The study included 1,255 adults: 251 people with lung cancer and 1,004 controls. Researchers fitted a multivariable logistic regression model and assessed internal performance with 10-fold cross-validation and 2,000 bootstrap resamples. The abstract reports that performance was consistent across the bootstrap resampling and cross-validation analyses.

The final model included eight predictors: age, sex, household solid fuel use, asbestos exposure, years of smoking, years of exposure to secondhand smoke, prior tuberculosis, and family history of cancer. Investigators also derived a simplified, point-based risk score intended to improve practical applicability. The abstract states that the score had discriminatory performance and classification accuracy comparable with the full model. The supplied abstract does not provide confidence intervals for the reported performance measures.

These findings describe a prediction model and its performance within the studied population, rather than evidence that using the tools improves clinical outcomes or produces benefits from screening or prevention interventions. The case-control design and internal validation limit the conclusions that can be drawn about performance in other populations. External validation is required to establish generalizability. The authors further state that recalibration to population-level incidence would be needed if the model is used to estimate absolute risk or define population-based clinical or public health thresholds. The article was published in BMC Cancer as open-access research; the version supplied was identified as subject to further edits before replacement by the final Version of Record.

Subject of Research: Internal validation of a locally developed lung cancer risk prediction model in Ethiopia

Article Title: Development and validation of a lung cancer risk prediction model in Ethiopia

Article References: Estifanos, N., Egata, G., Addissie, A., Yusuf, H., Bitew, A., Getachew, A., & Deyessa, N. (2026). Development and validation of a lung cancer risk prediction model in Ethiopia. BMC Cancer. https://doi.org/10.1186/s12885-026-16997-x

Image Credits: AI Generated

DOI: 10.1186/s12885-026-16997-x

Keywords: lung cancer, risk prediction, case-control study, internal validation, logistic regression, AUC, Ethiopia, risk score, Development, validation, lung, cancer

The cross-sectional analysis examined survey data from 27,651 U.S. adults whose average age was 53.1 years. Fifty-five point eight percent were female, 67.4% were non-Hispanic White, 6.3% had immunosuppression and 13.8% had one or more cancers. Researchers used descriptive and inferential statistics, including logistic regression analyses, with SAS version 9.4. They used 95% confidence intervals and a P-value threshold of .05 to determine statistical significance.

Adults with one or more cancers had an adjusted odds ratio of 1.31 for reporting influenza vaccination in 2021 compared with the comparison group (95% CI, 1.19–1.45; P < .001). Adults with a weak immune status also had higher adjusted odds, with an adjusted odds ratio of 1.47 (95% CI, 1.26–1.71; P < .001). Delay in care because an appointment was not available was associated with higher odds of vaccination, with an adjusted odds ratio of 1.13 (95% CI, 1.02–1.26; P = .02).

By contrast, reporting a delay in care because no doctor would accept the respondent’s health insurance was associated with lower adjusted odds of vaccination, with an adjusted odds ratio of 0.82 (95% CI, 0.70–0.97; P = .03). These findings describe adjusted associations in survey data and do not establish that cancer, immune status or healthcare access barriers caused vaccination decisions. The study’s reported scope is the U.S. adult population represented in the 2022 National Health Interview Survey, including people with immunosuppression and cancer; the supplied report does not provide results for additional populations or seasons. The authors concluded that the identified factors warrant intensified efforts to increase vaccination rates, particularly among vulnerable adults.

Subject of Research: Determinants of influenza vaccination among U.S. adults in the 2021/2022 season

Article Title: Determinants of influenza vaccination during the 2021/2022 influenza season in the United States: interactive roles of immunosuppression, cancer diagnosis, and healthcare access barriers

Article References: Adegbile, O. E., Kanmodi, K. K., Abib, O. M., & Brooks, B. (2026). Determinants of influenza vaccination during the 2021/2022 influenza season in the United States: interactive roles of immunosuppression, cancer diagnosis, and healthcare access barriers. BMC Public Health. https://doi.org/10.1186/s12889-026-29966-7

Image Credits: AI Generated

DOI: 10.1186/s12889-026-29966-7

Keywords: influenza vaccination, 2021/2022 influenza season, 2022 National Health Interview Survey, U.S. adults, cancer, immunosuppression, healthcare access, health insurance, logistic regression, Determinants, influenza, vaccination

The study enrolled patients hospitalized with pathologically confirmed intraductal breast lesions at one hospital from April 2025 to April 2026. The cohort included 132 benign cases and 68 malignant cases. Every participant underwent both conventional two-dimensional ultrasound and automated breast ultrasound, or ABUS.

Investigators compared ultrasonographic features between the benign and malignant groups, then applied univariate and multivariate logistic regression to identify independent risk factors and construct a combined diagnostic model. Investigators used receiver operating characteristic curves to compare the diagnostic efficiency of the combined model with the two imaging approaches evaluated separately. Malignant lesions differed from benign lesions in several ultrasound features.

The malignant group had significantly higher incidences of irregular shape, microcalcification, abundant blood flow, duct dilatation of at least 5 mm, a convergence sign, lesion extension toward the nipple, and marginal spiculation or angulation. Each of these seven associations was statistically significant at P < 0.05, and multivariate logistic regression identified them as independent risk factors for malignancy. The model’s imaging variables therefore captured several distinct aspects of lesion appearance: irregular morphology, the presence of microcalcification, blood flow classified as Adler grade II–III, duct dilatation of at least 5 mm, convergence of tissue features, extension toward the nipple, and spiculation or angulation at the margin.

The analysis linked their combined presence with the likelihood that a pathologically confirmed lesion was malignant. The results indicate that ABUS and conventional two-dimensional ultrasound together produced favorable diagnostic performance for preoperative benign–malignant differentiation of intraductal breast lesions. In the studied population, the combined approach performed better than either examination alone according to the reported area under the curve.

The study therefore supports the combined use of the two ultrasound methods as an imaging aid for this specific group of lesions, rather than establishing a general benefit across all types of breast abnormalities. Several features of the study limit how broadly the findings can be interpreted. It was conducted in a single hospital and included 200 hospitalized patients with pathologically confirmed intraductal breast lesions, including 68 malignant cases.

