Activating ESR1 mutations are a major mechanism of resistance to aromatase inhibitors in hormone receptor-positive, HER2-negative metastatic breast cancer (mBC). International guidelines, including those from ASCO, NCCN, and ESMO, recommend liquid biopsy as the preferred approach for ESR1 mutation testing at progression on endocrine therapy, with digital PCR (dPCR) and next-generation sequencing (NGS) as the preferred analytical platforms. Although elacestrant was approved by the U.S. Food and Drug Administration together with Guardant360® Dx as its companion diagnostic, European regulatory frameworks allow the use of validated in-house assays for ESR1 testing, which are increasingly being implemented across clinical laboratories. To support the clinical implementation of ESR1 testing and improve analytical standardization in routine practice, we performed a European multicentre analytical verification study using dPCR- and NGS-based liquid biopsy workflows. Six referral institutions participated in this study. All laboratories verified dPCR workflows and four also verified NGS-based assays using standardized reference materials containing clinically relevant ESR1 mutations. Limit of detection (LoD) and limit of blank (LoB) were determined in each laboratory according to locally validated workflows following CLSI-based verification procedures. Analytical sensitivity and specificity were assessed across platforms, focusing on the two most frequently tested ESR1 hotspot mutations, p.Y537S and p.D538G. Total DNA input ranged from 10 to 30 ng per reaction for dPCR assays, while NGS input followed platform-specific requirements. LoD values ranged from 0.01% to 0.07% variant allele frequency (VAF) for dPCR assays and from 0.01% to 0.1% for NGS-based workflows, depending on platform characteristics and DNA input. Limit of blank values were low across laboratories, with minimal background signal detected in plasma samples from healthy donors. Comparable analytical performance was observed across different platforms when assays were performed under validated laboratory conditions. These results demonstrate that both dPCR and NGS-based liquid biopsy workflows can be successfully implemented for ESR1 mutation testing in routine clinical practice using locally validated assays. This multicentre verification study provides practical guidance on assay verification, DNA input requirements, and key analytical parameters required to ensure reliable ESR1 mutation detection across different European laboratories. Robust analytical verification of ESR1 testing may improve diagnostic reliability and support personalized treatment strategies for patients with hormone receptor-positive, HER2-negative metastatic breast cancer.
The study of harmonized reference intervals (RIs) in Croatia was published in 2004. Since then, notable methodology and population characteristics changes have occurred in the Croatian population, highlighting the need for re-evaluation. This study aimed to verify harmonized RIs for hematology, coagulation, and clinical chemistry laboratory tests and assess their applicability in one clinical hospital center. This verification study included 100 apparently healthy adults (40 males, 60 females; aged 22 to 75 years), selected using both a priori and a posteriori exclusion criteria following CLSI EP28-A3c guidelines. Blood samples were collected under standardized preanalytical conditions. Hematology analysis was performed on the Sysmex XN1000 analyzer (Sysmex, Kobe, Japan), coagulation testing on the Sysmex CS2500 analyzer (Siemens Healthcare Diagnostics Inc., Marburg, Germany), and clinical chemistry analysis on the Beckman Coulter DxC700AU (Beckman Coulter, Brea, USA) and Radiometer ABL90 FLEX (Radiometer, Brønshøj, Denmark) analyzers. Reference intervals were considered verified if at least 90% of results (18/20) fell within the predefined RIs. All tested hematology and coagulation RIs were successfully verified. Among 27 tested biochemical analytes, 25 RIs were verified in the first sample set. Total calcium and alkaline phosphatase RIs required additional verification. International harmonized RIs were adopted and successfully verified for both tests using two independent sample sets. Most harmonized RIs in use in Croatia remain applicable to the adult population of one clinical hospital center. However, periodic re-evaluation is needed due to changes in analytical methods and population characteristics. Reference intervals from international harmonization projects can provide a valid alternative when local verification fails.
