The motivation is straightforward. In developing countries, expensive disease-specific tests are often out of reach for large portions of the population, while the CBC test, ranging from roughly 600 to 1000 Pakistani rupees, is widely affordable. Its twelve standard parameters encode a surprising amount of clinical signal: white blood cell counts point to infection and leukemia, hemoglobin, red blood cell count and hematocrit indicate anemia and polycythemia, platelet counts reveal bleeding and inflammatory disorders, and the differential leukocyte fractions help identify bacterial infection, bone marrow suppression and allergic responses. Yet in most hospitals these reports exist only as paper printouts, unread by any algorithm and difficult to store or analyze at scale.

To build their dataset, the researchers collected nearly 8,000 real CBC reports from 2023 and 2024 at a local hospital, all in PDF form with identical layouts. After applying inclusion and exclusion criteria, 6,500 patient records remained. Crucially, all patient demographics and personal identifiers were removed in compliance with privacy guidelines, and ethical clearance was obtained from the hospital. The team then benchmarked six OCR engines on the scanned pages, including PyMuPDF, OpenCV, PdfMiner, Tesseract, EasyOCR and PaddleOCR, measuring each with the character error rate, the standard metric that counts substitutions, insertions and deletions against a manually verified ground truth.

PaddleOCR emerged as the clear winner, achieving 99 percent accuracy with a character error rate of just 0.01, compared to 93 percent for EasyOCR and 84 percent for PyMuPDF and OpenCV. The winning pipeline involved converting PDFs to images, then applying Gaussian filtering to remove high-frequency noise, median filtering to suppress salt-and-pepper artifacts without blurring character edges, skew correction to align table boundaries, histogram equalization to enhance contrast, and min-max normalization to rescale pixel intensities. A detection model localized the table region and identified rows, columns and value positions, while a CRNN-based recognition model with a structure-preserving module ensured that each extracted value stayed aligned with its correct test name, preventing records from being mixed together.

With the table data digitized, the team faced a labeling problem: the extracted records had no disease annotations. They solved this with a rule-based algorithm built on clinically accepted reference ranges verified by medical experts at the source hospital. Each of the twelve parameters received a dictionary of thresholds with disease associations for values above and below normal. A hemoglobin below 13 g/dL, for example, flags anemia, while a value above 18 g/dL suggests polycythemia; an MCV under 75 fL indicates microcytic anemia, an MCHC above 35 g/dL indicates hyperchromic anemia, platelets below 150 per microliter signal thrombocytopenia, and elevated neutrophils point to acute bacterial infection. The algorithm processed every record against these rules, generating both a binary disease-or-normal label and, importantly, multiple simultaneous disease labels per patient.

The data was then cleaned through exploratory analysis, with outliers detected in the platelet feature capped using a Z-score threshold to preserve clinical relevance, and all features standardized before modeling. For the multi-disease task, the team used multi-label classification, decomposing the single multi-disease output column into seventeen separate binary target columns via a Multi-Label Binarizer, so that each record became a vector of yes-or-no flags for each condition. Five classifiers were trained: logistic regression, decision tree, random forest, gradient boosting and support vector machine, all implemented in scikit-learn on a 16 GB Windows machine running Python 3.12.2.

Hyperparameter tuning proved decisive. Using GridSearchCV integrated with 10-fold cross-validation on an 80:20 train-test split, the team searched parameter grids for each model, setting the decision tree maximum depth to 20 with the Gini criterion, constraining the random forest to a depth of three, six leaf nodes and square-root feature sampling to prevent overfitting, fixing gradient boosting at 100 estimators with a learning rate of 0.1, and configuring the support vector classifier with a regularization parameter C of 3, scaled gamma and a radial basis function kernel. Untuned models peaked at around 74 percent F1-score for binary classification, but after tuning the cross-validated results climbed substantially, with gradient boosting reaching an F1 of 89.4 percent and an ROC-AUC of 0.94.

On the held-out test set, random forest delivered the best binary performance with 91.7 percent precision, 90.9 percent recall and a 0.94 ROC-AUC, narrowly edging out gradient boosting. The multi-label results told a similar story: random forest achieved around 93 percent precision and an F1 of 92.0 percent on cross-validation, with gradient boosting close behind, while logistic regression and the decision tree lagged. The authors noted that misclassified cases tended to involve borderline values only slightly outside normal ranges, highlighting how sensitive threshold-based labeling can be when clinical signals overlap. Across both tasks, the ensemble methods that aggregate many decision trees proved most robust to this ambiguity.

Transparency was addressed with explainable AI. The team applied SHAP, or SHapley Additive Explanations, analysis independently to each disease label, ranking features by their marginal contribution to each prediction. The analysis identified white blood cells, red blood cells, hemoglobin, MCV, platelets, hematocrit, MCH, MCHC and neutrophils as the most influential parameters driving both binary and multi-disease classification. This patient-level interpretability, the researchers argue, is essential for clinician trust in high-stakes healthcare settings, allowing a doctor to see exactly which blood values pushed a given prediction.

The study is the first of its kind, the authors say, to tackle multi-disease prediction from CBC features using a large real-world dataset, and it arrives with clear caveats. The data comes from a single hospital region and only twelve parameters, so results cannot yet be generalized to broader populations, and rule-based labeling may miss complex diagnostic patterns that only appear in richer clinical data. The team is already working to integrate the system into an Android application so patients and doctors can interpret reports interactively, and they suggest that hospitals should maintain online repositories to make digitization easier in the first place. If validated across multi-center data, the approach could turn the world’s most routine blood test into a low-cost early warning system, catching anemia, infections and more serious hematological conditions before expensive, time-consuming confirmatory tests become necessary.

Subject of Research: AI-based multi-disease prediction from digitized complete blood count reports

Article Title: AI-enabled multi-disease diagnosis using low-cost CBC reports

Article References: Bibi, M., Saeed, W. K., Jabeen, S., Shah, S., Ali, G., ELAffendi, M., & Khan, M. (2026). AI-enabled multi-disease diagnosis using low-cost CBC reports. Frontiers in Imaging, 5, Article 1754419. https://doi.org/10.3389/fimag.2026.1754419

Image Credits: AI Generated

DOI: 10.3389/fimag.2026.1754419

Keywords: complete blood count, artificial intelligence, machine learning, OCR, PaddleOCR, multi-label classification, disease diagnosis, medical imaging, healthcare, anemia, random forest, gradient boosting

Non-suicidal self-injury, commonly abbreviated NSSI, refers to deliberate destruction of body tissue without suicidal intent, such as cutting, burning or hitting oneself. It is distressingly common in adolescence and ranks among the strongest known predictors of a later suicide attempt, which is precisely why researchers are hunting for accessible biological signals that might flag vulnerable young patients early. The neutrophil-to-lymphocyte ratio is an attractive candidate for exactly that role because it is essentially free: it requires nothing beyond the complete blood count that is routinely drawn in almost every emergency department on the planet. Neutrophils and lymphocytes are the two most abundant classes of white blood cells, and their balance shifts characteristically under physiological stress, with neutrophils rising as part of the innate inflammatory response while lymphocyte numbers fall under the influence of stress hormones such as cortisol.

The study, led by Azize Aydemir of Giresun University and child and adolescent psychiatrist Berkan Şahin, compared 89 participants aged 9 to 18 years. Thirty-six of them had received a DSM-5 diagnosis of non-suicidal self-injury after presenting to the paediatric emergency department of Giresun Maternity and Children’s Training and Research Hospital, while 53 healthy controls were recruited from routine outpatient wellness visits at the same institution. Rather than relying on automatically reported values, the team recomputed every index from the source cell counts, a methodological detail that matters because laboratory analysers calculate such ratios in slightly different ways and rounding errors can distort small differences.

The headline result was statistically robust. The median NLR in the self-injury group was 2.03, with an interquartile range running from 1.51 to 3.33, compared with a median of 1.42 and an interquartile range of 1.17 to 1.74 in the healthy comparison group. The difference was highly significant on the Mann-Whitney U test, with a p-value below 0.001 and a rank-biserial correlation of 0.47, a figure that indicates a moderately large effect. Notably, two closely related inflammatory indices, the platelet-to-lymphocyte ratio and the monocyte-to-lymphocyte ratio, showed no significant differences between the groups, suggesting the signal is specifically tied to the neutrophil-lymphocyte axis rather than to inflammation markers in general.

To test whether the association could be explained by age or sex, the researchers ran a binary logistic regression with self-injury status as the outcome and NLR, age and sex as covariates. After adjustment, each one-unit increase in NLR was associated with roughly 2.24 times higher odds of belonging to the self-injury group, with a 95 percent confidence interval stretching from 1.29 to 3.91 and a p-value of 0.005. Female sex emerged as the single strongest correlate in the model, carrying an adjusted odds ratio of 8.58, with a confidence interval from 2.79 to 26.36. That magnitude reflects both a genuine sex difference in adolescent self-injury epidemiology, which is well documented internationally, and the modest size of the sample, which inflates uncertainty around adjusted estimates. A sensitivity analysis using Firth penalised likelihood, a technique designed to stabilise regression estimates in small or imbalanced datasets, supported the main finding.