The model was developed and evaluated in that same single-center cohort, and the supplied results do not establish performance in other hospitals, screening populations, or patients with a wider range of breast findings. Because the study was limited to one institution and a relatively small malignant subgroup, the combined model’s diagnostic performance should be viewed as specific to the evaluated cohort. The reported AUC, sensitivity, and specificity quantify performance in this sample; they do not by themselves demonstrate that the model will maintain the same results in external populations.

Further evidence would be needed to determine how consistently the approach performs across different clinical settings. Overall, the study found that combining automated and conventional ultrasound produced the strongest reported discrimination between benign and malignant intraductal breast lesions. The analysis linked improved classification to seven ultrasound characteristics identified as independent risk factors in the cohort.

Its findings offer an imaging-based model for preoperative assessment in hospitalized patients with pathologically confirmed intraductal lesions, while the single-center design defines the boundary of the evidence.

Subject of Research: Preoperative ultrasound diagnosis of intraductal breast lesions

Article Title: Diagnostic value of combined automated breast ultrasound system and conventional two-dimensional ultrasound for intraductal breast lesions

Article References: Yang, Q., Yi, R., Guo, F., Gou, T., Cai, L., Wang, S., Zhang, Y., & Luo, Y. (2026). Diagnostic value of combined automated breast ultrasound system and conventional two-dimensional ultrasound for intraductal breast lesions. BMC Medical Imaging. https://doi.org/10.1186/s12880-026-02832-z

Image Credits: AI Generated

DOI: 10.1186/s12880-026-02832-z

Keywords: ABUS, breast ultrasound, intraductal lesions, breast cancer, diagnostic imaging, two-dimensional ultrasound, microcalcification, duct dilatation, logistic regression, diagnostic accuracy, Diagnostic, value

The analysis included 806 suspected patients whose records were reported between January 2023 and December 2024. PCR results classified 190 patients as positive. Researchers used univariate and multivariable logistic regression to evaluate which clinical and epidemiological factors were independently associated with laboratory-confirmed CCHF. This observational approach measures associations within the study population and does not establish that any factor caused infection.

Fever had the largest reported adjusted association with PCR-confirmed CCHF, producing an adjusted odds ratio of 13.70 and a 95% confidence interval of 3.25–58.82. Thrombocytopenia was also strongly associated with confirmation, with an adjusted odds ratio of 9.26 and a 95% confidence interval of 3.61–23.81. Both associations had P values below 0.001. The broad confidence interval for fever, particularly its lower bound of 3.25 and upper bound of 58.82, indicates uncertainty about the magnitude of the estimated association despite its statistical significance.

Bleeding at an injection site remained associated with confirmed CCHF after adjustment for other variables, although its adjusted odds ratio of 1.85 was smaller than those for fever and thrombocytopenia. Its 95% confidence interval was 1.05–3.25, with P = 0.002. Subcutaneous bleeding did not remain statistically significant after other variables were considered. Epidemiological exposure had an adjusted odds ratio of 8.13 (95% CI 5.15–12.66; P < 0.001), and urban residence had an adjusted odds ratio of 1.906 (95% CI 1.272–2.857; P = 0.002). These regional findings should not be treated as universal risk estimates for CCHF.

The investigators also evaluated a three-point screening score that assigned one point for each of three findings: fever, thrombocytopenia and injection-site bleeding. PCR positivity rose significantly with increasing score, from 0% among patients with a score of zero to 41.5% among those scoring three points (P < 0.001). Receiver operating characteristic analysis indicated modest discrimination, with an area under the curve of 0.643 and a 95% confidence interval of 0.601–0.684.

At a cutoff of two or more points, the score had a sensitivity of 98.4%, a specificity of 24.2% and a negative predictive value of 98.0% in the study population. These results show that nearly all PCR-positive patients were captured by the threshold, but the low specificity means that many patients who tested negative also met the two-point criterion. The negative predictive value was also conditional on the study population and its observed PCR-positive proportion, rather than representing a universal performance estimate.

The combined findings indicate that the three clinical signs were useful as indicators of PCR confirmation among people already suspected of having CCHF, but they were insufficient to distinguish confirmed infection reliably on their own. Epidemiological exposure was also strongly associated with confirmation in the multivariable analysis. The authors therefore characterized the proposed score as exploratory and stated that it requires further internal and external validation before routine clinical use.

Important limitations accompany the results. The study was retrospective and surveillance-based, drew on records from one Iraqi governorate, and included only suspected cases whose outcomes could be classified by PCR. The sample also included a substantial proportion of non-confirmed patients, with 190 positive results among 806 suspected patients. The modest area under the curve, low specificity and retrospective single-region setting constrain how directly the score can be transferred to other populations, healthcare settings or periods. The study therefore supports clinical prioritisation as a possibility for further evaluation, rather than demonstrating a validated stand-alone diagnostic rule.

Subject of Research: Clinical predictors of PCR-confirmed Crimean-Congo hemorrhagic fever among suspected cases in Thi-Qar, Iraq

Article Title: Beyond thrombocytopenia: clinical predictors of PCR-confirmed Crimean-Congo hemorrhagic fever in Thi-Qar, Iraq (2023–2024)

Article References: Hommod, S. A., Alasadi, R. A., Razzaq, A. K., & Roomi, A. B. (2026). Beyond thrombocytopenia: clinical predictors of PCR-confirmed Crimean-Congo hemorrhagic fever in Thi-Qar, Iraq (2023–2024). BMC Infectious Diseases. https://doi.org/10.1186/s12879-026-14617-4

Image Credits: AI Generated

DOI: 10.1186/s12879-026-14617-4

Keywords: Crimean-Congo hemorrhagic fever, CCHF, PCR confirmation, clinical predictors, thrombocytopenia, injection-site bleeding, screening score, surveillance study, Iraq, Beyond, clinical, predictors

The research team assembled a retrospective cohort of 255 patients with histologically confirmed non-small cell lung cancer, every one of whom had undergone systematic lymph node dissection, meaning that the true pathological status of their lymph nodes was known with certainty. This gold-standard ground truth is critical, because many previous prediction studies relied on sampling rather than complete dissection, leaving room for missed metastases to distort the results. The patients were randomly divided, with stratification by their pathological lymph node status to preserve the balance of positive and negative cases, into a training cohort of 178 patients and an internal held-out test cohort of 77 patients.

