Nevin Manimala

JAMA Netw Open. 2025 May 1;8(5):e2512941. doi: 10.1001/jamanetworkopen.2025.12941.

ABSTRACT

IMPORTANCE: Migrant and refugee populations are increasing globally, and children and adolescents in these populations are particularly susceptible to mental disorders. North Korean refugee (NKR) youths now living in South Korea share a culture, language, and history with South Korean (SK) youths, making these 2 groups suitable for studying the environmental factors in psychiatric health.

OBJECTIVE: To compare the risk of developing mental illness and individual psychiatric disorders among NKR youths and SK youths.

DESIGN, SETTING, AND PARTICIPANTS: This cohort study obtained data from Korea’s National Health Insurance Service claims database from 2005 to 2021. Participants were aged 1 to 18 years between 2007 and 2010 who had no prior psychiatric claims, whether as outpatients or inpatients. Children and adolescents of the SK general population matched 1:10 on sex and age were compared with the NKR youth population. Statistical analyses were performed from August 2024 to March 2025.

MAIN OUTCOMES AND MEASURES: Risk, presented as hazard ratio (HR), of developing psychiatric disorders in NKR youths compared with SK youths. This risk was assessed using a Cox proportional hazards regression model and Kaplan-Meier time-to-event probabilities.

RESULTS: In total, 1618 NKR youths (810 males [50.1%]; mean [SD] age, 9.48 [4.62] years) and 308 927 SK youths (194 331 females [62.9%]; mean [SD] age, 11.80 [4.72] years) were included in this study. NKR youths showed a significantly higher risk of developing psychiatric disorders (HR, 1.29; 95% CI, 1.17-1.43) compared with SK youths. Individual psychiatric disorder risks were higher for NKR youths, including posttraumatic stress disorder (HR, 2.33; 95% CI, 1.34-4.06; P = .003), attention-deficit/hyperactivity disorder (HR, 1.67; 95% CI, 1.32-2.11; P < .001), bipolar affective disorders (HR, 1.61; 95% CI, 1.20-2.15; P < .001), major depressive disorder (HR, 1.53; 95% CI, 1.33-1.75; P < .001), and anxiety (HR, 1.25; 95% CI, 1.11-1.42; P < .001).

CONCLUSIONS AND RELEVANCE: In this cohort study, compared with SK youths, NKR youths showed a significantly higher risk of developing psychiatric disorders overall and individual psychiatric disorders, such as posttraumatic stress disorder, attention-deficit/hyperactivity disorder, bipolar disorder, major depressive disorder, and anxiety. This finding is important for establishing medical and educational service plans and policies for the increasing number of refugee youths in South Korea.

PMID:40440014 | DOI:10.1001/jamanetworkopen.2025.12941

JAMA Netw Open. 2025 May 1;8(5):e2512947. doi: 10.1001/jamanetworkopen.2025.12947.

ABSTRACT

IMPORTANCE: Fair clinical prediction models are crucial for achieving equitable health outcomes. Intersectionality has been applied to develop algorithms that address discrimination among intersections of protected attributes (eg, Black women rather than Black persons or women separately), yet most fair algorithms default to marginal debiasing, optimizing performance across simplified patient subgroups.

OBJECTIVE: To assess the extent to which simplifying patient subgroups during training is associated with intersectional subgroup performance in emergency department (ED) admission models.

DESIGN, SETTING, AND PARTICIPANTS: This prognostic study of admission prediction models used retrospective data from ED visits to Beth Israel Deaconess Medical Center Medical Information Mart for Intensive Care IV (MIMIC-IV; n = 160 016) from January 1, 2011, to December 31, 2019, and Boston Children’s Hospital (BCH; n = 22 222) from June 1 through August 13, 2019. Statistical analysis was conducted from January 2022 to August 2024.

MAIN OUTCOMES AND MEASURES: The primary outcome was admission to an in-patient service. The accuracy of admission predictions among intersectional subgroups was measured under variations on model training with respect to optimizing for group level performance. Under different fairness definitions (calibration, error rate balance) and modeling methods (linear, nonlinear), overall performance and subgroup performance of marginal debiasing approaches were compared with intersectional debiasing approaches. Subgroups were defined by self-reported race and ethnicity and gender. Measures include area under the receiver operator characteristic curve (AUROC), area under the precision recall curve, subgroup calibration error, and false-negative rates.

