Category: Nevin Manimala

Phytopathology. 2026 Apr 6. doi: 10.1094/PHYTO-11-25-0360-R. Online ahead of print.

ABSTRACT

Basil (Ocimum basilicum) is one of the most widely grown herb crops for both fresh consumption, processing, and essential oil production. These markets are reliant upon high-quality pathogen-free material as disease renders fresh material unsellable and may alter taste and essential oil profiles for industrial producers. Meeting high-quality standards has become increasingly difficult with the introduction of basil downy mildew caused by Peronospora belbahrii into the US and reported globally. Basil downy mildew, is a highly destructive foliar disease of basil resulting in chlorosis, large lesions, grey-black spores and plant death. Fungicides have shown variable efficacy between field seasons, are expensive and require diligent and repeated applications for complete control. Repeated applications of high-risk fungicides provide selective pressure for pathogen evolution exemplified by the appearance of mefenoxam resistant populations of Peronospora belbahrii. Resistant sweet basil varieties, though initially successful, have begun to break down under rapid pathogen race differentiation. This adaptability makes understanding the population genetics of the pathogen critical. This study, utilizing simple sequence repeats, examined the genetic diversity of 92 Peronospora belbahrii isolates collected during 2018-2024 from Israel, Italy, Hawaii, Florida, Kansas, and the US Northeast. Genetic diversity statistics indicate a highly heterozygous, largely non-clonal population structure in both the USA and Israel. These two populations appear to be at least partially genetically distinct and indicate differentiation along geographic lines. Significant genetic differences were also detected between race 0 and race 1, though a distinct set of classifying markers was not found.

PMID:41941754 | DOI:10.1094/PHYTO-11-25-0360-R

JMIR Med Inform. 2026 Apr 6;14:e86533. doi: 10.2196/86533.

ABSTRACT

BACKGROUND: Electronic medical records are a vast and valuable source of information, useful for tasks such as estimating disease prevalence. However, in routine primary care, much of this information is in free-text format rather than in a structured form and, therefore, not readily amenable to analysis. Manual coding of this textual data is both time-consuming and resource-intensive, making it impractical for large datasets. Although powerful open-source language models offer new opportunities for automated coding, their use on short heterogeneous primary care notes, particularly in German-language settings, remains insufficiently studied.

OBJECTIVE: By providing hands-on guidance for applied health researchers, this study aims to demonstrate the effective and accurate automatic classification of free-text notes using a language model fine-tuned for automated International Statistical Classification of Diseases, Tenth Revision (ICD-10) coding.

METHODS: Building on the extensive Family Medicine Research Using Electronic Medical Records (FIRE) routine database from the Institute of Primary Care at the University Hospital Zurich and the University of Zurich, we trained a large language model-based multilabel classifier on a dataset of 38,728 free-text notes, which had been manually categorized into 47 classes using specific ICD-10 codes and code ranges or nondiagnostic/ad hoc labels (eg, “unclear diagnosis,” “status post”). We stratified the labeled data into training (70%), validation (15%), and posttraining test (15%) sets, ensuring similar label distributions across these sets. Using the Transformers Python library, we trained the model over 10 epochs and evaluated it on the posttraining test dataset.

RESULTS: Across 48 classes, the FIRE classifier achieved strong performance on the held-out posttraining set, with F1-scores of 0.85 (micro, overall across all predictions), 0.86 (macro, mean of per-class scores treating classes equally), and 0.84 (weighted, per-class scores weighted by class frequency).

CONCLUSIONS: This study demonstrates steps for training open-source large language models and highlights the potential to streamline and scale the extraction of diagnostic information for practical applications. Our model can be robustly deployed, for example, for prescreening and labeling of free-text information, thus potentially reducing the burden of repetitive and error-prone manual handling.

PMID:41941723 | DOI:10.2196/86533

JMIR Form Res. 2026 Apr 6;10:e90673. doi: 10.2196/90673.

ABSTRACT

BACKGROUND: In recent years, large language models (LLMs) have undergone swift cycles of refinement and iteration. However, in the realm of clinical medicine, different LLMs’ capability of logical reasoning and disease diagnosis needs further investigation.

OBJECTIVE: The aim of our study was to evaluate the performance of 4 different LLMs in the National Health Professional and Technical Qualification Examination in China.

