A multimodal neural network with gradient blending improves predictions of survival and metastasis in sarcoma.

The objective of this study is to develop a multimodal neural network (MMNN) model that analyzes clinical variables and MRI images of a soft tissue sarcoma (STS) patient, to predict overall survival and risk of distant metastases. We compare the performance of this MMNN to models based on clinical variables alone, radiomics models, and an unimodal neural network. We include patients aged 18 or older with biopsy-proven STS who underwent primary resection between January 1st, 2005, and December 31st, 2020 with complete outcome data and a pre-treatment MRI with both a T1 post-contrast sequence and a T2 fat-sat sequence available. A total of 9380 MRI slices containing sarcomas from 287 patients are available. Our MMNN accepts the entire 3D sarcoma volume from T1 and T2 MRIs and clinical variables. Gradient blending allows the clinical and image sub-networks to optimally converge without overfitting. Heat maps were generated to visualize the salient image features. Our MMNN outperformed all other models in predicting overall survival and the risk of distant metastases. The C-Index of our MMNN for overall survival is 0.77 and the C-Index for risk of distant metastases is 0.70. The provided heat maps demonstrate areas of sarcomas deemed most salient for predictions. Our multimodal neural network with gradient blending improves predictions of overall survival and risk of distant metastases in patients with soft tissue sarcoma. Future work enabling accurate subtype-specific predictions will likely utilize similar end-to-end multimodal neural network architecture and require prospective curation of high-quality data, the inclusion of genomic data, and the involvement of multiple centers through federated learning.

Protocol for the development and validation of a machine-learning based tool for predicting the risk of hypertriglyceridemia in critically-ill patients receiving propofol sedation.

Introduction Propofol is a widely used sedative-hypnotic agent for critically-ill patients requiring invasive mechanical ventilation (IMV). Despite its clinical benefits, propofol is associated with increased risks of hypertriglyceridemia. Early identification of patients at risk for propofol-associated hypertriglyceridemia is crucial for optimizing sedation strategies and preventing adverse outcomes. Machine learning (ML) models offer a promising approach to predict individualized patient risks of propofol-associated hypertriglyceridemia. Methods and analysis We propose the development of a ML model aimed at predicting the risk of propofol-associated hypertriglyceridemia in ICU patients receiving IMV. The study will utilize retrospective data from four Mayo Clinic sites. Nested cross-validation (CV) will be employed, with a 10-fold inner CV loop for model tuning and selection as well as an outer loop using leave-one-site-out CV for external validation. Feature selection will be conducted using Boruta and LASSO-penalized logistic regression. Data preprocessing steps include missing data imputation, feature scaling, and dimensionality reduction techniques. Six ML algorithms will be tuned and evaluated. Bayesian optimization will be used for hyperparameter selection. Global model explainability will be assessed using permutation importance, and local model explainability will be assessed using SHAP. Ethics and dissemination The proposed ML model aims to provide a reliable and interpretable tool for clinicians to predict the risk of propofol-associated hypertriglyceridemia in ICU patients. The final model will be deployed in a web-based clinical risk calculator. The model development process and performance measures obtained during nested cross-validation will be described in a study publication to be disseminated in a peer-reviewed journal. The proposed study has received ethics approval from the Mayo Clinic Institutional Review Board (IRB #23-007416).

Development of Machine Learning Models for Predicting the 1-Year Risk of Reoperation After Lower Limb Oncological Resection and Endoprosthetic Reconstruction Based on Data From the PARITY Trial.

BACKGROUND: Oncological resection and reconstruction involving the lower extremities commonly lead to reoperations that impact patient outcomes and healthcare resources. This study aimed to develop a machine learning (ML) model to predict this reoperation risk. METHODS: This study was conducted according to TRIPOD + AI. Data from the PARITY trial was used to develop ML models to predict the 1-year reoperation risk following lower extremity oncological resection and reconstruction. Six ML algorithms were tuned and calibrated based on fivefold cross-validation. The best-performing model was identified using classification and calibration metrics. RESULTS: The polynomial support vector machine (SVM) model was chosen as the best-performing model. During internal validation, the SVM exhibited an AUC-ROC of 0.73 and a Brier score of 0.17. Using an optimal threshold that balances all quadrants of the confusion matrix, the SVM exhibited a sensitivity of 0.45 and a specificity of 0.81. Using a high-sensitivity threshold, the SVM exhibited a sensitivity of 0.68 and a specificity of 0.68. Total operative time was the most important feature for reoperation risk prediction. CONCLUSION: The models may facilitate reoperation risk stratification, allowing for better patient counseling and for physicians to implement measures that reduce surgical risks.

