2D-DIGE based urinary proteomics and functional enrichment studies to reveal novel Mycobacterium tuberculosis and human protein biomarker candidates for pulmonary tuberculosis.

In the absence of a sensitive and specific diagnostic modality capable of detecting all forms of tuberculosis (TB), proteomics may identify specific Mycobacterium tuberculosis (M.tb) proteins in urine, with a potential as biomarkers. To identify candidate biomarkers for TB, proteome profile of urine from pulmonary TB patients was compared with non-disease controls (NDC) and disease controls (DC, Streptococcus pneumonia infected patients) using a combination of two-dimensional difference gel electrophoresis (2D-DIGE) and liquid chromatography tandem mass spectrometry (LCMS/MS). Eleven differentially expressed host proteins and Eighteen high abundant M.tb proteins were identified. Protein-protein interactome (PPI) and functional enrichment analyses like Gene Ontologies, Reactome pathway etc. demonstrated that the human proteins mainly belong to extracellular space and show physiological pathways for immune response and hematological disorders. Whereas, M.tb proteins belong to the cell periphery, plasma membrane and cell wall, and demonstrated catalytic, nucleotide binding and ATPase activities along with other functional processes. The study findings provide valuable inputs about the biomarkers of TB and shed light on the probable disease consequences as an outcome of the bacterial pathogenicity.

Metabolomic Profiling for Diagnosis and Prognostication in Surgery: A Scoping Review.

OBJECTIVE: This review assimilates and critically evaluates available literature regarding the use of metabolomic profiling in surgical decision-making. BACKGROUND: Metabolomic profiling is performed by nuclear magnetic resonance spectroscopy or mass spectrometry of biofluids and tissues to quantify biomarkers (ie, sugars, amino acids, and lipids), producing diagnostic and prognostic information that has been applied among patients with cardiovascular disease, inflammatory bowel disease, cancer, and solid organ transplants. METHODS: PubMed was searched from 1995 to 2019 to identify studies investigating metabolomic profiling of surgical patients. Articles were included and assimilated into relevant categories per PRISMA-ScR guidelines. Results were summarized with descriptive analytical methods. RESULTS: Forty-seven studies were included, most of which were retrospective studies with small sample sizes using various combinations of analytic techniques and types of biofluids and tissues. Results suggest that metabolomic profiling has the potential to effectively screen for surgical diseases, suggest diagnoses, and predict outcomes such as postoperative complications and disease recurrence. Major barriers to clinical adoption include a lack of high-level evidence from prospective studies, heterogeneity in study design regarding tissue and biofluid procurement and analytical methods, and the absence of large, multicenter metabolome databases to facilitate systematic investigation of the efficacy, reproducibility, and generalizability of metabolomic profiling diagnoses and prognoses. CONCLUSIONS: Metabolomic profiling research would benefit from standardization of study design and analytic approaches. As technologies improve and knowledge garnered from research accumulates, metabolomic profiling has the potential to provide personalized diagnostic and prognostic information to support surgical decision-making from preoperative to postdischarge phases of care.

Kinetic Isotope Effects on Electron Transfer Across Self-Assembled Monolayers on Gold.

Reactions requiring controlled delivery of protons and electrons are important in storage of energy in small molecules. While control over proton transfer can be achieved by installing appropriate chemical functionality in the catalyst, control of electron-transfer (ET) rates can be achieved by utilizing self-assembled monolayers (SAMs) on electrodes. Thus, a deeper understanding of the ET through SAM to an immobilized or covalently attached redox-active species is desirable. Long-range ET across several SAM-covered Au electrodes to covalently attached ferrocene is investigated using protonated and deuterated thiols (R-SH/R-SD). The rate of tunneling is measured using both chronoamperometry and cyclic voltammetry, and it shows a prominent kinetic isotope effect (KIE). The KIE is ∼2 (normal) for medium-chain-length thiols but ∼0.47 (inverse) for long-chain thiols. These results imply substantial contribution from the classical modes at the Au-(H)SR interface, which shifts substantially upon deuteration of the thiols, to the ET process. The underlying H/D KIE of these exchangeable thiol protons should be considered when analyzing solvent isotope effects in catalysis utilizing SAM.

Study of aqueous cytokines in patients with different patterns of diabetic macular edema based on optical coherence tomography.