Before any machine learning began, the investigators screened candidate predictors using univariable and multivariable logistic regression within the training cohort. This filtering step matters because feeding an algorithm hundreds of weakly informative variables invites overfitting, the statistical equivalent of memorizing noise. What survived the screen was strikingly parsimonious: central tumor location, elevated carcinoembryonic antigen, elevated squamous cell carcinoma antigen, and solid nodule status on imaging. Perhaps the most surprising finding was what did not survive. None of the quantitative PET metabolic parameters, such as standardized uptake values that measure how avidly a tumor consumes radioactive glucose, remained independently associated with lymph node metastasis.

That negative result challenges a widespread intuition. Positron emission tomography combined with computed tomography is a cornerstone of modern lung cancer staging, and many clinicians assume that the metabolic numbers extracted from these scans carry independent predictive power. In this dataset, once clinical location, serological tumor markers, and the morphological character of the nodule were accounted for, the quantitative metabolic measurements added nothing. The study’s title emphasizes morphological features for good reason: whether a nodule appears solid, as opposed to partly ground-glass in texture, turned out to be one of the four pillars of prediction, a visual quality that radiologists assess routinely.

With four predictors in hand, the team trained six different machine-learning algorithms, optimizing each through repeated nested five-fold cross-validation. Nested cross-validation is the rigorous way to tune hyperparameters while keeping an honest estimate of performance, since the inner loop selects settings and the outer loop evaluates them on data not used for that selection. Among the six candidates, XGBoost, a gradient-boosted decision tree method that builds an ensemble of weak learners sequentially, each correcting the errors of its predecessors, achieved the highest cross-validated area under the receiver operating characteristic curve in the training cohort: 0.735, with a 95 percent confidence interval of 0.653 to 0.813. It was therefore designated the primary model.

The decisive test came when all six models, frozen and unmodified, were applied to the held-out test cohort of 77 patients. XGBoost delivered an AUC of 0.747, with a 95 percent confidence interval of 0.631 to 0.846, confirming that its training performance was not an artifact of overfitting. An AUC in this range represents moderate discrimination, meaning the model separates patients with and without nodal spread considerably better than a coin flip but far short of diagnostic certainty. The authors also compared the test-set AUCs of all six algorithms using paired DeLong tests with Holm correction for multiple comparisons, and found no statistically significant differences among them, a humbling result for anyone hoping that a sophisticated algorithm would dramatically outperform its rivals.

To probe that point further, the researchers added a conventional multivariable logistic regression model using the same four predictors as a simple reference. On the test cohort, this old-fashioned statistical workhorse achieved an AUC of 0.733, with a confidence interval of 0.612 to 0.841, and a DeLong test found no significant difference from XGBoost, with a P value of 0.301. Even more telling, the logistic regression model posted a Brier score of 0.152, compared with 0.198 for XGBoost, indicating that the simpler model’s predicted probabilities were actually better calibrated to observed outcomes. In plain terms, a transparent equation with four terms matched the flagship machine-learning algorithm in discrimination and beat it in calibration.

The primary XGBoost model was additionally evaluated for calibration, clinical utility, and feature contribution using SHAP, or SHapley Additive exPlanations, a technique borrowed from cooperative game theory that assigns each feature a quantified contribution to every individual prediction. This interpretability layer is increasingly considered essential for clinical machine learning, because a black-box risk score that physicians cannot interrogate is unlikely to earn trust or survive regulatory scrutiny. The study also reported a high negative predictive value, meaning that when the model predicted an absence of lymph node metastasis, that reassurance was usually correct, a property with real surgical relevance for selecting patients who might safely avoid extended nodal dissection.

Yet the authors are admirably candid about the limits. The model’s positive predictive value was limited, so a positive prediction could not reliably confirm metastasis, and invasive pathological confirmation remains indispensable. The absence of significant differences among the six algorithms, combined with the near-identical performance of plain logistic regression, suggests that the predictive ceiling here is set by the information content of the four features themselves, not by the sophistication of the learner. Moderate discrimination, a modest sample of 255 patients from a single center, and an internal test cohort rather than a truly external one all argue for caution.

The study’s real contribution may be methodological discipline rather than algorithmic novelty. It demonstrates a standardized, transparent workflow, from predictor screening through nested cross-validation to held-out evaluation, calibration analysis, and SHAP-based explanation, using only parameters available before surgery. The authors conclude that external multicenter evaluation is required before any clinical implementation, and that the results support cautious interpretation. If future multi-institutional validation confirms these findings, the payoff could be meaningful: a simple, interpretable risk score, computable from a CT report and two blood tests, that helps surgeons stratify lymph node metastasis risk in non-small cell lung cancer and decide, with greater confidence, who needs the most aggressive nodal surgery and who does not.

Subject of Research: Predicting lymph node metastasis in non-small cell lung cancer using clinical, serological, and PET/CT-derived features with interpretable machine learning

Article Title: Predicting lymph node metastasis in non-small cell lung cancer using clinical and PET/CT-derived morphological features: An interpretable machine-learning study

Article References: Xu, Y., Hao, J., Dou, Y., & Zeng, B. (2026). Predicting lymph node metastasis in non-small cell lung cancer using clinical and PET/CT-derived morphological features: An interpretable machine-learning study. PLOS One, 21(10), e0361043. https://doi.org/10.1371/journal.pone.0361043

Image Credits: AI Generated

DOI: 10.1371/journal.pone.0361043

Keywords: non-small cell lung cancer, lymph node metastasis, machine learning, XGBoost, PET/CT, logistic regression, SHAP, tumor markers, carcinoembryonic antigen, cancer staging, predictive modeling, cross-validation

The research team, led by Zhi Xu, Yongqi Shao, Zhijun Zhang, and Yonggui Yuan, assembled a cohort of 721 patients diagnosed with major depressive disorder according to DSM-IV criteria, alongside 960 healthy controls. All participants were of Chinese ancestry, a detail that matters enormously in psychiatric genetics, where most large-scale genome-wide association studies have historically been conducted in European populations. Genetic differences between ancestry groups can distort risk predictions, so testing polygenic approaches in East Asian cohorts is an important step toward clinical tools that work globally.

