A Study on Intelligent Optical Bone Densitometry.

Osteoporosis is a prevalent chronic disease worldwide, particularly affecting the aging population. The gold standard diagnostic tool for osteoporosis is Dual-energy X-ray Absorptiometry (DXA). However, the expensive cost of the DXA machine and the need for skilled professionals to operate it restrict its accessibility to the general public. This paper builds upon previous research and proposes a novel approach for rapidly screening bone density. The method involves utilizing near-infrared light to capture local body information within the human body. Deep learning techniques are employed to analyze the obtained data and extract meaningful insights related to bone density. Our initial prediction, utilizing multi-linear regression, demonstrated a strong correlation (r = 0.98, p-value = 0.003**) with the measured Bone Mineral Density (BMD) obtained from Dual-energy X-ray Absorptiometry (DXA). This indicates a highly significant relationship between the predicted values and the actual BMD measurements. A deep learning-based algorithm is applied to analyze the underlying information further to predict bone density at the wrist, hip, and spine. The prediction of bone densities in the hip and spine holds significant importance due to their status as gold-standard sites for assessing an individual's bone density. Our prediction rate had an error margin below 10% for the wrist and below 20% for the hip and spine bone density.

Brain tumor grading diagnosis using transfer learning based on optical coherence tomography.

In neurosurgery, accurately identifying brain tumor tissue is vital for reducing recurrence. Current imaging techniques have limitations, prompting the exploration of alternative methods. This study validated a binary hierarchical classification of brain tissues: normal tissue, primary central nervous system lymphoma (PCNSL), high-grade glioma (HGG), and low-grade glioma (LGG) using transfer learning. Tumor specimens were measured with optical coherence tomography (OCT), and a MobileNetV2 pre-trained model was employed for classification. Surgeons could optimize predictions based on experience. The model showed robust classification and promising clinical value. A dynamic t-SNE visualized its performance, offering a new approach to neurosurgical decision-making regarding brain tumors.

Mild cognitive impairment estimation based on functional near-infrared spectroscopy.

Patients with mild cognitive impairment (MCI) are at a high risk of developing future dementia. However, early identification and active intervention could potentially reduce its morbidity and the incidence of dementia. Functional near-infrared spectroscopy (fNIRS) has been proposed as a noninvasive modality for detecting oxygenation changes in the time-varying hemodynamics of the prefrontal cortex. This study sought to provide an effective method for detecting patients with MCI using fNIRS and the Wisconsin card sorting test (WCST) to evaluate changes in blood oxygenation. The results revealed that all groups with a lower mini-mental state examination grade had a higher increase in HHb concentration during a modified WCST (MCST). The increase in the change in oxygenated hemoglobin concentration in the stroke group was smaller than that in the normal group due to weak cerebrovascular reactivity.

Rapid On-Site AI-Assisted Grading for Lung Surgery Based on Optical Coherence Tomography.

The determination of resection extent traditionally relies on the microscopic invasiveness of frozen sections (FSs) and is crucial for surgery of early lung cancer with preoperatively unknown histology. While previous research has shown the value of optical coherence tomography (OCT) for instant lung cancer diagnosis, tumor grading through OCT remains challenging. Therefore, this study proposes an interactive human-machine interface (HMI) that integrates a mobile OCT system, deep learning algorithms, and attention mechanisms. The system is designed to mark the lesion's location on the image smartly and perform tumor grading in real time, potentially facilitating clinical decision making. Twelve patients with a preoperatively unknown tumor but a final diagnosis of adenocarcinoma underwent thoracoscopic resection, and the artificial intelligence (AI)-designed system mentioned above was used to measure fresh specimens. Results were compared to FSs benchmarked on permanent pathologic reports. Current results show better differentiating power among minimally invasive adenocarcinoma (MIA), invasive adenocarcinoma (IA), and normal tissue, with an overall accuracy of 84.9%, compared to 20% for FSs. Additionally, the sensitivity and specificity, the sensitivity and specificity were 89% and 82.7% for MIA and 94% and 80.6% for IA, respectively. The results suggest that this AI system can potentially produce rapid and efficient diagnoses and ultimately improve patient outcomes.

Characterization of pathogenic factors for premenstrual dysphoric disorder using machine learning algorithms in rats.