PURPOSE: To compare aqueous angiogenic and inflammatory cytokine concentrations in different patterns of diabetic macular edema (DME) based on optical coherence tomography (OCT). METHODS: This prospective study was conducted between July 1, 2015, and March 31, 2016, for 9 months. Aqueous samples were obtained from 52 consecutive DME patients and 16 controls. DME patients were divided according to OCT patterns into diffuse retinal thickening (DRT; n = 17), cystoid macular edema (CME; n = 20) and serous retinal detachment (SRD; n = 15) groups. Interleukin (IL)-6, IL-8, vascular endothelial growth factor (VEGF) and tumor necrosis factor alpha (TNF-α) levels were measured by RayBio(R) Human ELISA Kit. RESULTS: IL-6, IL-8 and VEGF levels differed significantly between three DME groups (p < 0.001 in all cases), but the differences in TNF-α levels were not significant (p = 0.226). VEGF and IL-6 levels correlated with central foveal thickness in DRT and SRD groups, respectively. CONCLUSION: Aqueous cytokine levels vary with different morphological patterns of DME though the role of TNF-α needs to be studied further, and both anti-angiogenic and anti-inflammatory agents are required simultaneously for treatment of DME.

Effect on contrast sensitivity after clear, yellow and orange intraocular lens implantation.

The objective of this study is to evaluate contrast sensitivity function (CSF) after clear, yellow- and orange-tinted intraocular lens (IOL) implantation. This was a prospective randomized study of 98 patients with senile cataract for a period of 6 months from day 1 of August 2014 to day 31 of January 2015. After phacoemulsification, 33 patients were implanted with clear IOLs (AcrySof UV-filtering IOL, SA60AT), 32 patients were implanted with yellow coloured IOLs (AcrySof Natural blue-light-attenuating and UV-filtering IOL, SN60AT with IMPRUV(®) filter) and 33 patients were implanted with orange-tinted blue-filtering IOLs (PC440Y Optech). After 1 month, monocular CSF was done under photopic (85 cd/m(2)) and mesopic (3 cd/m(2)) illumination condition with CSV-1000 test. The best corrected visual acuity (BCVA) after 1 month was 0.021 ± 0.058 logMAR for clear lens, 0.022 ± 0.059 logMAR for yellow lens and 0.019 ± 0.065 logMAR for orange lens (p = 0.989). Uniocular average photopic contrast sensitivity was 1.36 ± 0.19, 1.43 ± 0.18 and 1.46 ± 0.15 log units for clear lens, yellow lens and orange lens, respectively (statistically not significant; p = 0.076). Average mesopic contrast sensitivity was 1.02 ± 0.21 log units for clear lens, 1.00 ± 0.17 log units for yellow lens and 0.99 ± 0.15 log units for orange lens (statistically not significant; p = 0.771). Yellow or orange coloured blue-filtering IOLs are comparable to clear IOLs in terms of photopic and mesopic contrast sensitivity.

The protonation state of thiols in self-assembled monolayers on roughened Ag/Au surfaces and nanoparticles.

The protonation state of thiols in self-assembled monolayers (SAMs) on Ag and Au surfaces and nanoparticles (NPs) has been an issue of contestation. It has been recently demonstrated that deuterating the thiol proton produces ostentatious changes in the Raman spectra of thiols and can be used to detect the presence of the thiol functional group. Surface enhanced Raman spectroscopy (SERS) of H/D substituted aliphatic thiols on Ag surfaces clearly shows the presence of S-H vibration between 2150-2200 cm(-1) which shifts by 400 cm(-1) upon deuteration and a simultaneous >20 cm(-1) shift in the C-S vibration of thiol deuteration. Large shifts (>15 cm(-1)) in the C-S vibration are also observed for alkyl thiol SAMs on Au surfaces. Alternatively, neither the S-H vibration nor the H/D isotope effect on the C-S vibration is observed for alkyl thiol SAMs on Ag/Au NPs. XPS data on Ag/Au surfaces bearing aliphatic thiol SAMs show the presence of both protonated and deprotonated thiols while on Ag/Au NPs only deprotonated thiols are detected. These data suggest that aliphatic thiol SAMs on Au/Ag surfaces are partially protonated whereas they are totally deprotonated on Au/Ag NPs. Aromatic PhSH SAMs on Ag/Au surfaces and Ag/Au NPs do not show these vibrations or H/D shifts as well indicating that the thiols are deprotonated at these interfaces.