Malignant Hyperthermia.

OBJECTIVES: A narrative expert review aiming to summarize the clinical epidemiology and management of critically ill patients with malignant hyperthermia (MH). DATA SOURCES: Medline searches were conducted to identify relevant articles describing the epidemiology, pathophysiology, and management of MH. Guidelines from key MH organizations were also incorporated into this review. STUDY SELECTION: Relevant studies regarding MH in both ICU and perioperative settings were reviewed. DATA EXTRACTION: Data from relevant studies were summarized and qualitatively assessed. DATA SYNTHESIS: MH is a severe reaction triggered by inhalational volatile anesthetics and succinylcholine in genetically susceptible patients. The condition is characterized by an early onset (min to hr) rise in temperature, hypercarbia, and muscular rigidity following exposure to triggering medications with potential complications of coagulopathy, rhabdomyolysis, and acute kidney injury. Acute management necessitates a coordinated multidisciplinary team approach with specific management using dantrolene, active cooling, and hyperventilation. A suspected MH reaction has important implications for future anesthetic exposure for both the patient and their family. All suspected reactions should be followed up at a specialized MH testing center using muscle contracture and genetic testing. CONCLUSIONS: Increasing use of inhalational anesthetics in the ICU underscores the need for enhanced education on the diagnosis and management of MH to ensure optimal patient sedation care and safety.

E3 ubiquitin ligase RNF180 mediates the ALKBH5/SMARCA5 axis to promote colon inflammation and Th17/Treg imbalance in ulcerative colitis mice.

SMARCA5, a protein in the SWI/SNF family, has been previously implicated in the development of ulcerative colitis (UC) through methylation. However, the specific molecular mechanisms by which SMARCA5 contributes to colonic inflammation and the imbalance between Th17 and Treg cells remain unclear. This study was designed to explore these molecular mechanisms. A UC mouse model was established using dextran sulfate sodium induction, followed by measurements of mouse weight, disease activity index (DAI) score, colon length, pathological changes in the colon, and FITC-dextran concentration. The levels of IL-17a, IFN-γ, IL-6, TNF-α, TGF-ß, and IL-10 were measured, along with the protein expression of ZO-1 and Occludin. Flow cytometry was used to assess the presence of IL-17 + CD4 + (Th17 +) cells and FOXP3 + CD25 + CD4 + (Treg +) cells in the spleen and mesenteric lymph nodes of UC mice. We observed that SMARCA5 and RNF180 were increased, while ALKBH5 was downregulated in UC mouse colon tissue. SMARCA5 or RNF180 knockdown or ALKBH5 overexpression ameliorated the colon inflammation and Th17/Treg cell imbalance in UC mice, shown by increased body weight, colon length, FOXP3 + CD25 + CD4 + T cells, and the levels of ZO-1, Occludin, TGF-ß, IL-10, and FOXP3. It decreased DAI scores, IL-17 + CD4 + T cells, and levels of IL-17a, IFN-γ, IL-6, TNF-α, and ROR-γt. ALKBH5 inhibited SMARCA5 expression via m6A modification, while RNF180 reduced ALKBH5 expression via ubiquitination. Our findings indicate that RNF180 aggravated the colon inflammation and Th17/Treg cell imbalance in UC mice by regulating the ALKBH5/SMARCA5 axis.

A Novel Instruction Driven 1-D CNN Processor for ECG Classification.

