the accuracy. However, challenges remain in maintaining diagnostic consistency and so forth. Continuous improvements are necessary to better integrate and interpret complex data for more accurate nutrient deficiency identification. To tackle these challenges, this research proposes the customised convolutional neural network‐transfer learning (CCNN‐TL) model for identifying multi‐nutrient deficiencies in paddy leaves. This model includes several key phases: image preprocessing, segmentation, feature extraction, data augmentation and identification. Initially, the paddy leaf images undergo preprocessing using the improved Wiener filtering (IWF) technique. Next, the modified U‐Net model is proposed for segmenting the preprocessed images. In the feature extraction phase, relevant features are identified from the segmented images. These features are then augmented through the data augmentation process. Finally, the CCNN‐TL model is used for multi‐nutrient deficiency identification. The model's effectiveness is demonstrated through comprehensive simulations and experimental evaluations. These evaluations showcase its enhanced performance, with improved accuracy, precision and specificity compared to traditional methods. The CCNN‐TL scheme attained the greatest accuracy of 0.982, precision of 0.975 and F‐measure of 0.973. The Nutrient‐Deficiency‐Symptoms‐in‐Rice dataset was employed for simulations and analysis, ensuring a solid foundation for the evaluations.

Estimation of the pattern of ocular manifestations, risk factors, and imaging of rhino-orbital-cerebral mucormycosis in COVID-19 patients

ABSTRACT Purpose: To estimate the pattern of ocular manifestations, as observed clinically and radiologically, and associated risk factors in cases of coronavirus disease 2019 (COVID-19)-related rhino-orbital cerebral mucormycosis (ROCM) during the second wave at presentation to the hospital in northern India. Materials and Methods: A total of 35 patients of ROCM associated with active cases of COVID-19 or recovered cases of COVID-19 presenting to the Ophthalmology outpatient department or admitted in the COVID ward of the hospital with clinical features suggestive of ROCM. They were confirmed as mucormycosis on KOH and fungal cultures of nasal scrapings and histopathology of debrided tissue. This is an ambispective study with retrospective medical records review for COVID-19 analysis and prospective assessment of ROCM-associated COVID-19 during the second outbreak of the COVID-19 pandemic. Results: The total patients included were 35, which ranged in the age group of 52.91+/- 11.93 years, and the male-to-female ratio was 24:11. The duration between the first positive COVID report and onset of COVID-19-associated ROCM was 9.46+/-11.63 days. The majority (82.9%; n-29) either were diabetics or had high blood glucose levels during the recent COVID-19 infection. 45.7% of patients gave a history of steroid therapy. The most common ocular clinical features were ptosis (80%) and loss of vision 48.7%, respectively. There was diffuse involvement of sinuses in 18 cases (51.4%). Staging based on magnetic resonance imaging scans showed that 7 (20%) had stage II, 20 (57.1%) had stage III, and 28.7% had stage IV disease. Ten patients had varied intracranial extension ranging from meningitis and cavernous sinus thrombophlebitis/thrombosis to brain abscess. Conclusions: COVID-associated ROCM was very rapidly spreading and more destructive. Ptosis and loss of vision related to third nerve involvement and CRAO, respectively, were alarming signs and bad prognostic indicators to the patient. Intracranial involvement and CRAO were poor prognostic features in this type of mucormycosis.

Compare Noise Robust Least-Squares Method with Other Methods for Estimation of the Parameters of Frechet Distribution and Neutrosophic Generalization

The Frechet distribution is a versatile probability distribution that is used within a loose range in many important statistical fields, such as image processing, data analysis, and pattern recognition. It aims to explore and study the estimation of the parameters of the Frechet distribution using the noise-robust least squares method, as in the research paper, and it also has uses. There are many real-world scenarios. It is known that there is a growing challenge in estimating the parameter because of the noisy data. Depending on rigorous simulations and experimental analysis, we provide a novel powerful way to estimate the parameters for the Frechet Distribution Robust Least Squares approach to be flexible. Also, the results approach of this work will be very helpful in estimating the Frechet distribution parameters for diverse statistical applications. Also, we generalize our results to include the generalized neutrosophic case of this distribution dealing with neutrosophic numbers.

