Type 2 diabetes (T2D) is a global health concern with increasing prevalence. Comorbid hypothyroidism (HT) exacerbates kidney, cardiac, neurological and other complications of T2D; these risks can be mitigated pharmacologically upon detecting HT. The current HT standard of care (SOC) screening in T2D is infrequent, delaying HT diagnosis and treatment. We present a first-to-date machine learning algorithm (MLA) clinical decision tool to classify patients as low vs. high risk for developing HT comorbid with T2D; the MLA was developed using readily available patient data from harmonized multinational datasets. The MLA was trained on data from NIH All of US (AoU) and UK Biobank (UKBB) (Combined dataset) and achieved a high negative predictive value (NPV) of 0.989 and an AUROC of 0.762 in the Combined dataset, exceeding AUROCs for the models trained on AoU or UKBB alone (0.666 and 0.622, respectively), indicating that increasing dataset diversity for MLA training improves performance. This high-NPV automated tool can supplement SOC screening and rule out T2D patients with low HT risk, allowing for the prioritization of lab-based testing for at-risk patients. Conversely, an MLA output that designates a patient to be at risk of developing HT allows for tailored clinical management and thereby promotes improved patient outcomes.
LLMs can accomplish specialized medical knowledge tasks, however, equitable access is hindered by the extensive fine-tuning, specialized medical data requirement, and limited access to proprietary models. Open-source (OS) medical LLMs show performance improvements and provide the transparency and compliance required in healthcare. We present OpenMedLM, a prompting platform delivering state-of-the-art (SOTA) performance for OS LLMs on medical benchmarks. We evaluated OS foundation LLMs (7B-70B) on medical benchmarks (MedQA, MedMCQA, PubMedQA, MMLU medical-subset) and selected Yi34B for developing OpenMedLM. Prompting strategies included zero-shot, few-shot, chain-of-thought, and ensemble/self-consistency voting. OpenMedLM delivered OS SOTA results on three medical LLM benchmarks, surpassing previous best-performing OS models that leveraged costly and extensive fine-tuning. OpenMedLM displays the first results to date demonstrating the ability of OS foundation models to optimize performance, absent specialized fine-tuning. The model achieved 72.6% accuracy on MedQA, outperforming the previous SOTA by 2.4%, and 81.7% accuracy on MMLU medical-subset, establishing itself as the first OS LLM to surpass 80% accuracy on this benchmark. Our results highlight medical-specific emergent properties in OS LLMs not documented elsewhere to date and validate the ability of OS models to accomplish healthcare tasks, highlighting the benefits of prompt engineering to improve performance of accessible LLMs for medical applications.
Benchmarking , Humans , Software
Background/Objective: Autism spectrum disorder (ASD) is a neurodevelopmental condition characterized by lifelong impacts on functional social and daily living skills, and restricted, repetitive behaviors (RRBs). Applied behavior analysis (ABA), the gold-standard treatment for ASD, has been extensively validated. ABA access is hindered by limited availability of qualified professionals and logistical and financial barriers. Scientifically validated, parent-led ABA can fill the accessibility gap by overcoming treatment barriers. This retrospective cohort study examines how our ABA treatment model, utilizing parent behavior technicians (pBTs) to deliver ABA, impacts adaptive behaviors and interfering behaviors (IBs) in a cohort of children on the autism spectrum with varying ASD severity levels, and with or without clinically significant IBs. Methods: Clinical outcomes of 36 patients ages 3-15 years were assessed using longitudinal changes in Vineland-3 after 3+ months of pBT-delivered ABA treatment. Results: Within the pBT model, our patients demonstrated clinically significant improvements in Vineland-3 Composite, domain, and subdomain scores, and utilization was higher in severe ASD. pBTs utilized more prescribed ABA when children initiated treatment with clinically significant IBs, and these children also showed greater gains in their Composite scores. Study limitations include sample size, inter-rater reliability, potential assessment metric bias and schedule variability, and confounding intrinsic or extrinsic factors. Conclusion: Overall, our pBT model facilitated high treatment utilization and showed robust effectiveness, achieving improved adaptive behaviors and reduced IBs when compared to conventional ABA delivery. The pBT model is a strong contender to fill the widening treatment accessibility gap and represents a powerful tool for addressing systemic problems in ABA treatment delivery.