RESULTS: The MIMIC-IV cohort included 160 016 visits (mean [SD] age, 53.0 [19.3] years; 57.4% female patients; 0.3% American Indian or Alaska Native patients, 3.7% Asian patients, 26.2% Black patients, 10.0% Hispanic or Latino patients, and 59.7% White patients; 29.5% admitted) and the BCH cohort included 22 222 visits (mean [SD] age, 8.2 [6.8] years; 52.1% male patients; 0.1% American Indian or Alaska Native patients, 4.0% Asian patients, 19.7% Black patients, 30.6% Hispanic or Latino patients, 0.2% Native Hawaiian or Pacific Islander patients, 37.7% White patients; 16.3% admitted). Among MIMIC-IV groups, intersectional debiasing was associated with a reduced subgroup calibration error from 0.083 to 0.065 (22.3%), while marginal fairness debiasing was associated with a reduced subgroup calibration error from 0.083 to 0.074 (11.3%; difference, 11.1%); among BCH groups, intersectional debiasing was associated with a reduced subgroup calibration error from 0.111 to 0.080 (28.3%), while marginal fairness debiasing was associated with a reduced subgroup calibration error from 0.111 to 0.086 (22.6%; difference, 5.7%). Among MIMIC-IV groups, intersectional debiasing was associated with lowered subgroup false-negative rates from 0.142 to 0.125 (11.9%), while marginal debiasing was associated with lowered subgroup false-negative rates from 0.142 to 0.132 (6.8%; difference, 5.1%). Fairness improvements did not decrease overall accuracy compared with baseline models (eg, MIMIC-IV: mean [SD] AUROC, 0.85 [0.00], both models). Intersectional debiasing was associated with lowered error rates in several intersectional subpopulations compared with other strategies.

CONCLUSIONS AND RELEVANCE: This study suggests that intersectional debiasing better mitigates performance disparities across intersecting groups than marginal debiasing for admission prediction. Intersectionally debiased models were associated with reduced group-specific errors without compromising overall accuracy. Clinical risk prediction models should consider incorporating intersectional debiasing into their development.

PMID:40440013 | DOI:10.1001/jamanetworkopen.2025.12947

JAMA Netw Open. 2025 May 1;8(5):e2512994. doi: 10.1001/jamanetworkopen.2025.12994.

ABSTRACT

IMPORTANCE: Large language models (LLMs) have not yet been compared with traditional diagnostic decision support systems (DDSSs) on unpublished clinical cases.

OBJECTIVE: To compare the performance of 2 widely used LLMs (ChatGPT, version 4 [hereafter, LLM1] and Gemini, version 1.5 [hereafter, LLM2]) with a DDSS (DXplain [hereafter, DDSS]) on 36 unpublished general medicine cases.

DESIGN, SETTING, AND PARTICIPANTS: This diagnostic study, conducted from October 6, 2023, to November 22, 2024, looked for the presence of the known case diagnosis in the differential diagnoses of the LLMs and DDSS after data from previously unpublished clinical cases from 3 academic medical centers were entered. The systems’ performance was assessed both with and without laboratory test data. Each case was reviewed by 3 physicians blinded to the case diagnosis. Physicians identified all clinical findings as well as the subset deemed relevant to making the diagnosis for mapping to the DDSS’s controlled vocabulary. Two other physicians, also blinded to the diagnoses, entered the data from these cases into the DDSS, LLM1, and LLM2.

EXPOSURES: All cases were entered into each LLM twice, with and without laboratory test results. For the DDSS, each case was entered 4 times: for all findings and for findings relevant to the diagnosis, each with and without laboratory test results. The top 25 diagnoses in each resulting differential diagnosis were reviewed.

MAIN OUTCOMES AND MEASURES: Presence or absence of the case diagnosis in the system’s differential diagnosis and, when present, in which quintile it appeared in the top 25 diagnoses.

RESULTS: Among 36 patient cases of various races and ethnicities, genders, and ages (mean [SD] age, 51.4 [16.4] years), in the version with all findings but no laboratory test results, the DDSS listed the case diagnosis in its differential diagnosis more often (56% [20 of 36]) than LLM1 (42% [15 of 36]) and LLM2 (39% [14 of 36]), although this difference did not reach statistical significance (DDSS vs LLMI, P = .09; DDSS vs LLM2, P = .08). All 3 systems listed the case diagnosis in most cases if laboratory test results were included (all findings DDSS, 72% [26 of 36]; LLM1, 64% [23 of 36]; and LLM2, 58% [21 of 36]).