We established a methodology using machine learning algorithms for determining the pathogenic factors for premenstrual dysphoric disorder (PMDD). PMDD is a disease characterized by emotional and physical symptoms that occurs before menstruation in women of childbearing age. Owing to the diverse manifestations and various pathogenic factors associated with this disease, the diagnosis of PMDD is time-consuming and challenging. In the present study, we aimed to establish a methodology for diagnosing PMDD. Using an unsupervised machine-learning algorithm, we divided pseudopregnant rats into three clusters (C1 to C3), depending on the level of anxiety- and depression-like behaviors. From the results of RNA-seq and subsequent qPCR of the hippocampus in each cluster, we identified 17 key genes for building a PMDD diagnostic model using our original two-step feature selection with supervised machine learning. By inputting the expression levels of these 17 genes into the machine learning classifier, the PMDD symptoms of another group of rats were successfully classified as C1-C3 with an accuracy of 96%, corresponding to the classification by behavior. The present methodology would be applicable for the clinical diagnosis of PMDD using blood samples instead of samples from the hippocampus in the future.

Intraoperative application of optical coherence tomography for lung tumor.

On-site instant determination of benign or malignant tumors for deciding the types of resection is crucial during pulmonary surgery. We designed a portable spectral-domain optical coherence tomography (SD-OCT) system to do real-time scanning intraoperatively for the distinction of fresh tumor specimens in the lung. A total of 12 ex vivo lung specimens from six patients were enrolled. Three patients were diagnosed with invasive adenocarcinoma (IA), while the others were benign. After OCT-imaged reconstruction, we compared the qualitative morphology of OCT and histology among malignant, benign, and normal tissues. In addition, through analysis of the quantitative data, a discrete difference in optical attenuation coefficients around the junctional surface was shown by our data processing. This study demonstrated a feasible OCT-assisted resection guide by a rapid on-site tumor diagnosis. The results indicate that future deep learning of OCT-captured image systems able to improve diagnostic and therapeutic efficiency is warranted.

Temporal mirror-symmetry in functional signals recorded from rat barrel cortex with optical coherence tomography.

Functional optical coherence tomography (fOCT) detects activity-dependent light scattering changes in micro-structures of neural tissue, drawing attention as in vivo volumetric functional imaging technique at a sub-columnar level. There are 2 plausible origins for the light scattering changes: (i) hemodynamic responses such as changes in blood volume and in density of blood cells and (ii) reorientation of dipoles in cellular membrane. However, it has not been clarified which is the major contributor to fOCT signals. Furthermore, previous studies showed both increase and decrease of reflectivity as fOCT signals, making interpretation more difficult. We proposed combination of fOCT with Fourier imaging and adaptive statistics to the rat barrel cortex. Active voxels revealed barrels elongating throughout layers with mini-columns in superficial layers consistent with physiological studies, suggesting that active voxels revealed by fOCT reflect spatial patterns of activated neurons. These voxels included voxels with negative changes in reflectivity and those with positive changes in reflectivity. However, they were temporally mirror-symmetric, suggesting that they share common sources. It is hard to explain that hemodynamic responses elicit positive signals in some voxels and negative signals in the other. On the other hand, considering membrane dipoles, polarities of OCT signals can be positive and negative depending on orientations of scattering particles relative to the incident light. Therefore, the present study suggests that fOCT signals are induced by the reorientation of membrane dipoles.

Using Signal Features of Functional Near-Infrared Spectroscopy for Acute Physiological Score Estimation in ECMO Patients.

Extracorporeal membrane oxygenation (ECMO) is a vital emergency procedure providing respiratory and circulatory support to critically ill patients, especially those with compromised cardiopulmonary function. Its use has grown due to technological advances and clinical demand. Prolonged ECMO usage can lead to complications, necessitating the timely assessment of peripheral microcirculation for an accurate physiological evaluation. This study utilizes non-invasive near-infrared spectroscopy (NIRS) to monitor knee-level microcirculation in ECMO patients. After processing oxygenation data, machine learning distinguishes high and low disease severity in the veno-venous (VV-ECMO) and veno-arterial (VA-ECMO) groups, with two clinical parameters enhancing the model performance. Both ECMO modes show promise in the clinical severity diagnosis. The research further explores statistical correlations between the oxygenation data and disease severity in diverse physiological conditions, revealing moderate correlations with the acute physiologic and chronic health evaluation (APACHE II) scores in the VV-ECMO and VA-ECMO groups. NIRS holds the potential for assessing patient condition improvements.