Electrocardiography (ECG) has emerged as a ubiquitous diagnostic tool for the identification and characterization of diverse cardiovascular pathologies. Wearable health monitoring devices, equipped with on-device biomedical artificial intelligence (AI) processors, have revolutionized the acquisition, analysis, and interpretation of ECG data. However, these systems necessitate AI processors that exhibit flexible configuration, facilitate portability, and demonstrate optimal performance in terms of power consumption and latency for the realization of various functionalities. To address these challenges, this study proposes an instruction-driven convolutional neural network (CNN) processor. This processor incorporates three key features: (1) An instruction-driven CNN processor to support versatile ECG-based application. (2) A Processing element (PE) array design that simultaneously considers parallelism and data reuse. (3) An activation unit based on the CORDIC algorithm, supporting both Tanh and Sigmoid computations. The design has been implemented using 110 nm CMOS process technology, occupying a die area of 1.35 mm2 with 12.94 µW power consumption. It has been demonstrated with two typical ECG AI applications, including two-class (i.e., normal/abnormal) classification and five-class classification. The proposed 1-D CNN algorithm performs with a 97.95% accuracy for the two-class classification and 97.9% for the five-class classification, respectively.

Efficacy and Safety of Hydrocortisone, Ascorbic Acid, and Thiamine Combination Therapy for the Management of Sepsis and Septic Shock: A Systematic Review and Meta-Analysis of Randomised Controlled Trials.

INTRODUCTION: This systematic review aimed to assess the efficacy and safety of hydrocortisone, ascorbic acid, and thiamine (HAT) combination therapy in patients with sepsis and septic shock. METHODS: We conducted a database search in MEDLINE, Embase, CENTRAL, Web of Science, and CNKI for randomised controlled trials (RCTs) comparing HAT against placebo/standard of care or against hydrocortisone in sepsis/septic shock patients. Outcomes included mortality, ICU/hospital length of stay (LOS), vasopressor durations, mechanical ventilation durations, change in SOFA at 72 h, and adverse events. RCT results were pooled in random-effects meta-analyses. Quality of evidence was assessed using GRADE. RESULTS: Fifteen RCTs (N = 2,594) were included. At 72 h, HAT reduced SOFA scores from baseline (mean difference [MD] -1.16, 95% confidence interval [CI]: -1.58 to -0.74, I2 = 0%) compared to placebo/SoC, based on moderate quality of evidence. HAT also reduced the duration of vasopressor use (MD -18.80 h, 95% CI: -23.67 to -13.93, I2 = 64%) compared to placebo/SoC, based on moderate quality of evidence. HAT increased hospital LOS (MD 2.05 days, 95% CI: 0.15-3.95, I2 = 57%) compared to placebo/SoC, based on very low quality of evidence. HAT did not increase incidence of adverse events compared to placebo/SoC. CONCLUSIONS: HAT appears beneficial in reducing vasopressor use and improving organ function in sepsis/septic shock patients. However, its advantages over hydrocortisone alone remain unclear. Future research should use hydrocortisone comparators and distinguish between sepsis-specific and comorbidity- or care-withdrawal-related mortality.

Coordination of oral anticoagulant care at hospital discharge (COACHeD): pilot randomised controlled trial.

OBJECTIVES: To evaluate whether a focused, expert medication management intervention is feasible and potentially effective in preventing anticoagulation-related adverse events for patients transitioning from hospital to home. DESIGN: Randomised, parallel design. SETTING: Medical wards at six hospital sites in southern Ontario, Canada. PARTICIPANTS: Adults 18 years of age or older being discharged to home on an oral anticoagulant (OAC) to be taken for at least 4 weeks. INTERVENTIONS: Clinical pharmacologist-led intervention, including a detailed discharge medication management plan, a circle of care handover and early postdischarge virtual check-up visits to 1 month with 3-month follow-up. The control group received the usual care. OUTCOMES MEASURES: Primary outcomes were study feasibility outcomes (recruitment, retention and cost per patient). Secondary outcomes included adverse anticoagulant safety events composite, quality of transitional care, quality of life, anticoagulant knowledge, satisfaction with care, problems with medications and health resource utilisation. RESULTS: Extensive periods of restriction of recruitment plus difficulties accessing patients at the time of discharge negatively impacted feasibility, especially cost per patient recruited. Of 845 patients screened, 167 were eligible and 56 were randomised. The mean age (±SD) was 71.2±12.5 years, 42.9% females, with two lost to follow-up. Intervention patients were more likely to rate their ability to manage their OAC as improved (17/27 (63.0%) vs 7/22 (31.8%), OR 3.6 (95% CI 1.1 to 12.0)) and their continuity of care as improved (21/27 (77.8%) vs 2/22 (9.1%), OR 35.0 (95% CI 6.3 to 194.2)). Fewer intervention patients were taking one or more inappropriate medications (7 (22.5%) vs 15 (60%), OR 0.19 (95% CI 0.06 to 0.62)). CONCLUSION: This pilot randomised controlled trial suggests that a transitional care intervention at hospital discharge for older adults taking OACs was well received and potentially effective for some surrogate outcomes, but overly costly to proceed to a definitive large trial. TRIAL REGISTRATION NUMBER: NCT02777047.