CONCLUSIONS AND RELEVANCE: In this diagnostic study comparing the performance of a traditional DDSS and current LLMs on unpublished clinical cases, in most cases, every system listed the case diagnosis in their top 25 diagnoses if laboratory test results were included. A hybrid approach that combines the parsing and expository linguistic capabilities of LLMs with the deterministic and explanatory capabilities of traditional DDSSs may produce synergistic benefits.

PMID:40440012 | DOI:10.1001/jamanetworkopen.2025.12994

Adv Ther. 2025 May 29. doi: 10.1007/s12325-025-03207-6. Online ahead of print.

ABSTRACT

INTRODUCTION: Survival outcomes associated with different androgen receptor pathway inhibitors (ARPIs) prescribed for the treatment of metastatic castration (hormone)-sensitive prostate cancer (mCSPC) have not been directly compared. The objective of this study was to compare overall survival (OS) by 24 months among ARPI-naïve patients with mCSPC initiating apalutamide or enzalutamide.

METHODS: A retrospective, causal longitudinal inverse probability of treatment weighted analysis was conducted to compare OS between patients initiating apalutamide or enzalutamide between December 2019 and December 2023 using de-identified linked US clinical and insurance claims data. Patients were excluded if they had prior exposure to ARPIs, had evidence of castration resistance, had < 12 months of database activity prior to ARPI initiation, were diagnosed with other primary cancers, or were treated with other advanced prostate cancer (PC)-related treatment (except docetaxel). Using an intention-to-treat approach, weighted Cox proportional hazards models were used to compare OS by 24 months between patients treated with apalutamide or enzalutamide (primary analyses; exploratory analyses used all available follow-up).

RESULTS: Overall, 1810 and 1909 ARPI-naïve patients who initiated apalutamide or enzalutamide, respectively, were included. Measured baseline characteristics between cohorts were well balanced after weighting (for both: mean age 73.0 years, ~ 60% white, ~ 23% black or African American, ~ 78% Medicare-insured, mean Quan-CCI 8.6, ~ 20% with visceral metastasis, 56.2% with de novo PC). At 24 months post index, there was a statistically significant 23% reduction in the risk of mortality among patients who initiated apalutamide compared with enzalutamide (hazard ratio [HR] 0.77; 95% confidence interval [CI] 0.62, 0.96; p = 0.019). Results remained consistent when using all available follow-up metrics (HR 0.77; 95% CI 0.64, 0.93; nominal p = 0.008).

CONCLUSION: In this head-to-head causal analysis among ARPI-naïve patients with mCSPC, treatment with apalutamide resulted in better survival outcomes at 24 months compared with enzalutamide. Longer follow-up studies are required to fully determine the therapeutic comparator impact of these agents.

PMID:40439959 | DOI:10.1007/s12325-025-03207-6

Adv Ther. 2025 May 29. doi: 10.1007/s12325-025-03229-0. Online ahead of print.

ABSTRACT

INTRODUCTION: Cushing’s syndrome (CS) is a rare, chronic condition caused by prolonged exposure to elevated levels of circulating cortisol, and characterized by high morbidity and mortality. The primary treatment option for CS is surgery; however, medical therapy may be useful when surgery is unsuitable, refused, or has not been curative, or a rapid control of hypercortisolism is required. While osilodrostat and metyrapone are two treatments for controlling cortisol levels, they have not been compared directly in a clinical trial. This study evaluated the comparative efficacy and tolerability of osilodrostat versus metyrapone for the treatment of CS using indirect treatment comparison methods.

METHODS: Unanchored matching-adjusted indirect comparison was used to synthesize relative treatment effects by reweighting patient-level data from two clinical trials for osilodrostat to published aggregate data for metyrapone. Efficacy endpoints included complete response (CR), defined as mean urinary free cortisol ≤ 1.0 × the upper limit of normal, at Weeks 12, 24, and 36. Tolerability endpoints included all-cause treatment discontinuation and treatment discontinuation due to lack of efficacy (LoE) or adverse events (AEs).

RESULTS: The base case analysis demonstrated that osilodrostat provides increased odds of CR versus metyrapone at Week 12 [odds ratio (OR) 2.75; 95% confidence interval (CI) 1.29, 5.88], Week 24 (OR 3.28; 95% CI 1.58, 6.84) and Week 36 (OR 10.50; 95% CI 1.84, 59.96), implying a greater proportion of patients experience normalized cortisol levels at these time-points. Although the base case analysis showed that the odds of all-cause discontinuation and discontinuation due to LoE or AEs were numerically lower for osilodrostat, the evidence was insufficient to show a statistically significant difference.

