New papers on Diagnostics & imaging
140 new papers on diagnostics & imaging in the last 7 days, within Medicine & health. These are the 50 Pipette rates most worth reading, with the main result in the authors' own words.
The best of the week
Large-scale esophageal cancer screening through noncontrast computed tomography and artificial intelligence
Here we developed the Esophageal AI-Guided malignant Lesion Evaluation (EAGLE) model to detect precancerous lesions and cancer from chest NC CT, a task historically considered impossible.
Peer-reviewed journalClaims a big stepReal-world useDifferentiating benign from malignant adnexal masses by biomarker-agnostic plasma proteomics using adaptive machine learning
Interpretation: A single plasma proteomic measurement interpreted using adaptive machine learning discriminates malignant from benign adnexal masses with good internal and external validation, and is complementary to ultrasound, warranting prospective evaluation as a panel-free diagnostic adjunct to imaging.
PreprintClaims a big stepReal-world useA world model simulates the latent dynamics of human health
Here we introduce HealthFlux, a pan-modal world model that learns the latent dynamics of health from 5,647 features across eleven data domains, spanning clinical records, blood tests, genetics, proteomics, metabolomics and MRI, in 502,166 UK Biobank participants.
PreprintBold claims, read criticallyReal-world usePopulation-scale plasma proteomics reveals pre-diagnostic biomarkers and shared pathways of cancer risk across 25 cancers
We identified 1,732 protein-cancer associations (FDR P < 0.05) across 25 cancer endpoints in multivariable-adjusted models, including 60 proteins associated with overall cancer risk.
PreprintClaims a big stepReal-world useGut microbiome alterations predictive of cancer risk
Finally, we demonstrate that the extent of microbiome alterations can effectively stratify both the patients with diagnosed cancer and individuals at elevated cancer risk.
PreprintReal-world useMultiscale mapping of venous remodelling in idiopathic pulmonary fibrosis
Collectively, our multimodal, multiscale approach connects previously unresolved three-dimensional venous architecture to its molecular endothelial correlate, establishing venous enlargement as a prognostically significant and clinically relevant feature of IPF.
PreprintReal-world useTransducin: an open-source pipeline recovering SNOMED-CT coded measurements from the undocumented Optopol .OPT and Zeiss Cirrus private-tag formats as DICOM Structured Reports
Transducin provides the first publicly documented specification of the Optopol .OPT format and the first open-source multivendor pipeline generating SNOMED-CT coded DICOM Structured Reports from both Optopol REVO and Zeiss Cirrus devices, closing a gap explicitly confirmed by both manufacturers' own documentation.
PreprintBold claims, read criticallyClaims a big stepReal-world useCode availableSelf-aggregated photo-heterostructures enhancing fluorescence for cancer diagnosis and treatment monitoring
Leveraging synergistic evanescent field coupling and self-aggregation of photo-heterostructures, PAEF sensor streamlines efficient isolation (~90%) and ultrasensitive profiling of tumor extracellular vesicles (EVs) in a single assay, achieving sensitivity <10 EVs/μL (104-fold better than ELISA) without additional amplification, outperforming most state-of-the-art EV fluorescence assays.
Peer-reviewed journalBold claims, read criticallyReal-world useGenMC: Real-Time Generative Monte Carlo Surrogate for Quantitative Photoacoustic Imaging
Trained on MC-generated synthetic datasets, GenMC produces high-fidelity fluence maps in under 30 ms per frame, a four-orders-of-magnitude speed-up over conventional MC simulation, and reaches peak signal-to-noise ratios of up to 36.24 dB in vivo, outperforming UNet and Pix2Pix baselines.
PreprintReal-world useMOLAR: MRI-based Opportunistic Localization and Recognition of teeth
We present MOLAR (MRI-based Opportunistic Localization And Recognition of teeth), an open pipeline that isolates the dental region (the "bite box") from T1w brain MRI and counts the teeth, without additional scanning or commercial software.
PreprintReal-world useAlphaGenome Atlas: in silico mutagenesis of the entire human genome improves prioritization and interpretation of non-coding variants
Thus, AlphaGenome Atlas improves the prioritization and molecular interpretation of non-coding variants with genetic and clinical significance.
