The scale of the problem motivating this reversal is stark. Despite decades of progress, cardiovascular disease remains the leading cause of death worldwide, and the attrition rate in cardiometabolic drug development remains punishingly high. Many agents that showed dramatic effects in animal models of atherosclerosis, heart failure or hypertension produced negligible benefit, or unacceptable toxicity, when they reached human trials. The authors of the new work contend that this is not a failure of chemistry or clinical execution but a failure of translational assumptions: animal models, however refined, cannot fully reproduce human cardiovascular physiology, human genetic variation, or the decades-long natural history of diseases that unfold over a lifetime in people.

The human-first proposal inverts the classical sequence. Rather than beginning with a hypothesis generated in a model system, investigators begin with comprehensive molecular measurements drawn directly from human populations — genomes, transcriptomes, proteomes, metabolomes, epigenomes and, increasingly, single-cell profiles of human cardiovascular tissue. These layers of ‘omics’ data are then integrated computationally to reveal which genes, pathways and cell types are genuinely perturbed in human disease. Only after such human-grounded signals are identified does the work move toward experimental validation, drug targeting and therapeutic design. In effect, the clinic and the population study become the source of hypotheses, and the laboratory becomes the venue for testing them.

Each omics layer contributes a different kind of evidence. Genome-wide association studies have already delivered hundreds of genetic loci linked to myocardial infarction, atrial fibrillation, cardiomyopathy and related traits, but most of these associations point to non-coding regions of the genome whose function is unknown. Transcriptomics converts those static genetic signals into dynamic statements about which genes are actively expressed in diseased hearts and vessels. Proteomics captures the actual effector molecules of biology — the proteins that drugs must engage — while metabolomics offers a real-time readout of cellular chemistry and environmental influence, including diet, microbiome activity and medication effects. Epigenomic profiling explains how risk is encoded not only in DNA sequence but in the regulation of gene activity across a lifetime.

The real power, the authors argue, emerges from integration. No single omics layer is sufficient, because each is noisy, incomplete and context-dependent, but convergent evidence across layers can separate true disease biology from statistical artifact. If a genetic variant associated with coronary artery disease falls in a regulatory region that is active specifically in human vascular smooth muscle cells, and if the gene it controls shows altered expression and altered protein abundance in diseased tissue, and if metabolites in the same pathway track with disease severity, the case for that pathway’s causal involvement becomes far stronger than any single measurement could provide. Multi-omics integration is thus a way of triangulating on human disease mechanisms with a confidence that single-technology studies rarely achieve.

Recent technological advances have made this vision practical in a way it was not a decade ago. Single-cell RNA sequencing can now resolve the cellular composition of human heart tissue cell by cell, revealing disease-specific states in cardiomyocytes, fibroblasts, endothelial cells and immune cells that bulk measurements average away. Spatial transcriptomics preserves the anatomical context of gene expression, showing not just which cells are involved but where they sit within the architecture of a plaque or an infarcted wall. Long-read sequencing is closing gaps in genome annotation. Mass spectrometry has pushed proteomics toward near-comprehensive coverage of the human proteome. Meanwhile, large biobanks linked to electronic health records — containing hundreds of thousands of participants with genetic data and longitudinal clinical outcomes — provide the population-scale foundation on which human-first discovery depends.

Human pluripotent stem cell technologies supply the experimental counterpart to these population resources. Induced pluripotent stem cell-derived cardiomyocytes and vascular cells allow investigators to model an individual’s genetic background in a dish, testing how specific risk variants alter cell behaviour under controlled conditions. When combined with CRISPR-based gene editing, these systems permit precise causal tests: take a human variant identified through population omics, introduce or correct it in human cells, and observe the consequences. This closes the loop of the reversed pipeline, in which human observation generates the hypothesis and human-derived experimental systems validate it before any animal model or clinical trial is considered.

The therapeutic implications are already visible in recent cardiovascular successes that followed this logic in reverse. PCSK9 inhibitors emerged from human genetics — people with loss-of-function variants in the gene had low LDL cholesterol and reduced heart attack risk — rather than from animal screening. The discovery that elevated lipoprotein(a) is causally linked to aortic stenosis and coronary disease came from human cohort genetics, and drugs targeting that protein are now in late-stage trials. Angiotensin-related pathways, inflammation-driven residual risk identified through human trial data with canakinumab, and genetic validation of targets for heart failure all illustrate the same principle: targets grounded in human evidence have repeatedly outperformed targets chosen on model-organism grounds alone.

The authors are careful to note the substantial challenges that remain. Multi-omics datasets are expensive, and most existing data come from populations of European ancestry, raising urgent questions about equitable generalisation. Integrating heterogeneous data types requires statistical and computational methods that are still maturing, and correlational signals at population scale do not automatically establish causation. Human tissue, particularly healthy and early-disease cardiac tissue, is difficult to obtain, and much of what is available comes from end-stage disease or organ donors, limiting the view of how cardiovascular disease begins. Privacy and consent frameworks for deeply characterised human data remain an active area of policy development. The human-first approach, in other words, demands infrastructure — biobanks, computational platforms, tissue networks and diverse cohorts — as much as it demands new biology.

Even so, the strategic argument is compelling and timely. The pharmaceutical industry has invested heavily in human genetics and real-world data precisely because traditional pipelines have underdelivered in cardiometabolic disease. Academic consortia assembling multi-omics atlases of the human heart and vasculature are generating public resources that any laboratory can interrogate. Artificial intelligence and machine learning, trained on integrated human datasets, are beginning to predict gene function, prioritize drug targets and identify patient subgroups that classical diagnostics lumped together. The human-first framework unifies these developments into a coherent discovery philosophy: measure human biology comprehensively, infer mechanism from the data, validate in human-derived systems, and only then build the therapeutic. If the approach continues to deliver, the pipeline that once flowed from bench to bedside may increasingly be understood as flowing the other way — with the patient, and the population, at its source.

Subject of Research: A human-first multi-omics strategy for cardiovascular disease discovery

Article Title: Reversing the pipeline: a ‘human-first’ multi-omics approach to cardiovascular discovery

Article References: Reversing the pipeline: a ‘human-first’ multi-omics approach to cardiovascular discovery. (n.d.). https://doi.org/10.1038/s12276-026-01835-8

Image Credits: AI Generated

DOI: 10.1038/s12276-026-01835-8

Keywords: cardiovascular disease, multi-omics, human genetics, drug discovery, genomics, proteomics, metabolomics, single-cell sequencing, precision medicine, translational research, Reversing, pipeline

The study, published in Nature Genetics, describes a clinically annotated, population-scale single-cell atlas built from bone marrow samples of 341 patients, spanning the full continuum of the disease from precursor conditions through newly diagnosed myeloma to relapsed and refractory disease after multiple lines of therapy. Using single-cell RNA sequencing enriched for plasma cells and the surrounding CD45-positive immune compartment, the researchers captured the transcriptomes of tens of thousands of individual cells, allowing them to dissect the malignant compartment cell by cell rather than averaging signals across bulk tumor tissue, which has historically masked the very heterogeneity that drives treatment failure.

Technically, the effort was formidable. Samples were processed using the MARS-seq platform and an updated version, MARS-seq2.0, with rigorous quality control on mitochondrial content, unique molecular identifier counts and detected genes. Cells were annotated through a computational pipeline combining scVI latent-space integration, uniform manifold approximation and projection embeddings, and inferCNV-based inference of copy number alterations to distinguish malignant plasma cells from their normal counterparts. The team then applied non-negative matrix factorization to decompose malignant gene expression into recurrent transcriptional programs, validated for stability through hundreds of bootstrapped iterations. This dual-layer analytical strategy allowed the investigators to separate two largely independent axes of tumor biology: what kind of myeloma a patient has, and how fast that myeloma is growing.

The first axis yielded five recurrent malignant transcriptional archetypes, designated MM1 through MM5, each corresponding to a stable pattern of gene expression anchored in distinct biological pathways. These archetypes align with known myeloma biology, including immunoglobulin heavy chain translocations such as t(11;14) with its cyclin D and BCL-2 dependencies, t(4;14) with NSD2 dysregulation, MAF and MAFB associated programs, and features reflecting unfolded protein response burden and bone marrow niche interactions. Crucially, the archetypes were not merely descriptive. They correlated with genomic features, therapeutic sensitivity patterns and clinical outcomes, and the team demonstrated that the classification could be ported to independent bulk RNA datasets, including the Blueprint cohort and the Multiple Myeloma Research Foundation’s CoMMpass cohort of treatment-naive patients, confirming that the single-cell-defined signatures retain prognostic power even when measured on standard clinical platforms.

The second axis, orthogonal to the archetypes, is a proliferative program. By scoring single-cell proliferation signatures and characterizing plasmablastic cells, the rapidly dividing precursors of antibody-secreting plasma cells, the researchers quantified the fraction of malignant cells actively cycling in each patient’s marrow. Proliferation has long been recognized as a poor prognostic marker in myeloma, measured historically by crude methods such as plasma cell labeling indices. The new work refines this concept at single-cell resolution, showing that the proportion of proliferating malignant plasma cells stratifies patients within every archetype, revealing intra-archetypal heterogeneity that earlier bulk approaches could not detect. In relapsed and refractory patients, higher proliferative fractions predicted shorter progression-free survival, and the effect persisted in multivariate Cox regression models adjusting for cytogenetic risk, age and prior treatment lines.