Interventions to Improve Adherence to Oral Pre-exposure Prophylaxis: A Systematic Review and Network Meta-analysis.

For people at risk of HIV infection, pre-exposure prophylaxis (PrEP) can reduce the risk of infection in anticipation of exposure to HIV. The effectiveness of PrEP relies upon a user's adherence to their PrEP regimen. We sought to assess the effect of PrEP adherence interventions compared to usual care or another intervention for people at risk of HIV. We searched electronic databases from 2010 onwards for randomized controlled trials (RCTs) involving persons at risk of HIV randomized to an adherence promoting intervention vs usual care or another intervention. We used network meta-analyses to compare PrEP adherence for all participant populations. Certainty of evidence was assessed using Confidence in Network Meta-Analysis (CINeMA). 21 trials (N = 4917) were included in qualitative analysis (19 in network meta-analyses (N = 4101)). HIV self-testing interventions with adherence feedback elements improved adherence compared to usual care (risk ratio (RR): 1.83, 95%CI 1.19, 2.82). In contrast, HIV self-testing alone was inferior to HIV self-testing with adherence feedback (RR: 0.58, 95%CI 0.37-0.92). Reminders alone also were inferior to HIV self-testing with adherence feedback on adherence (RR: 0.53, 95%CI 0.34-0.84) and had similar effects on adherence as usual care (RR: 0.98, 95%CI: 0.86-1.11). Interventions with only one component were inferior for adherence than those with two components (RR: 0.74, 95%CI 0.62-0.88) and those with three components (RR: 0.78, 95%CI 0.65-0.93). The certainty of evidence was moderate for HIV self-testing plus adherence feedback and interventions with two or three components. When designing future PrEP adherence interventions, we recommend strategies with more than one but no more than three components.

Proarrhythmic major adverse cardiac events with donepezil: A systematic review with meta-analysis.

BACKGROUND: Cholinesterase inhibitors (ChEIs) are regularly used in Alzheimer's disease. Of the three ChEIs approved for dementia, donepezil is among the most prescribed drugs in the United States with nearly 6 million prescriptions in 2020; however, it is classified as a "known risk" QT interval-prolonging medication (QTPmed). Given this claim is derived from observational data including single case reports, we aimed to evaluate high-quality literature on the frequency and nature of proarrhythmic major adverse cardiac events (MACE) associated with donepezil. METHODS: We searched Medline, Embase, International Pharmaceutical Abstracts, and Cochrane Central from 1996 onwards for randomized controlled trials (RCTs) involving patients age ≥18 years comparing donepezil to placebo. The MACE composite included mortality, sudden cardiac death, non-fatal cardiac arrest, Torsades de pointes, ventricular tachyarrhythmia, seizure or syncope. Random-effects meta-analyses were performed with a treatment-arm continuity correction for single and double zero event studies. RESULTS: Sixty RCTs (n = 12,463) were included. Twenty-five of 60 trials (n = 5886) investigated participants with Alzheimer's disease and 33 trials monitored electrocardiogram data. The mean follow-up duration was 31 weeks (SD = 36). Mortality was the most commonly reported MACE (252/331, 75.8% events), the remainder were syncope or seizures, with no arrhythmia events. There was no increased risk of MACE with exposure to donepezil compared to placebo (risk ratio [RR] 1.08, 95% CI 0.88-1.33, I2 = 0%) and this was consistent in the subgroup analysis of trials including participants with cardiovascular morbidities (RR 1.14, 95% CI 0.88-1.47). Subgroup analysis suggested a trend toward more events with donepezil with follow-up ≥52 weeks (RR: 1.32, 0.98-1.79). CONCLUSIONS: This systematic review with meta-analysis found donepezil may not be arrhythmogenic. Donepezil was not associated with mortality, ventricular arrhythmias, seizure or syncope, although longer durations of therapy need more study. Further research to clarify actual clinical outcomes related to QTPmed is important to inform prescribing practices.