To develop and validate a machine learning-based model for predicting the risk of breast cancer occurrence in BI-RADS 4 patients. 216 breast lesions from 212 patients were included for retrospective analysis. The training set (151 cases) and the validation set (65 cases) were randomly divided from the whole data set at a ratio of 7:3. Use logistic as well as LASSO regressions to identify independent risk factors. Subsequently, eight ML models were constructed. After a comprehensive comparison, the optimal model was selected and visualized using the SHAP algorithm. Through feature selection, six parameters were identified as independent risk factors. The Emax-2 shell cutoff (104.71 kPa) demonstrated the highest diagnostic efficacy for the 4a subgroup. Among the eight ML algorithms, the random forest model exhibited potential overfitting risks(AUC=1.00), whereas the LR model demonstrated superior stability. Consequently, the LR model was selected as the predictive model, and a nomogram was constructed based on it. In this study, the LR model enhanced the capacity to identify BI-RADS 4 lesions as BC. However, the study is a single-center study with a relatively small sample size, which brings certain restrictions to the clinical application of the model. The LR model is the best choice for predicting BC incidence in BI-RADS 4 lesions, which can help clinicians improve BC identification and treatment at an early stage.
Accurate complete blood count (CBC) measurements are fundamental to modern hematology practice, and the introduction of new automated hematology analyzers requires rigorous analytical validation to ensure reliability, comparability, and clinical safety. The Atellica HEMA 580 (Siemens Healthineers AG, Erlangen, Germany) is a high-throughput hematology analyzer for which independent Clinical and Laboratory Standards Institute (CLSI)-guided validation data remain limited. This study aimed to perform a comprehensive analytical validation of the Atellica HEMA 580 and to evaluate its analytical comparability with the Sysmex XN-3100 (Sysmex Corporation, Kobe, Japan). Analytical validation was conducted in accordance with CLSI guidelines. Precision was assessed following CLSI EP05-A3, linearity according to EP06-A, carryover using EP07-A2, and method comparison per EP09-A3. Precision was evaluated using 15 within-run and 15 inter-assay replicates across three quality control levels. Linearity was evaluated using proportional dilution from neat to 1:16. Forty-four paired routine clinical samples were analyzed in parallel on both analyzers using ordinary least squares, Passing-Bablok, and Deming regression, supplemented by Bland-Altman analysis. Within-run analytical imprecision was low, with coefficients of variation of 0.76% for white blood cells, 0.73% for red blood cells, 0.25% for hemoglobin, and 3.59% for platelets, together with stable inter-assay reproducibility. Linearity of red blood cell and hemoglobin measurements demonstrated high proportionality across the evaluated analytical ranges. Platelet method comparison demonstrated a modest negative proportional bias (Deming slope 0.95; 95% confidence interval, 0.91-0.99). High correlation coefficients across major hematological parameters further supported strong analytical agreement between the two analyzers. Bland-Altman analysis showed minimal systematic bias, and carryover remained low across all parameters. Atellica HEMA 580 met CLSI analytical validation criteria and demonstrated analytical agreement with the Sysmex XN platform, supporting its use in routine and high-throughput hematology laboratories.
Color Doppler mapping (DUSG) enables evaluation of parenchymal perfusion after clamping the renal artery. The aim of this study was to evaluate DUSG and compare its outcomes with ICG fluorescence imaging and with a control group without intraoperative imaging. We retrospectively analyzed 426 patients operated on at our institution between 2018 and July 2025. Ischemia was verified using DUSG (n = 174; 41%) and ICG (n = 29; 7%). A choice of method was non-systematic based on surgeon decision only. The control group included 223 patients (52%). Demographic, oncological and surgical parameters, selective clamping, clamp adjustments, and positive resection margins (pR1) were evaluated. ANOVA/Kruskal-Wallis and chi-square tests were used for comparison. No significant differences were found between groups in tumor size, BMI, blood loss or WIT. Selective clamping was more frequent with DUSG and ICG compared in controls (p < 0.001). Clamp adjustment was required in 24% of DUSG, 62% of ICG, and 3% of control cases (p < 0.001). The rate of pR1 was low across groups. Intraoperative perfusion verification is advantageous when ischemia is not clearly visible. Both DUSG and ICG are safe and reliable. DUSG represents an effective, accessible and economical alternative suitable for broader clinical use.