Combining the two axes produced an improved risk stratifier that outperformed existing molecular subtyping schemes. Patients could be placed into joint archetype-proliferation subgroups with meaningfully distinct progression-free and overall survival, and the framework added prognostic information beyond standard clinical variables including high-risk cytogenetics and chromosome 1p deletion. The validation in CoMMpass, one of the largest longitudinally followed myeloma cohorts in the world, demonstrated robustness and portability across sequencing platforms, an essential prerequisite for clinical translation. In principle, a myeloma patient’s tumor could one day be assigned to an archetype and proliferation state from a routine biopsy, guiding intensity of upfront therapy and informing decisions about transplantation, novel agents or early escalation.

Perhaps the most clinically electrifying result, however, came from the atlas’s use as a target-discovery engine. The team built a computational pipeline that ranked every protein-coding gene by a composite score integrating malignant enrichment, specificity for malignant plasma cells relative to normal plasma cells, and restriction across healthy tissues, the latter being critical to minimize off-tumor toxicity for any future immunotherapy. This screen surfaced FCRL2, an Fc receptor-like molecule with established roles in B cell biology, as a surface target expressed by malignant plasma cells but largely restricted to the B cell lineage elsewhere in the body. The atlas approach meant the researchers could verify not just that myeloma cells express FCRL2, but that expression is preserved across archetypes and proliferation states, addressing the antigen escape problem that plagues current myeloma immunotherapies.

The translational proof followed swiftly. The researchers engineered chimeric antigen receptor T cells directed against FCRL2 and tested them against myeloma cell lines with varying levels of target expression. In vitro, FCRL2-redirected CAR-T cells killed antigen-positive myeloma cells in an antigen-specific manner, with luciferase-based co-culture assays showing progressive suppression of tumor cell growth compared to non-transduced controls, and detailed immunophenotyping confirming proper CAR expression and memory-phenotype differentiation of the engineered cells. In mouse models, FCRL2-targeted CAR-T cells conferred a significant survival benefit. Given that existing myeloma immunotherapies targeting BCMA and GPRC5D eventually fail through antigen loss and relapse, a third lineage-restricted target backed by a genome-wide, single-cell-verified prioritization pipeline offers a credible path toward combination or sequential immunotherapy strategies.

For patients, the near-term significance is prognostic rather than therapeutic: an archetype and proliferation score could refine risk assessment well before relapse, when treatment decisions matter most. For the field, the study establishes a template for how population-scale single-cell atlases can move beyond description into actionable classification and target nomination. All of the underlying data, including the full scRNA-seq dataset deposited in the Gene Expression Omnibus and the analysis code released openly by the Amit lab, are publicly available, ensuring that other groups can interrogate, extend and challenge the framework. As single-cell sequencing costs fall and clinical grade assays mature, the line between research atlases and routine diagnostics grows thinner, and this myeloma atlas may be remembered as a turning point where that line was crossed for a historically intractable cancer.

Subject of Research: Single-cell transcriptomic atlas of multiple myeloma defining malignant archetypes, proliferative states, and the immunotherapy target FCRL2

Article Title: A single-cell atlas of multiple myeloma defines malignant archetypes and proliferative states

Article References: Zada, M., Kurilovich, A., Shapira, N., Wang, S.-Y., Sharet-Eshed, R., Kfir-Erenfeld, S., Schlossberg, M., Zorde, E., Asherie, N., Gur, C., Chalan, P., Shalita, R., Ben Yehuda, M., Zwicky, P., von Locquenghien, M., Ingelfinger, F., Mazuz, K., David, E., Gurevich-Shapiro, A., … Amit, I. (2026). A single-cell atlas of multiple myeloma defines malignant archetypes and proliferative states. Nature Genetics, 58(9), 2254-2269. https://doi.org/10.1038/s41588-026-02725-5

Image Credits: AI Generated

DOI: 10.1038/s41588-026-02725-5

Keywords: multiple myeloma, single-cell RNA sequencing, transcriptional archetypes, proliferation, FCRL2, CAR-T cell therapy, risk stratification, plasma cells, Nature Genetics, precision medicine, CoMMpass cohort, target discovery

The scale of the problem the guidelines address is stark. Approximately 15 to 25 percent of colorectal cancer patients already have liver metastases when their cancer is first diagnosed, and a further 15 to 25 percent develop such metastases after curative resection of the primary tumor. Critically, 80 to 90 percent of these liver metastases are initially unresectable with curative intent. Left untreated, patients with colorectal liver metastases have a median survival of only 6.9 months, and the five-year survival rate for unresectable disease falls below 5 percent. By contrast, patients who achieve complete resection of their liver metastases, or reach a status of no evidence of disease, experience median survival of 35 to 60 months and five-year overall survival rates of 40 to 57 percent. This dramatic gap between the untreated and the successfully treated is precisely what the new guidelines are designed to close.

On the diagnostic front, the 2025 edition codifies a tiered imaging strategy. For every newly diagnosed colorectal cancer patient, liver ultrasound and contrast-enhanced abdominal computed tomography are recommended as routine screening tools, alongside serum tumor markers such as carcinoembryonic antigen and CA19-9. When ultrasound or CT findings are suspicious but inconclusive, the guidelines endorse supplemental testing with serum alpha-fetoprotein, contrast-enhanced liver ultrasound, and contrast-enhanced liver MRI. Liver-specific contrast-enhanced MRI earns a particularly strong endorsement because it demonstrates higher accuracy for detecting lesions smaller than one centimeter, a critical threshold since tiny metastases are the ones most easily missed and most consequential for surgical planning. PET-CT and PET-MRI are explicitly not recommended for routine use, reserved instead for selected clinical scenarios where additional information is genuinely needed.

The guidelines also formalize a rigorous surveillance program after radical resection of the primary colorectal tumor. Patients should undergo history-taking, physical examination, liver ultrasound and tumor marker testing every three to six months for the first two years, then every six months until five years, and annually thereafter. For patients with stage II and III disease, annual contrast-enhanced CT of the chest, abdomen and pelvis is recommended for three to five years. Electronic colonoscopy should be performed within one year of surgery, with repeat examination at three years and then every five years if no abnormalities are found. After patients achieve no evidence of disease status for liver metastases, the surveillance intensifies further, with tumor markers checked every three months for two years and abdominal imaging every three months during the same period, reflecting the high risk of recurrence in this population.

Perhaps the most consequential section of the 2025 update concerns molecular profiling. The guidelines recommend mismatch repair and microsatellite instability testing for all colorectal cancer patients, a recommendation grounded in the transformative role of immune checkpoint inhibitors in tumors with deficient mismatch repair or high microsatellite instability. RAS genotyping, covering exons 2, 3 and 4 of both KRAS and NRAS, is recommended for all patients with colorectal liver metastases, both for its prognostic value and its role in predicting response to anti-EGFR therapy. BRAF V600E mutation testing is endorsed as both a prognostic indicator and a guide to treatment selection, while HER2 testing is recommended for metastatic patients to inform post-progression decisions. The guidelines also acknowledge emerging biomarkers including tumor mutational burden, POLE and POLD1 mutations, NTRK fusions, RET rearrangements and c-MET alterations, and highlight circulating tumor DNA-based minimal residual disease assessment as a promising but not yet fully validated tool.

On prevention, the guidelines are unequivocal that standardized radical treatment of the primary colorectal cancer remains the most effective strategy for reducing liver metastasis risk. For colon cancer, this means complete mesocolic excision with adequate proximal and distal margins and removal of the associated mesentery and lymphatic drainage. For rectal cancer, total mesorectal excision is mandatory for mid and lower tumors. The update also embraces neoadjuvant approaches aimed at eradicating micrometastases invisible to imaging. For patients with deficient mismatch repair or microsatellite instability-high rectal cancer, immune checkpoint inhibitors can achieve such favorable outcomes that many patients may avoid surgery altogether through a watch-and-wait approach. For proficient mismatch repair tumors staged T3 or higher, or with positive lymph nodes, neoadjuvant radiotherapy, chemoradiotherapy or chemotherapy is recommended, and total neoadjuvant treatment is endorsed as an option that increases complete response rates and facilitates organ preservation.

The multidisciplinary team model receives a Grade A recommendation for all patients with colorectal liver metastases. The guidelines specify that the team should include colorectal or gastrointestinal surgeons, hepatobiliary surgeons, medical oncologists, radiation oncologists, interventional radiologists, diagnostic radiologists, ultrasound specialists and pathologists. The documented advantages of this approach include more accurate molecular profiling, more precise staging, fewer treatment delays, more individualized treatment planning, improved coordination, enhanced quality of life, better survival outcomes and superior cost-effectiveness. The MDT framework is then used to stratify patients into distinct therapeutic pathways: those with initially resectable metastases, those with potentially convertible disease, and those whose metastases will never be resectable, each with tailored goals ranging from cure to disease control.