BACKGROUND: Applied behavioral analysis (ABA) is regarded as the gold standard treatment for autism spectrum disorder (ASD) and has the potential to improve outcomes for patients with ASD. It can be delivered at different intensities, which are classified as comprehensive or focused treatment approaches. Comprehensive ABA targets multiple developmental domains and involves 20-40 h/week of treatment. Focused ABA targets individual behaviors and typically involves 10-20 h/week of treatment. Determining the appropriate treatment intensity involves patient assessment by trained therapists, however, the final determination is highly subjective and lacks a standardized approach. In our study, we examined the ability of a machine learning (ML) prediction model to classify which treatment intensity would be most suited individually for patients with ASD who are undergoing ABA treatment. METHODS: Retrospective data from 359 patients diagnosed with ASD were analyzed and included in the training and testing of an ML model for predicting comprehensive or focused treatment for individuals undergoing ABA treatment. Data inputs included demographics, schooling, behavior, skills, and patient goals. A gradient-boosted tree ensemble method, XGBoost, was used to develop the prediction model, which was then compared against a standard of care comparator encompassing features specified by the Behavior Analyst Certification Board treatment guidelines. Prediction model performance was assessed via area under the receiver-operating characteristic curve (AUROC), sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). RESULTS: The prediction model achieved excellent performance for classifying patients in the comprehensive versus focused treatment groups (AUROC: 0.895; 95% CI 0.811-0.962) and outperformed the standard of care comparator (AUROC 0.767; 95% CI 0.629-0.891). The prediction model also achieved sensitivity of 0.789, specificity of 0.808, PPV of 0.6, and NPV of 0.913. Out of 71 patients whose data were employed to test the prediction model, only 14 misclassifications occurred. A majority of misclassifications (n = 10) indicated comprehensive ABA treatment for patients that had focused ABA treatment as the ground truth, therefore still providing a therapeutic benefit. The three most important features contributing to the model's predictions were bathing ability, age, and hours per week of past ABA treatment. CONCLUSION: This research demonstrates that the ML prediction model performs well to classify appropriate ABA treatment plan intensity using readily available patient data. This may aid with standardizing the process for determining appropriate ABA treatments, which can facilitate initiation of the most appropriate treatment intensity for patients with ASD and improve resource allocation.
Objective This study examines the implementation of a hybrid applied behavioral analysis (ABA) treatment model to determine its impact on autism spectrum disorder (ASD) patient outcomes. Methods Retrospective data were collected for 25 pediatric patients to measure progress before and after the implementation of a hybrid ABA treatment model under which therapists consistently captured session notes electronically regarding goals and patient progress. ABA treatment was streamlined for consistent delivery, with improved software utilization for tracking scheduling and progress. Eleven goals within three domains (behavioral, social, and communication) were examined. Results After the implementation of the hybrid model, the goal success rate improved by 9.7% compared to the baseline; 41.8% of goals showed improvement, 38.4% showed a flat trend, and 19.8% showed deterioration. Multiple goals trended upwards in 76% of the patients. Conclusion This pilot study demonstrated that enhancing the consistency with which ABA treatment is monitored/delivered can improve patient outcomes as seen through improved attainment of goals.
Mild cognitive impairment (MCI) is cognitive decline that can indicate future risk of Alzheimer's disease (AD). We developed and validated a machine learning algorithm (MLA), based on a gradient-boosted tree ensemble method, to analyze phenotypic data for individuals 55-88 years old (n = 493) diagnosed with MCI. Data were analyzed within multiple prediction windows and averaged to predict progression to AD within 24-48 months. The MLA outperformed the mini-mental state examination (MMSE) and three comparison models at all prediction windows on most metrics. Exceptions include sensitivity at 18 months (MLA and MMSE each achieved 0.600); and sensitivity at 30 and 42 months (MMSE marginally better). For all prediction windows, the MLA achieved AUROC ≥ 0.857 and NPV ≥ 0.800. With averaged data for the 24-48-month lookahead timeframe, the MLA outperformed MMSE on all metrics. This study demonstrates that machine learning may provide a more accurate risk assessment than the standard of care. This may facilitate care coordination, decrease healthcare expenditures, and maintain quality of life for patients at risk of progressing from MCI to AD.