Pharmacological prevention of bone loss and fractures following solid organ transplantations: Protocol for a systematic review and network meta-analysis.

INTRODUCTION: Solid organ transplant (SOT) recipients can experience bone loss caused by underlying conditions and the use of immunosuppressants. As a result, SOT recipients are at risk for decreased bone mineral density (BMD) and increased fracture incidences. We propose a network meta-analysis (NMA) that incorporates all available randomized control trial (RCT) data to provide the most comprehensive ranking of anti-osteoporotic interventions according to their ability to decrease fracture incidences and increase BMD in SOT recipients. METHODS: We will search MEDLINE, EMBASE, Web of Science, CINAHL, CENTRAL and CNKI for relevant RCTs that enrolled adult SOT recipients, assessed anti-osteoporotic therapies, and reported relevant outcomes. Title and full-text screening as well as data extraction will be performed in-duplicate. We will report changes in BMD as weighted or standardized mean differences, and fracture incidences as risk ratios. SUCRA scores will be used to provide rankings of interventions, and quality of evidence will be examined using RoB2 and CINeMA. DISCUSSIONS: To our knowledge, this systematic review and NMA will be the most comprehensive quantitative analysis regarding the management of bone loss and fractures in SOT recipients. Our analysis should be able to provide physicians and patients with an up-to-date recommendation for pharmacotherapies in reducing incidences of bone loss and fractures associated with SOT. The findings of the NMA will be disseminated in a peer-reviewed journal.

Outcomes of Vasopressin-Receptor Agonists Versus Norepinephrine in Adults With Perioperative Hypotension: A Systematic Review.

Consensus statements recommend the use of norepinephrine and/or vasopressin for hypotension in cardiac surgery. However, there is a paucity of data among other surgical subgroups and vasopressin analogs. Therefore, the authors conducted a systematic review of randomized controlled trials (RCTs) to compare vasopressin-receptor agonists with norepinephrine for hypotension among those undergoing surgery with general anesthesia. This review was registered prospectively (CRD42022316328). Literature searches were conducted by a medical librarian to November 28, 2023, across MEDLINE, EMBASE, CENTRAL, and Web of Science. The authors included RCTs enrolling adults (≥18 years of age) undergoing any surgery under general anesthesia who developed perioperative hypotension and comparing vasopressin receptor agonists with norepinephrine. The risk of bias was assessed by the Cochrane risk of bias tool for randomized trials (RoB-2). Thirteen (N = 719) RCTs were included, of which 8 (n = 585) enrolled patients undergoing cardiac surgery. Five trials compared norepinephrine with vasopressin, 4 trials with terlipressin, 1 trial with ornipressin, and the other 3 trials used vasopressin as adjuvant therapy. There was no significant difference in all-cause mortality. Among patients with vasoplegic shock after cardiac surgery, vasopressin was associated with significantly lower intensive care unit (N = 385; 2 trials; mean 100.8 v 175.2 hours, p < 0.005; median 120 [IQR 96-168] v 144 [96-216] hours, p = 0.007) and hospital lengths of stay, as well as fewer cases of acute kidney injury and atrial fibrillation compared with norepinephrine. One trial also found that terlipressin was associated with a significantly lower incidence of acute kidney injury versus norepinephrine overall. Vasopressin and norepinephrine restored mean arterial blood pressure with no significant differences; however, the use of vasopressin with norepinephrine was associated with significantly higher mean arterial blood pressure versus norepinephrine alone. Further high-quality trials are needed to determine pooled treatment effects, especially among noncardiac surgical patients and those treated with vasopressin analogs.