In recent years, extensive research has been conducted on avatars, and multiple studies have demonstrated their effectiveness as a medium for remote operation. While avatars are effective when teleoperated, they must also be capable of autonomous behavior in the absence of an operator. In particular, avatars whose appearance closely resembles that of a real individual need to possess conversational abilities that reflect the personality of the person being modeled. This paper presents the development of a speech generation system that produces personality-consistent utterances using a large language model (LLM) and speech synthesis technology. We call this system AvatarLLM. Through system evaluation, we examined the factors contributing to the perception of individuality. Experimental results indicated that the utterances generated by AvatarLLM were perceived as more likely reproducing the modeled individual than those of the actual person. Furthermore, we found that the perceived identity of the utterances could influence the perceived identity of the voice itself.
Methamphetamine (METH) is a highly addictive psychostimulant that alters gene expression in brain reward circuits. This study aimed to identify METH-associated transcriptional changes in the nucleus accumbens (NAc) and explore potential pharmacological interventions. We conducted an in silico analysis of publicly available microarray data (GSE46717) from the Gene Expression Omnibus (GEO). Differentially expressed genes (DEGs) were identified using limma and analyzed for functional enrichment via EnrichR. Protein-protein interaction (PPI) networks were constructed using STRING to identify hub genes, validated in silico with jackknife resampling. Adult male Wistar rats were injected with METH (10 mg/ kg, followed by 2.5 mg/kg after one month), and expression of selected hub genes was measured in NAc tissue using quantitative polymerase chain reaction (qPCR). Connectivity mapping was applied to identify candidate drugs reversing METH-induced transcriptional changes. We identified 280 DEGs (210 upregulated, 70 downregulated). Upregulated pathways included caffeine metabolism, long-term potentiation, and cocaine addiction, whereas GABAergic and glutamatergic synapse genes were downregulated. Network analysis highlighted Fos, Crh, Oprl1, and Slc17a6 as hub genes, validated both computationally and experimentally. Connectivity mapping identified D-64131 and Mebendazole as potential therapeutics. METH induces substantial transcriptional alterations in the NAc, affecting synaptic signaling and addiction pathways. Integrating in silico network analysis with experimental validation identified robust hub genes and suggested candidate compounds for therapeutic intervention.
High-dose-rate (HDR) brachytherapy provides a highly conformal cancer treatment modality by exploiting steep dose gradients, and achieving excellent tumour control while minimising radiation exposure to healthy tissues. In-vivo dosimetry (IVD) serves as an essential quality assurance tool, offering independent verification of delivered dose. However, its accuracy can be affected by several measurement-related uncertainties. This study aimed to characterise diode-based IVD for Co-60 HDR brachytherapy and quantify the uncertainties influencing detector performance. A Co-60 HDR afterloading system (SagiNova®) was used along with a diode-based in-vivo detector. Calibration was performed using a Polymethyl Methacrylate PMMA phantom. A custom-designed acrylic phantom was fabricated to ensure reproducible detector positioning and fixed geometry during irradiation. Detector linearity and uniformity were assessed by delivering known doses from 1 to 8 Gy in 1 Gy increments. Since conventional brachytherapy treatment planning systems do not account for tissue heterogeneity, additional measurements were performed by placing materials simulating bone (Teflon), lung (cork), and soft tissue (acrylic) of 1-3 cm thickness between the source and detector. The diode exhibited excellent stability, with repeatability showing < 2% relative standard deviation. Sensitivity across cumulative absorbed doses demonstrated < 2.5% variation, confirming strong consistency. A linear response was observed throughout the tested dose range. Heterogeneity analysis revealed notable dose perturbations: as expected, bone-equivalent material produced the highest attenuation, while lung-equivalent material resulted in the least, underscoring the importance of accounting for tissue density variations in IVD measurements. Although IVD offers valuable real-time dose verification in HDR brachytherapy, its widespread clinical adoption remains limited by challenges such as detector size and the steep dose gradients surrounding the source. Comprehensive commissioning-including evaluation of linearity, reproducibility, geometric dependence, and heterogeneity effects-is critical for understanding detector behaviour under clinical conditions. It is concluded that accurate characterisation of these uncertainties enhances the reliability of diode-based IVD systems and supports their integration into routine brachytherapy practice for improved patient safety and treatment precision.