Polygenic risk scores, or PRS, are built by summing the small effects of many single nucleotide polymorphisms across the genome, each weighted according to its association with the disease as estimated in a large external study. In this work, the researchers drew base statistics from the Psychiatric Genomics Consortium’s East Asian datasets for major depressive disorder and schizophrenia, and from BioBank Japan for insomnia and body mass index. The idea was to test not only whether a depression-specific score predicts depression, but whether genetically related traits — sleep disturbance and adiposity among them — carry detectable risk signal in the same individuals.

Before any scoring took place, the team applied stringent quality control using PLINK v1.90, filtering variants on criteria such as minor allele frequency and genotyping quality. They then used PRSice v2.3.5 with its PRSet function to generate optimized scores across a range of p-value thresholds, a standard technique that tunes how permissive the score is in including weakly associated variants. Statistical comparisons between patients and controls were carried out with independent two-sample t-tests or Mann-Whitney U tests in R v4.3.0, and logistic regression models adjusted for age, sex, and ancestry principal components quantified how well the scores discriminated case status.

The headline result was clear: the depression-specific polygenic risk score differed significantly between patients and controls, with p less than 0.001. The logistic prediction model built on this score was statistically significant, achieving an empirical p-value of 9.999 × 10⁻⁵, though the pseudo-R² of 2.1 percent underscores how modest the explanatory power remains. In plain terms, the genetic score carries real signal, but depression is shaped by far more than inherited common variants — environment, life stress, and gene-environment interplay all contribute, and no current PRS comes close to serving as a standalone diagnostic.

Perhaps more striking was what happened with the cross-phenotype scores. Models built from schizophrenia, insomnia, and body mass index statistics failed to reach statistical significance in this cohort. That null result is informative rather than disappointing. It suggests that, at least in this Chinese sample and with these particular base datasets, the shared genetic architecture between depression and those related traits is not strong enough to yield transferable predictive signal — a reminder that genetic correlations reported in European-ancestry studies do not always replicate cleanly across populations and phenotype definitions.

The study’s most innovative contribution lies in its pathway-based approach. Rather than treating the genome as an undifferentiated mass of variants, the researchers used Gene Ontology gene sets drawn from the Molecular Signatures Database to construct pathway-specific PRS — scores restricted to variants within genes belonging to defined biological functions. This yielded 16 significant pathway scores in the Biological Process category, 13 in Molecular Function, and 12 in Cellular Component. Notably, most of these pathway PRS models showed greater explanatory power than the single-phenotype whole-genome score, hinting that biological organization concentrates signal that is otherwise diluted across the full genome.

Among the prioritized pathways, the study highlighted genes involved in processes such as uridine diphosphate-glucuronosyltransferase activity, retinoic acid signaling, hypothalamic-pituitary-adrenal axis regulation, and endoplasmic reticulum function — biological themes with plausible connections to stress response, neurodevelopment, and cellular stress handling in depression. The authors are careful to frame these as statistically prioritized candidates for future functional investigation rather than confirmed causal mechanisms, which is the appropriate scientific posture at this stage.

Methodologically, the work demonstrates a template that other groups can follow: rigorous quality control, threshold optimization, covariate-adjusted regression, and pathway-level aggregation using established gene set resources. The pathway approach effectively performs a gene set analysis on top of polygenic scoring, trading some statistical breadth for biological interpretability. When a pathway-level score outperforms the genome-wide score, it suggests that the true signal is concentrated in coherent biological modules — a finding that can guide laboratory work toward specific molecular machinery rather than leaving researchers with a diffuse list of risk loci.

The limitations are worth stating plainly. A pseudo-R² of 2.1 percent means the depression PRS explains only a small fraction of case-control status, and the cross-phenotype null results temper expectations for borrowing signal from related traits in this population. Larger East Asian discovery cohorts will be needed to strengthen base statistics, and replication in independent Chinese samples will be essential before any clinical translation. Still, by validating the depression PRS in a non-European cohort and demonstrating that pathway-based scoring can extract more explanatory power than conventional single-phenotype scores, the study charts a practical route toward genetically informed risk stratification for one of psychiatry’s most burdensome disorders.

Subject of Research: Polygenic risk score analysis of major depressive disorder in a Chinese cohort using single-phenotype and Gene Ontology pathway-based approaches

Article Title: Polygenic risk score analysis for major depressive disorder in a Chinese cohort: single-phenotype and pathway-based approaches

Article References: Xu, Z., Shao, Y., Li, Y., Liu, R., Tang, H., Cai, Y., Dang, X., Sun, J., Zhang, B., Chen, B., Zhang, Z., & Yuan, Y. (2026). Polygenic risk score analysis for major depressive disorder in a Chinese cohort: single-phenotype and pathway-based approaches. BMC Psychiatry. https://doi.org/10.1186/s12888-026-08589-7

Image Credits: AI Generated

DOI: 10.1186/s12888-026-08589-7

Keywords: polygenic risk score, major depressive disorder, genetics, psychiatry, Gene Ontology, pathway analysis, Chinese cohort, GWAS, Psychiatric Genomics Consortium, BioBank Japan, logistic regression, depression risk

The study was led by researchers at San Diego State University and UC San Diego Health Moores Cancer Center, who surveyed people who had received care at the cancer center between January 2018 and December 2019. Of 5,901 eligible individuals invited to complete an online survey in 2022, 954 responded, and after exclusions for missing data and stage 0 diagnoses, 652 respondents formed the analytic sample. Crucially, the team weighted these responses to reflect the demographics of the full cancer center patient population, producing a weighted sample of 4,620 survivors. The weighted population was 58.6 percent non-Hispanic White, 58.5 percent female, and more than half were aged 65 or older, with breast cancer and genitourinary cancers the most common diagnoses.