The story grew more complicated when the team examined how well NLR actually discriminated between cases and controls. Receiver operating characteristic analysis, the standard method for evaluating binary classifiers, yielded an area under the curve of 0.736, with a 95 percent confidence interval of 0.630 to 0.843. An AUC of 0.736 sits in the conventional category of moderate discrimination, well short of the 0.8 to 0.9 range typically expected of a clinically useful screening tool. More striking still, NLR proved no better than the absolute neutrophil count alone, which achieved an AUC of 0.724; a formal DeLong test comparing the two curves returned a p-value of 0.80, meaning the small apparent advantage of the ratio could easily be chance.

The Youden index, which balances sensitivity and specificity to find the optimal cut-off on an ROC curve, suggested a threshold of 1.72 for flagging concerning cases. But internal validation punctured confidence in that number. A bootstrap analysis produced a 95 percent percentile interval for the optimal threshold stretching all the way from 1.22 to 2.19, an enormous range for a value that would have to be applied at the bedside, and when the model was tested through repeated cross-validation, its discrimination slipped from 0.736 to 0.717. These procedures simulate how a model performs on data it has never seen, and the drop indicates real optimism in the apparent performance of the original fit.

The researchers probed the robustness of the association from several angles, and the core signal held up in most of them. The elevated NLR persisted after the analysis excluded patients with a diagnosis of major depressive disorder, after excluding anyone with any psychiatric comorbidity, and after excluding patients who had presented following self-poisoning, a presentation that could itself perturb inflammatory markers. One exception was revealing: among patients who were unmedicated at the time of presentation, the association was attenuated and no longer statistically significant. Medication status appears to act as an effect modifier of some kind, though the study cannot determine whether psychotropic drugs, the underlying illness severity that prompts their prescription, or chance in a small subgroup explains the pattern.

The authors are explicit that the findings should be treated as hypothesis-generating rather than practice-changing. The cases were sampled during an acute emergency presentation, a moment when pain, fear, sleep deprivation and the stress response itself can all shift white cell dynamics, so the elevated NLR may reflect the physiological storm of the crisis rather than a stable trait of young people prone to self-injury. The case group also carried substantial psychiatric morbidity, and the comparison group consisted of healthy wellness-visit attendees rather than emergency department patients with other complaints, so the study cannot say whether NLR is specific to self-injury or simply a general marker of acute distress or illness. The proper next step, they argue, is research that incorporates both psychiatric comparison groups and emergency department controls.

Even with those caveats, the study matters for what it signals about the direction of biological psychiatry. Indices such as NLR cost nothing to compute, are already embedded in every admission workflow, and are being investigated across depression, psychosis and suicidal behaviour in adult populations; paediatric emergency data of this kind, and data from Türkiye specifically, had been lacking. If larger, prospective studies with appropriate comparison groups replicate and refine the association, a value printed automatically on every blood count report could eventually serve as one more data point in a multimodal risk assessment that combines clinical interview, validated screening instruments and, perhaps, inexpensive biomarkers. For now, no child should be flagged for self-harm risk on the basis of a blood ratio alone, and the Youden-optimal cut-off of 1.72 should be regarded as unstable until validated externally. What the Giresun team has delivered is a carefully validated but modest signal, honest about its own limits, that points toward a future in which the humble complete blood count does a little more work in protecting young people in crisis than anyone expected.

Subject of Research: The neutrophil-to-lymphocyte ratio as an inflammatory marker associated with non-suicidal self-injury in children and adolescents

Article Title: Neutrophil-to-lymphocyte ratio in children and adolescents presenting to a paediatric emergency department with non-suicidal self-injury: a retrospective case-control study with healthy comparison participants

Article References: Aydemir, A., & Şahin, B. (2026). Neutrophil-to-lymphocyte ratio in children and adolescents presenting to a paediatric emergency department with non-suicidal self-injury: a retrospective case-control study with healthy comparison participants. BMC Pediatrics. https://doi.org/10.1186/s12887-026-07669-7

Image Credits: AI Generated

DOI: 10.1186/s12887-026-07669-7

Keywords: non-suicidal self-injury, neutrophil-to-lymphocyte ratio, inflammation, adolescents, paediatric emergency, complete blood count, case-control study, ROC analysis, biomarkers, child and adolescent psychiatry, suicide risk, systemic inflammation

The study focused on the Lika Pramenka, a hardy Croatian breed traditionally managed under extensive farming conditions. The thirty ewes selected from a herd of roughly eight hundred animals were, on average, five years old and in their fourth lactation, with a lambing index of 1.2. Their diet was deliberately simple and consistent: 1.8 kilograms of meadow hay per day supplemented with 800 grams of oats. At the start of the trial the ewes averaged 53.36 kilograms in body weight and scored 3.75 on the standard five-point body condition scale. Blood was drawn from the same animals on four occasions — twenty days before lambing, and then on the 40th, 70th and 100th days of lactation — yielding 120 samples in total. This repeated-measures design allowed the team to compare each ewe against her own baseline across the demanding transition from pregnancy to milk production.

The laboratory work was exhaustive. Complete blood counts were determined the same day using an automated three-differential hematology analyzer, capturing white blood cells, red blood cells, platelets, hemoglobin, hematocrit and the erythrocyte constants known as MCV, MCH and MCHC. Differential leukocyte counts — the proportions of neutrophils, lymphocytes, basophils, eosinophils and monocytes — were read from blood smears stained with a Pappenheim stain. Serum was separated by centrifugation, frozen at minus 80 degrees Celsius, and later analyzed on an automatic chemistry platform for minerals including calcium, inorganic phosphorus and iron; metabolites including glucose, cholesterol fractions, triglycerides and non-esterified fatty acids; total protein, albumin and urea; and the activities of five enzymes associated with liver and tissue metabolism. Statistical analysis using general linear models and Tukey comparisons identified which changes across the four sampling points were genuinely significant rather than random variation.

The red blood cell results tell a story of oxygen transport under pressure. Compared with lactating ewes, sheep in late pregnancy carried significantly higher counts of red blood cells and higher concentrations of hemoglobin and hematocrit, along with a greater proportion of lymphocytes but fewer platelets and a lower share of neutrophils. Once lactation began, the picture reversed: red blood cell counts, hemoglobin and hematocrit fell significantly, with the lowest values recorded at the second measurement early in lactation, while platelet counts, MCH and neutrophil proportions rose. The authors note that these early-lactation declines could indicate anemia, and indeed the second measurement met commonly used diagnostic thresholds for slight anemia in ruminants, with hemoglobin at 67.70 grams per liter and hematocrit at 0.24 percent. In the other measurements, no such indications appeared.

Why would a milking ewe become temporarily anemic? The researchers point to a mechanism described in earlier literature: hemodilution driven by the physiological adaptations of early lactation, when large amounts of water are mobilized toward the mammary gland to support milk synthesis. Some authors have also argued that elevated red cell destruction in mammary tissue contributes to the reduced hematocrit. Early lactation is additionally a period of energy deficit in ruminants, because fat mobilization peaks precisely when milk production is highest, and this high-intensity metabolism generates oxidative stress and exposes animals to infection. Related work in dairy cattle has shown that cows in early lactation can have compromised immune systems that make them susceptible to bacterial infection and anemia, and that good feeding management can prevent these effects. The team also observed macrocytosis and hypochromia — larger-than-normal red cells with reduced hemoglobin concentration — relative to reference values for sheep.

The biochemical panel revealed an equally dramatic metabolic pivot. Late-pregnant ewes had significantly higher blood concentrations of glucose and triglycerides and a higher albumin-to-globulin ratio, but significantly lower concentrations of urea, total protein, albumin, globulins, calcium and iron, together with reduced activity of the enzymes GGT, CK and AST. During lactation the pattern flipped: glucose and the albumin-to-globulin ratio fell while urea, total protein, albumin, globulins, HDL-cholesterol, calcium, inorganic phosphorus and CK activity rose. The elevated glucose in late pregnancy aligns with earlier findings in sheep and cattle, where increased hepatic gluconeogenesis in the final weeks of gestation raises blood glucose, which then declines as the demands of milk synthesis take over after parturition. The highest concentrations of non-esterified fatty acids appeared in early lactation, a classic signature of body fat being mobilized to cover the energy deficit when glucose supplies fall short.

The protein and mineral data exposed a nutritional weak point in the flock’s diet. Urea concentrations in the first lactation measurement sat at the lower limit of reference values, which the authors interpret as a sign of insufficient protein in the feed — a conclusion reinforced by total protein and albumin concentrations that were at or below the lower limits of reference ranges in the first two lactation measurements. These changes indicated dysproteinemia, an imbalance in blood proteins. Meanwhile, the reduced calcium and iron concentrations in late-pregnant and lactating ewes reflect the intense consumption of these minerals during fetal growth and milk secretion; calcium is deployed from blood to milk, and iron is excreted in milk, explaining the significant drop in serum iron during lactation. Earlier studies have shown that reduced blood calcium during late pregnancy and early lactation results from the combined calcium demands of fetal and milk secretion.