Surgical treatment remains the gold standard for cure, and the 2025 edition reflects how far the boundaries of resectability have expanded. Surgical decisions are no longer restricted by tumor size, number or location alone. The guidelines now support resection with margins smaller than one centimeter, resection of hepatic pedicle lymph node metastases, and resection of resectable extrahepatic disease including pulmonary and peritoneal metastases. For patients with insufficient future liver remnant volume, the guidelines describe a sophisticated arsenal of techniques: portal vein embolization to induce compensatory hypertrophy, the ALPPS procedure for rapid remnant growth, liver venous deprivation as a less invasive alternative, and selective internal radiation therapy with yttrium-90 microspheres, which can shrink tumors while simultaneously stimulating growth of the untouched liver lobe. Liver transplantation combined with systemic therapy is cautiously endorsed for selected patients with liver-limited disease that cannot achieve no evidence of disease status despite multimodal treatment, an approach supported by the recent TransMet randomized trial showing significantly improved overall survival.

For unresectable disease, the guidelines detail an escalating therapeutic ladder. First-line chemotherapy combines fluoropyrimidines with oxaliplatin and/or irinotecan, with molecular targeted agents such as bevacizumab or cetuximab added to increase conversion to resectability. Triplet FOLFOXIRI regimens are endorsed for fit patients when targeted therapy is contraindicated. For tumors with deficient mismatch repair, immune checkpoint inhibitors such as pembrolizumab or the nivolumab-ipilimumab combination are the preferred first-line approach, having demonstrated significantly improved disease control and conversion rates compared with chemotherapy. Maintenance therapy with low-toxicity regimens after induction, sequential switching between FOLFOX and FOLFIRI upon progression, and later-line options including trifluridine/tipiracil, regorafenib and fruquintinib are all codified. Local therapies including radiofrequency ablation for lesions under three centimeters, microwave ablation for lesions under five centimeters, and stereotactic body radiotherapy for oligometastatic disease round out the comprehensive toolkit, ensuring that even patients whose metastases can never be removed can still expect meaningful prolongation of survival and quality of life.

Subject of Research: Clinical guidelines for the diagnosis and comprehensive treatment of colorectal cancer liver metastases

Article Title: Chinese guidelines for the diagnosis and comprehensive treatment of colorectal liver metastasis (2025 Edition)

Article References: Fan, J., Gu, J., Jia, B., Li, J., Qin, X., Wang, X., Xu, J., Xu, R., Ye, Y., Zhang, S., Zhang, Z., & Chinese College of Surgeons, Chinese Medical Doctor Association (CMDA); Gastrointestinal Surgery Group, Chinese Society of Surgery, Chinese Medical Association (CMA); Colorectal Surgery Group, Chinese Society of Surgery, CMA; Colorectal Cancer Professional Committee, Chinese Anti-Cancer Association; Colorectal Cancer Professional Committee, CMDA; Colorectal Cancer Expert Committee, Chinese Society of Clinical Oncology; Colorectal Surgeon Committee, Chinese College of Surgeons, CMDA; Metastasis Research Com (2026). Chinese guidelines for the diagnosis and comprehensive treatment of colorectal liver metastasis (2025 Edition). Clinical Cancer Bulletin, 5(1), Article 10. https://doi.org/10.1007/s44272-026-00062-6

Image Credits: AI Generated

DOI: 10.1007/s44272-026-00062-6

Keywords: colorectal cancer, liver metastases, clinical guidelines, immunotherapy, hepatic surgery, molecular profiling, chemotherapy, multidisciplinary team, precision medicine, radiofrequency ablation, liver transplantation, biomarkers

SALI is defined as an acute, secondary hepatic dysfunction that arises during sepsis and is observed in roughly 34 to 46 percent of patients. At the bedside it announces itself through abnormalities in standard liver tests, which can follow hepatocellular, cholestatic, or mixed patterns and range from modest enzyme elevations to profound liver failure. Crucially, the authors insist that SALI be treated as an operational clinical syndrome rather than the signature of a single underlying mechanism. Abnormal liver biochemistry in a septic patient may reflect inflammatory injury, cholestasis, hypoxic hepatitis caused by insufficient oxygen delivery, right-sided cardiac congestion, drug toxicity, or pre-existing conditions such as metabolic dysfunction-associated steatotic liver disease and occult fibrosis. The question, they argue, is not whether SALI has uniform biology, because it does not, but whether routinely available data can flag the patients most likely to deteriorate.

On the mechanistic side, one of the most striking recent findings concerns the gut. In a mouse model of sepsis, Murao and colleagues identified a pathway in which gut-primed neutrophils drive hepatic injury. Gut intraepithelial lymphocytes interact with neutrophils through the molecule CD112, facilitating the formation of neutrophil extracellular traps, the web-like DNA structures that neutrophils eject to ensnare pathogens. These primed neutrophils migrate through the portal vein into the liver, where they release their traps and activate Kupffer cells, the liver’s resident macrophages, triggering the secretion of interleukin-6 and tumor necrosis factor-alpha. Notably, portal vein neutrophils from septic mice produced significantly more neutrophil extracellular traps and induced greater Kupffer cell activation than systemic neutrophils, an effect that disappeared entirely in mice lacking PAD4, the enzyme essential for trap formation. The implication is provocative: the gut does not merely spill inflammatory mediators into the portal circulation, it actively educates immune cells that then inflict damage on distant organs.

Although the authors caution that translation to human disease requires care, the concept has clear clinical resonance. The liver receives most of its blood supply from the portal circulation and is therefore continuously bathed in gut-derived inflammatory signals. During sepsis, disruption of the intestinal barrier allows bacterial translocation and the spillover of pathogen-associated molecular patterns, which activate hepatic Toll-like receptors. Supporting this mechanistic bridge, human data from Czaikoski and colleagues have shown that neutrophil extracellular traps accumulate in organ tissue during experimental and clinical sepsis and correlate with damage. Together, these findings nominate trap formation and downstream Kupffer cell activation as candidate precision-medicine targets in SALI.

A second biological pillar concerns the loss of hepatic immune tolerance. In health, the liver is a strikingly tolerant organ, and Kupffer cells orchestrate that tolerance through antigen clearance and the induction of regulatory T cells. Recent work shows that during hepatic inflammation this tolerogenic phenotype collapses: Kupffer cells lose their signature tolerance markers, and antigen presentation shifts to infiltrating monocyte-derived macrophages. Activated Kupffer cells then recruit further immune cells to the liver, amplifying injury. Evidence from viral hepatitis research suggests that the transition from tolerance to inflammation involves dysregulation of inhibitory pathways, such as the Toll-like receptor pathway inhibitor SHIP, that normally restrain receptor signaling and keep Kupffer cells quiescent. Hepatic stellate cells, likewise, depend on inhibitory signals to remain dormant; when stimulated by microbial products or damage-associated molecular patterns, they produce extracellular matrix proteins and profibrogenic cytokines, and their contractile activation can raise sinusoidal resistance and portal pressure. Toll-like receptor 4-dependent crosstalk between Kupffer cells and stellate cells converts inflammatory signals into profibrogenic activation.

Within sepsis specifically, the inflammatory polarization of Kupffer cells toward the M1 phenotype has emerged as a hallmark of SALI. Extracellular cold-inducible RNA-binding protein, a damage-associated molecular pattern released during stress, promotes this M1 polarization through Toll-like receptor 4 signaling, driving overproduction of inflammatory cytokines. In mouse sepsis models the ratio of M1 to M2 Kupffer cells rises sharply, indicating a decisive shift toward proinflammatory function, and this polarization is not merely a byproduct of inflammation but an active driver of hepatocyte injury through reactive oxygen species, cytokines, and the recruitment of more neutrophils. In parallel, regulated forms of cell death, including apoptosis, necroptosis, pyroptosis, and ferroptosis, appear to contribute to hepatocyte dysfunction. These converging mechanisms point toward the restoration of hepatic immune tolerance as a promising future therapeutic strategy, though no SALI-targeted therapy has yet been established.

It is on the clinical side that the commentary delivers its most immediately practical message. In a retrospective cohort study spanning two large intensive care cohorts, Palmowski and colleagues examined how well routine biomarkers could stratify mortality risk among patients meeting operational criteria for SALI, defined as sepsis-associated liver-test abnormalities within seven days of sepsis onset, excluding pre-existing chronic liver disease. The criteria included alanine aminotransferase at five or more times the upper limit of normal, alkaline phosphatase at twice the upper limit, or elevated bilirubin combined with enzyme elevations. Their central finding was that the De Ritis ratio, the simple ratio of aspartate to alanine aminotransferase, outperformed both the conventional R-factor and alanine aminotransferase alone in predicting thirty-day mortality. A ratio of one or below indicated low risk, values between one and two indicated intermediate risk, and values of two or above flagged the highest risk, a pattern consistent across infection sources and admission types.