PreprintReal-world useDualStream-MTCA: A Hybrid Deep Learning Architecture for Joint Prediction and Early Detection of Sepsis and Heart Failure in Adult Intensive Care
Significance: To our knowledge, this is the first model built specifically to predict both conditions jointly, with calibration, decision-curve and external evaluation.
PreprintReal-world useInvestigating 24-Hour Urine Expected Ranges of Non-Kidney Stone Formers: an EAU Endourology Research Group Systematic Review and Meta-Analysis
Take Home MessageThis study establishes expected 24-hour urine values in a non-stone-forming population and provides empirical support for the derivation of reference ranges for several common urine analytes.
PreprintReal-world useDefining the Genetic and Phenotypic Landscape of Primary Immune Regulatory Disorders Using the ClinGen Validation Framework
CAPSULE SUMMARYUsing the ClinGen framework, this study describes the first global initiative to validate gene-disease relationships for primary immune regulatory disorders, identifying 49 relationships and three phenotypic clusters, providing a resource for genetic testing and diagnosis.
PreprintReal-world useInvestigating 24-Hour Urine Expected Ranges of Non-Kidney Stone Formers: an EAU Endourology Research Group Systematic Review and Meta-Analysis
Take Home MessageThis study establishes expected 24-hour urine values in a non-stone-forming population and provides empirical support for the derivation of reference ranges for several common urine analytes.
PreprintReal-world useHigh-fidelity genotyping and therapeutic oligonucleotides analysis by liquid-phase ion trap coupled with mass spectrometry
By eliminating laborious pretreatment, the LPIT-ESI-MS system establishes a robust platform for high-fidelity genotyping and the rigorous analysis of therapeutic oligonucleotides, accelerating both clinical diagnostics and biopharmaceutical quality control.
Peer-reviewed journalBold claims, read criticallyReal-world useArtificial Intelligence-Enhanced Electrocardiography for Detection and Prediction of Hypertrophic Cardiomyopathy across Monogenic and Polygenic Susceptibility
ConclusionsAI-ECG detects the HCM phenotype at baseline in sarcomere variant carriers, predicts development of HCM in G+/P- individuals, and complements PRS in the general population, supporting AI-ECG as a scalable tool to detect HCM and guide surveillance in individuals with monogenic or polygenic susceptibility.
PreprintReal-world useEffect of Screening with Multicancer Early-Detection Test on Late-Stage Cancer Diagnosis
CONCLUSIONS: In this trial, we did not observe a lower incidence rate of stage III or IV cancers across 12 prespecified cancers after three screening rounds with multicancer early-detection testing plus usual care than with usual care alone.
Peer-reviewed journalReal-world useMind the Gaps: Validating species presence under sparse genomic recovery in clinical metagenomics
Evaluated across 643 long-read metagenomic samples, including 280 clinical samples, MAVI reproduced independently supported organism calls, resolved sparse low-coverage ambiguities and exposed workflow-level collapse of stochastic dispersion, without relying on heuristic coverage thresholds.
PreprintReal-world useScalable Computational Phenomics of Nuclear Morphology
NucXplore thus unifies interpretable representation, high-performance computing, and dynamical modeling for cellular phenotyping from routine histology.
PreprintReal-world useErrors, Hallucinations, and Clinical Impact of General-Purpose Multimodal Large Language Models in Histopathology
Conclusion: Diagnostic accuracy alone substantially underestimates the safety limitations of general-purpose LLMs in histopathology.
PreprintReal-world useRedefining Non-Invasive Post-Transplant Surveillance: A Bayesian Meta-Analysis and Decision Curve Framework for Donor-Derived Cell-Free DNA in Heart Transplantation
Our tri-state, prevalence-aware framework integrating rule-out, indeterminate, and rule-in zones with selective repeat testing, more accurately reflects biomarker behavior and yields greater net benefit than any single cutoff because these anchors are pooled, population-level estimates rather than universal constants, programs should adopt this architecture and calibrate their own high-sensitivity rule-out and high-specificity rule-in thresholds to their local assay and population.
PreprintReal-world usePerspective independence, more than personas, drives LLM teams - and where they reverse
On the external benchmark the team beat the single call on both pre-specified recall endpoints (top-3 +3.0 points, p = 0.0079; top-5 +3.9, p = 3.8 x 10^-5;); a factorial attributes the gain to independent generation plus moderated synthesis, not the specialist roles.