Liver enzyme activity added a final layer of evidence for the metabolic burden. Significantly lower activity of GGT, CK and AST in late-pregnant ewes, followed by increases during lactation, indicated an overload of the organism and of the liver itself with protein deamination, linked to the increased demands of milk secretion. Elevated activities of ALT, AST and GGT in lactating sheep are associated with more intense hepatic metabolism and the body’s response to negative energy balance. The authors suggest these enzyme changes could be connected with reduced dry matter intake and increased fat mobilization around lambing, which can lead to hepatic lipidosis and disturb normal liver function. Notably, most enzyme activities remained within reference limits except GGT during lactation, suggesting the strain was significant but not yet pathological.

Beyond the individual measurements, the correlation analysis demonstrated how tightly these blood indicators are interwoven. The team found very strong positive correlations between red blood cells and hemoglobin (r = 0.98), red blood cells and hematocrit (r = 0.97) and hemoglobin and hematocrit (r = 0.95), which is physiologically expected since hematocrit measures the fraction of blood made up of red cells. A very strong negative correlation emerged between neutrophils and lymphocytes (r = −0.94), a ratio widely used to assess stress in animals. Among the biochemical parameters, very strong positive correlations linked cholesterol with HDL-cholesterol (r = 0.87), cholesterol with LDL-cholesterol (r = 0.77), total protein with albumin (r = 0.70) and total protein with globulins (r = 0.91), while a strong negative correlation appeared between globulins and the albumin-to-globulin ratio (r = −0.60). These interdependencies confirm that blood indicators respond as a coordinated system to the ewe’s metabolic state.

The practical message for shepherds and veterinarians is clear: blood values cannot be interpreted in a vacuum. The study demonstrates that late pregnancy and lactation significantly reshape most indicators of the metabolic profile and complete blood count, with the most pronounced changes occurring in late pregnancy and the earliest weeks of lactation, and that feed quality — particularly protein and mineral content — plays a decisive role in shaping those values. The authors recommend that any assessment of a sheep herd’s metabolic profile should account for the animals’ physiological status and the quality of their nutrition in order to obtain the most accurate picture of herd health, breeding quality, possible nutritional errors and welfare. For extensive farming systems like the one in Lika, where veterinary intervention is often minimal, a single blood draw interpreted without this context could mislabel a healthy, hard-working ewe as sick — or, more dangerously, miss a genuine problem hidden behind what looks like a normal result.

Subject of Research: Changes in the blood metabolic profile and complete blood count of ewes during late pregnancy and lactation

Article Title: Blood metabolic profile and complete blood count of sheep during late pregnancy and lactation

Article References: Antunović, Z., Kučan, J., Klir Šalavardić, Ž., Djidara, M., & Novoselec, J. (2026). Blood metabolic profile and complete blood count of sheep during late pregnancy and lactation. Archives Animal Breeding, 69(2), 337-346. https://doi.org/10.5194/aab-69-337-2026

Image Credits: AI Generated

DOI: 10.5194/aab-69-337-2026

Keywords: sheep, Lika Pramenka, blood metabolic profile, complete blood count, late pregnancy, lactation, anemia, non-esterified fatty acids, glucose, liver enzymes, calcium, animal nutrition

The scale of the analysis is what sets it apart. Studies of inflammatory biomarkers in infection have typically focused on a single organ system or a handful of infection types, comparing perhaps a few thousand patients against controls. By aggregating data across respiratory, digestive, urogenital, dermatological, and head-neck-orofacial infections, the researchers were able to look for patterns that recur across anatomical boundaries and, just as importantly, patterns that diverge. Respiratory infections emerged as the most prevalent category in the cohort, followed by urogenital infections, a distribution that mirrors the global burden of infectious disease. But prevalence was only the starting point. The heart of the study lay in whether the twelve hematological parameters, which include white cell differentials, red cell indices, and platelet measures, behaved differently depending on the site of infection.

They did, and the differences were biologically coherent. Patients with digestive infections showed elevated neutrophil counts, consistent with the robust innate immune recruitment that characterizes bacterial infections of the gastrointestinal tract and its associated organs. Respiratory infections exhibited increased hematocrit levels, a finding that may reflect hemoconcentration, hypoxic stress, or inflammatory effects on plasma volume in patients with pneumonia and other lower respiratory tract disease. Perhaps most distinctive were the signatures associated with fallopian tube infections, which demonstrated a decreased platelet-to-lymphocyte ratio alongside elevated basophil counts, a combination not seen in the other infection categories. These site-associated patterns, the authors emphasize, reflect non-specific systemic inflammatory responses rather than pathogen-specific mechanisms, but their consistency across hundreds of thousands of patients suggests that the peripheral blood does encode information about the anatomical theater of an infection.

To ensure that these patterns were not artifacts of confounding, the researchers applied rigorous statistical machinery. Propensity score matching was used to balance infection patients and healthy controls on baseline characteristics, with standardized mean differences below 0.1 after matching, indicating successful covariate balance across the major infection groups and dozens of subgroups. Missing laboratory data, an inevitable feature of any large retrospective cohort, was handled through multiple imputation, and the completeness of missingness was documented variable by variable for each source dataset. Differential expression analyses were corrected for multiple testing using false discovery rate control. This methodological discipline matters because observational comparisons of inflamed versus healthy populations are notoriously vulnerable to demographic skew, and the authors’ decision to match patients one-to-one on covariates before comparing blood parameters strengthens the causal interpretability of the observed signatures, even if the retrospective design still precludes definitive causal claims.

Beyond conventional differential analysis, the team built predictive models to test whether the hematological signatures could be harnessed for classification. Using the twelve blood parameters plus age and sex as inputs, they constructed multivariable logistic regression models and random forest models, the latter an ensemble machine learning method that aggregates hundreds of decision trees to capture non-linear interactions among variables. The models were evaluated with five-fold cross-validation and held-out test sets, with performance quantified using the area under the receiver-operating-characteristic curve, the area under the precision-recall curve, F1 score, sensitivity, specificity, and precision, each reported with 95 percent bootstrap confidence intervals. The random forest consistently outperformed both single biomarkers and logistic regression, a result that underscores how combinations of ordinary blood values can carry diagnostic signal that no individual parameter provides alone.

The system-level performance figures were striking. Random forest achieved an area under the curve of 0.965 for distinguishing respiratory infections from controls, 0.953 for digestive infections, 0.928 for dermatological infections, 0.901 for urogenital infections, and 0.894 for head-neck-orofacial infections. Values in this range indicate excellent discrimination between infected and healthy individuals at the system level, comfortably exceeding the performance of any single traditional biomarker such as the neutrophil-to-lymphocyte ratio or mean platelet volume. Principal component analysis, an unsupervised technique that projects high-dimensional data onto its axes of greatest variance, revealed clear separation between infection patients and healthy controls, providing visual confirmation that the multivariate hematological space occupied by infected individuals differs systematically from that of the uninfected.

Localizing infection to a specific site proved considerably harder. In an exploratory two-site classification task, distinguishing respiratory from digestive infections using blood parameters alone, the random forest model achieved an area under the curve of 0.795, which the authors characterize as moderate discriminative ability. This drop in performance is unsurprising given the biology: systemic inflammation produces overlapping hematological consequences regardless of where the inciting infection sits, and the twelve routine parameters capture only a coarse snapshot of the immune response. The authors are careful to frame the two-site result as exploratory, and they note that the supplementary analyses comparing each of the 29 organ-specific infection groups against matched controls are intended to distinguish infection from health rather than to directly localize infection sites. The distinction between detecting infection and pinpointing its location is central to interpreting the study’s contribution honestly.

That honesty extends to the study’s stated conclusion, which is notably measured. The authors write that the combined models, particularly random forest, improved discrimination but remain exploratory, and that the findings support the potential of routine blood tests as probabilistic adjuncts rather than standalone diagnostics in precision infection medicine. This framing matters clinically. A complete blood count costs pennies, returns results in minutes, and is already ordered in virtually every febrile patient, so even modest probabilistic information about infection presence and likely site could meaningfully accelerate triage, guide the choice of imaging or microbiological testing, and support antimicrobial stewardship decisions in settings where advanced diagnostics are unavailable. But no clinician should read a random forest output as a diagnosis, and the authors’ refusal to overstate the models’ localization ability is a welcome contrast to the hype that often accompanies machine learning studies in medicine.

The study also carries implications for how biomarker research is conducted. By demonstrating that site-associated patterns exist and can be quantified at unprecedented scale, the work establishes a reference framework against which future studies of more specific markers, such as C-reactive protein, procalcitonin, or transcriptomic signatures, can be benchmarked. The open-access publication, with extensive supplementary tables documenting propensity score matching, post-matching balance, missingness proportions, and model performance across all 29 infection groups, provides a level of methodological transparency that facilitates independent scrutiny and reuse. The research protocol was approved by the Institutional Review Board of the First Affiliated Hospital of Guangzhou Medical University, and the requirement for informed consent was waived given the retrospective design and the use of fully de-identified archival data, an approach consistent with the Declaration of Helsinki.