Convenient detection of the thiol functional group using H/D isotope sensitive Raman spectroscopy.

Raman spectra of several thiols (amino acids, peptides and organic) show that the C-S-H bending mode (ßCSH) shifts from ∼850 cm(-1) to ∼620 cm(-1) on deuteration of the thiol proton by simply dissolving them in D2O and CD3OD where detection by (1)H NMR is not possible. A nondestructive analytical tool for the detection of thiols in solid/neat and in solution is developed.

Effect of axial ligand, spin state, and hydrogen bonding on the inner-sphere reorganization energies of functional models of cytochrome P450.

Using a combination of self-assembly and synthesis, bioinspired electrodes having dilute iron porphyrin active sites bound to axial thiolate and imidazole axial ligands are created atop self-assembled monolayers (SAMs). Resonance Raman data indicate that a picket fence architecture results in a high-spin (HS) ground state (GS) in these complexes and a hydrogen-bonding triazole architecture results in a low-spin (LS) ground state. The reorganization energies (λ) of these thiolate- and imidazole-bound iron porphyrin sites for both HS and LS states are experimentally determined. The λ of 5C HS imidazole and thiolate-bound iron porphyrin active sites are 10-16 kJ/mol, which are lower than their 6C LS counterparts. Density functional theory (DFT) calculations reproduce these data and indicate that the presence of significant electronic relaxation from the ligand system lowers the geometric relaxation and results in very low λ in these 5C HS active sites. These calculations indicate that loss of one-half a π bond during redox in a LS thiolate bound active site is responsible for its higher λ relative to a σ-donor ligand-like imidazole. Hydrogen bonding to the axial ligand leads to a significant increase in λ irrespective of the spin state of the iron center. The results suggest that while the hydrogen bonding to the thiolate in the 5C HS thiolate bound active site of cytochrome P450 (cyp450) shifts the potential up, resulting in a negative ΔG, it also increases λ resulting in an overall low barrier for the electron transfer process.

Transformers and large language models in healthcare: A review.

With Artificial Intelligence (AI) increasingly permeating various aspects of society, including healthcare, the adoption of the Transformers neural network architecture is rapidly changing many applications. Transformer is a type of deep learning architecture initially developed to solve general-purpose Natural Language Processing (NLP) tasks and has subsequently been adapted in many fields, including healthcare. In this survey paper, we provide an overview of how this architecture has been adopted to analyze various forms of healthcare data, including clinical NLP, medical imaging, structured Electronic Health Records (EHR), social media, bio-physiological signals, biomolecular sequences. Furthermore, which have also include the articles that used the transformer architecture for generating surgical instructions and predicting adverse outcomes after surgeries under the umbrella of critical care. Under diverse settings, these models have been used for clinical diagnosis, report generation, data reconstruction, and drug/protein synthesis. Finally, we also discuss the benefits and limitations of using transformers in healthcare and examine issues such as computational cost, model interpretability, fairness, alignment with human values, ethical implications, and environmental impact.

Developing a fair and interpretable representation of the clock drawing test for mitigating low education and racial bias.

The clock drawing test (CDT) is a neuropsychological assessment tool to screen an individual's cognitive ability. In this study, we developed a Fair and Interpretable Representation of Clock drawing test (FaIRClocks) to evaluate and mitigate classification bias against people with less than 8 years of education, while screening their cognitive function using an array of neuropsychological measures. In this study, we represented clock drawings by a priorly published 10-dimensional deep learning feature set trained on publicly available data from the National Health and Aging Trends Study (NHATS). These embeddings were further fine-tuned with clocks from a preoperative cognitive screening program at the University of Florida to predict three cognitive scores: the Mini-Mental State Examination (MMSE) total score, an attention composite z-score (ATT-C), and a memory composite z-score (MEM-C). ATT-C and MEM-C scores were developed by averaging z-scores based on normative references. The cognitive screening classifiers were initially tested to see their relative performance in patients with low years of education (< = 8 years) versus patients with higher education (> 8 years) and race. Results indicated that the initial unweighted classifiers confounded lower education with cognitive compromise resulting in a 100% type I error rate for this group. Thereby, the samples were re-weighted using multiple fairness metrics to achieve sensitivity/specificity and positive/negative predictive value (PPV/NPV) balance across groups. In summary, we report the FaIRClocks model, with promise to help identify and mitigate bias against people with less than 8 years of education during preoperative cognitive screening.

