Latest Research
All publications from the Cancer3.AI database, newest first.
LobePrior segments lung lobes on computed tomography images in the presence of severe abnormalities.
Ribeiro JA, et al
Researchers developed LobePrior, an automated method for segmenting the lobes of the lungs from CT scans, designed to work reliably even when disease has severely distorted normal anatomy. The system combines deep neural networks with probabilistic anatomical models in a three-stage pipeline: a coarse processing step, a high-resolution segmentation step using specialized attention-based networks for each lobe, and a final refinement stage. LobePrior was tested on multiple datasets including patients with cancerous nodules and COVID-19 lung consolidations, achieving Dice scores of up to 0.978, outperforming competing methods with statistically significant improvements across all datasets. Accurate lobe segmentation is critical for clinicians planning surgery, monitoring disease progression, and measuring treatment response in conditions such as lung cancer and pneumonia. By automating this previously challenging task even in severely diseased lungs, LobePrior has the potential to streamline radiological workflows and improve the consistency of pulmonary disease assessment worldwide.
Scientific reports
Source →Discovery of potent kinase inhibitors with improved pharmacokinetics and safety potentials through structural optimization of dasatinib.
He K, et al
Researchers have developed PDD-87, a structurally optimized derivative of dasatinib, a widely used cancer drug that has been limited in clinical practice by safety concerns and poor pharmacokinetics. PDD-87 was found to potently inhibit key cancer-driving kinases — including ABL, BTK, and SRC family kinases — with IC₅₀ values below 1 nM, and it demonstrated strong antiproliferative effects against multiple leukemia and lymphoma cell lines. In animal studies, PDD-87 significantly reduced tumor growth in a mouse xenograft model in a dose-dependent manner, while showing markedly improved oral bioavailability (18.9% vs. 3.9% for dasatinib) and lower drug clearance in rats. Critically, PDD-87 accumulated far less in lung tissue than dasatinib — with lung-to-plasma ratios up to 5-fold lower — suggesting a substantially reduced risk of pulmonary toxicity, one of dasatinib's most serious side effects. The compound was well-tolerated in mice at doses up to 200 mg/day for two weeks with no overt toxicity observed. These findings position PDD-87 as a promising next-generation kinase inhibitor with a potentially superior safety and pharmacokinetic profile for treating leukemia and other cancers.
Communications medicine
Source →SenSet defines cell-type specific senescence signatures in the aged human lung.
Hasanaj E, et al
Researchers developed a computational tool called SenSet to identify genes specifically associated with cellular senescence — a state of irreversible cell growth arrest — in the aging human lung, using the largest available single-cell lung dataset, the Human Lung Cell Atlas. Cellular senescence is known to drive aging and diseases such as lung cancer, fibrosis, chronic obstructive pulmonary disease, and asthma, but studying it has been difficult because senescent cells are rare in healthy tissue and highly variable across cell types. The team validated SenSet in a sophisticated three-dimensional human lung tissue model exposed to established senescence-inducing agents, confirming the tool's accuracy and sensitivity. Using SenSet, the researchers uncovered distinct senescence signatures in specific lung cell populations linked to normal aging and environmental exposures such as pollution or tobacco smoke. These findings provide a comprehensive map of senescent cells in the healthy aging lung and offer a powerful new resource for understanding and potentially targeting senescence in serious lung conditions. For patients, this research opens the door to more precise diagnostic markers and therapeutic strategies aimed at slowing or reversing age-related lung disease.
The EMBO journal
Source →The utility of electronic frailty index in cancer patients undergoing chemotherapy.
Michael A, et al
A large UK study examined whether a computerized frailty scoring tool called SCARF could predict how well cancer patients tolerate chemotherapy, analyzing data from nearly 79,000 patients treated between 2015 and 2018 across three cancer types: breast, colorectal, and non-small-cell lung cancer. The SCARF index is based on a cumulative deficit model, meaning it tallies a patient's existing health problems to estimate overall frailty. Researchers found that severely frail colorectal cancer patients aged 70 or older faced twice the risk of dying within 30 days of chemotherapy compared to younger patients, while severely frail breast cancer patients in the same age group faced a risk six times higher. These findings are clinically significant because current fitness assessments used in oncology are time-consuming and frequently skipped in practice, whereas an electronic tool like SCARF could allow oncologists to quickly identify high-risk patients before initiating treatment. The study concludes that the SCARF index shows real promise for improving treatment decision-making in cancer care, though further validation in larger and more diverse populations is needed.
British journal of cancer
Source →Post-transcriptional activation of TGF-β signaling via SMAD mRNA stabilization by EIF4A3 promotes EMT and metastasis in gastric cancer.
Liu Y, et al
Researchers investigated the role of a protein called EIF4A3 in gastric cancer, a disease with high mortality rates largely due to late diagnosis and rapid spread to other organs. The study found that EIF4A3 is significantly elevated in gastric cancer tissues and cell lines compared to normal tissue, and its higher expression correlates with more advanced disease stages and worse patient outcomes. Mechanistically, the team discovered that EIF4A3 creates a dangerous positive feedback loop with the TGF-β/SMAD signaling pathway: TGF-β activates EIF4A3 production, and EIF4A3 in turn stabilizes the messenger RNAs of key SMAD proteins, keeping them active longer and amplifying signals that drive cancer cells to invade surrounding tissue and spread to distant organs such as the lungs. When EIF4A3 was silenced in laboratory models, tumor growth, cell migration, invasion, and lung metastasis were all markedly reduced. These findings identify EIF4A3 as a novel post-transcriptional amplifier of cancer-promoting signaling and a promising target for new diagnostic tools and therapies aimed at preventing metastatic progression in gastric cancer patients.
Cell biology and toxicology
Source →