Limitations remain, and readers should hold them in view. The cohort is retrospective and drawn from Chinese hospital populations, raising questions about generalizability to other demographics and healthcare settings. Hematological parameters are influenced by age, sex, comorbidities, medications, and circadian variation, and while propensity matching addressed measured confounders, unmeasured ones may persist. The models distinguish infected from healthy individuals, a task far easier than the clinically harder problem of distinguishing bacterial from viral infection or grading severity among infected patients. Still, the central message stands: the routine blood count, interpreted through the lens of large-scale pattern analysis and machine learning, contains a reproducible, site-associated signature of infection. As the authors conclude, these signatures offer auxiliary information for infection site classification and support the potential of routine blood tests as probabilistic adjuncts in the emerging discipline of precision infection medicine, a vision in which one of medicine’s oldest and cheapest tests acquires a new, computationally informed role.

Subject of Research: Site-associated blood-based hematological signatures for infection detection and classification across 29 infection types

Article Title: Systematic identification of site-associated blood-based signature patterns across 29 infection types: a large-scale analysis of 455,530 patients toward precision infection medicine

Article References: Qian, F., Chen, B., Cheng, J., Mo, J., Liang, Y., Sun, B., Zeng, G., & Chang, Z. (2026). Systematic identification of site-associated blood-based signature patterns across 29 infection types: a large-scale analysis of 455,530 patients toward precision infection medicine. BMC Infectious Diseases. https://doi.org/10.1186/s12879-026-14457-2

Image Credits: AI Generated

DOI: 10.1186/s12879-026-14457-2

Keywords: infection biomarkers, complete blood count, machine learning, random forest, propensity score matching, hematology, diagnostic markers, respiratory infections, digestive infections, precision medicine, BMC Infectious Diseases, retrospective study

The study, published in BMC Medicine by Lanfei Bi, Wenli Li and colleagues working across Nanjing Medical University and several partner institutions, describes a DNA methylation panel measured in peripheral blood that can help distinguish patients with esophageal squamous cell carcinoma from healthy individuals. DNA methylation refers to the addition of methyl groups to cytosine bases, typically at cytosine-guanine dinucleotide sites known as CpG sites. These epigenetic marks regulate gene expression without altering the underlying DNA sequence, and they are among the earliest and most consistent molecular changes that occur during cancer development. Because tumor cells shed DNA into the bloodstream and because immune and blood cell populations themselves respond to malignancy, methylation patterns circulating in blood can serve as a molecular fingerprint of disease occurring elsewhere in the body.

The researchers focused on three genes: HYAL2, which encodes hyaluronan glucosidase 2, an enzyme involved in the breakdown of hyaluronic acid, a component of the extracellular matrix; S100P, a calcium-binding protein implicated in cell proliferation and metastatic behavior; and DYRK4, or dual-specificity tyrosine phosphorylation regulated kinase 4, a signaling kinase. All three have been linked to cancer biology in prior work, but the team’s goal was not simply to confirm their involvement. Instead, they wanted to quantify methylation at specific CpG sites within these genes precisely enough to build a diagnostic classifier, using mass spectrometry to measure methylation levels quantitatively rather than merely detecting the presence or absence of marks.

The study design rested on two independent case-control investigations with a combined total of 601 patients with esophageal squamous cell carcinoma and 802 healthy controls, a sample size that lends considerable statistical weight to the findings. In the first study, the team observed significant hypomethylation, meaning a loss of methyl marks, at 11 of the 16 CpG sites examined across the three genes. The effect was substantial: each 10 percent decrease in methylation was associated with odds ratios ranging from 1.35 to 1.94 for having the disease, with P-values below 0.004. In the second study, which enrolled patients with earlier-stage disease, hypomethylation was confirmed at 9 CpG sites within HYAL2 and S100P, with odds ratios between 1.27 and 1.69 per 10 percent methylation decrease and P-values below 0.001. The replication across two independent cohorts is a critical strength, since many candidate biomarkers fail precisely at this validation stage.

Perhaps the most intriguing biological observation was that methylation differences between cases and controls grew larger in advanced or more aggressive tumors. This dose-response-like relationship suggests that the methylation changes are not random epiphenomena but track with the progression of the malignancy itself. It also raises the possibility that a methylation panel of this kind could eventually carry prognostic information, hinting at tumor stage or aggressiveness before imaging or endoscopy provides that answer. The authors are careful, however, to frame the panel as an auxiliary diagnostic tool, one that would complement rather than replace endoscopic examination and biopsy, which remain the gold standard for confirming esophageal cancer.

To translate the methylation measurements into a practical classifier, the researchers used logistic regression, a statistical method that models the probability of disease as a function of multiple predictor variables while adjusting for covariates such as age and sex. They evaluated performance with receiver operating characteristic curves, summarizing discriminatory ability through the area under the curve, or AUC, a value of 1.0 indicating perfect separation between cases and controls and 0.5 indicating performance no better than chance. In the first study, the methylation panel achieved AUC values between 0.68 and 0.78. In the second study, with earlier-stage cases, the panel reached AUC values of 0.70 to 0.86, a notable result given that early-stage disease is typically the hardest to detect and the most clinically valuable target.

The decisive methodological innovation came when the team incorporated complete blood count parameters into the model. A standard complete blood count measures the proportions and absolute numbers of different circulating cell types, including neutrophils, lymphocytes, monocytes and platelets, and these values are routinely altered in cancer patients through inflammation and immune modulation. When the researchers adjusted the methylation panel of the second study for these hematological parameters, the discriminatory ability improved markedly, yielding AUC values between 0.80 and 0.91 for detecting esophageal squamous cell carcinoma. The logic is elegant: methylation marks capture tumor-derived and tumor-responsive molecular signals, while blood count parameters capture the systemic physiological response to malignancy, and combining the two layers of information produces a classifier stronger than either alone.

The technical pipeline behind the measurements deserves attention as well. Quantitative methylation determination by mass spectrometry, following enzymatic processing steps that included shrimp alkaline phosphatase treatment, offers precision that older qualitative methylation assays lack. The researchers also controlled for potential confounders such as single nucleotide polymorphisms, which can interfere with methylation quantification at nearby CpG sites. Ethics approval was obtained from Nanjing Medical University and the Cancer Hospital of the Chinese Academy of Medical Sciences, and written informed consent was collected from all participants, with the authors declaring no competing interests.

The clinical implications are considerable, particularly for regions where esophageal squamous cell carcinoma is endemic. The disease shows striking geographic clustering, with high-incidence areas in parts of China, Iran and East Africa, where rates can be many times higher than the global average. In such settings, population-wide endoscopic screening is expensive, resource-intensive and uncomfortable for patients, which limits uptake. A blood-based test that could triage individuals for endoscopic follow-up would concentrate invasive procedures on those most likely to benefit. Because the panel relies on quantitative methylation signatures and standard blood count parameters, both of which can be measured in a conventional clinical laboratory, the authors argue that it represents a feasible and efficient strategy for in vitro auxiliary diagnosis.

Cautious interpretation remains essential. Case-control studies demonstrate association and classification performance but do not establish how the panel would perform in a true screening population, where the prevalence of disease is far lower than in a case-control design and false positives carry real costs. Prospective validation in independent cohorts, ideally including patients with premalignant lesions and other esophageal or inflammatory conditions, will be needed to define the panel’s specificity and its place in clinical guidelines. Nevertheless, the study adds to a rapidly growing body of work showing that epigenetic signatures in blood can reveal cancers that are otherwise invisible until late stages. If the findings hold up in larger, population-based trials, a simple blood test reading the methylation patterns of HYAL2, S100P and DYRK4, combined with routine blood count data, could become a powerful ally in catching one of the world’s deadliest cancers while it is still curable.

Subject of Research: A peripheral blood DNA methylation panel for auxiliary diagnosis of esophageal squamous cell carcinoma

Article Title: Peripheral blood-based DNA methylation panel for auxiliary diagnosis of esophageal squamous cell carcinoma

Article References: Bi, L., Li, W., Gu, W., Jin, D., Zhao, Z., Zhao, L., Wu, J., Zhang, J., Song, Y., Jiang, Y., & Yang, R. (2026). Peripheral blood-based DNA methylation panel for auxiliary diagnosis of esophageal squamous cell carcinoma. BMC Medicine. https://doi.org/10.1186/s12916-026-05209-9

Image Credits: AI Generated

DOI: 10.1186/s12916-026-05209-9

Keywords: DNA methylation, esophageal squamous cell carcinoma, epigenetics, biomarkers, HYAL2, S100P, DYRK4, blood test, cancer diagnosis, complete blood count, mass spectrometry, early detection

The biological problem the index addresses is the defining paradox of Kawasaki disease. The illness presents as a feverish, inflammatory condition of childhood, with mucocutaneous signs such as cracked lips, conjunctival injection, rash and swollen lymph nodes, yet its most consequential damage is vascular. In the acute phase, a diffuse inflammatory process targets the medium-sized arteries, and the coronary arteries are particularly vulnerable. When inflammation weakens the arterial wall, the vessel can dilate into an aneurysm, a lesion that may regress over time or persist for life, setting the stage for stenosis, thrombosis and ischemic heart disease decades later. Intravenous immunoglobulin, given within the first ten days of illness, dramatically reduces this risk, but a meaningful fraction of children still develop coronary abnormalities, and some fail to respond to first-line therapy at all. Identifying those children early, ideally at the first encounter, is the goal of every risk-stratification effort in the field.