Development of "Target Detection and Identification System" for Combat Military Vehicles via Artificial Intelligence and Decision Support Matrix

Modern military intelligence systems are improving day by day. In the context of ground forces, combat vehicle-type classification technology serves as a valuable supporting asset. Understanding the capabilities and limitations of different combat vehicle types is crucial for developing effective tactics and strategies. The implementation of artificial intelligence-supported military vehicle type classification technology has raised increasing concerns as image processing, pattern recognition, and deep learning have advanced. You Only Look Once (YOLO) has demonstrated numerous considerable advantages in object detection and image classification. This method accelerates target detection because it can predict objects in real-time. The high accuracy of detection and assessment assists operational personnel in the field. The YOLO prediction method produces accurate results with minimal background errors and facilitates the understanding of generalized object representations. This paper utilizes YOLO to demonstrate combat vehicle type detection, specifically employing the YOLOv8m model, which is well-suited for mobile deployments. This study focuses on a process that begins with detecting targets using images obtained through electro-optical systems, aimed at supporting the activities of identifying and defining targets, which are essential components of target management systems. The process involves developing a Target Detection and Identification (TDI) system based on deep learning, incorporating steps such as pre-processing, segmentation, feature extraction, classification, additional data extraction, and decision support. This system is designed to interpret the results obtained from these processes and provide actionable recommendations. Furthermore, it assists in addressing the weapon target assignment problem in the land environment.

Design and development augmented reality using OpenCV for primary education

Children are known to enter a developmental stage between the ages of three and five. It is during this time that they learn how to write letters by tracing them over and making connections between the dots in books. However, this method of involving kids isn't very successful. What if, instead of limiting children to paper and pencil, we allowed them to use the power of air itself to practice writing? Would that bring a magical element to learning? Would it allow children to begin their educational journey? What if education turns into a magical experience that captivates children and turns the ordinary into something fascinating? Computer vision (openCV) with augmented reality (AR). Writing in air has become a evolving yet significant area of research in the field of image processing and pattern recognition. It has led to the development of various applications which have human machine interface. With high quality computers, video cameras and techniques for automatic video analysis, Object tracking has become the most sought out area in the domain of Computer Vision. This study focuses on comparing virtual air canvas built with OpenCV and mediapipe and virtual air canvas built using OpenCV and numpy. Further into the paper, we shall present you our novel idea of integrating Augmented Reality with Computer vision and its impact on teaching and learning.

Assimilating vegetation indices and canopy cover derived from remote sensing into the AquaCrop model for estimating rice yield

AbstractAccurate forecasting of rice yield using contemporary data is a paramount management tool. To achieve this objective, the AquaCrop, in conjunction with remote sensing (RS) data, was employed to predict yield in the paddy fields of SANRU. The required AquaCrop data, including ground (200 locations) and RS (Sentinel‐2, MODIS) data, were collected in two seasons, namely, 2020 and 2021. Leveraging Sentinel‐2 imagery, various vegetation indices (VIs) were computed, encompassing the normalized difference vegetation index (NDVI), rice growth vegetation index (RGVI) and soil adjusted vegetation index (SAVI). Additionally, MODIS and spatio‐temporal fusion algorithms (Spatial and Temporal Adaptive Reflectance Fusion Model [STARFM] and Enhanced Spatial and Temporal Adaptive Reflectance Fusion Model [ESTARFM]) were used to create a regular time pattern in satellite images taken on cloudy days. The results revealed that the yields observed for 2020 and 2021, on the basis of field data, were 4450 and 4370 kg/ha, respectively. Moreover, by leveraging the AquaCrop, the forecasted yield was ascertained both with and without the assimilation of RS data. Notably, the findings underscored that the incorporation of RS data significantly enhanced the model's predictive precision, particularly in estimating yield. The model's efficacy was demonstrated by its ability to forecast the end‐of‐season yield for the years 2020 and 2021, which yielded maximum RMSEs of 400 and 470 kg/ha, respectively.

Clinical, laboratory, and orbital imaging features of giant cell arteritis in comparison to non-arteritic anterior ischemic optic neuropath: a single center case series.

Giant cell arteritis (GCA) is the most common vasculitis in patients older than 50 years and is considered a "do not miss" diagnosis. However, it remains a diagnostic challenge given overlapping clinical syndromes such as non-arteritic anterior ischemic optic neuropathy (NAION) and poorly explored imaging findings. In this retrospective study between the time period of January 2013 and December 2021, a total of 13 consecutive patients with a pathological diagnosis of GCA and 8 patients with clinical diagnosis of NAION were isolated. Demographic and clinical data for each patient were collected, including pertinent laboratory data. Pertinent physical exam data was also collected, including fundoscopic exam and visual acuity. Two neuroradiologist assessed the orbital MRI imaging findings of GCA and NAION for the presence and characterization of imaging abnormalities. Assessment for potential relationship between GCA orbital findings, laboratory and visual outcomes was performed. Finally, comparison between GCA and NAION imaging findings was performed. 13 GCA patients were assessed. 9 patients had abnormal orbital findings. Of these 8 patients had bilateral orbital involvement The most common imaging findings was perineuritis of the optic nerve sheath, present in 7 patients. In total, 8 NAION patients were assessed. All patients demonstrate optic nerve involvement. The Snellen test was converted to logmar, and visual acuity was assessed for both NAION and GCA for each eye at diagnosis and at the last follow-up. There was no statistical significance for either eye for both GCA and NAION at initial diagnosis and final follow-up. In the 4 GCA patients with normal MRI findings and 9 GCA patients with abnormal MRI findings, there was no statistical significance between initial presentation and final follow-up visual acuity. GCA and NAION are potentially overlapping clinical syndromes with different treatment approach and poorly explored imaging findings. Our case series assesses the orbital imaging findings of both syndromes while noting different imaging pattern of both on MRI, which can serve as a potential tool to aid in diagnosis of both. .