Multifunctional composite coatings composed of metal oxide nanoparticles dispersed in polymer matrices are an advanced solution to solve the problem of stone heritage deterioration. Their innovative design is meant to be stable, durable, transparent, easy to apply and remove, non-toxic, hydrophobic, and permeable. Coating formulations for the protection of buildings and monuments have been intensively researched lately. Such formulations are based on multifunctional composite coatings incorporating metal oxides. The present work aims to combine the hydrophobic properties of sodium polyacrylate (NaPAC16) with the antimicrobial effectiveness, with promising antimicrobial results even in the absence of light, and good compatibility of MgO (a safe to use, low cost and environmentally friendly material) and TiO2 (with antibacterial and antifungal properties), in order to develop coatings for stone materials protection. MgO (pure phase periclase) and TiO2 (pure phase anatase) nanopowders were prepared through sol-gel method, specifically routes. Aqueous dispersions of hydrophobically modified polymer (NaPAC16, polyacrylic acid sodium salt) and MgO/TiO2 nanopowders were deposited through layer-by-layer dip coating technique on glass slides and through immersion on stone fragments closely resembling the mosaic stone from the fourth century AD Roman Mosaic Edifice, from Constanta, Romania. The oxide nanopowders were characterized by: Thermal analysis (TG/DTA), scanning electron microscopy (SEM), X-ray diffraction (XRD), BET specific surface area and porosity, and UV-Vis spectroscopy for band gap determination. An aqueous dispersion of modified polyacrylate polymer and oxide nanopowders was deposited on different substrates (glass slides, red bricks, gypsum mortars). Film hydrophobicity was verified by contact angle measurements. The colour parameters were evaluated. Photocatalytic and antimicrobial activity of the powders and composite coatings were tested.
The nanofiltration composite membranes were obtained by incorporation of KIT-6 ordered mesoporous silica, before and after its functionalization with amine groups, into polyphenylene-ether-ether-sulfone (PPEES) matrix. The incorporation of silica nanoparticles into PPEES polymer matrix was evidenced by FTIR and UV-VIS spectroscopy. SEM images of the membranes cross-section and their surface topology, evidenced by AFM, showed a low effect of KIT-6 silica nanoparticles loading and functionalization. The performances of the obtained membranes were appraised in permeation of Chaenomeles japonica fruit extracts and the selective separation of phenolic acids and flavonoids. The obtained results proved that the PPEES with functionalized KIT-6 nanofiltration membrane, we have prepared, is suitable for the polyphenolic compound's concentration from the natural extracts.
The Ti/hierarchical zeolites Y were obtained by direct and post synthesis methods and loaded with Fe(III) by ion-exchange and impregnation resulting Fe-Ti/hierarchical zeolites Y photocatalysts. The synthesized materials were characterized by XRD, SEM microscopy, N2 physical adsorption, Raman, UV-Vis and XPS and EPR spectroscopy. XRD patterns evidenced the crystalline structure of the zeolite Y in all materials, excepting the samples with higher Fe content. The presence of anatase was evidenced by XRD and Raman spectroscopy in the samples obtained by impregnation while α-Fe2O3 was depicted in the Raman spectra of the samples with Ti and lower Fe loading. SEM images and N2 adsorption-desorption isotherms confirmed the formation of mesopores together with microporous crystals of zeolite Y. The UV-Vis spectra proved a red-shifted adsorption band for samples with iron. In all these samples XPS shows Fe3+ as oxide on the surface and EPR Fe3+ in tetrahedral coordination. Different variables such as hierarchical structure, amount of iron, catalyst loading, concentration of pollutant solution, pH value were studied to estimate their effects on performances of photocatalysts in degradation of amoxicillin from aqueous solution in UV and Visible light. A higher adsorption capacity and degradation efficiency of amoxicillin (100%) was noticed for hierarchical materials, especially for higher iron oxide loaded samples.