The authors of the commentary are careful to frame these strata correctly. The De Ritis ratio is not a liver-specific diagnostic marker or a mechanistic endotype, and elevated aspartate aminotransferase can also signal hypoxic hepatitis, shock, right-sided congestion, systemic inflammation, chronic kidney disease, alcohol-related injury, or cardiometabolic comorbidity. Its pragmatic value lies in risk enrichment among patients who already meet operational SALI criteria, complementing rather than replacing SOFA-bilirubin scoring. Interpreted alongside the SOFA score, lactate, hemodynamic status, cardiac context, and comorbidities, a rising ratio should trigger a structured reassessment: is infection control optimized, are hemodynamics adequate, is the lactate trajectory improving, is there occult congestion or biliary obstruction, are hepatotoxic drugs on board, and does the patient carry underlying fibrosis risk? In this framework, routine liver tests define the dominant biochemical injury pattern, whether hepatocellular, cholestatic, or mixed, and link prediction to the prevention of further hepatic and systemic deterioration and of iatrogenic harm.

The translational pathway forward, the authors suggest, will require prospective studies testing whether serial liver tests, the De Ritis ratio, SOFA scores, lactate, hemodynamic data, and immune readouts such as monocyte HLA-DR expression or ex vivo monocyte cytokine responses can identify reproducible SALI trajectories and clinically actionable phenotypes. Preclinical work has already nominated an unusually rich set of therapeutic targets, including neutrophil extracellular trap formation, Kupffer cell polarization, inflammasome activation, ferroptosis, necroptosis, and the restoration of hepatic immune tolerance. Until such approaches are validated, however, current clinical utility remains deliberately pragmatic: recognize SALI early, classify the dominant biochemical pattern, stratify mortality risk with the De Ritis ratio, hunt actively for reversible contributors, and intensify surveillance in high-risk patients.

For the authors, the larger significance of this work lies in adding an organ-specific decision layer to the 2026 Surviving Sepsis Campaign framework. New-onset liver-test abnormalities in septic adults, they argue, should no longer be treated as incidental laboratory noise. When detected, the humble ratio of two transaminases, a calculation older than modern critical care and available in every hospital on earth, may identify the patients who need intensified monitoring and protection from modifiable second hits, while the expanding immunobiology of the gut-liver axis steadily maps the routes toward genuine precision medicine for a complication that affects as many as one in two patients with sepsis.

Subject of Research: Sepsis-associated liver injury: immunobiology and bedside risk stratification with routine liver tests

Article Title: Sepsis-associated liver injury: from liver-test risk signals to immunobiology-guided precision medicine

Article References: Katsounas, A., Tsochatzis, E., & Rello, J. (2026). Sepsis-associated liver injury: from liver-test risk signals to immunobiology-guided precision medicine. Intensive Care Medicine. https://doi.org/10.1007/s00134-026-08593-1

Image Credits: AI Generated

DOI: 10.1007/s00134-026-08593-1

Keywords: sepsis-associated liver injury, De Ritis ratio, Kupffer cells, neutrophil extracellular traps, gut-liver crosstalk, hepatic immune tolerance, risk stratification, intensive care, Toll-like receptors, precision medicine, Sepsis-associated, liver

The scale of the problem is easy to underestimate. High-throughput sequencing and mass spectrometry now allow laboratories to measure the expression of every gene in the human genome across hundreds of samples at once. A dataset might record 10,000 genes while including fewer than a hundred patients. This imbalance creates what statisticians call the curse of dimensionality: the number of possible feature subsets grows as two to the power of n, so for a dataset with 10,000 genes the search space is astronomically larger than anything a brute-force enumeration could ever cover. Worse, adding features does not reliably improve a model. Extra genes can introduce redundancy and noise, causing classifiers to overfit the training data while performing poorly on patients they have never seen. Running times grow as well, because computational complexity rises steadily with the number of features examined.

Existing feature selection strategies fall into three broad families, each with well-known trade-offs. Filtering methods, which rank genes using statistical measures such as mutual information, are fast and scalable but blind to the interactions between features. Wrapper methods, which evaluate subsets by feeding them to a classifier, capture those nonlinear relationships but at a punishing computational cost. Embedded methods such as LASSO regression and tree-based models select features during training, but none of these approaches ask the deeper biological question: which genes actually work together, and which modules of co-regulated genes drive the disease being studied? The new framework was built precisely to fill that gap, treating the biology of gene cooperation as the starting point rather than an afterthought.

The first stage of CJWGA relies on Weighted Gene Co-expression Network Analysis, or WGCNA, a technique originally proposed by Zhang and Horvath that constructs a weighted network linking genes whose expression levels rise and fall together across samples. Genes are not loners; they participate in biological processes through intricate webs of interaction, and WGCNA captures those relationships from a systems perspective. The pipeline begins with Z-score normalization of expression values, followed by a Pearson correlation matrix that is then transformed into a weighted adjacency matrix using a soft thresholding exponent chosen so the network follows a scale-free topology, in which a few highly connected hub genes dominate while most genes have few connections. A Topological Overlap Measure, which accounts for shared neighbors, is then fed into hierarchical clustering to identify modules of functionally related genes, with module eigengenes derived by principal component analysis.

But the authors recognized that relying on a single eigengene per module throws away too much information. A lone principal component cannot reflect the diversity of functions within a module, and it can be biased by outlier expression patterns. Their answer is a preprocessing step called IMGCNet, which uses conditional mutual information to rank genes within each module by how much extra information they carry about the disease label, given the eigengene is already known. A higher conditional mutual information value means a gene retains a strong dependency on the phenotype even after controlling for what the module representative already explains. Larger modules are allowed to retain more genes and smaller modules fewer, through a descending allocation rule that preserves the biological representativeness of each module without letting small, specialized groups flood the analysis.

The second stage hands the modules to an enhanced version of NSGA-II, the classic multi-objective genetic algorithm that balances competing goals by evolving a population of candidate solutions toward a Pareto front. Here the two objectives are minimizing the number of selected genes and maximizing classification accuracy, formalized with a binary decision vector over features and evaluated with a K-Nearest Neighbor classifier on a 70-30 train-test split. Crucially, the researchers designed a hierarchical encoding scheme: the first layer of each chromosome encodes which modules are selected, and the second layer encodes which genes within each chosen module survive. This two-layer structure preserves the biological meaning of the modularization rather than flattening it back into a flat string of bits.

The heart of the contribution lies in two new operators. The Combined Information Entropy Crossover Operator, or CIECO, computes a joint mutual information score across the genes selected in both parents, those selected in neither, and those selected in only one. The resulting value, transformed through a probabilistic function, decides whether crossover should prune doubly-selected genes, promote single-selected ones, or hold steady. When the score is positive, unselected genes carry little information and conservative trimming is favored; when it is negative, redundant double selections are removed and a few unselected genes are introduced to seek greater information content. The Joint Adaptive Mutation Operator, or JAMO, then fine-tunes individual genes using an adaptive rate that depends on iteration progress, the proportion of genes already selected in the module, and the ratio of joint mutual information between selected and unselected genes, with an exponent parameter that keeps the balance under control.

The experimental evaluation covered eight publicly available gene expression datasets, including Brain_Tumor1, Brain_Tumor2, CNS, Leukemia, Leukemia1, Leukemia2, Lung_Cancer and Prostate_Tumor, all with more than 5,000 features and sample sizes between 50 and 203. Against three specialist algorithms, FQEISS, WMOSS and WQEISS, CJWGA achieved the lowest classification error on the CNS, Leukemia, Leukemia1, Leukemia2 and Prostate_Tumor datasets, while selecting the smallest feature subsets on six of the eight datasets. On Inverted Generational Distance, a standard measure of how well a computed Pareto front approximates the true optimum, CJWGA scored zero, meaning perfect overlap with the reference front, on six datasets. Ablation experiments confirmed that both new operators contribute measurably: removing the crossover operator or the mutation operator individually degraded either accuracy or subset compactness. Parameter sweeps established that a crossover proportion of 0.2 and a mutation exponent of 3 offered the most robust results. Because joint mutual information is computed only within compact modules rather than across the entire feature space, the framework retains reasonable scalability even as dataset dimensionality grows.

The implications reach beyond benchmark tables. A feature selection method that respects gene co-expression relationships can point clinicians toward biologically meaningful biomarkers rather than statistical artifacts, a prerequisite for precision medicine where a compact, interpretable gene panel must support diagnosis and treatment decisions. The authors caution, however, that systematic biological interpretation of the selected genes remains future work, and they note that the framework could be extended to dimensionality reduction problems well outside genomics. As high-throughput biology continues to generate data faster than medicine can absorb it, tools like CJWGA suggest that the path forward lies in algorithms that speak both languages fluently: the language of information theory and the language of biological networks. The study is available as open access, supported by the National Natural Science Foundation of China and several provincial research programs.