CONCLUSION: These analyses show that osilodrostat increases the odds of achieving CR at Weeks 12, 24, and 36 versus metyrapone, demonstrating that osilodrostat is a more efficacious treatment option for normalizing cortisol levels in CS patients.

PMID:40439958 | DOI:10.1007/s12325-025-03229-0

Adv Ther. 2025 May 29. doi: 10.1007/s12325-025-03238-z. Online ahead of print.

ABSTRACT

INTRODUCTION: An imbalanced diet is one of the leading causes of metabolic dysfunction-associated steatotic liver disease (MASLD) development. Diet modification remains the leading approach in the disease management. However, the role of minerals in MASLD development and treatment is poorly understood. In this retrospective study we compared minerals intake in patients with MASLD and age- and sex-matched controls, based on the data of a food frequency questionnaire.

METHODS: A retrospective database search was performed to identify eligible data of the nutritional assessment with software based on a food frequency questionnaire. The institutional medical records of the obtained cohort were then searched for medical conditions in accordance with the inclusion/exclusion criteria. On the basis of the presence of MASLD, the subjects were allocated to either MASLD or the control group. Sex- and age-matched pairs were formed for the analysis. Consumption of major minerals and trace elements was compared using non-parametric statistics.

RESULTS: Records of 15,862 subjects were screened, and the data of 226 sex- and age-matched pairs of patients with MASLD and controls were selected for the analysis. The absolute average daily intake of most of the minerals, except silicon, cobalt, molybdenum, nickel, and chromium, was greater in the MASLD group than in the control group. However, relative value (per 1000 kcal) analysis revealed that only boron intake was greater in the MASLD group (28.3 ± 38.5 vs 19.5 ± 24.7 μg/day, p = 0.013). Subjects with MASLD exceeded the recommended daily allowance (RDA) for sodium (241% of RDA), phosphorus (211%), vanadium (1576%), manganese (410%), and selenium (197%) intake, but consumed less than the recommended amounts of silicon (5% of the RDA), molybdenum (28%), fluorine (3%), zinc (91%), and chromium (37%).

CONCLUSION: Patients with MASLD consumed greater amounts of most minerals than did the control group due to overeating. When diet modification for patients with MASLD is planned, the intake of calcium, zinc, and boron needs to be controlled, and the diet may be modified with food supplements or specific foods rich in these minerals.

PMID:40439957 | DOI:10.1007/s12325-025-03238-z

Discov Oncol. 2025 May 29;16(1):944. doi: 10.1007/s12672-025-02819-2.

ABSTRACT

BACKGROUND: Cervical cancer remains a significant global health concern, with immune system regulation potentially playing a crucial role in disease development. This study investigates potential causal relationships between genetic variants associated with immune cell populations and cervical cancer risk.

METHODS: We employed multiple Mendelian randomization (MR) approaches inverse-variance weighted, MR-Egger, simple median, and weighted median methods to evaluate genetic instrumental variables linked to various T-cell and B-cell subtypes. Differential gene expression was analyzed using single-cell RNA sequencing, while forest plots, scatter plots, and funnel plots facilitated comprehensive MR analysis.

RESULTS: Forest plots consistently demonstrated odds ratios clustered tightly around 1.000 (range: 0.998-1.001), despite some variants reaching statistical significance (p < 0.05). MR analysis of CD69 + LGALS3A + regulatory T-cells, CD8 + T-cell populations, and CD20 + B-cells revealed only weak associations with cervical cancer susceptibility. Comparative analysis across different MR methodologies produced consistent results with minimal horizontal pleiotropy bias.

CONCLUSION: While immune cell genetic factors may contribute to cervical cancer development, their causal effects appear modest. These findings suggest that genetic predisposition through immune cell regulation likely plays a supplementary rather than primary role in cervical cancer pathogenesis.

PMID:40439950 | DOI:10.1007/s12672-025-02819-2

Environ Monit Assess. 2025 May 29;197(6):693. doi: 10.1007/s10661-025-14130-2.

ABSTRACT

To assess the environmental impact of training activities, surface soil was systematically collected at 3 training ranges (designated as A, B1, B2) located along the southeast coast of Guangdong, China. Analytical results of energetic compounds (ECs) in soil revealed that 2,4,6-trinitrotoluene (TNT) was the predominant ECs, with a mean concentration of 234 ng/g. The detection rate for Hexahydro-1,3,5-trinitro-1,3,5-triazine (RDX) was over 50%, peaking at 710 ng/g. Heavy metal analysis demonstrated significant enrichment of cadmium (Cd), mercury (Hg), lead (Pb), copper (Cu), and arsenic (As) in soil. Particularly noteworthy were the Cu and Pb concentrations in range B2, reaching maximum levels of 259 mg/kg and 801 mg/kg, respectively. Metal enrichment was predominantly attributed to metallic components and shrapnel fragments originating from explosive materials. Field investigations indicated that metallic fragments dispersed in surface soils exhibited limited weathering, with iron (Fe) constituting the primary metallic element, accompanied by trace amounts of chromium (Cr). The quantitative analysis of ECs and heavy metals concentrations suggests a clear trend of environmental accumulation. These findings underscore the necessity for implementing long-term environmental monitoring programs and conducting comprehensive risk assessments to evaluate potential ecological risks related to military training activities.