Current practice relies on a patchwork of tools. The American Heart Association, in its 2024 scientific statement on the diagnosis and management of Kawasaki disease, emphasizes echocardiographic surveillance and clinical judgment, and standardized Z-score formulas are used to express a child’s coronary artery dimensions relative to body surface area, allowing clinicians to distinguish dilation from aneurysm and to track progression. Predictive models have also been developed, including a widely discussed risk model validated in a North American population by Son and colleagues, which combines demographic and laboratory variables to estimate aneurysm probability before treatment begins. Saad and colleagues themselves have contributed to this effort, publishing a predictive modeling approach for early risk stratification in Pediatric Research earlier in 2026. Yet each tool has limitations: models built in one population may not transfer to another, Z-scores depend on accurate imaging and appropriate formula selection, and no single laboratory value has proven sufficiently discriminating on its own.

Into this gap steps the Aggregate Index of Systemic Inflammation. The AISI belongs to a family of composite inflammatory cell indexes that have attracted growing attention across medicine because they can be derived at no additional cost from parameters already reported on every differential blood count. In its standard formulation, the index multiplies the platelet count, the neutrophil count and the monocyte count, and divides by the lymphocyte count. Each component captures a facet of the inflammatory response: neutrophils and monocytes are the cellular vanguard of innate inflammation, platelets are acute-phase reactants whose numbers climb as interleukin-6 drives thrombopoiesis, and lymphocytes tend to fall under the suppressive weight of stress hormones and inflammatory cytokines. A high AISI therefore reflects a convergence of myeloid activation, thrombocytosis and relative lymphopenia, a pattern that in many diseases tracks the intensity of systemic inflammation more faithfully than any single cell line.

The appeal of such an index in Kawasaki disease is grounded in the disease’s own biology. Recent reviews, including a 2024 analysis in Nature Reviews Rheumatology by Noval Rivas and colleagues, have highlighted platelets not as passive bystanders but as active mediators of vascular inflammation in the illness. Platelets interact directly with activated endothelium, release pro-inflammatory and pro-thrombotic mediators, and contribute to the remodeling of the vessel wall that underlies aneurysm formation. Neutrophils, meanwhile, have been implicated in endothelial injury through the release of neutrophil extracellular traps, and monocytes infiltrate the inflamed arterial wall. The idea that a product of these very cell counts might correlate with coronary outcomes is therefore not a statistical accident in search of a mechanism; it is a hypothesis with a plausible pathophysiological backbone. The same logic has propelled AISI research in adult cardiovascular medicine, where a 2024 retrospective study in the Journal of Inflammation Research reported associations between the index and clinical outcomes in patients with acute myocardial infarction.

The immediate trigger for the commentary is a study by Liu, Peng and Zhou, published online in Pediatric Research in September 2026, which examined the association between the AISI and coronary artery aneurysm in patients with Kawasaki disease. That investigation reported that a higher AISI was associated with the presence of coronary artery aneurysm, suggesting that the index could help identify children who require closer coronary surveillance. Saad and his co-authors, writing from institutions spanning Egypt, Jordan, Russia, Saudi Arabia, the United Kingdom and Pakistan, use their commentary to appraise what this finding means for practice. Their assessment, summarized in the article’s impact statements, is measured: the AISI may serve as an additional inflammatory marker for identifying children with Kawasaki disease who need closer coronary monitoring, and AISI-based risk stratification may ultimately contribute to personalized approaches in disease management.

The commentary’s caution is as instructive as its enthusiasm. The authors state plainly that current evidence supports further prospective validation before the AISI can be included in treatment decision pathways. This is a crucial distinction in pediatric cardiology, where treatment intensification, such as the addition of corticosteroids or other adjunctive agents for high-risk children, carries its own risks and where misclassifying a low-risk child as high-risk can expose an infant to unnecessary immunosuppression. A retrospective association, however biologically plausible, does not establish that the index adds predictive information beyond what is already captured by age, sex, days of fever at diagnosis, baseline laboratory values and serial echocardiography. Establishing incremental value requires prospective cohorts, ideally across multiple populations, in which the AISI is measured at diagnosis and tested against coronary outcomes in models that already include established predictors.

Several technical considerations will shape that validation. The AISI’s components are not static: platelet counts in Kawasaki disease typically rise in the second week of illness, so the timing of the blood draw relative to fever onset materially changes the value. Thrombocytosis may be absent in the first days, precisely when treatment decisions must be made, which raises the question of whether an early AISI retains discriminative power. Related composite indexes built from the same count, such as ratios pairing platelets, neutrophils or monocytes with lymphocytes, have been studied in other inflammatory and surgical settings, and head-to-head comparisons would clarify whether the full four-component product offers genuine added sensitivity. Albumin, a negative acute-phase reactant that has been linked to coronary lesion progression in retrospective cohorts, and other conventional markers such as C-reactive protein, provide further benchmarks against which any new index must prove itself. Standardization of cutoffs, which vary across studies and populations, is another unresolved hurdle.

The broader significance of the commentary lies in its vision of low-cost, widely accessible risk stratification. The complete blood count is among the cheapest and most universally available tests in medicine, including in the low- and middle-income settings where Kawasaki disease is often diagnosed late and where advanced imaging may be scarce. An index that extracts additional prognostic information from data already in hand, without new reagents, equipment or expense, aligns squarely with the needs of global pediatric practice. If prospective studies confirm that the AISI meaningfully refines aneurysm risk prediction, it could be incorporated into the early triage of children at diagnosis, flagging those who warrant intensified anti-inflammatory therapy, more frequent echocardiography or referral to specialized centers. Conversely, if validation fails, the episode will still have sharpened the field’s understanding of how cellular inflammation relates to coronary pathology.

For now, the message from Saad and colleagues is one of disciplined optimism. The Aggregate Index of Systemic Inflammation is not a new test but a new lens on an old one, and its candidacy in Kawasaki disease rests on a coherent mechanistic rationale and an emerging body of evidence rather than on settled proof. The authors declare no competing interests and received no external funding for the commentary, and they frame the index explicitly as a potential complement to, not a substitute for, established clinical and echocardiographic assessment. Children with Kawasaki disease will continue to be managed according to current guidelines, with intravenous immunoglobulin, aspirin and serial coronary imaging. But the commentary signals that the next chapter of aneurysm risk prediction may be written not with expensive biomarkers or novel imaging, but with a few numbers multiplied together from a blood count that every febrile child already receives.

Subject of Research: Evaluation of the Aggregate Index of Systemic Inflammation as a marker of coronary aneurysm risk in Kawasaki disease

Article Title: Inflammatory burden and coronary aneurysm risk in Kawasaki disease: appraising the Aggregate Index of Systemic Inflammation

Article References: Saad, K., Aly, S. E., Zeid, S. S. K., Ahmad, A. R., Elfarargy, M. S., Elkossi, M., Mumtaz, M. H., & Temsah, M.-H. (2026). Inflammatory burden and coronary aneurysm risk in Kawasaki disease: appraising the Aggregate Index of Systemic Inflammation. Pediatric Research. https://doi.org/10.1038/s41390-026-05590-4

Image Credits: AI Generated

DOI: 10.1038/s41390-026-05590-4

Keywords: Kawasaki disease, coronary artery aneurysm, Aggregate Index of Systemic Inflammation, AISI, inflammation, pediatric cardiology, risk stratification, complete blood count, platelets, echocardiography, intravenous immunoglobulin, biomarkers

Clonal haematopoiesis occurs when a blood stem cell acquires a somatic mutation and its descendants expand to form a detectable clone, often without any symptoms. Population sequencing shows it in as many as one in five adults over 65, yet transformation to an overt myeloid neoplasm such as acute myeloid leukaemia or myelodysplastic syndrome happens at only 0.5 to 1.0 percent per year. The clinical risk landscape is starkly uneven. According to the clonal haematopoiesis risk score, or CHRS, only about one percent of carriers detected in population-scale sequencing fall into the high-risk category, facing a 10-year blood cancer incidence of 52 percent, while roughly 90 percent are low risk with a 10-year incidence below one percent. Sequencing everyone, therefore, would generate enormous volumes of clinically meaningless results.