Molecular biomarkers involved in the progression of gallbladder inflammatory lesions to invasive cancer: A proteomic approach.

The progression of gallbladder inflammatory lesions to invasive cancer remains poorly understood, necessitating research on biomarkers involved in this transition. This study aims to identify and validate proteins associated with this progression, offering insights into potential diagnostic biomarkers for gallbladder cancer (GBC). Label-free liquid chromatography assisted tandem mass spectrometry (LC-MS/MS) proteomics was performed on samples from 10 cases each of GBC and inflammatory lesions, with technical duplicates. Validation was conducted through the enzyme-linked immunosorbent assay (ELISA) using 80 samples (40 GBC and 40 inflammatory lesions). Bioinformatics tools analyzed protein-protein interaction (PPI) networks and pathways. Statistical correlations with clinicopathological variables were assessed. Prognostic evaluation utilized Kaplan-Meier survival analysis and Cox regression analyses. mRNA expressions were studied using real time-polymerase chain reaction (RT-PCR). Out of 5,714 proteins analyzed, 621 were differentially expressed. Three upregulated (the S100 calcium-binding protein P [S100P], polymeric immunoglobulin receptor [PIGR], and complement C1q-binding protein [C1QBP]) and two downregulated (transgelin [TAGLN] and calponin 1 [CNN1]) proteins showed significant expression. Pathway analysis implicated involvement of proteoglycans in cancer and glycosaminoglycan metabolism. Significant correlations were observed between protein concentrations and clinicopathological variables. Prognostic factors such as tumor size, lymph node metastasis, and preoperative bilirubin levels were associated with overall survival. Protein-based assays demonstrated higher resolution compared to mRNA analysis, suggesting their utility in GBC risk stratification. S100P, PIGR, C1QBP, TAGLN, and CNN1 emerge as potential protein-based biomarkers involved in the progression from gallbladder inflammatory lesions to invasive cancer. These findings hold promise for improved diagnostic and prognostic strategies in GBC management.

A Network Analysis of Digital Clock Drawing for Command and Copy Conditions.

Graphomotor and time-based variables from the digital Clock Drawing Test (dCDT) characterize cognitive functions. However, no prior publications have quantified the strength of the associations between digital clock variables as they are produced. We hypothesized that analysis of the production of clock features and their interrelationships, as suggested, will differ between the command and copy test conditions. Older adults aged 65+ completed a digital clock drawing to command and copy conditions. Using a Bayesian hill-climbing algorithm and bootstrapping (10,000 samples), we derived directed acyclic graphs (DAGs) to examine network structure for command and copy dCDT variables. Although the command condition showed moderate associations between variables (µ|ßz|= 0.34) relative to the copy condition (µ|ßz| = 0.25), the copy condition network had more connections (18/18 versus 15/18 command). Network connectivity across command and copy was most influenced by five of the 18 variables. The direction of dependencies followed the order of instructions better in the command condition network. Digitally acquired clock variables relate to one another but differ in network structure when derived from command or copy conditions. Continued analyses of clock drawing production should improve understanding of quintessential normal features to aid in early neurodegenerative disease detection.

APRICOT-Mamba: Acuity Prediction in Intensive Care Unit (ICU): Development and Validation of a Stability, Transitions, and Life-Sustaining Therapies Prediction Model.