The methodological core of the study is straightforward but carefully executed. Participants answered a single question about whether they had used cannabis at any time since their cancer diagnosis, with use defined broadly to include THC-containing and CBD-containing products such as marijuana, concentrates, CBD-only products, prescription cannabinoids, edibles, and topical preparations. The researchers then examined three self-rated exposures: physical health, mental health, and overall quality of life, each captured on a scale from poor to excellent and dichotomized into above average versus average or below average. Because the three measures were moderately to strongly correlated, each was modeled separately using survey-weighted multivariable logistic regression, adjusted for age, sex at birth, race and ethnicity, cancer type, cancer stage, treatment status, and pre-diagnosis cannabis use.

The headline result concerns mental health. Survivors who rated their mental health as average or below average had significantly higher odds of post-diagnosis cannabis use, with an adjusted odds ratio of 1.88 and a 95 percent confidence interval of 1.13 to 3.11. In other words, after accounting for demographic and clinical differences, struggling mentally was associated with roughly an 88 percent increase in the odds of having used cannabis since diagnosis. The associations pointed in the same direction for the other two health domains but did not reach statistical significance: average or below average physical health yielded an odds ratio of 1.38, and average or below average quality of life an odds ratio of 1.24, both with confidence intervals crossing one.

The researchers also probed whether these relationships changed depending on clinical context, stratifying by cancer stage at diagnosis, treatment status, and whether the survivor had used cannabis before cancer. None of the formal interaction tests reached significance, meaning the team cannot claim the associations truly differ across these subgroups. Still, the stratified analyses were striking. Among survivors diagnosed with non-advanced disease, stages I or II, poorer mental health was associated with more than double the odds of cannabis use. The same held for survivors who had completed treatment, with an odds ratio of 2.01. Most eye-opening was the finding among people who had never used cannabis before their cancer diagnosis: those with average or below average mental health had more than four times the odds of starting cannabis afterward.

That last result raises an intriguing possibility that the authors address directly. A cancer diagnosis may serve as an initiation point for cannabis use among some survivors who had never previously touched the drug, potentially driven by a search for symptom relief. The authors caution, however, that this estimate was imprecise and could reflect other explanations, including how lifetime pre-diagnosis use was measured, whether some former users quit before diagnosis, simple reporting error, or residual confounding by factors the survey did not capture. Even with those caveats, the pattern fits a broader picture in which cannabis functions less as a wellness upgrade and more as a coping mechanism for distress.

The numbers behind who is using cannabis are themselves notable. In the weighted sample, 51.9 percent of survivors reported using cannabis before their cancer diagnosis, 32.0 percent reported using it after diagnosis, and another 14.6 percent said they had considered using it. Meanwhile, two-thirds of survivors rated their mental health and quality of life as above average, while only 47.1 percent said the same of their physical health. These figures sit within a larger American context: an estimated 18.6 million cancer survivors were living in the United States as of 2025, and cannabis use has grown steadily among adults with chronic health conditions as legalization has spread and perceptions of safety have softened.

There is an uncomfortable gap between enthusiasm and evidence. In previous work by the same team, more than 80 percent of surveyed survivors perceived benefits from cannabis, most commonly pain relief, reduced stress, anxiety, or depression, and better management of treatment-related side effects. Yet clinical guidelines tell a more reserved story. The American Society of Clinical Oncology’s guideline on cannabis and cannabinoids in adult cancer care concluded that the evidence is insufficient to recommend for or against cannabis for most cancer-related symptoms. The new study adds a useful observational data point: whatever survivors hope cannabis will do, those drawn to it are disproportionately those reporting mental distress, which may signal unmet symptom needs rather than proven therapeutic benefit.

The authors frame their findings through established behavioral science frameworks, including Symptom Management Theory and the Self-Medication Hypothesis. The former emphasizes that patients’ behavior is organized around the experience of symptoms and the pursuit of relief, while the latter proposes that people use substances to cope with persistent physical or psychological distress. Both fit the data: mental health emerged as the health domain most consistently tied to cannabis use, and the researchers suggest it may represent a target for clinical intervention, particularly because oncology clinicians may not know whether their patients are using cannabis at all. The pattern also aligns with the team’s earlier finding that survivors who perceived any of fifteen listed cannabis benefits had more than five-fold higher odds of use, while perceiving any risks cut the odds by 59 percent.

The study’s limitations deserve honest weight. Because the design is cross-sectional, cannabis use and self-rated health were measured at the same moment, so the direction of the relationship cannot be established and may well run in both directions. Cannabis use was self-reported and not verified through medical records or biological testing, and THC and CBD were lumped together despite being pharmacologically distinct compounds, obscuring any differential effects. Response rates were modest, selection bias is possible, and the single-institution setting in a state with legal medical and recreational cannabis may limit generalizability to regions with different laws. Still, the weighted sampling, the large clinical population, and the separation of physical health, mental health, and quality of life into distinct domains represent genuine advances over earlier work that treated quality of life as a single composite. The authors call for longitudinal studies to clarify timing, qualitative work to capture how survivors view cannabis within their coping strategies, and finer-grained measurement of products, frequency, and route of administration. As cannabis becomes ever more embedded in cancer culture, the question is shifting from whether survivors use it to why some do, some do not, and what their choices reveal about the gaps in supportive cancer care.

Subject of Research: Post-diagnosis cannabis use and its association with self-rated health status and quality of life among cancer survivors

Article Title: Post-diagnosis cannabis use and its association with health status and quality of life in cancer survivors

Article References: Kasiri, N., McDaniels-Davidson, C., Doran, N., Patel, S. P., Strong, D. R., & Parada, H., Jr (2026). Post-diagnosis cannabis use and its association with health status and quality of life in cancer survivors. Supportive Care in Cancer, 34(10), Article 931. https://doi.org/10.1007/s00520-026-11185-w

Image Credits: AI Generated

DOI: 10.1007/s00520-026-11185-w

Keywords: cannabis, cancer survivors, mental health, quality of life, physical health, THC, CBD, survivorship care, self-rated health, Supportive Care in Cancer, medical cannabis, logistic regression

The study, conducted by Marlena Jolanta Piekut of Warsaw University of Technology, draws on the Polish Household Budget Survey compiled by Statistics Poland, combining annual aggregated statistics spanning 2004 through 2024 with detailed household-level microdata covering more than 28,000 households in each of 2023 and 2024. Poland is a particularly revealing case: Eurostat data show Polish households devote one of the highest shares of total spending to energy in the European Union, at 7.6 percent, reflecting both the country’s aging housing stock and its continued reliance on coal and other solid fuels for heating.