Subject of Research: Gene feature selection in high-dimensional medical gene expression data using co-expression networks and genetic algorithms

Article Title: Advancing Gene Feature Selection: A Synergistic Approach with Co-expression Networks and Genetic Algorithms

Article References: Wangy, Z., Ding, W., Zhang, J., Heidari, A. A., Wang, M., & Chen, H. (2026). Advancing Gene Feature Selection: A Synergistic Approach with Co-expression Networks and Genetic Algorithms. Journal of Advanced Research. https://doi.org/10.1016/j.jare.2026.08.064

Image Credits: AI Generated

DOI: 10.1016/j.jare.2026.08.064

Keywords: gene feature selection, co-expression networks, WGCNA, genetic algorithms, multi-objective optimization, NSGA-II, mutual information, bioinformatics, cancer classification, precision medicine, high-dimensional data, biomarker discovery

The research, led by Xiaoyu Xu and Zaixu Cui of the Chinese Institute for Brain Research in Beijing, together with colleagues at Peking University Sixth Hospital, Stanford University and other institutions, took aim at a fundamental problem in pediatric psychiatry. ADHD affects a substantial share of school-age children worldwide, yet no validated biomarkers exist for tracking symptom trajectories or guiding treatment selection in youth. Clinicians must largely rely on trial and error when choosing between medications, and families often wait weeks or months to learn whether a prescription is working. The team asked whether the developing brain’s structural connections could supply the missing prognostic information.

Their approach rested on the idea of normative growth charts, familiar from pediatrics, where a child’s height and weight are compared against population curves to flag unusual development. The researchers applied the same logic to the brain’s wiring diagram. Using diffusion magnetic resonance imaging, which traces the bundles of nerve fibers that connect distant brain regions, they built normative age-related trajectories of white-matter structural connectivity from a large longitudinal developmental cohort comprising 6,687 scans from typically developing youths and 1,114 scans from youths with ADHD. They then quantified, for every individual with ADHD, how much each connection deviated from the trajectory expected for that person’s age. An independent replication cohort of 355 typically developing and 477 ADHD participants allowed the team to confirm that their findings were not an artifact of a single dataset.

The first major result was that youths with ADHD showed pronounced deviations in structural connectivity, and those deviations were not distributed randomly across the brain. Instead, they clustered overwhelmingly at the association end of what neuroscientists call the sensorimotor–association connectional axis, a gradient that runs from regions devoted to basic sensation and movement to the higher-order association cortices that support attention, executive control and self-regulation. These association networks are precisely the circuits implicated in ADHD symptoms, and they are also the slowest-maturing parts of the brain, continuing to develop well into adolescence and early adulthood. The findings echo an influential earlier report that ADHD involves a delay in cortical maturation, but extend it from the gray matter of the cortex to the white-matter highways that link cortical networks together.

The study then probed how these deviations evolve. A subset of higher-order association connections showed ADHD-specific reductions in deviation with age, changes that went beyond typical developmental patterns and could not be explained by ordinary maturation. Critically, these converging trajectories statistically mediated the age-related decline in ADHD symptoms observed across development, suggesting a mechanistic account of why many children appear to grow out of the disorder. When the researchers followed individuals across two years, they found that within-person decreases in deviation tracked symptom improvement over the same interval, linking individual brain maturation to individual clinical course in a way that cross-sectional group comparisons never could.

The most clinically provocative findings concerned treatment. Using data from youths treated with either atomoxetine or methylphenidate, the two most widely prescribed ADHD medications, the team tested whether baseline structural connectivity deviations could predict response to a 12-week course of treatment. The answer was strikingly specific. Deviations predicted response to atomoxetine, a norepinephrine reuptake inhibitor whose effects are concentrated in prefrontal association circuits, but not to methylphenidate, a stimulant whose primary mechanism centers on dopamine signaling in striatal reward pathways. Follow-up imaging further revealed that treatment itself was associated with reductions in deviation, hinting that effective medication may nudge wayward white-matter development back toward the normative curve. Together, these results identify structural connectivity deviation as a developmental biomarker with prognostic relevance, supporting precision care through symptom monitoring and treatment stratification.

Technically, the study represents a synthesis of several modern neuroimaging and statistical methods. Diffusion MRI data were preprocessed and reconstructed with tools including QSIPrep and MRtrix3, with anatomically constrained tractography and multi-tissue constrained spherical deconvolution used to estimate the strength of each white-matter connection. Cortical parcellations derived from functional connectivity provided a common map of brain regions organized along the sensorimotor–association axis. Normative trajectories were modeled with generalized additive models for location, scale and shape, the same statistical machinery used to construct World Health Organization child growth standards, and deviation was quantified as the distance between an individual’s connectivity and the population curve. Longitudinal scanner effects were harmonized with longitudinal ComBat, and mediation analysis, mixed-effects models and structural equation modeling tied the deviations to symptom change.

The scale of the evidence base deserves emphasis. Prior studies of white matter in ADHD have often compared groups of a few dozen participants and produced inconsistent results, a pattern documented in meta-analyses of more than one hundred diffusion imaging studies. By anchoring deviation estimates in a normative cohort of thousands and replicating them in an independent cohort, the researchers sidestepped the case-control designs that have long limited interpretation. The normative modeling framework they used was developed specifically to understand heterogeneity in clinical cohorts, recognizing that each patient’s brain tells an individual story that average group differences obscure. The method also parallels the construction of lifespan brain charts published in recent years, extending that approach from brain volume to the connectome and from typically developing populations to clinical prediction.

The implications reach beyond ADHD. The sensorimotor–association axis has emerged in recent work as a general organizing principle of cortical development and function, and deviations along this axis have been linked to autism and other neurodevelopmental conditions. If individual deviation from normative development can forecast symptoms and treatment response in ADHD, the same logic may apply to other childhood psychiatric disorders that similarly lack biomarkers. The researchers have released their analysis code publicly, and the ABCD dataset underlying much of the work is available to qualified investigators, which should accelerate independent validation. Limitations remain: the medication analyses were observational, deviations were measured from diffusion imaging with inherent biases in tractography, and clinical deployment would require streamlined acquisition and standardized norms across scanner platforms.

Still, the study sketches a plausible near future in which a child newly diagnosed with ADHD undergoes a brief MRI session, their white-matter wiring is compared against a growth chart of the developing connectome, and the resulting deviation profile informs whether atomoxetine is likely to succeed, how their symptoms are likely to evolve over adolescence, and whether their brain is already converging toward the normative trajectory. For a disorder that has been defined almost entirely by behavior since it was first described more than a century ago, the prospect of a measurable, mechanistic, individualized biomarker drawn from the brain’s structural wiring marks a genuine turning point, one that could move pediatric psychiatry from reactive adjustment of prescriptions toward genuinely predictive, precision-guided care.

Subject of Research: Developmental deviations of association-network structural connectivity as predictive biomarkers of ADHD symptoms and treatment response in youth

Article Title: Developmental deviations of association-network structural connectivity in youths with ADHD predict symptom and treatment outcomes

Article References: Developmental deviations of association-network structural connectivity in youths with ADHD predict symptom and treatment outcomes. (n.d.). https://doi.org/10.1038/s41551-026-01779-4

Image Credits: AI Generated

DOI: 10.1038/s41551-026-01779-4

Keywords: ADHD, structural connectivity, white matter, diffusion MRI, normative modeling, brain development, association networks, atomoxetine, methylphenidate, biomarkers, precision medicine, pediatric psychiatry

The central problem the researchers set out to solve is one of incomplete explanation. Pathogenic or likely pathogenic variants in sarcomere-encoding genes, the category the team abbreviates as SARC-HCM-P/LP, account for only about one-third of hypertrophic cardiomyopathy cases. The remaining majority of patients have no detectable single-gene culprit, and even among confirmed carriers, the condition displays notoriously variable penetrance: some people harboring a dangerous variant develop severe disease early in life, while others reach old age with hearts that function normally. This uneven expressivity has long hinted that something beyond the rare variant itself, most plausibly the cumulative influence of thousands of common genetic variants scattered across the genome, helps determine who actually becomes ill and how sick they become.

Earlier attempts to capture that polygenic contribution ran into a stubborn equity problem. Existing polygenic risk scores for hypertrophic cardiomyopathy were built almost exclusively from genome-wide association studies of European-ancestry populations, and when applied to individuals of African, East Asian, Hispanic, or other ancestries, their predictive performance deteriorated sharply. This limitation mirrors a broader and well-documented weakness of genomics: because most large genetic datasets over-represent people of European descent, risk scores trained on them often fail the very populations that already bear disproportionate burdens of cardiovascular disease and face greater barriers to genetic diagnosis. The new study was designed from the ground up to confront that disparity rather than treat it as an afterthought.

To construct the score, the team drew on genome-wide association summary statistics from three complementary sources: the BioBank Japan, the Million Veteran Program, and a meta-analysis of seven European-ancestry cohorts. By combining association signals from Japanese, multiethnic American, and European datasets, the investigators built a score intended to capture disease-relevant variants across the genetic ancestry continuum rather than within a single population. The statistical machinery behind such scores involves weighting millions of single nucleotide polymorphisms according to their measured association with disease risk, then summing those weighted contributions for each individual to yield a single number representing inherited polygenic susceptibility. Methods of this kind, including Bayesian shrinkage approaches refined in recent years, allow researchers to distill a usable clinical signal from noisy, genome-scale data while guarding against overfitting.