Exposure to tobacco advertising at tobacco retail outlets (TROs) is associated with smoking initiation among youth. There is limited geospatial evidence on the density of TROs in Lao People's Democratic Republic (PDR), a country with high prevalence of tobacco smoking. This study examined the density and proximity of TROs around schools in two urban districts and one rural district of Vientiane Capital, Lao PDR, using geographic information systems. We audited 233 TROs around 27 schools between January 19 and February 18, 2024. TROs were mapped within 250 m and 500 m buffers in two urban districts (Chanthabuly and Sissatanak), and 500 m and 1,000 m buffers in one rural district (Naxaithong). Buffer analysis and network analysis estimated TRO density and median walking distances between TROs and schools. We used the Kruskal-Wallis test to determine if TRO density varied significantly in urban districts within two buffers and the chi-square test to examine differences in TRO characteristics based on proximity to schools. TRO density was defined as the number of TROs mapped within a 250 m radius of urban schools and those mapped within 1,000 m radius of rural schools. TRO density was higher within the buffer of 250-500 m in Chanthabuly (median = 12), followed by Sissatanak (median = 3) (p = 0.01). Comparing the two urban districts, the median distance between TROs and schools within the buffer of 250-500 m was significantly less in Sissatanak compared to Chanthabuly (p = 0.04). The shortest distance between an urban school and any TRO (i) without age verification signage was 21.58 m, (ii) with outside cigarette advertisements was 9.95 m, and (iii) selling "single" cigarettes was 311.36 m. In the rural district, the TRO density was higher within 500-1,000 m (median = 2) compared to within 500 m (median = 0.5). Within the context of the Lao PDR, our study provides the first geospatial evidence of tobacco retail outlet density in both urban and rural districts of Vientiane Capital, revealing a substantially higher concentration of outlets in urban districts.We also quantified walking distances between schools and outlets violating tobacco control measures, including lack of age verification signage and outdoor cigarette advertising. These findings suggest that stricter regulation of tobacco retail outlets could strengthen tobacco control policy implementation in Lao PDR.
Methotrexate (MTX) is a cornerstone non-surgical therapy for ectopic pregnancy, but its narrow therapeutic window necessitates strict dose verification to prevent toxicity. We report a case of severe MTX intoxication in a 31-year-old woman following salpingostomy, resulting from a dosing error. After receiving an excessive intramuscular dose, the patient developed severe gastrointestinal symptoms and acute kidney injury (AKI). Initial management included prompt leucovorin rescue, urine alkalinization, and intensive supportive care; however, renal dysfunction progressed with impaired MTX clearance. In the absence of glucarpidase, the patient was transferred to the intensive care unit (ICU) and treated with continuous renal replacement therapy (CRRT), which facilitated sustained MTX removal and was followed by gradual clinical and biochemical recovery. This case highlights the critical importance of dose verification in MTX administration and illustrates that, when conventional measures are insufficient and glucarpidase is unavailable, CRRT may serve as a viable salvage therapy for life-threatening MTX toxicity complicated by AKI.