METHODS: A total of 398 multiple-choice questions of 5 different question types were integrated within the examination with respect to the diagnosis or care of cases. These questions were categorized into different cardiology subspecialties and different clinical disciplines. DeepSeek V3 and R1 were accessed through an application programming interface, while ChatGPT 4o and o1 were queried via its public chat-based interface. We offered the same prompts instructing LLMs to assume the role of a physician and provide answers with explanations at the beginning of each conversation. We assessed different LLMs’ performance by the accuracy in the responses to the multiple-choice questions. For the first 3 examination sections, McNemar test was used to compare the accuracy among the models, with post hoc pairwise comparisons performed using partitions of chi-square method and Bonferroni correction (significance set at P<.008). For the fourth section involving partially credit scoring, one-way ANOVA was performed to compare the mean scores among the models, with statistical significance set at P<.05.

RESULTS: Both DeepSeek V3 and R1 showed superior performance in the first 3 sections of the Chinese National Health Professional and Technical Qualification Examination, achieving an overall performance of 93% and 93.6%, respectively. ChatGPT 4o and o1 achieved accuracies of 73.3% and 69%, respectively (all P<.001 compared with DeepSeek V3). For the fourth section, the performance of all 4 LLMs markedly declined compared to their results in the preceding sections. Particularly, in the section of gastroenterology and hematology, DeepSeek V3 achieved the highest accuracy, while R1 ranked first in cardiology and neurology. ChatGPT o1 achieved the highest accuracy in the topic of coronary artery disease, with no statistical significance.

CONCLUSIONS: DeepSeek V3 and R1 showed remarkable potential in facilitating clinical decision-making in the Chinese professional examination, with both outperforming ChatGPT 4o and o1. Nonetheless, future research should continue evaluating their economic efficiency and susceptibility to hallucination.

PMID:41941721 | DOI:10.2196/90673

JMIR Dermatol. 2026 Apr 6;9:e86402. doi: 10.2196/86402.

ABSTRACT

BACKGROUND: The increasing incidence and financial burden of skin cancer place immense pressure on the UK’s National Health Service (NHS). Systemic challenges, including dermatologist shortages and long waiting lists, complicate timely assessment of skin lesions for patients under the urgent suspected cancer pathway. While teledermatology offers an innovative solution compared to traditional face-to-face appointments, standard teledermatology models still face limitations in addressing health care access barriers. Community-based decentralized models may reduce such barriers, but the cost and operational impact of such specific models remain largely underresearched.

OBJECTIVE: This study evaluated the differences in financial cost to the NHS and patient waiting times at the Northern Care Alliance NHS Foundation Trust by comparing a community-based teledermatology model using Pathpoint eDerma against the Trust’s standard-of-care for patients in the urgent suspected skin cancer pathway.

METHODS: This study used an ambidirectional design involving 2 distinct analyses. The cost comparison analysis (CCA) compared costs incurred under the teledermatology model (intervention arm, n=563) against the Trust’s standard care, represented by a synthetic comparator arm (n=4011). The discrete event simulation (DES) modeled the operational impact on patient waiting times over a 1-year period. Data for the intervention arm were collected prospectively from December 2022 to May 2023 for CCA and up to November 2023 for DES, while comparator data were collected retrospectively from September 2021 to December 2022. Publicly available resource costs were incorporated to ensure the robustness of the analyses.

RESULTS: The community-based teledermatology model was associated with significant improvements in both cost to the NHS and patient waiting times. The CCA revealed a mean cost saving of £45 (£1=US $1.24) per referral (95% CI £22-£60; P<.001). This cost saving was associated with a 26% reduction in the proportion of patients requiring a full diagnostic biopsy, falling from 48% (1925/4011) in standard care to 22% (124/563) in the teledermatology model as well as time savings in face-to-face clinics and administration. Furthermore, the DES demonstrated that, on average, the teledermatology pathways decreased the time to reach a clinical diagnosis by 9.90 (95% CI 9.64-10.16) days; to communicate a diagnosis to patients by 54.18 (95% CI 50.76-57.61) days; and to reach a histopathological diagnosis by 62.8 (95% CI 59.76-65.83) days compared to standard care.