Differentiating Sinonasal Tumor Entities with Fluorescein-Enhanced Confocal Laser Endomicroscopy: A Step Forward in Precision Diagnostics.

Background/Objectives: Sinonasal malignancies are rare and highly diverse cancers that pose significant diagnostic challenges due to their variable histological features and complex anatomical locations. Accurate diagnosis is critical for guiding treatment, yet conventional methods often require multiple biopsies. This study aimed to evaluate the potential of confocal laser endomicroscopy (CLE) for real-time imaging of sinonasal tumors to characterize specific features of different entities and improve diagnostic precision. Methods: Ten patients with various sinonasal malignancies, including squamous cell carcinoma, adenocarcinoma, sinonasal undifferentiated carcinoma, olfactory neuroblastoma, sinonasal mucosal melanoma, and endonasal lymphoma, were examined using CLE during diagnostic endoscopy. CLE images were compared descriptively with histopathological cross-sections to identify unique imaging patterns for each tumor type. Results: CLE was feasible across all cases, with high-quality images obtained despite anatomical challenges in some cases. Characteristic features, such as vascular clusters in undifferentiated carcinoma, mucin-filled bubbles in adenocarcinoma, and small round cells in neuroblastoma, were identified and corresponded well with histopathological findings. CLE also helped guide biopsies by revealing areas with diagnostic relevance. Conclusions: CLE demonstrates promise as an adjunct diagnostic tool in sinonasal malignancies, offering real-time imaging that correlates with histopathological findings and aids in targeted biopsies. While this study provides preliminary insights into the utility of CLE, further research with larger cohorts and statistical validation is necessary to establish its diagnostic reliability and broader clinical application.

Transferable deep learning with coati optimization algorithm based mitotic nuclei segmentation and classification model

Image processing and pattern recognition methods have recently been extensively implemented in histopathological images (HIs). These computer-aided techniques are aimed at detecting the attentive biological markers for assisting the final cancer grading. Mitotic count (MC) is a significant cancer detection and grading parameter. Conventionally, a pathologist examines the biopsy image physically by employing higher-power microscopy. The MC cells have been marked physically at every analysis, and total MC must be utilized as a major aspect for the cancer ranking and considered as the initiative of cancers. Numerous pattern recognition algorithms for cell-sized objects in HIs depend upon segmentation to assess features. The correct description of the segmentation has been difficult, and feature outcomes can be highly complex to the segmentation. The MC cells are an essential element in many cancer grading methods. Extraction of the MC cell from the HI is a highly challenging assignment. This manuscript proposes the Coati Optimization Algorithm with Deep Learning-Driven Mitotic Nuclei Segmentation and Classification (COADL-MNSC) technique. The major aim of the COADL-MNSC technique is to utilize the DL model to segment and classify the mitotic nuclei (MN). In the preliminary stage, the COADL-MNSC approach implements median filtering (MF) for pre-processing. Besides, the COADL-MNSC approach utilizes the Hybrid Attention Fusion U-Net (HAU-UNet) model to segment the MN. Moreover, the capsule network (CapsNet) model is employed for the feature extraction method, and its hyperparameters are adjusted by utilizing the COA model. At last, the classification procedure is performed using the bidirectional long short-term memory (BiLSTM) model. Extensive simulations are performed under the MN image dataset to exhibit the excellent performance of the COADL-MNSC methodology. The experimental validation of the COADL-MNSC methodology portrayed a superior accuracy value of 98.89% over existing techniques under diverse measures.