The technical core of the analysis rests on three complementary methods. Trend analysis and income elasticity calculations tracked how the energy burden evolved over twenty years, revealing that elasticity exceeded one during two crisis periods, 2008 to 2011 and 2021 to 2022, peaking at 1.77. In those windows, energy spending grew faster than income, meaning households could not easily cut back on an essential good. During calmer years, elasticity hovered near zero or turned negative, consistent with energy behaving as a necessity whose relative budgetary weight shrinks as incomes rise.

Cluster analysis of the microdata then sorted households into distinct consumption profiles. Two dominant groups emerged in both survey years: an electricity-and-gas profile covering roughly 52 to 58 percent of households, skewed toward rural areas, lower incomes, and lower educational attainment, and a district-heating profile covering about 42 percent, concentrated in cities among wealthier, better-educated households. In 2023 only, a small third cluster appeared, comprising 6 percent of households whose energy spending was 83.8 percent solid fuel, overwhelmingly coal and firewood, and these households were poorer, more rural, and less educated than the rest.

The most consequential methodological move was the construction of a multidimensional Energy Poverty Index combining four binary dimensions: an energy cost burden exceeding 10 percent of disposable income, low household income, inability to keep the home adequately warm, and arrears on energy bills. Households meeting at least two criteria were classified as energy poor. The results were striking: 11.4 percent of Polish households met the multidimensional definition, while nearly 30 percent more exhibited a single symptom, placing them at elevated risk. In total, more than 40 percent of households showed at least one manifestation of energy vulnerability.

Logistic regression identified the structural drivers behind this risk. Reliance on non-labor income nearly doubled the odds of multidimensional energy poverty, with an odds ratio of 1.78, and heating with a solid-fuel stove increased the odds by 43 percent. Education showed a steep protective gradient: relative to households whose head had at most lower-secondary schooling, basic vocational education cut the odds to 0.65, secondary education to 0.39, and higher education to 0.27. Household composition mattered sharply as well, with single mothers facing roughly double the odds and couples with four or more children about 2.1 times the odds of the reference group.

Disposable income declined monotonically with energy poverty severity, from an average of 4,359 zloty per capita among households with no symptoms to 2,718 zloty for those with one and just 1,588 zloty for the multidimensionally poor. The at-risk group sat closer to the energy-poor than to the secure on nearly every characteristic, sharing the same risk factors, including solid-fuel heating and dependence on non-labor income, which the author argues makes them a priority target for preventive intervention before deprivations accumulate.

A leave-one-dimension-out sensitivity analysis tested whether these findings were artifacts of the index’s construction. Each of the four dimensions was excluded in turn and the regression re-estimated. The core predictors, employment status, education, household composition, and marital status, remained stable in direction and magnitude across all four specifications, with model fit actually improving in some cases, indicating that no single dimension was disproportionately driving the conclusions.

Perhaps the study’s most important warning concerns interpretation. A declining expenditure share, the author cautions, should not be read as improving welfare. Energy-poor households may respond to rising costs not by spending more but by underheating their homes and forgoing essential energy services, so a low energy share can signal unmet needs rather than comfort. Expenditure-based indicators, including the single income-cost measure adopted in Poland’s draft Social Climate Plan, systematically underestimate the problem by overlooking material deprivation and housing conditions.

The policy implications point toward integration rather than any single lever. Translating the odds ratios into population effects, the study estimates that shifting all solid-fuel households to alternative heating would reduce multidimensional energy poverty prevalence by only about 0.4 percentage points, and no single-factor intervention exceeds roughly 0.6 points. Meaningful reduction, the analysis concludes, requires coordinated action combining targeted social protection with investments in residential energy efficiency, housing modernization, low-carbon heating, and labor-market activation, so that the energy transition does not deepen the inequalities it is meant to resolve.

Subject of Research: Multidimensional energy poverty and household energy expenditure patterns in Poland

Article Title: Household energy expenditures and multidimensional energy poverty in Poland: patterns, determinants, and policy implications

Article References: Piekut, M. J. (2026). Household energy expenditures and multidimensional energy poverty in Poland: patterns, determinants, and policy implications. Frontiers in Environmental Science, 14, Article 1950882. https://doi.org/10.3389/fenvs.2026.1950882

Image Credits: AI Generated

DOI: 10.3389/fenvs.2026.1950882

Keywords: energy poverty, Poland, household energy expenditure, multidimensional index, logistic regression, cluster analysis, solid-fuel heating, energy affordability, Household Budget Survey, energy transition, thermal deprivation, social policy

The research, published in Frontiers in Allergy, examined perimenopausal women aged 45 to 60 years and compared them against patients with chronic non-fungal sinusitis, a control group that shares many symptoms but lacks the fungal mass at the heart of the disease. The team focused on two metabolic markers that have become central to modern cardiovascular and endocrine research: high-density lipoprotein cholesterol, the so-called good cholesterol, and a composite index known as TyG-BMI, which combines the triglyceride-glucose index with body mass index to capture insulin resistance and overall metabolic burden. Both markers are inexpensive to measure, which makes any association with a fungal sinus disease immediately interesting from a screening standpoint.

Fungal ball sinusitis, sometimes called fungus ball or aspergilloma of the paranasal sinuses, belongs to the non-invasive category of fungal sinus disease. Unlike invasive fungal infections that threaten immunocompromised patients, a fungal ball grows slowly within a blocked sinus cavity, forming a compact tangle of hyphae that surgeons must physically remove. The condition typically affects one sinus, most often the maxillary sinus, and produces chronic nasal obstruction, foul discharge, facial pressure and, in some cases, the sensation of a foreign body deep in the cheek. Because its symptoms mimic ordinary chronic rhinosinusitis, diagnosis frequently arrives only when computed tomography reveals characteristic calcifications or when the removed tissue is examined under a microscope.