The validation stage took place in a genuinely diverse national cohort: participants in the United States-based All of Us Research Program, one of the largest and most ancestrally diverse biomedical datasets ever assembled. The results were unambiguous. Individuals whose polygenic score placed them in the top quintile of the population had a 2.11-fold increased risk of developing hypertrophic cardiomyopathy compared with the rest of the population. That effect size, for a common-variant composite score alone, is clinically meaningful and rivals the discriminatory power that polygenic scores have achieved for more common conditions such as coronary artery disease. The score also improved overall risk stratification and showed trends toward improved ancestry-specific prediction, suggesting that the multiancestry training strategy partially, if not perfectly, mitigated the performance cliff that plagues European-derived scores.

The most striking finding, however, emerged when the researchers layered the polygenic score on top of the rare-variant picture. Among carriers of pathogenic or likely pathogenic sarcomere variants, individuals in the highest polygenic score quintile faced nearly a seventy-fold higher risk of actually developing hypertrophic cardiomyopathy compared with low-scoring counterparts. This is precisely the kind of result that cardiologists and genetic counselors have been waiting for. Variable penetrance among variant carriers has made counseling agonizingly uncertain: telling a young person they carry a disease-causing mutation without being able to say whether that mutation will ever manifest is of limited clinical use. A polygenic measure that helps distinguish the carriers likely to develop disease from those likely to remain unaffected converts a static genetic diagnosis into a dynamically graded risk estimate.

Beyond prediction of who develops disease, the score carried prognostic weight. Among individuals already diagnosed with hypertrophic cardiomyopathy, a higher polygenic score was associated with adverse cardiovascular outcomes, indicating that the same aggregate burden of common variants that raises disease susceptibility also shapes disease severity and trajectory. This suggests that polygenic information could eventually inform not only screening decisions but also surveillance intensity and management priorities for diagnosed patients. Extended analyses reinforced the pattern: among carriers of predicted deleterious variants, disease prevalence rose steadily across polygenic score quintiles, with the highest quintile showing roughly 2.4-fold higher penetrance than the lowest, and hazard ratios climbing in a graded fashion as score category increased.

The technical infrastructure supporting the study reflects the maturing standards of the polygenic risk score field. The investigators performed careful quality control, addressed the statistical pitfall known as Winner’s Curse that inflates effect estimates in discovery samples, and evaluated performance using measures including odds ratios per standard deviation of score change and area under the receiver operating characteristic curve across ancestry groups. In a notable commitment to transparency and reproducibility, the team made its analysis code publicly available on GitHub and deposited the underlying genome-wide association summary statistics in the GWAS Catalog, enabling other researchers to replicate, extend, or adapt the score for their own populations. Individual-level participant data remain accessible through the All of Us Researcher Workbench under its data use agreements.

The clinical implications reach well beyond hypertrophic cardiomopathy itself. Hypertrophic cardiomyopathy affects roughly one in several hundred people and remains a leading cause of sudden cardiac death in young athletes, yet many cases go undiagnosed until a catastrophic event occurs. A validated, ancestry-fair risk score could be deployed to flag individuals who warrant echocardiographic screening, genetic testing, or closer longitudinal follow-up, potentially catching disease before it strikes. For gene-positive family members of affected patients, combining sarcomere variant status with a polygenic score could personalize the schedule of cardiac imaging and the timing of preventive interventions, including the newer generation of cardiac myosin inhibitors that have transformed pharmacologic management of the disease.

The authors and observers of the field alike caution that scores of this kind are not yet ready to replace clinical judgment or guideline-based testing. The ancestry-specific performance gains, while encouraging, remained trends rather than definitive demonstrations, and further validation in independent, prospective cohorts will be essential before polygenic information enters routine cardiology practice. Questions about how best to communicate a seventy-fold relative risk to a worried variant carrier, and how insurers and employers might use such information, also demand careful attention. Still, the study marks a turning point: it demonstrates that the variable penetrance puzzle of hypertrophic cardiomyopathy is, at least in part, quantitatively solvable, and that the solution can be built to serve populations of all ancestries rather than a genetically privileged few. As multiancestry genomic resources continue to grow, the integration of rare variant status and polygenic background promises to become a standard pillar of precision cardiovascular medicine, reshaping how inherited heart disease is predicted, counseled, and ultimately prevented.

Subject of Research: Development and validation of a multiancestry polygenic risk score for hypertrophic cardiomyopathy risk stratification

Article Title: A multiancestry polygenic risk score improves stratification in patients with hypertrophic cardiomyopathy

Article References: Bal, H. S., Pampana, A., Nayak, A., Gaonkar, M., Patel, S., Yerabolu, K., Vekariya, N., Patel, N., Kalra, R., Li, P., Arora, G., & Arora, P. (2026). A multiancestry polygenic risk score improves stratification in patients with hypertrophic cardiomyopathy. Nature Cardiovascular Research, 5(9), 891-903. https://doi.org/10.1038/s44161-026-00866-8

Image Credits: AI Generated

DOI: 10.1038/s44161-026-00866-8

Keywords: hypertrophic cardiomyopathy, polygenic risk score, multi-ancestry genomics, sarcomere variants, variable penetrance, All of Us Research Program, cardiovascular genetics, risk stratification, precision medicine, genome-wide association study, genetic counseling, sudden cardiac death

The article, the product of a consensus workshop at an international meeting of the Organization for the Study of Sex Differences in May 2019 and a subsequent expert meeting in Ottawa sponsored by the Institute of Gender and Health of the Canadian Institutes of Health Research, brings together a strikingly interdisciplinary authorship. Neuroscientists, physiologists, historians of science, internists, nephrologists, nurses, social workers, epidemiologists, and implementation scientists, led by co-first authors Gillian Einstein of the University of Toronto and Louise Pilote of McGill University, along with senior author Cara Tannenbaum of the Université de Montréal, collectively propose a name and a vision for what they call Sex and Gender Science. They describe it as a transdisciplinary, integrative, bio-social-medical field that combines diverse methods to investigate how sex and gender dynamically interact to shape disease outcomes, illness experiences, and the very conduct of science itself.

The conceptual core of the proposal rests on the claim that sex and gender are interactional and deeply interwoven. Within sex, the authors situate hormonal, genetic, physiological, and morphological processes and characteristics, from chromosomal complements and gonadal steroid fluctuations to organ-level physiology and morphology. Within gender, they include roles, relations, norms, institutional imperatives, and identities, along with their downstream consequences for behavior, environment, access to resources, and exposure to stress. Crucially, the authors argue, these domains are not merely additive; they operate on each other. Gendered expectations influence hormonal and physiological responses, while biological traits shape how institutions and individuals respond to a person. In this framing, gender literally gets under the skin, while sex expresses itself inescapably through social worlds.

Concrete clinical examples illustrate the stakes. Acute coronary syndrome presents differently across patients in ways that reflect both vascular biology and gendered patterns of symptom recognition, care-seeking, and triage in emergency departments. In Alzheimer’s disease, two-thirds of affected individuals are women, and research discussed in the article points to interactions between reproductive histories, such as surgical menopause, stress physiology involving corticotropin-releasing factor, and gendered lifecourse exposures, rather than a single biological cause. In nephrology, sex differences in kidney physiology intersect with gendered differences in access to dialysis, transplantation, and medication adherence. Even occupational health reflects the interplay, as shown by analyses of personal protective equipment designed around male body norms during the COVID-19 pandemic, where ill-fitting gear compromised the safety of women in frontline roles. The authors also cite emerging concerns about automatic gender recognition technologies and gendered patterns of cannabis use disorder as domains where simplistic binary assumptions fail patients and obscure mechanisms.

Methodologically, Sex and Gender Science is defined by its refusal to choose between reductionist and constructionist approaches. The authors assume complexity, question established stereotypes and binaries, and support evolving methods that can capture interactional effects over time. That means experimental work in animal models attentive to sex as a biological variable can be paired with qualitative and quantitative human studies of gendered environments; intersectional epidemiology can be integrated with endocrine and genetic measurement; and implementation science can test whether sex- and gender-informed findings actually change clinical practice. The transdisciplinary ambition is deliberate: rather than creating another silo, the field aims to weave together the humanities, social sciences, and biomedical sciences, drawing on decades of feminist scholarship, gender studies, and basic biology that have until now developed along separate tracks.

The historical context for this synthesis is instructive. Biomedical research long treated male bodies as the default experimental subject, from clinical trials to preclinical animal studies, on the grounds that hormonal cycles made female subjects inconveniently variable. Landmark policy changes in the United States, Canada, and Europe over the past three decades required the inclusion of women in clinical research and the accounting for sex as a biological variable in grant applications. Simultaneously, scholars in gender and health demonstrated that social positions assigned by gender produce measurable health inequities, from differential exposure to violence and occupational hazards to differences in healthcare access and treatment. What was missing, the authors argue, was a systematic framework linking these two streams, so that researchers could ask not simply whether men and women differ, but how biological and social variables interact to produce particular outcomes in particular individuals.