On average, more than 5 million patients are admitted to intensive care units (ICUs) in the US, with mortality rates ranging from 10 to 29%. The acuity state of patients in the ICU can quickly change from stable to unstable, sometimes leading to life-threatening conditions. Early detection of deteriorating conditions can assist in more timely interventions and improved survival rates. While Artificial Intelligence (AI)-based models show potential for assessing acuity in a more granular and automated manner, they typically use mortality as a proxy of acuity in the ICU. Furthermore, these methods do not determine the acuity state of a patient (i.e., stable or unstable), the transition between acuity states, or the need for life-sustaining therapies. In this study, we propose APRICOT-M (Acuity Prediction in Intensive Care Unit-Mamba), a 1M-parameter state space-based neural network to predict acuity state, transitions, and the need for life-sustaining therapies in real-time among ICU patients. The model integrates ICU data in the preceding four hours (including vital signs, laboratory results, assessment scores, and medications) and patient characteristics (age, sex, race, and comorbidities) to predict the acuity outcomes in the next four hours. Our state space-based model can process sparse and irregularly sampled data without manual imputation, thus reducing the noise in input data and increasing inference speed. The model was trained on data from 107,473 patients (142,062 ICU admissions) from 55 hospitals between 2014-2017 and validated externally on data from 74,901 patients (101,356 ICU admissions) from 143 hospitals. Additionally, it was validated temporally on data from 12,927 patients (15,940 ICU admissions) from one hospital in 2018-2019 and prospectively on data from 215 patients (369 ICU admissions) from one hospital in 2021-2023. Three datasets were used for training and evaluation: the University of Florida Health (UFH) dataset, the electronic ICU Collaborative Research Database (eICU), and the Medical Information Mart for Intensive Care (MIMIC)-IV dataset. APRICOT-M significantly outperforms the baseline acuity assessment, Sequential Organ Failure Assessment (SOFA), for mortality prediction in both external (AUROC 0.95 CI: 0.94-0.95 compared to 0.78 CI: 0.78-0.79) and prospective (AUROC 0.99 CI: 0.97-1.00 compared to 0.80 CI: 0.65-0.92) cohorts, as well as for instability prediction (external AUROC 0.75 CI: 0.74-0.75 compared to 0.51 CI: 0.51-0.51, and prospective AUROC 0.69 CI: 0.64-0.74 compared to 0.53 CI: 0.50-0.57). This tool has the potential to help clinicians make timely interventions by predicting the transition between acuity states and decision-making on life-sustaining within the next four hours in the ICU.

Wearable sensors in patient acuity assessment in critical care.

Acuity assessments are vital for timely interventions and fair resource allocation in critical care settings. Conventional acuity scoring systems heavily depend on subjective patient assessments, leaving room for implicit bias and errors. These assessments are often manual, time-consuming, intermittent, and challenging to interpret accurately, especially for healthcare providers. This risk of bias and error is likely most pronounced in time-constrained and high-stakes environments, such as critical care settings. Furthermore, such scores do not incorporate other information, such as patients' mobility level, which can indicate recovery or deterioration in the intensive care unit (ICU), especially at a granular level. We hypothesized that wearable sensor data could assist in assessing patient acuity granularly, especially in conjunction with clinical data from electronic health records (EHR). In this prospective study, we evaluated the impact of integrating mobility data collected from wrist-worn accelerometers with clinical data obtained from EHR for estimating acuity. Accelerometry data were collected from 87 patients wearing accelerometers on their wrists in an academic hospital setting. The data was evaluated using five deep neural network models: VGG, ResNet, MobileNet, SqueezeNet, and a custom Transformer network. These models outperformed a rule-based clinical score (Sequential Organ Failure Assessment, SOFA) used as a baseline when predicting acuity state (for ground truth we labeled as unstable patients if they needed life-supporting therapies, and as stable otherwise), particularly regarding the precision, sensitivity, and F1 score. The results demonstrate that integrating accelerometer data with demographics and clinical variables improves predictive performance compared to traditional scoring systems in healthcare. Deep learning models consistently outperformed the SOFA score baseline across various scenarios, showing notable enhancements in metrics such as the area under the receiver operating characteristic (ROC) Curve (AUC), precision, sensitivity, specificity, and F1 score. The most comprehensive scenario, leveraging accelerometer, demographics, and clinical data, achieved the highest AUC of 0.73, compared to 0.53 when using SOFA score as the baseline, with significant improvements in precision (0.80 vs. 0.23), specificity (0.79 vs. 0.73), and F1 score (0.77 vs. 0.66). This study demonstrates a novel approach beyond the simplistic differentiation between stable and unstable conditions. By incorporating mobility and comprehensive patient information, we distinguish between these states in critically ill patients and capture essential nuances in physiology and functional status. Unlike rudimentary definitions, such as equating low blood pressure with instability, our methodology delves deeper, offering a more holistic understanding and potentially valuable insights for acuity assessment.

Resonance Raman and electrocatalytic behavior of thiolate and imidazole bound iron porphyrin complexes on self assembled monolayers: functional modeling of cytochrome P450.