The study design was straightforward but carefully targeted. The researchers enrolled fungal ball sinusitis patients treated at their single center between 2015 and 2025, then narrowed the analysis to perimenopausal women aged 45 to 60, using patients with chronic non-fungal sinusitis as the comparison group. Biochemical indicators were compared between the two groups, participants were sorted into quartiles based on their TyG-BMI values, and logistic regression models were used to identify which factors independently predicted fungal ball disease. Receiver operating characteristic curves, a standard tool for judging diagnostic performance, were then applied to test whether the metabolic markers could actually predict the condition.

The demographic findings alone tell a compelling story. Across the decade of cases, fungal ball sinusitis showed a steadily increasing trend, and the patient population skewed heavily toward middle-aged and elderly individuals. Women made up 63.96 percent of cases compared with 36.04 percent of men, and the peak age of female illness clustered tightly between 45 and 60 years, precisely the window of perimenopause. This concentration is exactly the kind of pattern that makes researchers suspect hormonal and metabolic mechanisms rather than simple exposure, since perimenopause is accompanied by falling estrogen levels, redistribution of body fat, rising insulin resistance and deteriorating lipid profiles.

When the researchers compared blood values between the two groups, the differences were clear. Women with fungal ball sinusitis had significantly lower HDL cholesterol levels and significantly higher TyG and TyG-BMI values than women with chronic non-fungal sinusitis. The relationship held up in a dose-response pattern: as TyG-BMI quartiles rose, HDL cholesterol fell progressively, and the prevalence of fungal ball sinusitis climbed in step. In statistical terms, the trend was significant at P less than 0.05, meaning the probability that these gradients arose by chance is small.

The logistic regression analysis then isolated the two markers as independent risk factors, each holding its predictive power after accounting for the other. Higher HDL cholesterol was protective: each unit increase was associated with an odds ratio of 0.482, with a 95 percent confidence interval of 0.282 to 0.823 and a P value of 0.008, meaning women with better HDL levels had roughly half the odds of fungal ball disease. Higher TyG-BMI, by contrast, carried risk, with an odds ratio of 1.192 per quartile step, a 95 percent confidence interval of 1.008 to 1.409 and a P value of 0.04. In plain language, the more insulin-resistant and metabolically burdened a woman was, the more likely she was to harbor a fungal ball in her sinuses, and the healthier her cholesterol profile, the lower that likelihood became.

One finding deserves careful interpretation. When the researchers combined TyG-BMI and HDL cholesterol into a single predictive model and tested it with ROC analysis, the combined model showed only limited predictive value for fungal ball sinusitis. This means the markers are statistically associated with the disease and may illuminate its biology, but they cannot yet serve as a reliable diagnostic test in the clinic. Association is not prediction, and a screening tool built on these numbers would miss too many cases and flag too many healthy women to be useful on its own. The authors instead frame their results as a call for targeted metabolic monitoring and early intervention, positions that could plausibly lower fungal ball risk if the underlying mechanisms hold up in future work.

What might those mechanisms be? The study was observational and retrospective, so it cannot prove causation, but the biology offers several plausible threads. Insulin resistance, which TyG-BMI captures, is known to alter mucosal immunity, promote low-grade systemic inflammation and change the composition of airway secretions, all of which could make the sinus environment more hospitable to fungal colonization. HDL cholesterol, meanwhile, does far more than ferry lipids; it modulates inflammation and immune cell function, and low HDL is a recognized hallmark of metabolic syndrome. Falling estrogen during perimenopause worsens both insulin resistance and lipid profiles, which could explain why the disease concentrates in this demographic window. A clogged, inflamed, metabolically altered sinus cavity may simply be easier for ambient fungal spores to colonize and harder for the immune system to clear.

The clinical takeaway is that fungal ball sinusitis should no longer be viewed as a purely local problem of the nose. For perimenopausal women with stubborn unilateral sinus symptoms, especially those with known metabolic risk, the study suggests that clinicians may want to look at the whole metabolic picture rather than treating the sinus in isolation. The rising annual incidence recorded at this single center over ten years adds urgency, because a disease that grows quietly in midlife women is a disease that health systems will increasingly encounter. Larger multicenter prospective studies will be needed to confirm the association, test whether metabolic intervention actually reduces risk, and determine whether TyG-BMI and HDL cholesterol can eventually earn a place in clinical prediction models. For now, the message is that the sinuses and the metabolism are more connected than anyone suspected, and the blood panel drawn for a routine midlife checkup may one day help flag the women who need a closer look up their nasal passages.

Subject of Research: Metabolic risk factors for fungal ball sinusitis in perimenopausal women

Article Title: Metabolic risk factors for fungal ball sinusitis in perimenopausal women

Article References: Wang, S.-Y., Xu, W., Liu, Y.-X., Jiang, L., Xu, Y., Teng, F.-Y., Hu, X.-Y., & Li, H.-X. (2026). Metabolic risk factors for fungal ball sinusitis in perimenopausal women. Frontiers in Allergy, 7, 1900638. https://doi.org/10.3389/falgy.2026.1900638

Image Credits: AI Generated

DOI: 10.3389/falgy.2026.1900638

Keywords: fungal ball sinusitis, perimenopausal women, TyG-BMI, HDL cholesterol, insulin resistance, chronic rhinosinusitis, metabolic syndrome, Frontiers in Allergy, retrospective study, sinus disease, estrogen, logistic regression

Roads are among the most pervasive drivers of habitat fragmentation on Earth, and amphibians suffer disproportionately. Because these animals depend on both aquatic and terrestrial habitats, move slowly, and often undertake explosive breeding migrations that force mass road crossings, a single stretch of pavement can slice through the entire annual cycle of a population. Conservationists have long known that proximity to water predicts amphibian roadkill, but most studies stop there, relying on coarse, generalized predictors applied to all species at once. The new research, led by a team analyzing five understudied frog species in one of Taiwan’s richest amphibian hotspots, shows just how much that approach misses.