The consensus process behind the article gives its claims unusual weight. By convening basic scientists who study sexually dimorphic neural circuits and stress pathways alongside clinicians, historians, and social scientists, the meeting produced a shared vocabulary and a shared commitment. The resulting highlights, as stated by the authors, include forging new ways of knowing about wellness, health, and disease; assuming complexity and questioning binaries; providing a more comprehensive picture of human health when sex and gender are studied together; and starting from the particularity of individuals with the ultimate goal of improving the well-being of all. That final point is significant politically as well as scientifically: the field explicitly encompasses transgender and gender-diverse populations, recognizing that gender identity itself interacts with biology in ways that mainstream medicine has historically ignored or pathologized.

For clinical practice, the implications are immediate. Diagnostic criteria, drug dosing, device design, and screening schedules built on averages derived from unrepresentative populations can be reexamined through the interactional lens that Sex and Gender Science provides. A patient is never just a sex and never just a gender; each person embodies a unique configuration of hormonal milieu, genetic background, organ physiology, gendered labor and caregiving burdens, institutional exposure, and identity. Precision medicine that ignores either domain is, on this argument, not precise at all. The authors point to the growth of national organizations, such as the Canadian Organization for Sex and Gender, and to curriculum reform in medical schools as evidence that institutional infrastructure is beginning to catch up with the science.

The article also acknowledges open questions and future directions. Measuring gender as rigorously as sex is measured remains an active methodological challenge, with validated instruments still evolving. Distinguishing interaction effects from confounding in observational data requires statistical sophistication and large, diverse cohorts. And the field must navigate political controversies surrounding sex and gender discourse without retreating from empirical evidence. The authors frame these challenges as generative rather than paralyzing, arguing that the very friction of integrating methods across disciplines produces better science, sharper questions, and more honest uncertainty. They call for funding structures, training programs, and publication venues that reward transdisciplinary work rather than penalizing it in favor of narrow, single-domain studies.

If Sex and Gender Science succeeds on its own terms, the payoff could be a medicine that finally matches the complexity of the patients it serves: one in which a heart attack, a dementia diagnosis, a kidney transplant, or a workplace injury is understood through the full entanglement of chromosomes, hormones, organs, identities, relationships, and institutions that shaped it. The article’s message to researchers is blunt in its simplicity. It is not either sex or gender; it is both, together. After decades of separation, the biological and the social are being brought back into a single scientific frame, and the authors argue that the health of everyone stands to gain from what this emerging field discovers next.

Subject of Research: Sex and Gender Science, an emerging transdisciplinary field integrating biological sex and social gender in health research

Article Title: Sex & Gender Science: integrating the social with the biological

Article References: Einstein, G., Pilote, L., De Vries, G., Richardson, S., Forger, N., Perović, M., Ahmed, S., Bauer, G., Graham, I. D., Greaves, L., Klinge, I., Logie, C. H., McKay, D., McMurtry, M. S., Oliffe, J. L., & Tannenbaum, C. (2026). Sex & Gender Science: integrating the social with the biological. Biology of Sex Differences. https://doi.org/10.1186/s13293-026-00977-8

Image Credits: AI Generated

DOI: 10.1186/s13293-026-00977-8

Keywords: sex and gender science, biological sex, gender and health, transdisciplinary research, Biology of Sex Differences, health inequities, sex differences, gender identity, precision medicine, Alzheimer's disease, cardiovascular disease, Canadian Institutes of Health Research

Cellular senescence is a state in which cells stop dividing but do not die. First described by Leonard Hayflick and Paul Moorhead in 1961, when they observed that cultured human fibroblasts could divide only a finite number of times, senescence was long regarded as a simple curiosum of the tissue culture dish. Over the following decades, however, it became clear that senescent cells accumulate in tissues throughout the body as organisms age, and that they are not merely passive bystanders. They secrete a potent cocktail of inflammatory cytokines, growth factors, and matrix-remodeling enzymes known collectively as the senescence-associated secretory phenotype, or SASP, which can remodel tissue microenvironments, drive chronic inflammation, and disturb the function of neighboring healthy cells.

That inflammatory shadow is now implicated in an astonishing range of age-related conditions, from osteoarthritis and atherosclerosis to neurodegeneration and frailty. Animal studies have added weight to the argument: in mice, the genetic or pharmacological removal of senescent cells can delay tissue dysfunction and extend health span. A growing class of drugs, termed senolytics, selectively eliminates senescent cells, while senomorphics aim to suppress their harmful secretions without killing them. Early-phase clinical trials are underway in several disease areas, and the prospect of intervening directly in the biology of aging has moved from the fringe to the center of geroscience.

Yet as the Coimbra participants emphasized, the field’s momentum is not matched by clinical readiness. One of the most fundamental problems is definitional. Senescence is not a single entity. Cells can enter the state through telomere shortening, DNA damage, oncogene activation, mitochondrial dysfunction, or other stresses, and the resulting senescent cells differ profoundly depending on the trigger, the tissue of origin, and the duration of the state. Some senescent cells are transient and beneficial, orchestrating wound healing, embryonic development, and tissue repair, while others persist for months or years and become quietly destructive. The roadmap therefore calls for a functional classification of senescent states, a systematic taxonomy that would distinguish which senescent cells are doing harm, which are doing good, and in what contexts.

Without such a classification, the field risks repeating mistakes that have hampered other areas of medicine. The single most cited obstacle to clinical translation is the absence of standardized, clinically actionable biomarkers. Researchers currently identify senescent cells through combinations of markers, including the cell cycle inhibitor p16INK4a, lysosomal enzyme activity measured by senescence-associated beta-galactosidase, DNA damage foci, and SASP profiling. No single marker is both specific and sensitive, and protocols vary widely between laboratories, making it difficult to compare results across studies, to design clinical trials with reliable endpoints, or to know whether an intervention has actually changed the senescent cell burden in a patient’s tissues. The roadmap prioritizes the development of agreed-upon biomarker panels that can be measured reproducibly, ideally in accessible samples such as blood, and validated as predictors of clinical outcomes.

The vision that emerges from the SENESCENCE2030 network is one of precision senescence medicine, an approach modeled on the way oncology moved from blunt chemotherapy to molecularly targeted therapies matched to a tumor’s specific profile. In this vision, a clinician would one day characterize a patient’s senescent cell landscape, determining which senescent cell types are present, in which tissues, driving which pathologies, and select a senolytic or senomorphic intervention accordingly. Achieving this requires not only biomarkers but also a deeper understanding of senescent cell heterogeneity at the single-cell level, including the application of transcriptomic, epigenomic, and proteomic technologies to map senescent states in human tissues across the life course.

The roadmap is equally clear that scientific discovery alone will not be enough. Translational and regulatory frameworks must be strengthened if senotherapeutics are ever to reach the clinic. Because aging itself is not an approved indication for drug approval, clinical trials must target specific age-related diseases, which raises questions about trial design, patient stratification, and endpoints that reflect biological aging rather than a single symptom. Regulators will need validated surrogate markers to judge whether a senotherapeutic is working, and the field must agree on safety standards, particularly for senolytic drugs that remove cells which may still be performing useful functions in some tissues. The Coimbra consensus explicitly calls for dialogue between researchers, industry, and regulatory agencies to define these standards before large trials begin.

International collaboration emerges as the connective tissue holding the roadmap together. The SENESCENCE2030 network itself is a COST Action, CA23119, funded by the European Cooperation in Science and Technology, and it spans dozens of institutions across Europe, from Naples and Barcelona to Exeter, Graz, Groningen, and beyond, with participants contributing expertise ranging from cardiology and toxicology to oncology and tissue regeneration. The consensus document argues that the challenges ahead, including biomarker standardization, data sharing, trial harmonization, and training of a new generation of geroscientists, are too large for any single laboratory, company, or country. Shared biobanks, open datasets, and cross-border clinical networks are framed as prerequisites rather than aspirations. The meeting’s industry–academia workshop format was itself a deliberate exercise in bridging the gap between discovery science and product development, ensuring that company perspectives on scalability, manufacturing, and regulatory pathways informed the research agenda from the outset.

The stakes are considerable. Populations across the world are aging rapidly, and the burden of chronic age-related disease threatens health systems and economies alike. If senescence-targeting interventions can be made safe, targeted, and effective, they would represent a fundamentally new form of medicine, one that treats upstream biological drivers shared by many diseases rather than each condition in isolation. The authors of the roadmap, led by Marco Demaria of the European Research Institute for the Biology of Ageing in Groningen together with Aniello Cerrato and a broad consortium of co-authors, are candid that the field is at an inflection point. The biology is compelling and the first clinical experiments have begun, but without the shared definitions, validated markers, regulatory clarity, and coordinated networks the roadmap describes, senescence medicine risks stalling in a haze of irreproducible results and failed trials. What the Coimbra consensus offers is a collectively agreed plan, and a reminder that the transition from laboratory insight to patient benefit is a discipline in its own right, demanding as much rigor and cooperation as the discoveries that set it in motion.