Electrodes bearing thiolate and imidazole coordinated iron porphyrin catalysts are constructed and characterized using resonance Raman spectroscopy, absorption spectroscopy, and electrochemistry. The cyclic voltammetry data and their pH dependences are used to establish the nature of the exchangeable trans ligands in both of these cases. In situ monitoring of partially reduced oxygen species (PROS) produced during O(2) reduction using rotating ring disc electrochemistry (RRDE) experiments provide direct insight into the "push-effect" of the thiolate ligand. The thiolate bound iron porphyrin electrode generates highly oxidizing species on the electrode during electrocatalytic O(2) reductions which are very reactive. These surfaces can utilize these oxidants to catalytically hydroxylate strong C-H bonds using molecular O(2) with turnover numbers as high as 200.

Identification of potential biomarkers to predict organ morbidity in COVID-19: A repository based proteomics perspective.

SARS-CoV-2 causes substantial extrapulmonary manifestations in addition to pulmonary disease. Some of the major organs affected are cardiovascular, hematological and thrombotic, renal, neurological, and digestive systems. These types of muti-organ dysfunctions make it difficult and challenging for clinicians to manage and treat COVID-19 patients. The article focuses to identify potential protein biomarkers that can flag various organ systems affected in COVID-19. Publicly reposited high throughput proteomic data from human serum (HS), HEK293T/17 (HEK) and Vero E6 (VE) kidney cell culture were downloaded from ProteomeXchange consortium. The raw data was analyzed in Proteome Discoverer 2.4 to delineate the complete list of proteins in the three studies. These proteins were analyzed in Ingenuity Pathway Analysis (IPA) to associate them to various organ diseases. The shortlisted proteins were analyzed in MetaboAnalyst 5.0 to shortlist potential biomarker proteins. These were then assessed for disease-gene association in DisGeNET and validated by Protein-protein interactome (PPI) and functional enrichment studies (GO_BP, KEGG and Reactome pathways) in STRING. Protein profiling resulted in shortlisting 20 proteins in 7 organ systems. Of these 15 proteins showed at least 1.25-fold changes with a sensitivity and specificity of 70%. Association analysis further shortlisted 10 proteins with a potential association with 4 organ diseases. Validation studies established possible interacting networks and pathways affected, confirmingh the ability of 6 of these proteins to flag 4 different organ systems affected in COVID-19 disease. This study helps to establish a platform to seek protein signatures in different clinical phenotypes of COVID-19. The potential biomarker candidates that can flag organ systems involved are: (a) Vitamin K-dependent protein S and Antithrombin-III for hematological disorders; (b) Voltage-dependent anion-selective channel protein 1 for neurological disorders; (c) Filamin-A for cardiovascular disorder and, (d) Peptidyl-prolyl cis-trans isomerase A and Peptidyl-prolyl cis-trans isomerase FKBP1A for digestive disorders.

Explainable semi-supervised deep learning shows that dementia is associated with small, avocado-shaped clocks with irregularly placed hands.

The clock drawing test is a simple and inexpensive method to screen for cognitive frailties, including dementia. In this study, we used the relevance factor variational autoencoder (RF-VAE), a deep generative neural network, to represent digitized clock drawings from multiple institutions using an optimal number of disentangled latent factors. The model identified unique constructional features of clock drawings in a completely unsupervised manner. These factors were examined by domain experts to be novel and not extensively examined in prior research. The features were informative, as they distinguished dementia from non-dementia patients with an area under receiver operating characteristic (AUC) of 0.86 singly, and 0.96 when combined with participants' demographics. The correlation network of the features depicted the "typical dementia clock" as having a small size, a non-circular or "avocado-like" shape, and incorrectly placed hands. In summary, we report a RF-VAE network whose latent space encoded novel constructional features of clocks that classify dementia from non-dementia patients with high performance.

Label-Free Proteomics of Oral Mucosa Tissue to Identify Potential Biomarkers That Can Flag Predilection of Precancerous Lesions to Oral Cell Carcinoma: A Preliminary Study.

