Tag: pubmed

Philos Trans A Math Phys Eng Sci. 2026 Aug 27;384(2327):20250067. doi: 10.1098/rsta.2025.0067.

ABSTRACT

As artificial intelligence (AI) systems are being widely deployed in safety-critical and high-stakes applications (e.g. medical diagnosis, autonomous vehicles, financial risk assessment), there is a growing demand for providing reliable and trustworthy machine predictions. However, since AI models become more complex in structure (a prominent example being deep neural networks) and bigger in size (e.g. large language model systems such as ChatGPT and Gemini), being able to understand, explain and quantify confidence in their predictions are ongoing challenges. Therefore, this special issue is dedicated to exploring the forefront of reliable uncertainty quantification in AI systems, using conformal prediction (CP), a leading statistical framework that offers predictions with valid coverage guarantees under minimal assumptions. The issue comprises the most recent, most novel and most practical developments of CP-based methods in cutting-edge AI applications, highlighting improvements over traditional methods.

PMID:42656163 | DOI:10.1098/rsta.2025.0067

Philos Trans A Math Phys Eng Sci. 2026 Aug 27;384(2327):20250080. doi: 10.1098/rsta.2025.0080.

ABSTRACT

Conformal predictive systems are sets of predictive distributions with theoretical out-of-sample calibration guarantees. The calibration guarantees are typically that the set contains a forecast distribution whose prediction intervals exhibit the correct marginal coverage at all levels. Conformal predictive systems are constructed using conformity measures that quantify how well possible outcomes conform with historical data. However, alternative methods have been proposed to construct predictive systems with more appealing theoretical properties. We study an approach to construct predictive systems that we term residual distribution predictive systems (RDPSs). In the split conformal setting, this approach nests conformal predictive systems with a popular class of conformity measures, providing an alternative perspective on the classical approach. In the full conformal setting, the two approaches differ, and the new approach has the advantage that it does not rely on a conformity measure satisfying fairly stringent requirements to ensure that the predictive system is well-defined; it can readily be implemented alongside any point-valued regression method to yield predictive systems with out-of-sample calibration guarantees. The empirical performance of this approach is assessed using simulated data, where it is found to perform competitively with conformal predictive systems. However, the new approach offers considerable scope for implementation with alternative regression methods. This article is part of the theme issue ‘Advancing uncertainty quantification in AI systems’.

PMID:42656161 | DOI:10.1098/rsta.2025.0080

Philos Trans A Math Phys Eng Sci. 2026 Aug 27;384(2327):20250069. doi: 10.1098/rsta.2025.0069.

ABSTRACT

Current experimental scientists have been increasingly relying on simulation-based inference (SBI) to invert complex nonlinear models with intractable likelihoods. However, posterior approximations obtained with SBI are often miscalibrated, causing credible regions to undercover true parameters. We develop CP4SBI, a model-agnostic conformal calibration framework that constructs credible sets with local Bayesian coverage. Our two proposed variants, namely local calibration via regression trees and cumulative distribution function CDF-based calibration, enable finite-sample local coverage guarantees for any scoring function, including highest posterior density (HPD), symmetric and quantile-based regions. Experiments on widely used SBI benchmarks demonstrate that our approach improves the quality of uncertainty quantification for neural posterior estimators (NPEs) using both normalizing flows and score-diffusion modelling. This article is part of the theme issue ‘Advancing uncertainty quantification in AI systems’.

PMID:42656156 | DOI:10.1098/rsta.2025.0069

Sheng Wu Yi Xue Gong Cheng Xue Za Zhi. 2026 Aug 25;43(4):845-852. doi: 10.7507/1001-5515.202502034.