The study combined nearly two years of standardized roadkill surveys — 33 motorbike and walking surveys conducted at speeds under 10 kilometres per hour — with citizen science records from Taiwan’s Roadkill Observation Network, yielding 234 amphibian roadkill records along a road that climbs from 240 to 360 metres in elevation, flanked by steep forested hills on one side and the Ruizi Creek on the other. Five target species were selected for modeling, including the vulnerable Rana sauteri, whose mass mortality events were documented by volunteer observers, and Buergeria robusta, a stream-breeding frog whose May and June breeding migrations are well known to local frog-rescue volunteers.

The methodological innovation lies in the analysis. Rather than pooling all species together, the team built ten separate models: one multispecies model, five single-species models, and four temporal models for Buergeria robusta split across its breeding season. Environmental predictors were measured within buffer zones radiating from 25 to 250 metres around each roadkill point, and — crucially — the buffers were split into uphill and downhill halves, because a road running along the base of a steep valley presents fundamentally different environments on each side. The predictors ranged from stream order and forest cover to fine-scale features rarely examined before: the number of streetlights, the length of Jersey barriers and W-beam guardrails, the condition of open drainage channels, and even the presence of bus stops and leaf litter.

To ensure the results were robust, the researchers used a dual-model consensus approach, requiring that predictors be retained by both a Random Forest model, which captures non-linear interactions, and a Logistic Regression model, which yields interpretable linear relationships. Class imbalance was handled with synthetic over-sampling and class weighting, and predictor preselection using harmonic mean p-values, correlation filtering and variance inflation factors trimmed the variable set before modeling. Only predictors surviving both algorithms were accepted as consensus findings — 24 in total, evenly split between habitat-related and road-related factors.

The performance gap between model types was striking. The multispecies model achieved test AUC values of only 0.552 to 0.610, barely better than chance, while single-species models reached 0.746 to 0.827 and the temporal models of Buergeria robusta peaked at 0.851. In other words, lumping species together destroys the very signal that matters. Each species responded to a distinct set of environmental drivers, and even within a single species the relevant predictors and their spatial scales shifted dramatically across the breeding season. The Hylarana latouchii model, notably, produced no consensus predictors at all, likely reflecting that species’ sedentary, generalist ecology and small sample size.

Water features emerged as the most consistent drivers, but with a directional twist. Downhill rivers and larger stream networks channel amphibian movements toward the road and raised mortality risk, while high-quality uphill streams were associated with fewer crossings — suggesting that when suitable habitat lies upslope, animals simply do not need to reach the pavement. For Odorrana swinhoana, a forest-dwelling species, uphill broadleaf forest increased risk, while for Buergeria robusta, downhill cropland proved deadly during the breeding season. Streetlights told an even more nuanced story: light density was positively associated with roadkill for Rana sauteri and for Buergeria robusta in June, yet negatively associated with roadkill for the same frog species in May, possibly because brighter ambient light in downhill foraging zones reduced the need to approach the road at all. This is the first documentation of artificial lighting shaping the spatial concentration of mass-migration crossings in these species.

Structural features of the road itself proved equally consequential, and their effects could flip depending on design and season. Long Jersey barriers and tall uphill retaining walls reduced roadkill by blocking access to the road, echoing findings for common toads in Portugal. But long downhill retaining walls and W-beam guardrails increased mortality, apparently by trapping migrating frogs or funneling them through gaps into traffic. Drainage channels played dual roles as well: vegetated channels acted as ecological traps that attracted animals, while longer waterlogged channels served as barriers that prevented crossings entirely. These details matter because they are exactly the features engineers control when building or retrofitting a road.

What elevates the study beyond a regional case study is its translation into practice. The authors present a hierarchical mitigation framework spanning the full life cycle of infrastructure projects. At the planning stage, road alignments should avoid critical habitats, maintaining buffers of at least 50 metres from large downhill streams and 25 metres from uphill forests, with underpasses and drift fences prioritized near stream crossings. At the design stage, streetlights can be fitted with amber or red spectra, shielding, and seasonal dimming schedules aligned with migration peaks; retaining walls can be retrofitted with escape ramps and permeable sections; drainage channels can be engineered with sloped sides so trapped animals can climb out. During predictable mass migrations, dynamic signage, temporary speed reductions, short-term road closures and volunteer-run bucket-trap fencing offer immediate relief, while ongoing maintenance — clearing drains, inspecting underpasses, adjusting lighting schedules — keeps permanent measures functional.

The framework’s real promise is transferability. Because it is organized around mechanisms — directional hydrology, structural permeability, landscape composition, illumination effects — rather than species labels, the authors argue it can be adapted to roads, railways, canals and energy corridors from arid to subtropical environments, and even extended to pedestrian safety mapping. The study’s limitations are acknowledged: the datasets span less than two years, dynamic variables such as rainfall and traffic volume were not included, and interannual variation remains unmeasured. Yet the core message is unambiguous and, for conservation planners, quietly revolutionary. One-size-fits-all hotspot maps and generalized mitigation guidelines are insufficient. Effective protection of the world’s most imperiled vertebrates demands models tuned to individual species, individual seasons, and individual roadside structures — down to the streetlight, the wall, and the ditch.

Subject of Research: Environmental and seasonal drivers of amphibian roadkill on a Taiwanese mountain road

Article Title: Precise roadkill environmental factors identification for sustainable planning and design through spatial and temporal modelling

Article References: Karanasios, P. I., Schmeller, D. S., Hsu, Z.-T., & Lin, Y.-P. (2026). Precise roadkill environmental factors identification for sustainable planning and design through spatial and temporal modelling. Nature Conservation, 61, 1-27. https://doi.org/10.3897/natureconservation.61.171624

Image Credits: AI Generated

DOI: 10.3897/natureconservation.61.171624

Keywords: amphibian roadkill, habitat fragmentation, Random Forest, logistic regression, streetlights, drainage channels, breeding migration, Taiwan, multiscale modeling, road ecology, conservation planning, citizen science