PreprintReal-world useClinical trajectories and genetic architecture across the neurological-psychiatric boundary
Clinical trajectories and genetic architecture thus draw the same diagnostic boundary and break it in the same places.
PreprintPsychological factors contribute more to chronic low back pain than spine pathology: An LLM-based analysis of radiology reports
Psychological comorbidities show the strongest associations with cLBP, while anatomical burden is mostly driven by demographics, reinforcing the psychosocial nature of cLBP.
PreprintReal-world useTowards Sustainable Magnetic Resonance Imaging: Insights from long-term, high-resolution energy recordings across an entire scanner fleet
Lastly, linear regression analysis showed that energy consumption was more strongly associated with scan duration than with average power, suggesting that scan duration may be an important factor influencing MRI energy consumption. ...
PreprintReal-world useKaiser Permanente National Cross-Vendor Validation of Mammography Artificial Intelligence Computer-Aided Diagnosis Algorithms in a US-Representative Population
Conclusions AI CAD algorithms differed meaningfully in performance on the same large, diverse screening examination cohort, and an open-source academic model performed comparably to two of the three commercial products.
PreprintReal-world useDeep learning-based assessment of ulcerative colitis activity from full-length endoscopic videos with spatial characterisation and histological correlation
ConclusionsAI-based analysis enables automated assessment of UC activity from full-length endoscopic videos, including remission detection and estimation of histological healing.
PreprintReal-world useUltrasound-enhanced exosome secretion and antibody-free SERS profiling for Alzheimer’s disease via iPSC-derived cortical organoids
Ultrasound-based stimulation and label-free exosome profiling reveal genotype-specific pathologies, providing a precision-diagnostic framework for Alzheimer's disease.
Peer-reviewed journalBold claims, read criticallyReal-world useAdvancing sarcoma diagnostics with expanded DNA methylation-based classification
ConclusionAdding new sarcoma types and expanding tumor sample numbers in each methylation class in the new sarcoma classifier decisively increased the number of diagnostic predictions and improved match with histologic evaluation.
PreprintReal-world useScalable Causal-Interpretable Machine Learning for Cancer Prescreening Using Electronic Health Records
To address this limitation, we propose UPEBNL, a scalable framework for causal-interpretable cancer prescreening based on parallel CBN learning.
PreprintReal-world usePrivacy-Aware Distillation of Large Language Models for Enhanced Multimorbidity Scoring
We introduce and evaluate a privacy-preserving knowledge distillation framework for LLM-based clinical modeling, using multimorbidity scoring as a healthcare task.
PreprintReal-world useDiagnostic Accuracy of WHO IMCI Visual Inspection for Neonatal Jaundice: A Prospective Multi-Country Cohort
Visual inspection had a sensitivity of 16%, specificity of 96%, positive percent agreement of 56%, and negative percent agreement of 86% at 3-5 days postnatal age, pooled across all sites and gestational ages, versus transcutaneous bilirubin defined by NICE nomograms.
PreprintReal-world useAutomated language impairment screening in acute stroke using connected speech
An ensemble of embedding-based classifiers distinguished patients with and without LI with 90% balanced accuracy (79% sensitivity, 100% specificity), outperforming independent embedding and discrete-linguistic-based classifiers, showing distinct LLMs contributed complementary information.
PreprintReal-world useQuantifying the genetic separability of disease subtypes
We introduce GenSep, a statistical framework that estimates the oracle case-case AUC attainable if the true additive genetic values of both subtypes were known, and the proportion of this genetic separation recovered by subtype-specific PRSs.
PreprintPerformance and safety of a multi-cancer early detection test: the PATHFINDER 2 study
These results provide insights into performance and safety of the MCED test in an intended-use population.
Peer-reviewed journalReal-world useValidity Gates Expose Limits of Longitudinal Multimodal Parkinson Disease Prediction
The framework reveals state support, stale or missing evidence, and unjustified prognostic, uncertainty or transport claims.