CONCLUSIONS: The implementation of the community-based teledermatology model appears to be a highly effective, cost-efficient strategy associated with shortened patient journeys. The intervention showed a faster initial triage phase, but the study identified the histopathology process as the next major systemic constraint that could deter further pathway efficiency. Achieving timely diagnosis for all patients, including those requiring diagnostic biopsies, will necessitate continued strategic investment in innovative technologies to accelerate this downstream process.

PMID:41941718 | DOI:10.2196/86402

JCO Clin Cancer Inform. 2026 Apr;10(2):e2500154. doi: 10.1200/CCI-25-00154. Epub 2026 Apr 6.

ABSTRACT

PURPOSE: In 2018, the US Preventive Services Task Force (USPSTF) updated its prostate cancer screening recommendations for men age 55-69 years from grade D, discouraging screening, to grade C, supporting individualized decision making based on clinician judgment and patient preference. Although one study reported increased screening after the 2017 draft recommendation, findings were based on privately insured populations. This study assessed changes in screening after USPSTF revisions in Colorado, examining variation by payer and area-level social determinants of health.

METHODS: Using Colorado’s All-Payer Claims Database, we included men age 55-69 years with continuous annual enrollment between 2014 and 2023 and measured prostate screening procedures. Negative binomial regression models with population offsets and spline knots in 2016 and 2018 estimated changes in screening, adjusting for age and year, and clustering at the Zip Code Tabulation Area level. Data from 2020 were not used in the model because of pandemic-related disruptions. Analyses were stratified by payer. To assess geographic variation, we estimated random-effects negative binomial models and examined associations between predicted screening changes and the social deprivation index.

RESULTS: In a sample of over 2 million person-years, 595,107 (27.8%) were screened during the study period. Screening rates increased steadily across all payers, rising from 23.1% in 2014 to 32.1% in 2023, with the largest gains among Medicare Advantage and commercial enrollees. Increases varied geographically, with smaller gains in areas with higher deprivation levels, in particular, areas with poverty, low educational attainment, and crowded housing.

CONCLUSION: Prostate cancer screening increased after recommendation revisions across all payers, but the magnitudes and trajectories varied. Our results suggest barriers to screening among eligible and insured men associated with socioeconomic factors.

PMID:41941708 | DOI:10.1200/CCI-25-00154

JCO Clin Cancer Inform. 2026 Apr;10(2):e2500162. doi: 10.1200/CCI-25-00162. Epub 2026 Apr 6.

ABSTRACT

PURPOSE: Digital patient portals (DPPs) provide patients a direct electronic link with their care teams allowing secure, rapid symptom reporting and the capture of electronic patient-reported outcome questionnaires. A proportion of patients face barriers to DPP adoption. We conducted a prospective observational cohort study to quantify the proportion of DPP nonadopters and identify factors associated with DPP nonadoption.

METHODS: All patients with prostate cancer undergoing radical radiotherapy at a Canadian Provincial Cancer Program were given the opportunity to install a DPP on their mobile device/tablet before starting radiotherapy. Each patient completed a self-reported questionnaire assessing their willingness and ability to use the DPP and quantified variables associated with their decision. A DPP adopter was defined as a patient who logged into the application and accessed their radiotherapy educational materials. Descriptive variables were tabulated by DPP adoption status. Differences in distributions of baseline characteristics were assessed using standard parametric and nonparametric statistics. Univariable and multivariable logistic regression analyses were performed to identify variables associated with DPP nonadoption.

RESULTS: Between July 2022 and January 2023, 118 patients with prostate cancer were enrolled with a median age of 71 years (range, 50-86). Most patients (56.8%) had a college diploma, 29.6% lived rurally, and 95.8% owned a smartphone. Twenty-eight patients (23.8%) were DPP nonadopters. On stepwise multivariable logistic regression analysis, poor self-reported ability to use mobile devices was associated with DPP nonadoption (P < .01).

CONCLUSION: Among patients with prostate cancer undergoing radical radiotherapy, poor self-reported ability to use smartphone technology was identified as the primary barrier to DPP adoption. Interventions geared at mitigating this barrier to DPP adoption are needed.

PMID:41941707 | DOI:10.1200/CCI-25-00162

Anesthesiology. 2026 Apr 3. doi: 10.1097/ALN.0000000000006078. Online ahead of print.