Migraine classification by machine learning with functional near-infrared spectroscopy during the mental arithmetic task.

Migraine is a common and complex neurovascular disorder. Clinically, the diagnosis of migraine mainly relies on scales, but the degree of pain is too subjective to be a reliable indicator. It is even more difficult to diagnose the medication-overuse headache, which can only be evaluated by whether the symptom is improved after the medication adjustment. Therefore, an objective migraine classification system to assist doctors in making a more accurate diagnosis is needed. In this research, 13 healthy subjects (HC), 9 chronic migraine subjects (CM), and 12 medication-overuse headache subjects (MOH) were measured by functional near-infrared spectroscopy (fNIRS) to observe the change of the hemoglobin in the prefrontal cortex (PFC) during the mental arithmetic task (MAT). Our model shows the sensitivity and specificity of CM are 100% and 75%, and that of MOH is 75% and 100%.The results of the classification of the three groups prove that fNIRS combines with machine learning is feasible for the migraine classification.

Differentiation of primary central nervous system lymphoma from glioblastoma using optical coherence tomography based on attention ResNet.

Significance: Differentiation of primary central nervous system lymphoma from glioblastoma is clinically crucial to minimize the risk of treatments, but current imaging modalities often misclassify glioblastoma and lymphoma. Therefore, there is a need for methods to achieve high differentiation power intraoperatively. Aim: The aim is to develop and corroborate a method of classifying normal brain tissue, glioblastoma, and lymphoma using optical coherence tomography with deep learning algorithm in an ex vivo experimental design. Approach: We collected tumor specimens from ordinal surgical operations and measured them with optical coherence tomography. An attention ResNet deep learning model was utilized to differentiate glioblastoma and lymphoma from normal brain tissues. Results: Our model demonstrated a robust classification power of detecting tumoral tissues from normal tissues and moderate discrimination between lymphoma and glioblastoma. Moreover, our results showed good consistency with the previous histological findings in the pathological manifestation of lymphoma, and this could be important from the aspect of future clinical practice. Conclusion: We proposed and demonstrated a quantitative approach to distinguish different brain tumor types. Using our method, both neoplasms can be identified and classified with high accuracy. Hopefully, the proposed method can finally assist surgeons with decision-making intraoperatively.

Degree of polarization uniformity for dental calculus visualization.

During the treatment for periodontitis, the removal of dental calculus is essential. Previously, we have proposed the DAM algorithm for intuitive identification of the site of lesion, enabling the non-contact assessment during the operation. Nonetheless, the delineation of dental calculus was still imperfect. To this end, here we utilized the power of polarization-sensitive optical coherence tomography and evaluated the contrast called degree of polarization uniformity for dental calculus visualization. The result showed that the selected index demonstrated excellent contrast of dental calculus from other normal dental hard tissues. The proposed contrast is promising for accurate dental calculus delineation.

Cultural Adaptation and Validation of Mullen Scales of Early Learning in Taiwanese children with Autism Spectrum Disorder, Global Developmental Delay, and Typically Developing Children.

BACKGROUND: The Mullen Scales of Early Learning (MSEL) is a standardized comprehensive developmental assessment tool for children aged 0-68 months. However, few Asia-based studies have explored cultural and linguistic adaptations of the MSEL or investigated its psychometric properties in populations with autism spectrum disorder (ASD). AIMS: This study evaluated the reliability and validity of the MSEL-Taiwan version (MSEL-T) for Taiwanese children with ASD, global developmental delay (GDD), and typical development (TD). METHODS AND PROCEDURES: The MSEL items were translated and modified according to the language and culture in Taiwan. In total, 191 children (ASD, 69; GDD, 36; and TD, 86) aged 19-68 months were assessed using the MSEL-T and Peabody Developmental Motor Scales 2 (PDMS-2) at enrollment, followed by the assessments of Vineland Adaptive Behavior Scale-Chinese version (VABS-C) at the age of 36 months or later. OUTCOMES AND RESULTS: All subscales were verified to have good interrater reliability and internal consistency, and subscale scores indicated moderate to high correlations with PDMS-2 and VABS-C scores. Significant differences in MSEL-T scores were observed between same-aged pairs of children with TD and GDD and between pairs of children with TD and ASD. CONCLUSIONS AND IMPLICATIONS: The findings provide evidence of validity and reliability of the MSEL-T. And it is suggested that the culturally and linguistically adapted MSEL-T is a good tool for the clinical assessment of children with and without ASD.

