Latest Research
All publications from the Cancer3.AI database, newest first.
Intervention mapping to develop a patient-advocate enhanced program to reduce sedentary behavior in endometrial cancer survivors.
Bates-Fraser LC, et al
Researchers developed a structured intervention called the Sitting Time Elimination Program (STEP) after Endometrial Cancer, designed to reduce the high levels of sedentary behavior observed in endometrial cancer survivors, who spend more than half their waking hours sitting. Using a six-step intervention mapping framework, the team systematically identified behavioral targets, evidence-based behavior change techniques, and implementation strategies tailored to this population. Patient advocates were engaged through one-on-one interviews to refine the program, and their input shaped a multi-component intervention incorporating wearable self-monitoring devices, small group health coaching, and a dedicated smartphone application. Notably, patient advocates identified more facilitators than barriers to the proposed program, suggesting strong potential for acceptance and engagement. The STEP intervention is now prepared for feasibility and effectiveness testing in clinical studies, offering a promising new direction for improving long-term health outcomes in endometrial cancer survivors where previous physical activity interventions have shown limited success.
Journal of cancer survivorship : research and practice
Source →Lung cancer risk in relation to indicative radon atlas metrics in Northern Ireland: a population-based case-control study using secondary data.
Delargy CM, et al
A population-based case-control study conducted in Northern Ireland investigated whether a nationwide indicative radon atlas could accurately reflect lung cancer risk in the general population. Researchers compared 1,687 primary lung cancer cases diagnosed in 2006 and 2014 with over 8,000 non-lung cancer controls, linking participants to radon exposure data based on their postcode and adjusting for factors such as age, sex, smoking, deprivation, and air pollution. The key finding was that individuals living in areas classified as high radon zones — where more than a certain percentage of homes exceed 200 becquerels per cubic metre — faced more than twice the risk of developing lung cancer compared to those in lower-exposure areas. Importantly, this association held even after controlling for fine particulate matter (PM2.5) air pollution, strengthening the evidence that radon itself contributes independently to lung cancer risk. These results support the use of area-based radon atlases as practical public health tools for identifying communities at elevated cancer risk, which has direct relevance for the design of lung cancer screening programmes currently being planned in Northern Ireland.
Environmental geochemistry and health
Source →Bacterial taxa associated with lung cancer cases in Southeast Asians: a pilot case-control study.
Low A, et al
A pilot case-control study investigated the association between specific bacterial taxa and lung cancer in Southeast Asian populations, a group that often develops lung cancer without the classical risk factor of heavy tobacco use. Researchers compared the microbial profiles of lung cancer patients against healthy controls to identify bacteria that were enriched or depleted in cancer cases. The study identified distinct bacterial taxa that were significantly associated with lung cancer diagnosis, suggesting that the lung or gut microbiome may play a role in cancer development or progression in this demographic. These findings are particularly relevant given that Southeast Asians show unique epidemiological patterns of lung cancer, including higher rates in non-smokers and women. While preliminary, this research opens a new avenue for understanding lung cancer biology and may eventually contribute to microbiome-based biomarkers or therapeutic strategies tailored to Asian populations.
Cellular oncology (Dordrecht, Netherlands)
Source →A scalable multimodal graph neural network for drug combination response prediction.
Saeed D, et al
Researchers developed a new artificial intelligence framework called the Multimodal Molecular Drug Graph Neural Network (MMDGNN) to predict which combinations of cancer drugs will work together more effectively than either drug alone. The system was designed to address a critical problem in oncology: cancer cells frequently develop resistance to targeted therapies, such as osimertinib used in EGFR-mutant lung cancer, making single-drug treatments insufficient over time. MMDGNN integrates two types of molecular data — chemical fingerprints and SMILES string representations — within an advanced graph neural network that can model complex, heterogeneous molecular interactions. When tested on four benchmark datasets, the model achieved a mean squared error of 16.18 and a Pearson correlation of 0.85, outperforming the previous leading model MGAE-DC by 6.8% in prediction error reduction. The framework is also designed for large-scale computing environments, making it practical for real-world deployment across diverse cancer types and large drug libraries. By enabling more accurate prediction of synergistic drug combinations, MMDGNN could help clinicians design more effective treatment regimens and reduce the trial-and-error process in precision oncology.
Molecular diversity
Source →S100A14 in Tumor-Derived EVs Targets PIAS3 to Reprogram Astrocytes and Induce Immunosuppressive Microenvironment Promoting Brain Metastasis and Germacrone Reversal Effect.
Feng Q, et al
Researchers investigated how cancer cells spread to the brain — a complication known as brain metastasis that carries the highest mortality among cancer patients — focusing on tiny particles called extracellular vesicles (EVs) shed by tumors. Using advanced proteomics, they discovered that a protein called S100A14 is highly enriched in EVs from brain-metastatic lung and breast cancer cells and in samples from clinical patients, and that EVs carrying S100A14 dramatically increased the rate of brain metastasis in mouse models. Mechanistically, S100A14 delivered by these vesicles binds to a regulatory protein called PIAS3 inside brain support cells (astrocytes), triggering a cascade that activates STAT3 signaling and causes astrocytes to release immune-suppressing chemokines that recruit myeloid-derived suppressor cells, effectively shielding tumors from immune attack. Critically, the team identified germacrone, a natural plant-derived compound, as a therapeutic agent that physically binds S100A14 in astrocytes, disrupts the S100A14-PIAS3 interaction, shuts down STAT3 activation, and reduces immunosuppressive cell recruitment — all without significant toxicity. These findings reveal a novel molecular mechanism driving brain metastasis and position germacrone as a promising candidate for preventing or treating brain metastasis in lung and breast cancer patients.
Advanced science (Weinheim, Baden-Wurttemberg, Germany)
Source →