ABSTRACT

BACKGROUND: The energy transferred from the mechanical ventilator to the respiratory system over time, so-called mechanical power, may cause lung injury even with protective ventilation. We hypothesized that protective ventilation aiming at moderate lung expansion results in less mechanical power and lung injury than aiming at permissive atelectasis or maximum lung expansion.

METHODS: Twenty-four anesthetized pigs with acute lung injury induced by saline lung lavage were randomly assigned to ventilation according to either Acute Respiratory Distress Syndrome Clinical Network´s low positive end-expiratory pressure (PEEP) table (LowPEEP), or high PEEP table (HighPEEP), or Open Lung Approach with PEEP titrated to the highest respiratory system compliance and periodic recruitment maneuvers (OLA) (n=8/group). Mechanical power was calculated from pressure-volume curves, and physiological variables were measured. We assessed pulmonary inflammation as the tissue-normalized uptake rate of 2-deoxy-2-[18F]fluoro-D-glucose measured by positron-emission computed tomography (KiS) and determined the gradient between randomization and 24h thereafter (∆KiS).

RESULTS: The median (IQR) PEEP was 5(0.1), 12(0.2), and 12(0.2) cmH2O during LowPEEP, HighPEEP, and OLA, respectively. ΔKiS was higher in LowPEEP than OLA (0.0183±0.0109 vs. 0.0049±0.0088 min-1; P=0.024; d=0.47), but did not differ significantly from HighPEEP (0.0080±0.0073 min-1; P=0.104; d=0.85). ΔKiS also did not differ between HighPEEP and OLA (P=0.876; d=0.60). The median (IQR) mechanical power in LowPEEP [9.5(1.5) J/min] was higher than in HighPEEP [7.5(2.3) J/min; P=0.008; d=4.28] and OLA [6.8(2.1) J/min; P=0.002; d=4.69], but did not statistically differ between HighPEEP and OLA (P=0.886; d=0.199). Mechanical power correlated positively with ΔKiS across groups (ρ=0.425, P=0.038).

CONCLUSIONS: In this porcine model of acute lung injury, protective ventilation with individualized higher PEEP aiming at lung expansion resulted in less pulmonary inflammation and lower mechanical power compared to protective ventilation with permissive atelectasis. A strategy with higher PEEP but without individualization did not differ significantly from the other two strategies regarding inflammation.

PMID:41941706 | DOI:10.1097/ALN.0000000000006078

Neurology. 2026 Apr 28;106(8):e214834. doi: 10.1212/WNL.0000000000214834. Epub 2026 Apr 6.

ABSTRACT

OBJECTIVES: Industry payments to clinicians influence prescribing, limiting use of lower-cost generics, and leading to nonadherence. This study of prescribing for multiple sclerosis examined whether payments from brand-name drug manufacturers were associated with prescribing of brand-name glatiramer and dimethyl fumarate after generics became available.

METHODS: This cross-sectional study used 2021-2022 Open Payments data and 2022-2023 Medicare Part D prescription data. Payments from Teva (glatiramer) and Biogen (dimethyl fumarate) were categorized as none, <$1,000, or ≥$1,000. The outcome was the proportion of brand-name prescriptions per clinician, categorized as low (<20%), medium (20%-79%), or high (≥80%). Associations were assessed using multinomial logistic regression, adjusting for prescriber type, prescription volume, and geographic region.

RESULTS: Among 2,675 glatiramer and 2,138 dimethyl fumarate prescribers, 1,026 (38.4%) and 1,238 (57.9%) received payments, respectively. Receiving ≥$1,000 in payments was associated with a greater odds of high brand-name prescribing for glatiramer (adjusted odds ratio [aOR] 4.21, 95% CI 1.81-9.81, p < 0.001) and dimethyl fumarate (aOR 2.53, 95% CI 1.57-4.07, p < 0.001). Payments <$1,000 were also associated with higher brand-name prescribing.

DISCUSSION: Payments to clinicians were associated with lower uptake of generic versions of 2 MS drugs, resulting in higher spending by patients and the US health care system.

PMID:41941704 | DOI:10.1212/WNL.0000000000214834

Anesthesiology. 2026 Apr 3. doi: 10.1097/ALN.0000000000006080. Online ahead of print.