Characteristics of brain connectivity during verbal fluency test: Convolutional neural network for functional near-infrared spectroscopy analysis.

Human connectome describes the complicated connection matrix of nervous system among human brain. It also possesses high potential of assisting doctors to monitor the brain injuries and recoveries in patients. In order to unravel the enigma of neuron connections and functions, previous research has strived to dig out the relations between neurons and brain regions. Verbal fluency test (VFT) is a general neuropsychological test, which has been used in functional connectivity investigations. In this study, we employed convolutional neural network (CNN) on a brain hemoglobin concentration changes (ΔHB) map obtained during VFT to investigate the connections of activated brain areas and different mental status. Our results show that feature of functional connectivity can be identified accurately with the employment of CNN on ΔHB mapping, which is beneficial to improve the understanding of brain functional connections.

Deep-Learning Approach to Predict Survival Outcomes Using Wearable Actigraphy Device Among End-Stage Cancer Patients.

Survival prediction is highly valued in end-of-life care clinical practice, and patient performance status evaluation stands as a predominant component in survival prognostication. While current performance status evaluation tools are limited to their subjective nature, the advent of wearable technology enables continual recordings of patients' activity and has the potential to measure performance status objectively. We hypothesize that wristband actigraphy monitoring devices can predict in-hospital death of end-stage cancer patients during the time of their hospital admissions. The objective of this study was to train and validate a long short-term memory (LSTM) deep-learning prediction model based on activity data of wearable actigraphy devices. The study recruited 60 end-stage cancer patients in a hospice care unit, with 28 deaths and 32 discharged in stable condition at the end of their hospital stay. The standard Karnofsky Performance Status score had an overall prognostic accuracy of 0.83. The LSTM prediction model based on patients' continual actigraphy monitoring had an overall prognostic accuracy of 0.83. Furthermore, the model performance improved with longer input data length up to 48 h. In conclusion, our research suggests the potential feasibility of wristband actigraphy to predict end-of-life admission outcomes in palliative care for end-stage cancer patients. Clinical Trial Registration: The study protocol was registered on ClinicalTrials.gov (ID: NCT04883879).

Signal properties of split-spectrum amplitude decorrelation angiography for quantitative optical coherence tomography-based velocimetry.

Split-spectrum amplitude-decorrelation angiography (SSADA) is a noninvasive and three-dimensional angiographic technique with a microscale spatial resolution based on optical coherence tomography. The SSADA signal is known to be correlated with the blood flow velocity and the quantitative velocimetry with SSADA has been expected; however, the signal properties of SSADA are not completely understood due to lack of comprehensive investigations of parameters related to SSADA signals. In this study, phantom experiments were performed to comprehensively investigate the relation of SSADA signals with flow velocities, time separations, particle concentrations, signal-to-noise ratios, beam spot sizes, and viscosities, and revealed that SSADA signals reflect the spatial commonality within a coherence volume between adjacent A-scans.

Optimization of extracorporeal membrane oxygenation therapy using near-infrared spectroscopy to assess changes in peripheral circulation: A pilot study.

Near-infrared spectroscopy (NIRS) has been proposed as a noninvasive modality for detecting complications in patients undergoing extracorporeal membrane oxygenation (ECMO), and it can simultaneously reveal the global circulatory status of these patients. We optimized ECMO therapy on the basis of real-time peripheral NIRS probing. Three patients underwent venoarterial (VA) ECMO and one patient underwent venovenous (VV) ECMO. All patients received peripheral ECMO cannulation with routine distal perfusion catheter placement. We designed an experimental protocol to adjust ECMO blood flow over 1 hour. Hemodynamic responses were measured using NIRS devices attached to the calf at approximately 60% of the distance from the ankle to the knee. HbO2 levels change substantially with adjustments in ECMO flow, and they are more sensitive than HHb levels and the tissue saturation index (TSI) are. NIRS for optimizing ECMO therapy may be reliable for monitoring global circulatory status.

Intelligent classification of platelet aggregates by agonist type.