Assessing the research landscape and clinical utility of large language models: a scoping review.

IMPORTANCE: Large language models (LLMs) like OpenAI's ChatGPT are powerful generative systems that rapidly synthesize natural language responses. Research on LLMs has revealed their potential and pitfalls, especially in clinical settings. However, the evolving landscape of LLM research in medicine has left several gaps regarding their evaluation, application, and evidence base. OBJECTIVE: This scoping review aims to (1) summarize current research evidence on the accuracy and efficacy of LLMs in medical applications, (2) discuss the ethical, legal, logistical, and socioeconomic implications of LLM use in clinical settings, (3) explore barriers and facilitators to LLM implementation in healthcare, (4) propose a standardized evaluation framework for assessing LLMs' clinical utility, and (5) identify evidence gaps and propose future research directions for LLMs in clinical applications. EVIDENCE REVIEW: We screened 4,036 records from MEDLINE, EMBASE, CINAHL, medRxiv, bioRxiv, and arXiv from January 2023 (inception of the search) to June 26, 2023 for English-language papers and analyzed findings from 55 worldwide studies. Quality of evidence was reported based on the Oxford Centre for Evidence-based Medicine recommendations. FINDINGS: Our results demonstrate that LLMs show promise in compiling patient notes, assisting patients in navigating the healthcare system, and to some extent, supporting clinical decision-making when combined with human oversight. However, their utilization is limited by biases in training data that may harm patients, the generation of inaccurate but convincing information, and ethical, legal, socioeconomic, and privacy concerns. We also identified a lack of standardized methods for evaluating LLMs' effectiveness and feasibility. CONCLUSIONS AND RELEVANCE: This review thus highlights potential future directions and questions to address these limitations and to further explore LLMs' potential in enhancing healthcare delivery.

Outcomes of Acute Respiratory Failure in Patients With Cancer in the United States.

OBJECTIVE: To determine the epidemiological effect-magnitude and outcomes of patients with cancer vs those without cancer who are hospitalized with acute respiratory failure (ARF). PATIENTS AND METHODS: We reviewed hospitalizations within the National Inpatient Sample (NIS) database between January 1, 2016, and December 31, 2018. Patients were classified based on a diagnosis of solid-organ cancer, hematologic cancer, or no cancer. Noninvasive positive pressure ventilation (NIPPV) failure was defined as patients who initially received NIPPV and had progression to invasive mechanical ventilation. Weighted samples were used to derive population estimates. RESULTS: During the study period, there were an estimated 8,837,209 admissions with ARF in the United States, 8.9% (783,625) of which had solid-organ cancer and 2.0% (176,095) had hematologic cancers. Annually, 319,907 patients with cancer are admitted with ARF, with 27.3% (87,302) requiring invasive mechanical ventilation and 10.0% (31,998) requiring NIPPV. In-hospital mortality was higher in patients with cancer vs those without cancer (24.0% [76,813] vs 12.3% [322,465]; P<.001), and this proprotion persisted when stratified by the highest method of oxygen delivery. Patients with cancer had longer hospital length of stay (7.0 days [3.0 to 12.0 days] vs 5.0 days [3.0 to 10.0 days]; P<.001) and were more likely to have NIPPV failure (14.9% [3,992] vs 12.8% [41,875]). Compared with those with solid-organ cancer, patients with hematologic cancers experienced worse outcomes. The association between underlying cancer diagnosis and outcomes remained consistent when adjusted for age, sex, and comorbidities. CONCLUSION: In the United States, patients with cancer account for over 10% of ARF hospital admissions (959,720 of 8,837,209). They experience an approximately 2-fold higher mortality versus those without cancer. Those with hematologic cancers appear to experience worse outcomes than patients with solid-organ cancers.

Automated Paper Screening for Clinical Reviews Using Large Language Models: Data Analysis Study.