ABSTRACT

BACKGROUND: A substantial proportion of patients report no pain after surgery, resulting in an excess of zero values that pose challenges for analysis using traditional statistical models. The present study was designed to test the hypothesis that a two-part model, commonly used in healthcare expenditures research, would demonstrate superior performance in predicting postsurgical pain when compared to traditional models, and would secondarily better identify predictors of this clinically important outcome.

METHODS: This study analyzed a prospectively collected single-center dataset (n=3925) of chronic postsurgical pain to compare a novel two-part modeling framework with logistic and linear regression. The two-part model first estimated the probability of experiencing postsurgical pain at 3 months (binary outcome), followed by the severity of pain among those affected (on a 1-10 numeric rating scale). To obtain an unbiased assessment of model performance, the data were randomly split into training (n=3000) and testing (n=925) datasets. Models were trained on the training dataset and evaluated on the testing set. This process was repeated 400 times to compute average performance estimates. As a secondary aim, the study assessed the associations of 15 baseline factors, including validated measures of patient-reported pain, functional and psychological measures, comorbidities, and surgical details, across the different modeling approaches.

RESULTS: The two-part model demonstrated superior predictive performance compared to linear regression alone, with a higher mean R² value (0.075 vs. 0.050, p<0.0001), lower root mean square error (RMSE: 1.466 vs. 1.485, p<0.0001), and lower mean absolute error (MAE: 1.020 vs. 1.030, p<0.0001). Statistical comparison with logistic regression alone was not possible due to the shared binary component. The two-part model identified 7 baseline preoperative covariates as independently associated with postsurgical pain that were missed by either logistic or linear models, or both: self-reported race, education level, ASA classification, overall body pain, widespread pain, symptom severity, and anxiety level. Significant baseline factors identified in all three models were patient sex, surgical type, and surgical site pain.

CONCLUSION: A two-part model may offer a superior statistical approach to linear and logistic regression, better identifing patient- and clinical care-related risk factors of chronic post-surgical pain, primarily by distinguishing factors that drive the occurrence of chronic pain from those that are associated with pain severity.

PMID:41941700 | DOI:10.1097/ALN.0000000000006080

JCO Oncol Pract. 2026 Apr 6:OP2501211. doi: 10.1200/OP-25-01211. Online ahead of print.

ABSTRACT

PURPOSE: Despite strong evidence and guidelines supporting prophylactic anticoagulation for ambulatory patients with cancer starting systemic cancer-directed therapy who are at high risk of venous thromboembolism (VTE), uptake in practice is limited. We evaluated the real-world impact of prophylactic anticoagulation in such patients receiving guideline-based care.

METHODS: We conducted an observational cohort study of patients assessed as part of a multidisciplinary VTE prevention program (the Vermont model) from 2016 to 2021. For this study, we included outpatients at high risk of VTE based on a Khorana risk score or Protecht risk score of ≥3. Based on the individualized decision making, patients either received or did not receive prophylactic anticoagulation. The primary outcome was VTE at 6 months after risk assessment. The secondary outcome was all-cause mortality at 6 months.

RESULTS: Of 573 high-risk patients assessed during the study period, 340 (59%) received thromboprophylaxis and 233 (41%) did not. Eleven (3.2%) on thromboprophylaxis developed a VTE within 6 months, compared with 18 (7.7%) not on thromboprophylaxis. After adjusting for age, sex, BMI, cancer stage, chemotherapy, immunotherapy, distance from center, and history of VTE, thromboprophylaxis reduced VTE (adjusted odds ratio [OR], 0.36 [95% CI, 0.16 to 0.80]) compared with no thromboprophylaxis. Of deaths within 6 months, 57 (16.8%) occurred in the thromboprophylaxis group versus 73 (31.3%) in the no thromboprophylaxis group (adjusted OR, 0.46 [95% CI, 0.30 to 0.71]).

CONCLUSION: Ambulatory patients with cancer at high risk of VTE who received prophylactic anticoagulation had improved clinical outcomes with fewer VTE events and less mortality. Our real-world data support available clinical trial data and underscore the importance of the provision of guideline-directed care in the cancer outpatient setting.

PMID:41941694 | DOI:10.1200/OP-25-01211