PreprintClinician Perspectives on Automated Facial Analysis for Pain Assessment: A Cross-Sectional Survey of Physicians, Advanced Practice Providers, and Nurses
These findings suggest clinicians view automated facial-analysis systems as potentially valuable adjuncts for pain monitoring, particularly in communication-impaired populations, while emphasizing the importance of accuracy, fairness, explainability, and workflow integration during system development and implementation.
PreprintReal-world useEasy to readGenomic Research as a Critical Pathway to Diagnosis for Individuals with Rare Disease
RGP demonstrates that a research-based, patient-driven remote sequencing study can effectively bridge a diagnostic gap for rare disease families underserved in current clinical care.
PreprintReal-world useLipoprotein remodelling and elevated serum apolipoprotein B distinguish adults with haemophagocytic lymphohistiocytosis from sepsis
Validation of a potential HLH-specific lipid signature using ELISA technology identified ApoB (p=0.007) and ApoB:ApoA1 (p=0.0002) as significantly different between HLH and HLH-mimics, including sepsis, supporting their utility as clinically translational biomarkers to distinguish HLH from a key mimic.
PreprintReal-world useAlpha-Mannosidosis: Quantitative Oligosaccharide Biomarker Analysis in Dried Blood Spots for Diagnosis, Newborn Screening, and Treatment Monitoring
ConclusionWe present a simple and rapid method for quantitative analysis of the specific AM oligosaccharide biomarker Man2GlcNAc, which is remarkably stable compared to enzymatic assays.
PreprintReal-world useContinuous Diagnostic Indices from BanffNET Automated Lesion Scores for Kidney Transplant Pathology: Development and Evaluation of a Diagnostic Prediction Model
ResultsThe BanffNET TCMR/TI Index, AMR/MVI Index and Activity Index showed excellent discrimination of Banff TCMR, Banff MVI and Banff Any Diagnosis respectively, with validation AUCs ranging from 0.80 to 0.92.
PreprintReal-world useGenomic Research as a Critical Pathway to Diagnosis for Individuals with Rare Disease
RGP demonstrates that a research-based, patient-driven remote sequencing study can effectively bridge a diagnostic gap for rare disease families underserved in current clinical care.
PreprintReal-world useEcho intensity is not an accurate surrogate marker of intramuscular fat and fibrosis: A systematic review with meta-analyses
Conclusions: Echo intensity is weakly-to-moderately associated with intramuscular fat, and this varied substantially across studies and muscles, whereas evidence for intramuscular fibrous tissue is limited and uncertain.
PreprintReal-world useAssessment of Fairness and Bias of an Image-based Surgical Site Infection Detection AI Model
ConclusionsThe model behaved fairly across age, gender, race, and skin color, however geographical differences persisted.
PreprintReal-world useBenchmarking open-source automated thigh muscle MRI segmentation algorithms
We address this gap by evaluating eight tools, six of which run independently, and the other two of which are meant to be auxiliary algorithms once muscles have been somehow initially labeled.
PreprintReal-world useWhole-Blood Transcriptome Profiling Revealed Gene Expression Signatures of Tourette Disorder
These results suggest that peripheral blood transcriptional alterations mirror biologically processes relevant to TD pathophysiology and provide a framework for the development of transcriptome-based biomarkers and functional studies of tic disorders.
PreprintMuscle stability deficits and musculoskeletal complaints in football players: Adaptive Force Ratio outperforms conventional strength parameters--a cross-sectional study with preliminary follow-up
ConclusionThe findings suggest that muscle stability assessment outperforms conventional strength parameters in discriminating players with and without complaints, with preliminary follow-up data providing tentative support for predictive value.
PreprintReal-world useSoft Temporal Scoring Using a Foundation Model: Optimal Frame Selection for Improved ONSD Measurement in Ultrasound Videos
To enable the use of ONSD in low resource settings by medical personnel of all levels, we present a sparsely supervised AI frame-work that scores each frame of an ONSD ultrasound video and selects the optimal frame for ONSD assessment.
PreprintReal-world usePerformance of a multi-cancer early detection test in the randomized controlled NHS-Galleri trial
Across rounds, specificity ranged from 99.50% to 99.60%; episode sensitivity ranged from 26.7% to 37.2% for all cancers and from 47.6% to 63.4% for 12 prespecified types; and CSO accuracy ranged from 91.1% to 93.6%.
Peer-reviewed journalReal-world use