ABSTRACT

Systemic lupus erythematosus (SLE) frequently involves the central nervous system, inducing abnormal alterations in brain functional and structural networks and resulting in cognitive and psychological dysfunction in patients. To identify abnormal brain functional network patterns associated with SLE, this study adopted a dynamic threshold strategy to detect key functional connections and construct sparse brain functional networks. A graph transformation network (GTNet) was utilized to model the optimized networks, capturing local topological features and global dependencies to effectively identify abnormal patterns of SLE-related brain functional networks. Experimental results showed that the proposed model achieved an average classification accuracy of (87.48 ± 6.77)% on the resting-state functional magnetic resonance imaging dataset consisting of 107 SLE patients and 107 healthy controls. Further analysis showed that the difference in small-world properties between the two groups was statistically significant ( t = -2.96, P < 0.01). In conclusion, the model constructed in this study provides a new scheme for the auxiliary diagnosis of SLE, and its automated classification architecture has potential application value.

PMID:42656117 | DOI:10.7507/1001-5515.202502034

Sheng Wu Yi Xue Gong Cheng Xue Za Zhi. 2026 Aug 25;43(4):704-710. doi: 10.7507/1001-5515.202602025.

ABSTRACT

Spikes, as a key electrophysiological marker of abnormal synchronized neuronal discharges, can be identified and detected by scalp electroencephalography (EEG), and are of great value for localizing epileptic foci and investigating the mechanisms of epilepsy. However, spike propagation is strongly influenced by the skull, making it difficult to infer the precise intracranial origin of spikes from scalp EEG alone. Therefore, based on simultaneous intracranial and scalp EEG recordings, this study investigated the propagation characteristics of spikes from the intracranial space to the scalp. The study mainly focused on four aspects: the characteristics of simultaneous intracranial and scalp spikes, the scalp visibility of intracranial spikes, the spatial distribution of scalp spikes, and the influence of epileptogenic and non-epileptogenic regions on spike propagation. The results showed that spikes detectable on the scalp differed significantly in energy-related features from those not detectable on the scalp ［ MD = 2.6 × 10 ^-6 (2.5 × 10 ^-6, 2.6 × 10 ^-6), P < 0.01］. When the intracranial spike activation area was less than 10 cm ², the scalp detection rate was 18.9%; when the area reached 30～40 cm ², the detection rate increased to 62.5%, showing an increasing trend. After spikes propagated to the scalp, the detection rate decreased by 63.5% when the horizontal distance of their spatial distribution exceeded 4 cm. In addition, there was no statistically significant difference in scalp detection rate between spikes originating from epileptogenic and non-epileptogenic regions ( P > 0.05). In summary, these findings are important for elucidating the propagation characteristics of spikes and may provide quantitative evidence for electrophysiological analysis and focus localization in epilepsy.

PMID:42656101 | DOI:10.7507/1001-5515.202602025

Vet Dermatol. 2026 Aug 27. doi: 10.1111/vde.70122. Online ahead of print.

ABSTRACT

BACKGROUND: Topical glucocorticoid (GC) is used to manage inflammatory skin diseases in dogs. However, contact times associated with detectable topical GC effects have not been clarified in dogs. Uncertainty regarding contact time may lead to premature removal or overuse, compromising efficacy and safety.

OBJECTIVES: To determine contact times at which detectable GC effects are observed after topical difluprednate lotion application in dogs.

ANIMALS: Six male Beagle dogs.

METHODS: Suppression of the histamine-induced vascular reaction (HVR), based on the principles of the vasoconstrictor assay, was used as an indicator of detectable GC effects. Difluprednate lotion was applied to a circular area (2 cm in diameter) on the lateral trunk 5, 10 and 30 min before intradermal histamine injection. Fluorescein was administered intravenously 5 min before histamine injection to visualise HVR. HVR images were obtained 15 min after histamine injection under UV illumination, and the fluorescent area was measured using ImageJ. Statistical analysis was performed using Dunnett’s multiple comparison test.

RESULTS: When difluprednate application was performed 5, 10 and 30 min before histamine injection, HVR areas were smaller than untreated control sites (p < 0.01). No significant differences were detected among sites treated 5, 10 and 30 min before histamine injection.

CONCLUSIONS AND CLINICAL RELEVANCE: Difluprednate lotion suppressed the HVR when applied 5, 10 and 30 min before histamine injection, indicating detectable dermal GC activity at the evaluated time points. Information regarding contact times at which topical GC effects can be detected may help clinicians to provide contact-time instructions and improve treatment management.