To assess the diagnostic performance of computed tomography (CT) for predicting main pancreatic duct injury in surgically explored patients with distal pancreatic trauma. We conducted a prospective single-center diagnostic accuracy study from August 2024 to March 2026. Consecutive patients with distal pancreatic trauma underwent contrast-enhanced abdominal CT before surgery. On CT, main pancreatic duct injury was defined by either a deep pancreatic parenchymal laceration involving at least 50% of pancreatic thickness or complete pancreatic transection. CT findings were compared with intraoperative assessment as the reference standard. Forty-two patients were included; 39 (92.9%) had intraoperatively confirmed main pancreatic duct injury. CT correctly classified 37 patients (36 true positives and 1 true negative), with 3 false-negative and 2 false-positive results. Sensitivity, positive predictive value, specificity, negative predictive value, and overall accuracy were 92.3% (95% CI: 79.1-98.4%), 94.7% (95% CI: 82.3-99.4%), 33.3% (95% CI: 0.8-90.6%), 25.0% (95% CI: 0.6-80.6%), and 88.1% (95% CI: 73.6-96.5%), respectively. All false-negative results occurred after early CT, performed 3-7 h after trauma. No clear pancreatic laceration was seen on the initial scan, but the injuries became more apparent on repeat CT. In surgically explored patients with distal pancreatic trauma, CT criteria based on a deep parenchymal laceration involving at least 50% of gland thickness or complete transection showed high sensitivity for predicting main pancreatic duct injury. However, specificity and negative predictive value were imprecise because very few duct-intact patients underwent surgical verification. All diagnostic estimates should therefore be interpreted in the context of selection and partial verification bias. Early CT can miss duct injury, and repeat CT should be considered when clinical suspicion remains high.
In invasion science, risk analysis tools are widely used to support the prevention and management of non-native species introductions. Many decision-support frameworks, including those derived from Weed Risk Assessment methodologies such as the Invasiveness Screening Kit (ISK) family of tools, require calibration of outcome scores to distinguish between higher- and lower-risk species. A prerequisite for such calibration is the a priori categorisation of screened species as invasive or non-invasive based on evidence from their introduced range, when applicable. This paper provides a curated global dataset comprising 1,926 taxa identified from 209 applications of the ISK tools across 266 risk assessment areas worldwide, spanning over 21 years. The taxa were assigned de novo a priori invasion-status categorisations through a standardised verification workflow using authoritative databases and scientific literature. All verification steps and categorisation-evidence sources are recorded in the dataset, ensuring reproducibility of the evidence supporting each final taxon-level categorisation. This dataset provides a standardised reference resource to support calibration and comparative analyses of invasion risk screening applications.
The white-backed planthopper, Sogatella furcifera (Horváth, 1899), is a destructive rice pest that causes severe yield losses across Asia. Here, we manually curated a class of olfactory receptors: ionotropic receptors (IRs) genes using genomic and transcriptomic data, characterized tissue-specific expression profiles, and explored the functions of 3 IRco genes (SfurIR8a, SfurIR25a, SfurIR76b) via RNA interference (RNAi) and two-choice behavioral assays. We identified 14 IR genes and 3 IRco genes in S. furcifera, and the 3 IRco genes were predominantly expressed in the antennae of both sexes. qPCR verification confirmed that the expression of the 3 IRco genes was significantly downregulated after RNAi, with no obvious lethal effects on the nymphs. Behavioral assays showed that silencing any of the 3 IRco genes drastically impaired the host location ability of S. furcifera. Compared with the control, the relative response rate to rice seedlings dropped from positive to negative values at 2.5 and 3 h, the lack-of-response percentage increased notably, and the precise positioning rate on rice seedlings decreased significantly. These results demonstrate that SfurIR8a, SfurIR25a, and SfurIR76b are essential for olfactory perception and host-seeking behavior in S. furcifera. This study reveals the molecular basis of IRco-mediated olfactory processes in this pest and provides promising gene targets for developing novel olfactory-disruption strategies for rice pest management.
Systematic reviews and meta-analyses underpin clinical guidelines and health policy, yet their validity may be compromised by limited access to underlying datasets and associated analytical code.Reliance on incomplete or inconsistently reported summary statistics forces researchers to use imputation and unverifiable assumptions, which can distort effect estimates and mislead clinical decision-making.The consequences extend beyond methodology: flawed evidence synthesis can influence treatment recommendations, healthcare spending, and patient safety, as illustrated by historical cases such as hormone replacement therapy.Despite widespread data-sharing policies, compliance remains low, enforcement weak, and monitoring almost non-existent, with many datasets remaining unavailable or inaccessible.This Policy Forum argues for strengthening enforceable data-sharing mechanisms, including clearer enforcement and pragmatic verification approaches within editorial workflows.