PMID:40439947 | DOI:10.1007/s10661-025-14130-2

Neuroradiology. 2025 May 29. doi: 10.1007/s00234-025-03654-6. Online ahead of print.

ABSTRACT

BACKGROUND: Thalassemia is an inherited hematological disorder characterized by chronic anemia and multisystem involvement, including potential impacts on the central nervous system. Research has suggested a relationship between thalassemia and glymphatic system dysfunction, which may contribute to neurocognitive impairments. However, the exact mechanisms linking thalassemia to changes in glymphatic function remain unclear.

OBJECTIVE: To investigate the relationship between glymphatic system function and thalassemia severity, particularly in patients with transfusion-dependent and non-transfusion-dependent thalassemia.

METHODS: The study included a population of 40 thalassemia patients (TM, n = 40) confirmed through genetic testing, consisting of 19 patients who were transfusion-dependent (TDT, n = 19) and 21 patients who were non-transfusion-dependent (NTDT, n = 21), along with 39 healthy controls (HC, n = 39) matched for age and sex. Glymphatic function was assessed using the index for diffusivity along the perivascular space (ALPS-index), with regions of interest selected from the medullary veins and crossing fibers in the lateral ventricles. Statistical analyses were performed using Analysis of Variance (ANOVA) for intergroup comparisons, supplemented by Bonferroni correction for multiple comparisons. Pearson correlation was utilized to explore the relationships between the diffusion tensor image analysis along the perivascular space (DTI-ALPS) index and demographic factors, while multiple linear regression was employed to adjust for confounding variables. A significance threshold of P less than 0.05 was established for all statistical tests.

RESULTS: The mean DTI-ALPS Index for thalassemia patients were significantly lower than those of healthy controls, indicating impaired glymphatic function. No significant differences in DTI-ALPS Index were found between transfusion-dependent and non-transfusion-dependent groups. Correlation analysis showed minimal influence of age and education on glymphatic function in the study population.

CONCLUSION: This study demonstrates significant impairment in glymphatic system function in thalassemia patients, potentially related to chronic anemia and iron overload. These findings contribute to understanding the mechanisms underlying cognitive dysfunction in thalassemia.

PMID:40439942 | DOI:10.1007/s00234-025-03654-6

Neurosurg Rev. 2025 May 29;48(1):455. doi: 10.1007/s10143-025-03597-9.

ABSTRACT

BACKGROUND: Artificial Intelligence (AI) has emerged as a transformative tool in medicine, particularly addressing neurosurgical challenges such as complex anatomical delineation and intraoperative decision-making. Despite advancements in diagnostic and prognostic algorithms, obstacles including algorithmic bias, data privacy, and model interpretability continue to limit its widespread clinical adoption.

OBJECTIVE: This systematic review aims to evaluate the current applications of AI in neurosurgery, compare the performance of various AI models, and examine the ethical challenges associated with their integration into clinical practice.

METHODS: A systematic literature search was conducted in PubMed, Scopus, and Web of Science databases, following PRISMA guidelines. Studies from 2015 to 2025 focusing on AI applications in diagnostic, prognostic, surgical, and intraoperative neurosurgical contexts were included. Statistical outcomes, model performance metrics, and ethical considerations were analyzed.

RESULTS: Thirteen studies met the inclusion criteria. AI models, particularly ML and DL, demonstrated superior diagnostic accuracy (AUC > 0.90) and improved prognostic predictions by up to 15%. AI-assisted surgical planning enhanced precision and reduced complication rates by 10-20%. However, algorithmic bias, limited transparency, and lack of external validation remain key barriers to clinical adoption.

CONCLUSION: AI improves diagnostic accuracy, prognostic predictions, and surgical precision while reducing complication rates. However, challenges such as bias, limited interpretability, and the need for external validation must be addressed for widespread clinical integration.

PMID:40439939 | DOI:10.1007/s10143-025-03597-9