PMID:42656084 | DOI:10.1111/vde.70122

J Adv Nurs. 2026 Aug 27. doi: 10.1111/jan.70744. Online ahead of print.

ABSTRACT

AIM: To determine the level of obstetric violence among pregnant women and identify associated sociodemographic and obstetric factors in Türkiye.

DESIGN: A descriptive cross-sectional study.

METHODS: The study was conducted with 505 pregnant women attending antenatal clinics in Training and Research Hospitals in the Black Sea region of Türkiye between July and December 2025. Data were collected using a self-administered sociodemographic form and the validated 18-item Obstetric Violence Scale for Pregnant Women (OVS-Pregnant Women). Total scale scores range from 0 to 72 (higher scores indicate higher perceived obstetric violence) across three subdimensions: Supportive Care and Information Support (0-28), Achieving Professional Standards of Care and Effective Communication (0-28), and Health Promotion and Encouragement (0-16). Data were analysed using descriptive statistics, and univariate and multivariate linear regression.

RESULTS: The mean age of participants was 30.02 ± 5.46 years. The mean total OVS score was 16.681 ± 13.961 (out of 72). Subdimension mean scores were 9.853 ± 8.428 for Supportive Care and Information Support, 4.310 ± 4.314 for Achieving Professional Standards of Care and Effective Communication, and 2.516 ± 3.470 for Health Promotion and Encouragement. Multivariate analysis identified two factors independently associated with lower obstetric violence: family monthly wage above the minimum wage and receiving antenatal care at a training and research hospital.

CONCLUSION: Perceived obstetric violence scores were toward the lower end of the possible range, with socioeconomic status and healthcare institution type influencing women’s care experiences.

IMPLICATIONS FOR THE PROFESSION AND/OR PATIENT CARE: Obstetric violence should be recognized as a structural issue, necessitating standardized measurement tools and observational studies. Future studies should examine how exposure to obstetric violence shapes women’s care preferences, informing patient-centered interventions.

IMPACT: The findings provide actionable evidence for maternity care providers, hospital administrators, and healthcare policymakers to foster respectful, patient-centered care.

REPORTING METHOD: STROBE.

PATIENT OR PUBLIC CONTRIBUTION: No patient or public contribution.

PMID:42656050 | DOI:10.1111/jan.70744

Cancer Med. 2026 Sep;15(9):e72206. doi: 10.1002/cam4.72206.

ABSTRACT

BACKGROUND: Neurological immune-related adverse events (N-irAEs) from immune checkpoint inhibitors (ICIs) are rare but potentially fatal. This study aimed to characterize their real-world profile to support early detection and management.

METHODS: We analyzed ICI-associated individual case safety reports (ICSRs) in the FDA Adverse Event Reporting System (FAERS) from April 2011 to June 2023. Disproportionality analyses used the reporting odds ratio (ROR) and the Bayesian confidence propagation neural network (BCPNN). Associations between serious N-irAEs and sex or age group (< 60 vs. ≥ 60 years) were assessed using chi-squared tests. We also evaluated the co-occurrence of myasthenia gravis with myositis and/or myocarditis and its association with mortality.

RESULTS: A total of 3486 ICSRs involving N-irAEs were identified, with 550 cases reporting fatal outcomes. A total of 4139 N-irAE signals were detected; among these, 2560 signals corresponded to the top 10 Preferred Terms (PTs) with the highest reporting frequencies. The median time from ICI initiation to the onset of N-irAEs was 82 days (IQR: 29-219 days). Of the 4139 N-irAE signals, 1925 (46.5%) were classified as serious adverse events; no statistically significant differences were observed in their distribution by sex (χ² = 2.492, p = 0.114) or age (χ² = 2.946, p = 0.086). A total of 461 cases of myasthenia gravis were identified, of which 219 (47.5%) were concomitant with myocarditis and/or myositis. Among the 137 fatal cases of myasthenia gravis, 70 (51.1%) were concomitant with myocarditis and/or myositis.