Platelets are anucleate cells in blood whose principal function is to stop bleeding by forming aggregates for hemostatic reactions. In addition to their participation in physiological hemostasis, platelet aggregates are also involved in pathological thrombosis and play an important role in inflammation, atherosclerosis, and cancer metastasis. The aggregation of platelets is elicited by various agonists, but these platelet aggregates have long been considered indistinguishable and impossible to classify. Here we present an intelligent method for classifying them by agonist type. It is based on a convolutional neural network trained by high-throughput imaging flow cytometry of blood cells to identify and differentiate subtle yet appreciable morphological features of platelet aggregates activated by different types of agonists. The method is a powerful tool for studying the underlying mechanism of platelet aggregation and is expected to open a window on an entirely new class of clinical diagnostics, pharmacometrics, and therapeutics.

Platelets are small cells in the blood that primarily help stop bleeding after an injury by sticking together with other blood cells to form a clot that seals the broken blood vessel. Blood clots, however, can sometimes cause harm. For example, if a clot blocks the blood flow to the heart or the brain, it can result in a heart attack or stroke, respectively. Blood clots have also been linked to harmful inflammation and the spread of cancer, and there are now preliminary reports of remarkably high rates of clotting in COVID-19 patients in intensive care units. A variety of chemicals can cause platelets to stick together. It has long been assumed that it would be impossible to tell apart the clots formed by different chemicals (which are also known as agonists). This is largely because these aggregates all look very similar under a microscope, making it incredibly time consuming for someone to look at enough microscopy images to reliably identify the subtle differences between them. However, finding a way to distinguish the different types of platelet aggregates could lead to better ways to diagnose or treat blood vessel-clogging diseases. To make this possible, Zhou, Yasumoto et al. have developed a method called the "intelligent platelet aggregate classifier" or iPAC for short. First, numerous clot-causing chemicals were added to separate samples of platelets taken from healthy human blood. The method then involved using high-throughput techniques to take thousands of images of these samples. Then, a sophisticated computer algorithm called a deep learning model analyzed the resulting image dataset and "learned" to distinguish the chemical causes of the platelet aggregates based on subtle differences in their shapes. Finally, Zhou, Yasumoto et al. verified iPAC method's accuracy using a new set of human platelet samples. The iPAC method may help scientists studying the steps that lead to clot formation. It may also help clinicians distinguish which clot-causing chemical led to a patient's heart attack or stroke. This could help them choose whether aspirin or another anti-platelet drug would be the best treatment. But first more studies are needed to confirm whether this method is a useful tool for drug selection or diagnosis.

Intelligent frequency-shifted optofluidic time-stretch quantitative phase imaging.

Optofluidic time-stretch quantitative phase imaging (OTS-QPI) is a powerful tool as it enables high-throughput (>10,000 cell/s) QPI of single live cells. OTS-QPI is based on decoding temporally stretched spectral interferograms that carry the spatial profiles of cells flowing on a microfluidic chip. However, the utility of OTS-QPI is troubled by difficulties in phase retrieval from the high-frequency region of the temporal interferograms, such as phase-unwrapping errors, high instrumentation cost, and large data volume. To overcome these difficulties, we propose and experimentally demonstrate frequency-shifted OTS-QPI by bringing the phase information to the baseband region. Furthermore, to show its boosted utility, we use it to demonstrate image-based classification of leukemia cells with high accuracy over 96% and evaluation of drug-treated leukemia cells via deep learning.

Virtual-freezing fluorescence imaging flow cytometry.

By virtue of the combined merits of flow cytometry and fluorescence microscopy, imaging flow cytometry (IFC) has become an established tool for cell analysis in diverse biomedical fields such as cancer biology, microbiology, immunology, hematology, and stem cell biology. However, the performance and utility of IFC are severely limited by the fundamental trade-off between throughput, sensitivity, and spatial resolution. Here we present an optomechanical imaging method that overcomes the trade-off by virtually freezing the motion of flowing cells on the image sensor to effectively achieve 1000 times longer exposure time for microscopy-grade fluorescence image acquisition. Consequently, it enables high-throughput IFC of single cells at >10,000 cells s-1 without sacrificing sensitivity and spatial resolution. The availability of numerous information-rich fluorescence cell images allows high-dimensional statistical analysis and accurate classification with deep learning, as evidenced by our demonstration of unique applications in hematology and microbiology.