BACKGROUND: The systematic review of clinical research papers is a labor-intensive and time-consuming process that often involves the screening of thousands of titles and abstracts. The accuracy and efficiency of this process are critical for the quality of the review and subsequent health care decisions. Traditional methods rely heavily on human reviewers, often requiring a significant investment of time and resources. OBJECTIVE: This study aims to assess the performance of the OpenAI generative pretrained transformer (GPT) and GPT-4 application programming interfaces (APIs) in accurately and efficiently identifying relevant titles and abstracts from real-world clinical review data sets and comparing their performance against ground truth labeling by 2 independent human reviewers. METHODS: We introduce a novel workflow using the Chat GPT and GPT-4 APIs for screening titles and abstracts in clinical reviews. A Python script was created to make calls to the API with the screening criteria in natural language and a corpus of title and abstract data sets filtered by a minimum of 2 human reviewers. We compared the performance of our model against human-reviewed papers across 6 review papers, screening over 24,000 titles and abstracts. RESULTS: Our results show an accuracy of 0.91, a macro F1-score of 0.60, a sensitivity of excluded papers of 0.91, and a sensitivity of included papers of 0.76. The interrater variability between 2 independent human screeners was κ=0.46, and the prevalence and bias-adjusted κ between our proposed methods and the consensus-based human decisions was κ=0.96. On a randomly selected subset of papers, the GPT models demonstrated the ability to provide reasoning for their decisions and corrected their initial decisions upon being asked to explain their reasoning for incorrect classifications. CONCLUSIONS: Large language models have the potential to streamline the clinical review process, save valuable time and effort for researchers, and contribute to the overall quality of clinical reviews. By prioritizing the workflow and acting as an aid rather than a replacement for researchers and reviewers, models such as GPT-4 can enhance efficiency and lead to more accurate and reliable conclusions in medical research.

The effects of pks+ Escherichia coli and bile acid in colorectal tumorigenesis among people with cholelithiasis or cholecystectomy.

BACKGROUND AND AIM: Patients with cholelithiasis (CL) or cholecystectomy (CE) would have more chances of getting colorectal adenoma (CRA) or cancer (CRC). We aimed to figure out the effects of gut microbiota and bile acid on colorectal neoplasm in CL and CE patients. METHODS: This was a retrospective observational study that recruited 514 volunteers, including 199 people with normal gallbladders (normal), 152 CL, and 163 CE patients. Discovery cohort was established to explore the difference in gut microbiota through 16S rRNA and metagenomics sequencing. Validation cohort aimed to verify the results through quantitative polymerase chain reaction (qPCR). RESULTS: Significant enrichment of Escherichia coli was found in patients with cholelithiasis or cholecystectomy both in the discovery cohort (16S rRNA sequencing, PNormal-CL = 0.013, PNormal-CE = 0.042; metagenomics sequencing, PNormal-CE = 0.026) and validation cohort (PNormal-CL < 0.0001, PNormal-CE < 0.0001). Pks+ E. coli was found enriched in CL and CE patients through qPCR (in discovery cohort: PNormal-CE = 0.018; in validation cohort: PNormal-CL < 0.0001, PNormal-CE < 0.0001). The differences in bile acid metabolism were found both through Tax4Fun analysis of 16S rRNA sequencing (Ko00120, primary bile acid biosynthesis, PNormal-CE = 0.014; Ko00121, secondary bile acid biosynthesis, PNormal-CE = 0.010) and through metagenomics sequencing (map 00121, PNormal-CE = 0.026). The elevation of serum total bile acid of CE patients was also found in validation cohort (PNormal-CE < 0.0001). The level of serum total bile acid was associated with the relative abundance of pks+ E. coli (r = 0.1895, P = 0.0012). CONCLUSIONS: E. coli, especially pks+ species, was enriched in CL and CE patients. Pks+ E. coli and bile acid metabolism were found associated with CRA and CRC in people after cholecystectomy.

The use of large language models in medicine: proceeding with caution.

Large language models, like ChatGPT and Bard, have potential clinical applications due to their ability to generate conversational responses and encode medical knowledge. However, their clinical adoption faces challenges including hallucinations, lack of transparency, and lack of consistency. Ethicolegal concerns surrounding patient consent, legal liability, and data privacy further complicate matters. Despite their promise, an optimistic but cautious approach is essential for the safe integration of large language models into clinical settings.