Classification of diabetic retinopathy grades using CNN feature extraction to segment the lesion

Diabetes's microvascular aftereffect, diabetic retinopathy (DR), is the primary cause of eyesight loss in the globe. In order to prevent vision impairment and to intervene promptly, early detection and precise classification of DR severity are essential. Using standard methods for diagnosing DR requires ophthalmologists to grade cases by hand, a process that can be laborious, subjective, and subject to observer error. In supervised learning task of classification, data instances are classified into predefined classes based on features. The relation between the traits and the classes can be found from the labelled data. After the training is completed, the classes of the unseen data. The frequent reason found for the loss of vision in diabetic retinopathy (DR) is found to be diabetes. Visual damage can be prevented by identifying the degree of DR at right time. For the grading of the DR, deep learning techniques are found to be very effective with maximum possible accuracy. The proposed model is useful in accurately classifying the DR images using the feature extraction with lesion segmentation, by implementing the patterns in the DR images. ReLU activation function is used in the proposed model. CNN feature extraction is used for the important feature extraction by applying the Convolution layers, and edges, textures, and forms are identified. As the model proceeds layer by layer, complicated patterns in the photos can be learned by the model, and can be analysed better. The features of the photos were extracted and found useful in segmentation and classification. ReLU is helpful in improving the convergence and also found useful in learning the patterns. Among the other activation functions, ReLU has higher computational efficiency and therefore is used in the model, which suits well for the DR application. A strong framework is proposed for the classification of the DR grade, for the lesion segmentation and CNN feature extraction. DR categorization using the proposed model is evaluated by data visualization of the important calculated metrics and found to be very effective.

Surface plasmon resonance to trace and measure cancer cell apoptosis using morphological and refractive index changes

Morphological changes of cancer cells are often used as an important indicator within efficiency studies of anticancer drugs. Morphological cell analysis on cell size and shape distribution is typically performed using microscopic methods, which are time consuming and require skilled personnel. Recently, more advanced image processing and pattern recognition have enabled identification and quantitative analysis of the cell’s abnormality and classification in an automated way. However, these methods usually involve multiple staining steps. In addition to computational complexity, the processes greatly compromise real-time applications of the system. Therefore, a non-invasive, real-time method allowing for assessment of living cells’ reactions to a death inducer is very much needed. Here, we present an SPR biosensor that measures the changes in cancer cells’ size and detachment, relating the cell confluency with the changes of the refractive index on the cell-substrate interface. As a proof-of-concept, we chose HeLa cell and hydrogen peroxide (H2O2) induced apoptosis as the model system to study the morphological changes of the cell. The results show that the SPR response to cell apoptosis agreed with the cellular morphological changes observed via microscopy. Interestingly, we observed simultaneous apoptosis and necrosis at high H2O2 concentrations. This simultaneous occurrence was verified using a mathematical model which incorporated other important factors such as cell thickness and intercellular refractive index. This model helped resolve the disagreement between SPR signal and cell confluency at high H2O2 concentrations. Our results show the potential of SPR as a label free and real time monitoring method for morphological changes and surface detachment of cancer cells. This method can be fully expanded to other cell-based sensing applications.

MR perfusion characteristics of pseudoprogression in brain tumors treated with immunotherapy - a comparative study with chemo-radiation induced pseudoprogression and radiation necrosis.

Pseudoprogression is an atypical imaging pattern of response to immunotherapy in patients with brain tumors. MR perfusion studies in this field are limited. The purpose of our study is to compare the perfusion features between pseudoprogression lesions in malignant gliomas and brain metastases treated with immunotherapy (iPsP) and the pseudoprogression after chemo-radiation therapy and radiation necrosis after radiation treatment (ChR-PsP & RN). We retrospectively reviewed 25 iPsP lesions in 16 brain tumor patients and 48 ChR-PsP & RN lesions in 35 patients. The cerebral blood volume (CBV) of MR dynamic susceptibility contrast (DSC) perfusion weighted imaging (PWI) was analyzed, and the mean and maximal values of the ratio of CBV (rCBVmean and rCBVmax) of iPsPs and ChR-PsP & RNs were calculated and compared between these two groups using the Mann-Whitney U test. A receiver operating characteristic curve analysis was conducted, and the optimal cutoff of perfusion parameters were determined using the area under the curve, sensitivity, and specificity. The medians of rCBVmean and rCBVmax in iPsP group were significantly higher (0.94 and 1.39 respectively) than ChR-PsP & RN group (0.67, p < 0.01 and 1.1, p = 0.01 respectively). The rCBVmean value of 0.69 can differentiate the iPsP from ChR-PsP & RN with the area under the curve of 0.71, sensitivity of 0.72, and specificity of 0.56. These findings may suggest immunotherapy-induced higher perfusion in the iPsP lesions compared to ChR-PsP & RN lesions in primary and metastatic brain tumors.