Zeolites , Amoxicillin , Ferric Compounds , Oxides , Titanium
BACKGROUND: Although the peptidyl-prolyl isomerase, cyclophilin-A (peptidyl-prolyl isomerase, PPIA), has been studied for decades in the context of its intracellular functions, its extracellular roles as a major contributor to both inflammation and multiple cancers have more recently emerged. A wide range of activities have been ascribed to extracellular PPIA that include induction of cytokine and matrix metalloproteinase (MMP) secretion, which potentially underlie its roles in inflammation and tumorigenesis. However, there have been conflicting reports as to which particular signaling events are under extracellular PPIA regulation, which may be due to either cell-dependent responses and/or the use of commercial preparations recently shown to be highly impure. METHODS: We have produced and validated the purity of recombinant PPIA in order to subject it to a comparative analysis between different cell types. Specifically, we have used a combination of multiple methods such as luciferase reporter screens, translocation assays, phosphorylation assays, and nuclear magnetic resonance to compare extracellular PPIA activities in several different cell lines that included epithelial and monocytic cells. RESULTS: Our findings have revealed that extracellular PPIA activity is cell type-dependent and that PPIA signals via multiple cellular receptors beyond the single transmembrane receptor previously identified, Extracellular Matrix MetalloPRoteinase Inducer (EMMPRIN). Finally, while our studies provide important insight into the cell-specific responses, they also indicate that there are consistent responses such as nuclear factor kappa B (NFκB) signaling induced in all cell lines tested. CONCLUSIONS: We conclude that although extracellular PPIA activates several common pathways, it also targets different receptors in different cell types, resulting in a complex, integrated signaling network that is cell type-specific.
In this paper we consider chemotherapy in a spatial model of tumor growth. The model, which is of reaction-diffusion type, takes into account the complex interactions between the tumor and surrounding stromal cells by including densities of endothelial cells and the extra-cellular matrix. When no treatment is applied the model reproduces the typical dynamics of early tumor growth. The initially avascular tumor reaches a diffusion limited size of the order of millimeters and initiates angiogenesis through the release of vascular endothelial growth factor (VEGF) secreted by hypoxic cells in the core of the tumor. This stimulates endothelial cells to migrate towards the tumor and establishes a nutrient supply sufficient for sustained invasion. To this model we apply cytostatic treatment in the form of a VEGF-inhibitor, which reduces the proliferation and chemotaxis of endothelial cells. This treatment has the capability to reduce tumor mass, but more importantly, we were able to determine that inhibition of endothelial cell proliferation is the more important of the two cellular functions targeted by the drug. Further, we considered the application of a cytotoxic drug that targets proliferating tumor cells. The drug was treated as a diffusible substance entering the tissue from the blood vessels. Our results show that depending on the characteristics of the drug it can either reduce the tumor mass significantly or in fact accelerate the growth rate of the tumor. This result seems to be due to complicated interplay between the stromal and tumor cell types and highlights the importance of considering chemotherapy in a spatial context.
Antineoplastic Agents/pharmacology , Models, Immunological , Neoplasms/immunology , Neovascularization, Pathologic/immunology , Taxoids/pharmacology , Vascular Endothelial Growth Factor A/immunology , Antineoplastic Agents/therapeutic use , Computer Simulation , Docetaxel , Endothelial Cells/immunology , Humans , Neoplasms/drug therapy , Taxoids/therapeutic use , Vascular Endothelial Growth Factor A/antagonists & inhibitors
We have developed glucose and lactate ultramicroelectrode (UME) biosensors based on glucose oxidase and lactate oxidase (with enzymes immobilized onto Pt UMEs by either electropolymerization or casting) for scanning electrochemical microscopy (SECM) and have determined their sensitivity to glucose and lactate, respectively. The results of our evaluations reveal different advantages for sensors constructed by each method: improved sensitivity and shorter manufacturing time for hand-casting, and increased reproducibility for electropolymerization. We have acquired amperometric approach curves (ACs) for each type of manufactured biosensor UME, and these ACs can be used as a means of positioning the UME above a substrate at a known distance. We have used the glucose biosensor UMEs to record profiles of glucose uptake above individual fibroblasts. Likewise, we have employed the lactate biosensor UMEs for recording the lactate production above single cancer cells with the SECM. We also show that oxygen respiration profiles for single cancer cells do not mimic cell topography, but are rather more convoluted, with a higher respiration activity observed at the points where the cell touches the Petri dish. These UME biosensors, along with the application of others already described in the literature, could prove to be powerful tools for mapping metabolic analytes, such as glucose, lactate, and oxygen, in single cancer cells.