The research team, led by investigators drawing on the UK Biobank, hypothesised that the molecular behaviour of dangerous clones leaves fingerprints in ordinary blood measurements. Mutant stem cells alter the production, size, and morphology of red cells, white cells, and platelets, and those changes are captured in the two dozen parameters reported with every complete blood count. To test the idea, they assembled an enormous training dataset of 369,260 UK Biobank participants aged 40 to 69 who had both whole-exome sequencing and blood counts drawn on the same day, then held back 92,316 participants for internal testing.

Defining the target was itself a technical exercise. The team computed a molecular clonal haematopoiesis score for each person, summing points for high-risk mutations in genes such as TP53, JAK2, SRSF2, and RUNX1, for a solitary DNMT3A mutation, for multiple mutations, and for a variant allele fraction of 20 percent or more. A score above 4 captured nearly everyone who later developed myeloid neoplasia or who was high or intermediate risk by CHRS. Participants were then labelled positive if they scored above the threshold and developed blood cancer, negative if they scored at or below it, and equivocal if they scored high but remained cancer-free during follow-up.

The model itself is a balanced random forest, a tree-based classifier chosen after comparing nine architectures on a resampled training set of 3,000 individuals, 1,000 per class. Synthetic minority over-sampling and random under-sampling corrected the extreme class imbalance, since true positives made up less than one tenth of one percent of the cohort. Feature selection was aggressive: highly correlated parameters such as white cell count and haematocrit were removed, the Boruta algorithm discarded sex and several less informative measures, and features inconsistently annotated across datasets were dropped. The final inputs are strikingly mundane: age plus absolute counts of red cells, lymphocytes, monocytes, neutrophils, eosinophils, reticulocytes, and platelets, along with haemoglobin, red cell distribution width, platelet crit, mean corpuscular volume, and giant platelets.

Performance in the UK Biobank test set was strong where it mattered most. The area under the receiver operating characteristic curve for the positive class reached 0.89, and sensitivity for high-risk clonal haematopoiesis was 0.73, with the three missed cases all under 60 with entirely normal blood counts. More strikingly, positive classification captured 100 percent of the 42 individuals deemed high risk by CHRS, and the overall negative predictive value was 0.993, meaning a negative result reliably rules out clinically significant clonal haematopoiesis. The model was intentionally permissive of false positives, accepting that sequencing would resolve them, and achieved one detected high-risk case for every 247 people flagged, compared with one per 2,500 under a screen-everyone strategy.

External validation demonstrated genuine portability. In a clinical cohort of 2,233 patients evaluated in Dana-Farber haematology clinics, where cytopenias were far more common, PRISM correctly classified 16 of 18 positive-label individuals, an AUC of 0.74 and sensitivity of 0.89, with a negative predictive value of 0.968. In the All of Us Research Program cohort of 14,674 Americans, the positive-class AUC was 0.87 with perfect sensitivity across ten cases and a negative predictive value of 0.998. Across all three cohorts, every CHRS-defined high-risk individual received a positive PRISM label, and intermediate-risk carriers almost never received a negative one. Benchmarking against the clonal cytopenia risk score showed that 18 of 19 high-risk clonal cytopenia cases were also flagged as positive.

The downstream consequences were equally telling. Among UK Biobank participants with clonal haematopoiesis who later developed myeloid neoplasia, 63 percent had been labelled positive and 25 percent equivocal, and the misclassified negatives were overwhelmingly low risk by post-sequencing CHRS. Ten-year cumulative incidence of myeloid neoplasia was 1.44 percent in the positive class versus 0.11 percent in the negative class, a hazard ratio of 8.81 that outperformed age, cytopenia, mean corpuscular volume, and red cell distribution width in multivariable models. Because clonal haematopoiesis also drives cardiovascular disease, the team checked whether negative labels would exclude people at cardiac risk; reassuringly, 10-year cardiovascular risk was lowest, at 4.6 percent, in the PRISM-negative group, compared with 15.7 percent in the positive class.

The authors are careful about scope. PRISM is a pre-sequencing triage tool, not a diagnostic test, and its equivocal class, which includes people with abnormal counts but no dangerous clones, cannot be resolved from a single blood draw; serial sampling and clinical correlation remain essential. The UK Biobank’s predominantly European ancestry limits generalisability, and specialised settings such as chemotherapy exposure or inherited bone marrow failure would need context-specific recalibration. Still, if prospective validation succeeds, the implications are considerable: haematology clinics could reassure worried patients with a negative label, transplant programmes could screen donors cheaply before sequencing, and health systems could cut unnecessary molecular testing by 14 to 55 percent while sparing low-risk individuals the psychological and financial burden of overdiagnosis. A model built from the humble blood count may become the gatekeeper for the genomic era of blood cancer prevention.

Subject of Research: A machine learning model using complete blood count parameters to prioritise sequencing for high-risk clonal haematopoiesis

Article Title: Derivation of a blood count-based machine learning model to prioritise sequencing for high-risk clonal haematopoiesis: a retrospective cohort study

Article References: Nandi, R., Zhao, K., Pershad, Y., Khondoker, N. N., Vlasschaert, C., Uddin, M. M., Natarajan, P., DeAngelo, D. J., Luskin, M. R., Lindsley, R. C., Uno, H., Shimony, S., Bick, A. G., & Weeks, L. D. (2026). Derivation of a blood count-based machine learning model to prioritise sequencing for high-risk clonal haematopoiesis: a retrospective cohort study. eClinicalMedicine, 100, Article 104238. https://doi.org/10.1016/j.eclinm.2026.104238

Image Credits: AI Generated

DOI: 10.1016/j.eclinm.2026.104238

Keywords: clonal haematopoiesis, machine learning, complete blood count, myeloid neoplasia, leukaemia risk, CHRS, UK Biobank, risk prediction, sequencing, premalignant conditions, cardiovascular disease, precision medicine

The commentary, published on 28 September 2026, centers on a study by Li and colleagues that examined the systemic immune-inflammation index, or SII, in more than a thousand young children. The SII is deceptively simple: it is calculated by multiplying the neutrophil count by the platelet count and dividing by the lymphocyte count, all values that already appear on every standard blood test ordered in a pediatric clinic. The index was originally developed in oncology, where it predicted prognosis in patients undergoing curative resection for hepatocellular carcinoma, on the logic that the relative balance of neutrophils, platelets, and lymphocytes captures the body’s overall inflammatory state better than any single cell type alone.

In their retrospective case-control study of 1,028 children aged zero to six years, Li and colleagues found that children with allergic rhinitis had significantly higher SII values than controls. The prevalence of allergic rhinitis rose steadily across quartiles of the index, and SII showed a positive association with total serum IgE, the antibody class that sits at the heart of allergic sensitization. Taken together, these findings suggest that even in preschool-aged children, allergic rhinitis is accompanied by a measurable shift in circulating immune cells that extends well beyond the nasal lining.

The most provocative result, however, is what the study did not find. SII correlated with IgE but not with eosinophil count or eosinophil percentage. This is a striking dissociation, because eosinophils have long been treated as the signature cell of allergic disease, the granulocyte whose recruitment to tissues and appearance in blood has anchored decades of allergy research and drug development. If a systemic inflammatory index can track the allergic state in young children without tracking eosinophils, it implies that other circulating players, including neutrophils, platelets, and lymphocytes, carry information about the disease that the eosinophil count alone misses.

Building on that dissociation, the authors of the study proposed a high-SII, low-eosinophil subgroup of pediatric allergic rhinitis and went on to evaluate an index combining IgE and SII as a way of identifying this phenotype. The commentary’s authors highlight this as the study’s most conceptually important contribution: the suggestion that pediatric allergic rhinitis is not one immunological entity but a set of phenotypes, some of which may be driven more by broad systemic inflammatory activity than by the canonical type 2, eosinophil-centered pathway. If validated, such subtyping could eventually help clinicians decide which children might respond to which therapies, from antihistamines to biologics targeting IgE or type 2 cytokines.

The idea that allergic rhinitis has systemic consequences is not new. More than two decades ago, Togias argued that the condition produces circulating eosinophilia and basophilia, persistent IgE responses after allergen provocation, and mechanistic links between inflammation in the upper and lower airways, helping to explain the tight epidemiological relationship between rhinitis and asthma. Since then, advances in allergic disease biology have pushed the field further away from a purely local mucosal model, implicating epithelial barrier dysfunction, innate immune activation, adaptive T cell skewing, and humoral IgE responses in a network of pathways that engage the whole organism rather than a single organ.

What is genuinely new is the prospect of capturing that systemic dimension cheaply and non-invasively. Allergic rhinitis is one of the most common chronic diseases of childhood, and its diagnosis remains anchored in clinical history, physical examination, and evidence of allergic sensitization. A biomarker derived from an ordinary blood count would not replace that clinical assessment, but it could offer an objective, quantitative dimension to a diagnosis that is otherwise symptom-based. It could also enable longitudinal tracking: if SII falls as disease activity falls, or rises before symptoms flare, clinicians might one day use it to monitor response to treatment in the same way inflammatory markers are used in other chronic conditions.