Traditional cigarette smoking has declined markedly over recent decades due to increased health awareness, regulatory action, and public health campaigns. At the same time, alternative nicotine products, particularly e-cigarettes and oral nicotine pouches, have gained popularity, especially in the United States and Europe and among younger populations. While cigarettes remain the most widely used tobacco product globally, tobacco and nicotine consumption now spans a broad range of combustible and non-combustible products, with smokeless tobacco predominating in some regions. Definitions of a "smoker" vary significantly between clinical practice and insurance medicine. Clinical definitions are nuanced and health-focused, whereas insurance classifications are typically binary and time-bound, often defining a smoker as anyone who has used a nicotine-containing product within the past 12 months. The growing use of alternative nicotine products, alongside evolving cannabis legislation, has exposed limitations in traditional definitions when it comes to accurately capturing associated risk. The expansion of alternative nicotine products challenges traditional smoker classifications used in insurance underwriting. While combustible tobacco remains the highest risk exposure, growing evidence supports the existence of a heterogeneous medium risk group associated with nicotine, noncombustible products. Although generally lower risk than cigarettes, these products are not risk-free, particularly considering early cardiovascular and metabolic effects, dual or poly-use patterns, and growing uptake at younger ages. For insurers, this evolving landscape highlights the benefit of a more nuanced underwriting approach, including enhanced disclosure of product type and use patterns, improved verification strategies, and potential pricing differentiation that better aligns risk with exposure.
The TCP transcription factor family is a key regulator of plant growth, development and stress adaptation. The TCP gene family in the stress-tolerant cereal Chenopodium quinoa has not been systematically characterized. In this study, 20 non-redundant CqTCP members were identified genome-wide in quinoa and classified into three subfamilies (PCF, CIN and CYC/TB1) based on phylogenetic relationships. CqTCPs possess few introns, and their promoter regions are enriched in various cis-acting elements related to abiotic stress, hormone and light responses. They exhibit tissue-specific expression, and the expression of multiple members is significantly regulated by salt and drought stresses. Functional verification showed that CqTCP11 is localized in the nucleus. Heterologous overexpression of this gene in Arabidopsis thaliana significantly improved seed germination rate under salt and drought stresses, and enhanced the antioxidant capacity and osmotic adjustment ability of seedlings. This study systematically characterized the features of the quinoa TCP family and its functions in stress responses, clarified the key role of CqTCP11 in stress resistance during seed germination, and provided candidate genes and theoretical support for the genetic improvement of stress resistance in quinoa.
As the histopathology workforce continues to struggle and service demand continues to increase, it has become prudent to consider viable avenues to try to alleviate diagnostic workload burden. One such avenue is computer-based technologies (CBTs). Breast cancer (BC) is the most common malignant neoplasm in the United Kingdom and requires additional testing for estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor-2 (HER2) status at the time of histological diagnosis. This makes BC diagnostics a promising candidate for the application of an efficient CBT. However, for clinical acceptance, these technologies must prove that they work within a real-life diagnostic environment. We present a study protocol for a prospective clinical service evaluation aimed to validate a UK Conformity Assessed-marked CBT's ability to provide ER, PR, and HER2 results for invasive BCs from scanned hematoxylin and eosin-stained whole slide images. This protocol has been designed to use and mimic a preexisting digital pathology workflow within a National Health Service tertiary referral cancer center without disrupting normal patient care. Eligible cases are identified prospectively through the laboratory information management system, and their whole slide images are extracted from the clinical digital workflow. After verification of national data opt-out status and the exclusion of appropriate cases (N=400 analyzable cases), these cases are analyzed on a dedicated computer in parallel to the existing clinical workflow by a UK Conformity Assessed-marked deep learning-based CBT in a separate environment, providing results for ER, PR, and HER2 status. These results are compared to the ER, PR, and HER2 status reported on the corresponding pathology report. To evaluate the CBT's performance, a range of accepted concordance measures will be applied, including specificity, sensitivity, false-positive rate, false-negative rate, positive predictive value, and negative predictive value. Moreover, time stamps representing the duration of image analysis will also be collected. This study started in April 2025. There are no results to present, as this paper focuses on study design, and results have yet to be generated. As of March 2026, overall, 366 potentially analyzable cases have been collected. The anticipated end date of the study is May 2026 (400-case target). Results will be presented in a separate publication. This design assesses a CBT within a clinical environment while effectively eliminating any unwanted effects on patient care. This type of service evaluation provides a useful step to establish confidence in a CBT before trialing its effect on patient care. It also offers the opportunity to support interventional randomized controlled trials, health economic evaluations, and usability studies. This protocol will hopefully prove useful to others who wish to conduct a similar service evaluation at their own institution. DERR1-10.2196/76785.