Biosensing Techniques/methods , Enzymes, Immobilized/chemistry , Glucose Oxidase/chemistry , Glucose/analysis , Lactic Acid/analysis , Mixed Function Oxygenases/chemistry , Animals , Cell Line, Tumor , Cells, Cultured , Electrochemistry/methods , Fibroblasts/chemistry , Fibroblasts/metabolism , Glucose/metabolism , Glucose Oxidase/metabolism , Humans , Lactic Acid/metabolism , Mice , Microelectrodes , Microscopy/methods , Mixed Function Oxygenases/metabolism , Oxidation-Reduction , Platinum/chemistry
We have developed a multiwalled carbon nanotube/dihydropyran (MWCNT/DHP) composite sensor for the electrochemical detection of insulin in a microfluidic device. This sensor has been employed for physiological measurements of secreted insulin from pancreatic islets in a Cytosensor previously modified to be a multianalyte microphysiometer (MAMP). When compared with other established electrochemical insulin sensors, the MWCNT/DHP composite film sensor presented improved resistance to fluidic shear forces, while achieving enhanced electrode kinetics. In addition, the preparation of the composite film is straightforward and facile with a self-polymerizing monomer, DHP, used to add mechanical stability to the film. The sensor film was able to detect insulin concentrations as low as 1muM in the MAMP during calibration experiments. The MWCNT/DHP composite sensor has been successfully used for the direct detection of insulin secreted by islets in the microphysiometer.
Insulin/analysis , Nanotubes, Carbon/chemistry , Pyrans/chemistry , Calibration , Electrodes , Oxidation-Reduction
The electrochemical and chemical oxidation of a series of C8-arylamine adducts of 2'-deoxyguanosine has been examined. The oxidations were found to be reversible by cyclic and square-wave voltammetry in both aqueous buffer and aprotic organic solvent. The mechanism of the oxidation in protic media was either one- or two-electron, depending on the aryl group. The chemical oxidation resulted in guanidinohydantoin and spiroiminodihydantoin rearrangement products similar to those observed for 8-oxo-7,8-dihydro-2'-deoxyguanosine.
Amines/chemistry , Deoxyguanosine/analogs & derivatives , Deoxyguanosine/chemistry , Amines/metabolism , Deoxyguanosine/metabolism , Electrochemistry , Oxidation-Reduction
We have developed a process to incorporate an integral membrane protein, Photosystem I (PSI), into an organic thin film at an electrode surface and thereby insulate the protein complex on the surface while mimicking its natural environment. The PSI complex, which is primarily more hydrophobic on the exterior than interior, is hydrophobically confined in vivo within the thylakoid membrane. To mimic the thylakoid membrane and entrap PSI on an electrode, we have designed a series of steps using a thin self-assembled monolayer (SAM) to adsorb and orient PSI followed by exposures to longer-chained methyl-terminated alkanethiols that place exchange with components of the original SAM in the interprotein domains. In this process, PSI is first adsorbed onto a HOC(6)S/Au substrate through a short exposure to a dilute solution of the protein to achieve a protein coverage of approximately 25%. The PSI/HOC(6)S/Au substrate is then placed into a solution containing one of various longer-chained alkanethiols including C(22)SH or C(18)OC(19)SH. Changes in thickness, interfacial capacitance, infrared spectra, and surface wettability were used to assess the extent of backfilling by the long-chained thiols. The coverage of the protein layer and the solvent used for backfilling affected the rate and quality of the SAM formed in the interprotein regions. After exposure of the PSI layer to solvents containing alkanethiols, there was only minor loss of protein on the surface and no real change in protein secondary structure as evidenced by reflectance absorption infrared spectroscopy.
Membranes, Artificial , Organic Chemicals/chemistry , Photosystem I Protein Complex/chemistry , Adsorption , Electrodes , Gold/chemistry , Hydrophobic and Hydrophilic Interactions , Protein Structure, Secondary , Sulfhydryl Compounds/chemistry , Surface Properties , Time Factors
We report the first directed adsorption of Photosystem I (PSI) on patterned surfaces containing discrete regions of methyl- and hydroxyl-terminated self-assembled monolayers (SAMs) on gold. SAM and PSI patterns are characterized by scanning electrochemical microscopy (SECM). The insulating protein complex layer blocks the electron transfer of the SECM mediator, thereby reducing the electrochemical current significantly. Uniformly and densely packed adsorbed protein layers are observed with SECM. Pattern images correlate with our previous studies where we showed that low-energy surfaces (e.g., CH3-terminated) inhibit PSI adsorption in the presence of Triton X-100, whereas high-energy surfaces (e.g., OH-terminated) enable adsorption. Therefore, a SAM pattern with alternating methyl and hydroxyl surface regions allows PSI adsorption only on the hydroxyl surface, and this is demonstrated in the resulting SECM images.