The commentary’s authors are careful to frame these findings as a beginning rather than an endpoint. The underlying study was retrospective and case-controlled, which establishes association but not causation or prediction over time. An elevated SII is not specific to allergic rhinitis; the index reflects general systemic inflammation and can be influenced by infections, obesity, and other conditions common in young children. Whether the high-SII, low-eosinophil phenotype represents a stable biological subgroup, a transient inflammatory state, or a marker of disease severity remains to be determined through prospective cohort studies that follow children over time and correlate the index with symptoms, treatment response, and the development of comorbidities such as asthma.

Nevertheless, the conceptual shift the commentary describes is significant. Recent years have seen a wave of interest in immune-inflammation markers in allergic rhinitis, with other groups tracking systemic indices as predictors of disease activity and treatment response in both adults and children. The convergence of this work points toward a future in which the humble complete blood count, an investigation performed millions of times a day in hospitals around the world, becomes a tool for stratifying allergic children by their underlying inflammatory biology. That would be a striking example of old data yielding new insight, extracted not through a novel assay but through a new way of reading numbers clinicians already possess.

For the parents of the millions of children who sneeze and sniff their way through allergy season, the immediate practical impact of this research is limited, and no one is suggesting a blood test in place of a clinical evaluation. But the scientific message is clear and increasingly hard to ignore: allergic rhinitis in young children is a systemic inflammatory disorder, its biology cannot be reduced to eosinophils, and the tools to see that bigger picture may already be sitting in the laboratory information system. As the Toronto commentary makes plain, the field is moving beyond the nose, and beyond the eosinophil, toward an integrated view of childhood allergy as a whole-body condition with measurable, and potentially actionable, systemic signatures.

Subject of Research: Systemic immune-inflammation index and IgE associations in pediatric allergic rhinitis

Article Title: Beyond eosinophils: systemic inflammation in pediatric allergic rhinitis

Article References: Medeleanu, M. V., Fairfield, O., & Upton, J. E. M. (2026). Beyond eosinophils: systemic inflammation in pediatric allergic rhinitis. Pediatric Research. https://doi.org/10.1038/s41390-026-05524-0

Image Credits: AI Generated

DOI: 10.1038/s41390-026-05524-0

Keywords: allergic rhinitis, pediatrics, systemic immune-inflammation index, IgE, eosinophils, inflammation biomarkers, complete blood count, allergic sensitization, immune phenotyping, Pediatric Research, childhood allergy, neutrophils

The research, led by Mirko Di Rosa and Luca Soraci of IRCCS INRCA in Cosenza, Italy, together with colleagues at the University of Calabria and partner institutions, pooled individual-level data from four prospective European cohorts of older adults. The cohorts were deliberately heterogeneous: 812 community-dwelling older people from the ECHA-PC-MUSA studies, 813 nursing home residents from the SADEL cohort, 579 long-lived individuals and area-matched controls from the Genetics of Healthy Ageing project, and 4,400 hospitalized patients from the REPORTAGE project. In total, 6,604 people were followed for up to five years, and the mean age of the pooled sample was 84.8 years, with a standard deviation of 9.2 years. That breadth of settings is precisely what makes the analysis unusual; most previous studies of RDW and mortality have examined a single clinical population, leaving open the question of whether the marker truly travels across the diverse worlds of home, hospital, and long-term care.

Technically, RDW is reported as a coefficient of variation, the standard deviation of red cell volume divided by the mean corpuscular volume, multiplied by one hundred. It is generated automatically by automated haematology analysers at no additional cost, which is part of its appeal as a potential screening tool. In this study the primary exposure was an RDW-CV of 15 percent or higher, a threshold commonly used to flag anisocytosis, the technical term for increased variability in red cell size. Among the European participants, 41.4 percent met this criterion, and over the five-year follow-up period 69.9 percent of the cohort died, a striking figure that reflects the advanced age and clinical vulnerability of the populations studied.

To test whether elevated RDW was an independent predictor of death rather than a bystander marker of known risk factors, the investigators used multilevel Cox proportional hazards regression, treating the cohort of origin as a random effect. The models adjusted for a formidable battery of geriatric prognostic factors: age, sex, the presence of anaemia, serum albumin, the neutrophil-to-lymphocyte ratio as an index of systemic inflammation, the number of chronic diseases and medications, and impairment in activities of daily living. The models also pre-specified stratification at age 85, allowing the association to be examined separately in the oldest-old, a group in which mortality prediction is notoriously difficult because standard risk factors often lose discriminatory power.

The headline result was unambiguous. Older adults with RDW-CV at or above 15 percent had a 58 percent higher hazard of death during follow-up than those below the threshold, with a pooled adjusted hazard ratio of 1.58 and a 95 percent confidence interval running from 1.48 to 1.68, a finding highly statistically significant at p less than 0.001. Just as important as the magnitude was the consistency: between-cohort heterogeneity, quantified by the theta statistic, was minimal at 0.06, and the direction of the association was concordant in three of the four cohorts. In epidemiological terms, this suggests that the relationship is not an artefact of any single population or care setting but a robust phenomenon that persists whether the blood was drawn in a nursing home, a hospital ward, or a community clinic.

Replication is where many biomarker studies falter, so the team took the additional step of validating their findings externally in an entirely different healthcare system. They analysed 2,471 participants from the 2016 wave of the Health and Retirement Study, a large, nationally representative survey of older Americans. Once again, the elevated-RDW group showed significantly higher mortality, with a hazard ratio whose 95 percent confidence interval extended from 1.23 to 2.19 and a p value below 0.001. The fact that the association survived in a transatlantic sample, with different ancestry profiles, laboratory platforms, and care structures, strengthens the case that RDW captures something fundamental about biological risk rather than a local idiosyncrasy of Italian geriatric medicine.

Why should the variability of red blood cell size carry information about survival? The most widely discussed mechanism implicates chronic, low-grade systemic inflammation and its effects on erythropoiesis, the production of red cells in the bone marrow. Inflammatory cytokines disturb iron metabolism, shorten red cell lifespan, and disrupt the careful regulation of cell volume during maturation, producing a wider distribution of cell sizes. Elevated RDW has also been linked to oxidative stress, nutritional deficiencies, renal dysfunction, and subtle derangements of haematopoietic stem cell dynamics. In this sense, the authors argue, RDW may act as an integrative readout of physiological dysregulation, a single number that summarises the accumulated wear on multiple organ systems. Its independence from the neutrophil-to-lymphocyte ratio and albumin in the adjusted models suggests it is not simply proxying inflammation or malnutrition but adds a distinct dimension of prognostic information.

The association was strongest in the oldest-old, participants aged 85 and above, which is where the clinical implications become most interesting. Predicting mortality in very old patients is a persistent challenge for geriatric medicine; chronological age alone is a blunt instrument, and comprehensive geriatric assessment, while valuable, is time-consuming and requires trained personnel. A marker that costs nothing extra, is already reported on every blood count, and performs at a fixed cut-off across community, nursing home, and hospital populations could help clinicians stratify risk quickly at the bedside. It might, for example, inform decisions about the intensity of investigations, the appropriateness of aggressive interventions, or the prioritisation of patients for closer monitoring. The researchers are careful, however, to frame RDW as a prognostic indicator, not a deterministic verdict; a hazard ratio of 1.58 describes average risk at the population level, and individual outcomes will vary widely.

Several caveats temper the enthusiasm. This is an observational analysis, so elevated RDW identifies association rather than proven causation, and there is as yet no evidence that lowering RDW would change survival. The exposure was measured at a single time point, leaving open questions about the value of tracking RDW trajectories over time, and although the adjustment set was unusually comprehensive for a biomarker study, residual confounding can never be fully excluded. The predominance of Italian cohorts in the pooled European sample also invites replication in other national contexts, even given the American validation. Nonetheless, the study clears a high bar: a fixed, pre-specified threshold, individual-level pooling across markedly different care settings, adjustment for the major geriatric prognostic factors including anaemia, functional status, inflammation, and nutrition, minimal heterogeneity, and external replication in a second continent. Few routinely available laboratory measures can claim a comparable evidentiary pedigree.

The broader lesson may be about where medicine should look for its next generation of risk markers. As automated analysers quietly compute dozens of derived indices with every routine sample, some of the most informative signals may already be sitting in electronic health records, unexamined. Red cell distribution width, a number most patients have never heard of, now has credible evidence behind it as an independent predictor of five-year mortality in older adults across the full spectrum of care, from independent living to intensive hospital treatment. For a rapidly ageing global population, and for clinicians searching for cheap, scalable tools to guide geriatric care, a modest column of digits on a standard blood report may prove to be one of the most consequential numbers in the chart.