Subject of Research: A consensus roadmap for translating cellular senescence research into precision senescence medicine

Article Title: A consensus roadmap from the SENESCENCE2030 network towards precision senescence medicine

Article References: Cerrato, A., Farsetti, A., Bordoni, L., Martins, R. R., Bengoetxea de Tena, I., Vrhovac Madunic, I., Mammadova, M., Raviola, S., Rima, M., Ozturk, M., Spinelli, R., Moisoi, N., Nicoli, F., Pangrazzi, L., Wouters, A., Albrakati, A., Abdellatif, M., Harries, L. W., Martini, G., … Demaria, M. (2026). A consensus roadmap from the SENESCENCE2030 network towards precision senescence medicine. Nature Aging. https://doi.org/10.1038/s43587-026-01222-y

Image Credits: AI Generated

DOI: 10.1038/s43587-026-01222-y

Keywords: cellular senescence, SENESCENCE2030, precision medicine, biomarkers, senolytics, aging, geroscience, Nature Aging, SASP, clinical translation, COST Action, senomorphic drugs

The review identifies three core challenges that have defined traditional pain management for decades. First, assessment depends on patients describing their own suffering, a process vulnerable to emotional state, cultural background and cognitive function, and effectively unusable for infants, dementia patients, the critically ill and postoperative patients who cannot self-report. Second, conventional imaging lacks the sensitivity to detect many pain-related structural changes: plain X-rays miss early osteoarthritis and soft tissue lesions, while MRI, despite excellent soft tissue resolution, struggles with functional pain and is expensive and time-consuming. Third, treatment selection relies on clinical experience and guidelines without individualized prediction, leaving roughly 30 to 40 percent of patients failing to respond adequately to their initial regimen. The result is prolonged suffering, repeated medication adjustments, rising costs and heightened risk of adverse drug reactions. Deep learning, the authors contend, offers an end-to-end pathway from symptom identification to mechanism analysis, extracting latent pain biomarkers from multi-source heterogeneous data.

The most technically rich portion of the review concerns objective pain assessment, where deep neural networks are being trained to quantify suffering from signals that patients cannot suppress. Computer vision models analyze facial micro-expressions such as frowning and squinting; speech systems capture changes in vocal tone, pitch jitter and spectral energy; and physiological pipelines integrate electroencephalography, skin conductance, heart rate variability and respiration. The performance figures are striking. A neonatal convolutional neural network recognizing pain from infant facial expressions achieved 91 percent accuracy with an area under the curve of 0.93, while a three-branch network analyzing newborn cries reached 96.77 percent accuracy in binary classification, using only about 2.6 percent of the parameters of VGG16. In adults, a spatial-temporal attention LSTM network classified postoperative pain into three levels from facial landmarks with 86.6 percent accuracy, and an autoencoder-LSTM model fused motion capture with surface electromyography to detect protective behaviors, improving over single-modality baselines by 38.5 percent.

Physiological signals have proven equally fertile ground. A framework called PainAttnNet, built on transformer architectures with multiscale feature extraction, classified pain intensity from electrodermal activity with 85.56 percent accuracy on the BioVid dataset. A dual-branch spatiotemporal model processing scalp EEG in children distinguished pain from non-pain states with 87.83 percent accuracy, and, notably, visualization of electrode contributions showed that accuracy remained at 84 percent even when only nine electrodes were retained, a finding that could dramatically simplify data collection in pediatric settings. A hybrid BiLSTM-support vector machine pipeline classified postoperative pain intensity from electrocardiographic signals at 84.14 percent validation accuracy, while bidirectional LSTMs applied to functional near-infrared spectroscopy achieved 90.6 percent accuracy across four pain intensity categories, outperforming unidirectional variants by 5 to 8.4 percentage points. Resting-state frontal EEG biomarkers have likewise been proposed for grading chronic neuropathic pain severity, moving the field closer to objective clinical translation.

Multimodal fusion, however, emerges as both the field’s greatest promise and its most sobering cautionary tale. Because any single signal can be lost in real clinical environments, obscured by oxygen masks, sedation, motion artifacts or equipment failure, fusing facial, vocal and physiological streams offers redundancy and robustness. In neonatal postoperative pain assessment, a decision-level voting fusion of facial expressions, body movements and crying maintained strong performance even when a quarter of each modality’s data was randomly removed, with the fused area under the curve of 0.868 clearly surpassing the best single modality at 0.774. Yet the review is candid that fusion is not a universal win: in real postoperative wards, single-modality models, particularly those using respiratory rate at 88.24 percent balanced accuracy, consistently outperformed multimodal fusion, which was degraded by motion artifacts, asynchronous acquisition and environmental noise rarely encountered in laboratory datasets. The authors call for cross-modal pretraining, medical knowledge graphs and event-driven fusion strategies to close this gap.

The second pillar of the review concerns intelligent structural identification, where convolutional and transformer-based models automatically segment the anatomical landscape of pain. Deep learning systems now detect lumbar spondylolisthesis from X-rays, quantify vertebral fractures through anchor-free keypoint detection with expert-level localization error of 0.92 millimeters and an AUC of 0.96, and grade intervertebral disc degeneration from MRI in real time using YOLOv5 architectures with over 95 percent classification accuracy. On the cervical spine, where vertebral similarity and complex anatomy make segmentation notoriously difficult, a 2D U-Net framework with superior-inferior labeling achieved Dice coefficients above 94 percent even on pathological data, and a transformer-based model reduced radiologist interpretation time for degenerative cervical MRI from up to 490 seconds to as little as 90 seconds, with the greatest benefit accruing to residents.

Nerve and needle localization extend this vision into interventional precision. Mask R-CNN-based systems segment the median nerve at the carpal tunnel from ultrasound without manual region selection, while U-Net variants track the vagus nerve in real time with over 90 percent recognition accuracy even in low-quality images, trained from mere bounding-box annotations. The dorsal root ganglion, a structure implicated in neuropathic pain but historically too small to segment automatically, has now been delineated in MRI using a meta-optimized nnU-Net framework, revealing genotype-related volume changes in a Fabry disease model. For ultrasound-guided nerve blocks, deep networks locate needle tips that are frequently invisible at steep angles: time-aware LSTMs combined with dynamic background subtraction recover weak tip echoes, and an optical-flow-enhanced YOLO variant tracks speckle dynamics of entirely invisible needles while cutting model parameters by 98 percent for real-time deployment, reaching sub-millimeter localization accuracy in robotic settings.

The third pillar maps AI onto specific pain conditions. For shoulder disorders, multimodal models fusing X-rays with clinical data rule out rotator cuff tears with 97.3 percent sensitivity, and 3D networks trained on more than 11,000 MRI studies classify full-thickness tears with AUCs as high as 0.99, outperforming experienced radiologists. In osteoarthritis, deep stacked ensembles grade knee severity at up to 99.71 percent accuracy, automated systems measure hip-knee-ankle angles 126.7 times faster than manual workflows, and a model called DeepKOA predicts structural and symptomatic progression over 24 to 48 months from multimodal MRI. Multiomic deep clustering has even identified three molecular subtypes of knee osteoarthritis that predict post-arthroplasty pain outcomes with AUCs of 0.84 to 0.88. For trigeminal neuralgia, machine learning on brain morphology predicted gamma knife surgery efficacy with 96.7 percent accuracy, and radiomics models now identify which patients will achieve durable relief from percutaneous balloon compression, lifting three-year pain-free survival in favorable subgroups from 51.1 percent to 86.4 percent. Machine learning models predicting postherpetic neuralgia from 23,326 real-world electronic health records, and LSTM networks forecasting cancer pain exacerbations hours before onset, illustrate the shift toward preemptive intervention.

The review closes with a bracing reality check. Most studies remain small, single-center and internally validated; when tested externally, performance routinely collapses, as when a sacroiliitis model’s sensitivity plummeted from near-expert levels to 56 percent. Data imbalance, annotation inconsistency, demographic bias and unmodeled anatomical variation pervade the literature, and explainable AI outputs often misalign with what clinicians actually need. Privacy risks from biometric pain data, unresolved liability frameworks, absent reimbursement mechanisms and the economic burden of deployment all stand between laboratory success and bedside reality. The authors argue that pain AI must now pivot from a method-driven race for benchmark accuracy to an evaluation-driven, utility-driven paradigm in which human-AI collaboration, longitudinal outcomes and patient-centered benefit define success. If that transition succeeds, the era in which suffering could only be described, rather than measured, understood and preempted, may finally be drawing to a close.

Subject of Research: Artificial intelligence applications in objective pain assessment, medical image analysis and treatment decision-making for chronic pain conditions

Article Title: Current state of research and future developments of artificial intelligence in pain diagnosis and treatment

Article References: Current state of research and future developments of artificial intelligence in pain diagnosis and treatment. (n.d.). https://doi.org/10.1186/s12967-026-08529-9

Image Credits: AI Generated

DOI: 10.1186/s12967-026-08529-9

Keywords: artificial intelligence, chronic pain, deep learning, pain assessment, multimodal data fusion, medical imaging, trigeminal neuralgia, osteoarthritis, postherpetic neuralgia, cancer pain, explainable AI, precision medicine