CONCLUSION: This study offers a comprehensive characterization of N-irAEs, highlights the fatality risk when myasthenia gravis co-occurs with myocarditis and/or myositis, and underscores the importance of early recognition, multi-system evaluation, and timely intervention to improve patient safety and optimize the risk-benefit balance of immunotherapy.

PMID:42656039 | DOI:10.1002/cam4.72206

Int J Gynaecol Obstet. 2026 Aug 26. doi: 10.1002/ijgo.71304. Online ahead of print.

ABSTRACT

BACKGROUND: The maternal continuum of care (CoC) refers to an integrated sequence of health services delivered across time and space, which women should access without interruption. It is a vital strategy for reducing maternal morbidity and mortality. This study assessed the level and determinants of maternal CoC utilization among women in the Gamo Zone, South Ethiopia.

METHODS: A community-based cross-sectional study was conducted from February 1 to 30, 2025, among women who had given birth within the 12 months preceding the data collection. A multi-stage sampling technique was used to select the study population. A total of 1294 women were selected using simple random sampling. Data were collected using Kobo Collect and analyzed in SPSS version 25.0. A multivariable model was fitted, and adjusted odds ratios (AOR) with 95% confidence intervals (CI) were reported. Statistical significance was set at P < 0.05.

RESULTS: The overall rate of maternal CoC utilization (defined as receiving four or more antenatal care visits, institutional delivery, and postnatal care within 2 days after birth) was 16.1%. Key factors significantly associated with full CoC utilization included knowledge of pregnancy danger signs (AOR = 12.59; 95% CI: 7.84, 20.02), perceived shorter distance to health facility (AOR = 2.12; 95% CI: 1.28, 3.53), urban residency (AOR = 12.00; 95% CI: 7.57, 20.00), maternal educational status (AOR = 2.93; 95% CI: 1.33, 6.45), and having a birth preparedness and complication readiness (BPCR) plan (AOR = 4.96; 95% CI: 2.89, 8.53).

CONCLUSION: The overall CoC utilization in the study area remains low, indicating that women are not getting the maximum possible health benefits from the existing healthcare services. Place of residence, distance to reach health facility, maternal education, knowledge of pregnancy danger signs, and BPCR were the factors affecting CoC utilization. Therefore, policymakers should consider strategies for the existing health system that focus on empowering women’s education, counseling women about pregnancy danger signs, and implementing the BPCR, as well as ensuring the physical accessibility of health facilities within a reasonable distance to enhance maternal CoC utilization.

PMID:42656019 | DOI:10.1002/ijgo.71304

Ophthalmic Genet. 2026 Aug 26:1-9. doi: 10.1080/13816810.2026.2720681. Online ahead of print.

ABSTRACT

INTRODUCTION: This prospective study evaluated the pharmacologic effect of oral acetazolamide on structural and functional outcomes in patients with X-linked retinoschisis (XLRS) using a cross-over design.

METHODS: Five male participants from the prior AXIS control arm received oral acetazolamide 250 mg twice daily for 32 weeks, with dose adjustments based on structural response and follow-up to week 64. The primary endpoint was the central subfield thickness (CST). Secondary measures included best-corrected visual acuity (BCVA), low-luminance visual acuity (LLVA), retinal sensitivity on microperimetry (MP), cystoid fluid collection volume (CFCV), and patient-reported outcomes via the Michigan Retinal Degeneration Questionnaire (MRDQ).

RESULTS: A significant difference in CST of -72.8 µm (p = 0.001; 95% CI – 117.2 to -28.4) was observed between the control and treatment periods on average over time with responses varying among patients. Statistically significant but modest improvements in BCVA, LLVA and retinal sensitivity on MP were also observed. Common treatment-related side effects included fatigue, paresthesia, dysgeusia, and tinnitus, leading one patient to discontinue treatment.

CONCLUSIONS: While the REAXIS crossover study demonstrated statistically significant improvements in CST, BCVA, and LLVA, the effect sizes were modest and highly variable across individuals. These findings support further exploration of personalized treatment strategies and predictive biomarkers in XLRS.

PMID:42655987 | DOI:10.1080/13816810.2026.2720681