Oral squamous cell carcinomas are mostly preceded by precancerous lesions such as leukoplakia and erythroplakia. Our study is aimed at identifying potential biomarker proteins in precancerous lesions of leukoplakia and erythroplakia that can flag their transformation to oral cancer. Four biological replicate samples from clinical phenotypes of healthy control, leukoplakia, erythroplakia, and oral carcinoma were annotated based on clinical screening and histopathological evaluation of buccal mucosa tissue. Differentially expressed proteins were delineated using a label-free quantitative proteomic experiment done on an Orbitrap Fusion Tribrid mass spectrometer in three technical replicate sets of samples. Raw files were processed using MaxQuant version 2.0.1.0, and downstream analysis was done via Perseus version 1.6.15.0. Validation included functional annotation based on biological processes and pathways using the ClueGO plug-in of Cytoscape. Hierarchical clustering and principal component analysis were performed using the ClustVis tool. Across control, leukoplakia, and cancer, L-lactate dehydrogenase A chain, plectin, and WD repeat-containing protein 1 were upregulated, whereas thioredoxin 1 and spectrin alpha chain, nonerythrocytic 1 were downregulated. Across control, erythroplakia, and cancer, L-lactate dehydrogenase A chain was upregulated whereas aldehyde dehydrogenase 2, peroxiredoxin 1, heat shock 70 kDa protein 1B, and spectrin alpha chain, nonerythrocytic 1 were downregulated. We found that proteins involved in leukoplakia were associated with alteration in cytoskeletal disruption and glycolysis, while in erythroplakia, they were associated with alteration in response to oxidative stress and glycolysis across phenotypes. Hierarchical clustering subgrouped half of precancerous samples under the main branch of the control and the remaining half under carcinoma. Similarly, principal component analysis identified segregated clusters of control, precancerous lesions, and cancer, but erythroplakia phenotypes, in particular, overlapped more with the cancer cluster. Qualitative and quantitative protein signatures across control, precancer, and cancer phenotypes explain possible functional outcomes that dictate malignant transformation to oral carcinoma.

FaIRClocks: Fair and Interpretable Representation of the Clock Drawing Test for mitigating classifier bias against lower educational groups.

The clock drawing test (CDT) is a neuropsychological assessment tool to evaluate a patient's cognitive ability. In this study, we developed a Fair and Interpretable Representation of Clock drawing tests (FaIRClocks) to evaluate and mitigate bias against people with lower education while predicting their cognitive status. We represented clock drawings with a 10-dimensional latent embedding using Relevance Factor Variational Autoencoder (RF-VAE) network pretrained on publicly available clock drawings from the National Health and Aging Trends Study (NHATS) dataset. These embeddings were later fine-tuned for predicting three cognitive scores: the Mini-Mental State Examination (MMSE) total score, attention composite z-score (ATT-C), and memory composite z-score (MEM-C). The classifiers were initially tested to see their relative performance in patients with low education (<= 8 years) versus patients with higher education (> 8 years). Results indicated that the initial unweighted classifiers confounded lower education with cognitive impairment, resulting in a 100% type I error rate for this group. Thereby, the samples were re-weighted using multiple fairness metrics to achieve balanced performance. In summary, we report the FaIRClocks model, which a) can identify attention and memory deficits using clock drawings and b) exhibits identical performance between people with higher and lower education levels.

Dynamic Delirium Prediction in the Intensive Care Unit using Machine Learning on Electronic Health Records.

Delirium is a syndrome of acute brain failure which is prevalent amongst older adults in the Intensive Care Unit (ICU). Incidence of delirium can significantly worsen prognosis and increase mortality, therefore necessitating its rapid and continual assessment in the ICU. Currently, the common approach for delirium assessment is manual and sporadic. Hence, there exists a critical need for a robust and automated system for predicting delirium in the ICU. In this work, we develop a machine learning (ML) system for real-time prediction of delirium using Electronic Health Record (EHR) data. Unlike prior approaches which provide one delirium prediction label per entire ICU stay, our approach provides predictions every 12 hours. We use the latest 12 hours of ICU data, along with patient demographic and medical history data, to predict delirium risk in the next 12-hour window. This enables delirium risk prediction as soon as 12 hours after ICU admission. We train and test four ML classification algorithms on longitudinal EHR data pertaining to 16,327 ICU stays of 13,395 patients covering a total of 56,297 12-hour windows in the ICU to predict the dynamic incidence of delirium. The best performing algorithm was Categorical Boosting which achieved an area under receiver operating characteristic curve (AUROC) of 0.87 (95% Confidence Interval; C.I, 0.86-0.87). The deployment of this ML system in ICUs can enable early identification of delirium, thereby reducing its deleterious impact on long-term adverse outcomes, such as ICU cost, length of stay and mortality.