Subject of Research: Red cell distribution width as an independent predictor of five-year all-cause mortality in older adults across diverse care settings

Article Title: Red cell distribution width and 5-year mortality across older adult care settings: a pooled European cohort analysis with US validation

Article References: Di Rosa, M., Aceto, M. A., Paparazzo, E., Serra Cassano, T., Mercatante, P., De Rango, F., Muglia, L., Cozza, A., Chinigò, C., Passarino, G., Cherubini, A., Corsonello, A., Sarzani, R., Antonicelli, R., Pelliccioni, G., Galeazzi, R., Gembillo, G., Santoro, D., Soraci, L., … Lattanzio, F. (2026). Red cell distribution width and 5-year mortality across older adult care settings: a pooled European cohort analysis with US validation. Journal of Translational Medicine. https://doi.org/10.1186/s12967-026-08997-z

Image Credits: AI Generated

DOI: 10.1186/s12967-026-08997-z

Keywords: red cell distribution width, RDW, mortality, aging, geriatrics, complete blood count, oldest old, prognosis, cohort study, anaemia, inflammation, biomarker

The research, conducted by Junxiu Liu of the Department of Child Health Care at Qianjiang Central Hospital of Chongqing, Yue Peng of the Department of Cardiovascular Medicine at the Children’s Hospital of Chongqing Medical University, and Dan Zhou of the Department of Respiratory Medicine at Qianjiang Central Hospital of Chongqing, enrolled 345 consecutive children with Kawasaki disease and examined them during the acute phase of illness. Among this cohort, 64 patients went on to develop a coronary artery aneurysm, a rate that reflects the stubborn reality that even with standard intravenous immunoglobulin therapy, a meaningful minority of affected children still sustain vascular injury. The team’s goal was not to invent a new biomarker from scratch but to test whether a composite index derived from ordinary laboratory parameters could outperform intuition and serve as a practical risk stratification instrument.

The Aggregate Index of Systemic Inflammation, commonly abbreviated AISI, is calculated by multiplying the neutrophil, platelet, and monocyte counts and dividing the product by the lymphocyte count. Each of these cellular components captures a different facet of the inflammatory response that defines Kawasaki disease. Neutrophils rise early in the illness and are implicated in the degradation of the extracellular matrix that weakens arterial walls; monocytes and their derived macrophages infiltrate the coronary vasculature and fuel the vasculitis at the heart of the disease; platelets surge in the convalescent phase and participate in the endothelial activation and prothrombotic state that characterize the disease’s vascular complications. Lymphocytes, by contrast, tend to be depleted during vigorous systemic inflammation, so a low lymphocyte count amplifies the index. By combining all four parameters, the AISI effectively integrates the intensity of innate immune activation, the prothrombotic tendency, and the immune dysregulation visible in a single routine blood draw.

Using univariable comparisons followed by multivariable logistic regression, the investigators identified three independent factors associated with coronary artery aneurysm formation. Younger age emerged as protective against the complication per unit increase in years, with an odds ratio of 0.980 and a 95 percent confidence interval of 0.966 to 0.985, a statistically significant association with a p value of 0.008. Lower serum albumin was also linked to aneurysm risk, with an odds ratio of 0.937 per unit increase and a confidence interval spanning 0.880 to 0.997, significant at p equal to 0.040. Albumin declines when inflammation becomes severe and vascular permeability increases, so hypoalbuminemia serves as a chemical mirror of inflammatory intensity and capillary leak. Most notably, the AISI itself, analyzed per 100-unit increment, carried an odds ratio of 1.045 with a 95 percent confidence interval of 1.003 to 1.089 and a p value of 0.036, indicating that each step up in the index independently raised the odds of an aneurysm even after accounting for age and nutritional markers.

To translate the continuous index into a clinically usable threshold, the team performed an exploratory analysis that yielded a cutoff value of 711.58. When the cohort was divided at this point into a low AISI group and a high AISI group, the difference in outcomes was striking: children in the high AISI group experienced coronary artery aneurysms at a significantly higher rate than those in the low AISI group, with a p value below 0.001. In practical terms, this means a pediatrician staring at a standard complete blood count on the day of admission could compute the index, compare it against the exploratory threshold, and obtain an early, data-driven estimate of the child’s likelihood of developing coronary involvement before echocardiographic changes become apparent.

The biological plausibility of the finding rests on the pathogenesis of Kawasaki disease itself. The condition is an acute, systemic vasculitis that preferentially targets medium-sized arteries, and the coronary arteries are particularly vulnerable, especially in infants and toddlers. Histopathological studies have long described panvasculitis with edema of the intima, infiltration of inflammatory cells, and fragmentation of the internal elastic lamina, processes that together permit the outward expansion of the vessel wall into an aneurysm. Oxidative stress, endothelial dysfunction, and platelet-mediated vascular inflammation have all been implicated in the progression from acute inflammation to structural arterial damage. Because neutrophils, monocytes, and platelets each participate in these cascades, a composite index that multiplies their concentrations provides a running tally of the inflammatory and thrombotic forces acting on the coronary wall, whereas lymphopenia signals the severity of the systemic immune disturbance.

The study builds on a growing literature linking derived inflammatory indices to cardiovascular outcomes in adult medicine. The AISI and its cousins, such as the systemic immune-inflammation index and the systemic inflammation response index, have been associated with clinical outcomes in acute myocardial infarction, with coronary artery calcification, with acute coronary syndrome in patients with and without diabetes, and with all-cause and cardiovascular mortality in female cancer patients, and the index has even been proposed as a predictor of mortality in stroke. Extending this framework to pediatric vasculitis represents a conceptual bridge: the same arithmetic that quantifies inflammatory burden in adult atherosclerotic disease may also capture the vasculitic intensity that drives coronary dilatation in children. Previous risk models for coronary aneurysms in Kawasaki disease, including a validated North American risk score, have relied on combinations of demographic, clinical, and laboratory variables; the appeal of the AISI lies in its simplicity and its availability at no additional cost.

The clinical implications, according to the authors, are potentially significant. The study’s impact statement notes that early identification based on the Aggregate Index of Systemic Inflammation may assist clinicians in implementing more aggressive treatment strategies, such as intravenous immunoglobulin combined with corticosteroids, thereby improving outcomes for Kawasaki disease patients. This matters because the standard regimen of intravenous immunoglobulin and aspirin fails to prevent coronary aneurysms in a substantial fraction of patients, and children who resist the first dose face difficult decisions about second-line agents, including repeat immunoglobulin, infliximab, or corticosteroids. Trials such as the KIDCARE study, which compared infliximab with a second dose of intravenous immunoglobulin in resistant cases, have refined the therapeutic arsenal, but those agents are typically deployed only after treatment failure is evident. A biomarker available on day one that flags high-risk children could shift that decision earlier, before the vascular injury has begun to accrue, which is precisely the window in which intensified therapy is believed to be most effective.

The authors are careful to frame their findings within the study’s design. The work was retrospective, drawing on records from consecutive patients at their institutions, and the cutoff of 711.58 is explicitly described as exploratory, meaning it will require validation in independent, ideally prospective, and ethnically diverse cohorts before it can be embedded in clinical guidelines. Albumin and age also entered the final model, suggesting that the AISI is best understood as one component of a broader risk picture rather than a stand-alone verdict. Nevertheless, the strength of the association, the significance across multiple independent variables, and the stark difference in aneurysm rates between high and low index groups give the finding real weight. The datasets generated in the study are available from the corresponding author on reasonable request, and the work was approved by the hospital’s Research Ethics Committee with waived informed consent, with no competing interests declared.

For a disease that was first described more than half a century ago and that continues to puzzle researchers with an unclear cause, the prospect of extracting actionable risk information from a routine blood count is a reminder that progress does not always require new molecules or expensive assays. Sometimes it requires looking harder at the data already in hand. If future validation confirms the exploratory threshold identified by Liu, Peng, and Zhou, the Aggregate Index of Systemic Inflammation could become a standard part of the acute Kawasaki disease assessment, helping clinicians worldwide decide within hours which febrile child needs the most aggressive anti-inflammatory therapy, and which can be managed with confidence under the conventional regimen. In the meantime, the study adds Kawasaki disease to the expanding list of inflammatory cardiovascular conditions for which a four-cell arithmetic snapshot of the immune system carries genuine prognostic power.

Subject of Research: The association between the Aggregate Index of Systemic Inflammation and coronary artery aneurysm formation in children with acute Kawasaki disease.

Article Title: Association of the Aggregate Index of Systemic Inflammation with coronary artery aneurysm in Kawasaki disease patients

Article References: Liu, J., Peng, Y., & Zhou, D. (2026). Association of the Aggregate Index of Systemic Inflammation with coronary artery aneurysm in Kawasaki disease patients. Pediatric Research. https://doi.org/10.1038/s41390-026-05500-8

Image Credits: AI Generated

DOI: 10.1038/s41390-026-05500-8

Keywords: Kawasaki disease, coronary artery aneurysm, Aggregate Index of Systemic Inflammation, AISI, pediatric cardiology, vasculitis, inflammation biomarker, complete blood count, intravenous immunoglobulin, serum albumin, risk prediction, children