Deep learning has become integral to medical imaging, but its tendency to memorize training data poses serious risks for patient privacy. Machine unlearning offers a potential remedy by revoking sensitive information, yet existing approaches face three key limitations: (1) they often achieve only output-level changes while residual feature representations remain; (2) they rely on batch retraining, making real-time removal of individual patient images infeasible; and (3) they lack rigorous metrics to verify forgetting in feature space. We propose AdaptForget, a domain-adaptive feature-level unlearning framework for privacy-preserving medical image analysis. AdaptForget introduces out-of-distribution (OOD) guidance to disentangle forgotten data from retained data in the feature manifold, supported by a theoretical feature-level unlearning bound. To prevent feature collapse, we design an OOD-driven feature-output disentanglement loss that enforces structured removal of forgotten data. To enable timely revocation, we formalize the task of single-entry forgetting, allowing immediate erasure of individual patient records. For objective auditing, we propose the isolation verification distance, a novel metric that quantifies feature separation and provides interpretable evidence of forgetting. Extensive experiments on four medical imaging benchmarks (histopathology, retinal fundus, dermatology, and OCT) as well as complementary healthcare record datasets demonstrate that AdaptForget achieves state-of-the-art privacy protection while preserving model utility. Code is publicly available at https://github.com/wangbrav/AdaptForget.
Synthesis route attribution enables accurate source tracing of chemical warfare agents and effective discrimination between distinct synthetic pathways. In the current study, we report for the first time the identification of trace process-related impurities derived from the synthesis of the vesicant ethyl sesquimustard, which is listed in Schedule 1.A.04 of the Chemical Weapons Convention (CWC). Using 1,2-ethanedithiol and 2-mercaptoethanol as precursors, 22 different synthetic routes were designed, and 88 batches of samples were produced through micro-synthesis. Gas chromatography-high resolution mass spectrometry (GC-HRMS) coupled with a non-targeted screening strategy was employed to analyze route-specific compounds. The obtained dataset was further used to train the orthogonal partial least squares discriminant analysis algorithm and generate a classification model consisting of eight sub-models. The verification results showed that the overall classification accuracy of the model was 21/22 (about 95%).
Rice is a fundamental food source for more than half of the global population, making stable yields and quality improvements vital for food security and sustainable agricultural development. Early infections of rice leaf diseases often exhibit subtle symptoms, while conventional control methods based on empirical judgment and routine pesticide application result in both yield losses and environmental pollution. A Multi-scale closed-loop tuning via spatial frequency collaborative sensitivity (MCCA-YOLO) model has been proposed in this paper with a multiscale closed-loop tuning and spatial frequency collaborative attention mechanism for the early detection and classification of rice crop diseases. MCCA-YOLO incorporates a closed-loop tuning compound network architecture that combines a dual-backbone feature extractor with a spatial frequency enhancement module to achieve system self-verification feedback, reducing transmission errors and enhancing the texture features of leaves. The framework implements a cross-scale weighted fusion and a deformable spatial hybrid attention enhanced bidirectional feature pyramid fusion network for dynamic feature adaptation, effectively accommodating the complex morphology of rice leaf lesions. By conducting comprehensive ablation studies and comparative experiments with existing techniques on the rice plant diseases v8 dataset, the proposed approach achieves a mean average precision (mAP) of 92.2%, outperforming well-established methods, while delivering superior precision (0.915) and recall (0.900). Extensive empirical validation of additional v9 and Rice Leaf Spot Disease (RLSD) datasets for rice plant diseases further demonstrates the model's outstanding performance.