Latest Research
All publications from the Cancer3.AI database, newest first.
Artificial Intelligence in Ocular Surface Tumors: Current Advances, Challenges, and Future Directions.
Ghanbari H, et al
This comprehensive review examines the emerging role of artificial intelligence in the diagnosis and management of ocular surface tumors, a group of rare but potentially life-threatening eye cancers that includes squamous neoplasia, conjunctival melanoma, and lymphoma. Currently, accurate diagnosis of these conditions demands highly specialized ophthalmologists and pathologists, with invasive excisional biopsy remaining the gold standard, creating barriers to timely and accessible care. The authors surveyed a broad range of AI models—encompassing machine learning and deep learning approaches—applied across multiple ophthalmic imaging modalities, showing that these technologies hold substantial promise for enabling earlier and less invasive diagnosis. Key challenges identified include the scarcity of large, well-annotated datasets for rare tumor types, variability in imaging protocols across clinical centers, and unresolved ethical and regulatory questions surrounding AI deployment in medical practice. The review concludes that bridging AI innovation with clinical ophthalmology practice could meaningfully enhance diagnostic accuracy, reduce reliance on invasive procedures, and ultimately improve outcomes for patients with these challenging conditions.
Diagnostics (Basel, Switzerland)
Source →The global landscape of cancer burden attributable to tobacco smoking in 2022: a population-based systematic analysis.
Zhu Q, et al
This large-scale population-based systematic analysis estimated the global cancer burden attributable to tobacco smoking in 2022, drawing on national smoking prevalence surveys, published relative risks, and GLOBOCAN 2022 incidence and mortality data to calculate population attributable fractions (PAFs) across sexes, regions, and cancer types. Researchers found that approximately 2.72 million new cancer cases and 1.80 million cancer deaths worldwide — representing roughly 15% and 19% of all cancers, respectively — could be attributed to tobacco use. Men bore a dramatically greater burden than women, with incidence and mortality PAFs more than four times higher in males, though the peak burden for men occurred in high Human Development Index regions while for women it peaked in very-high HDI regions. While laryngeal cancer had the highest smoking-attributable fraction, lung cancer dominated in absolute numbers, accounting for about half of all smoking-related cancer cases and deaths globally, and contributing up to 63% of the smoking-related burden in higher-income regions. A particularly notable finding was that never-smokers in Eastern Asia had lung cancer incidence and mortality rates exceeding those in any other geographical region, underscoring the urgent need to investigate non-tobacco risk factors driving this anomaly. These findings call for context-specific, aggressive tobacco control policies integrated into national cancer prevention strategies, especially in lower-HDI countries where oral, esophageal, and cervical cancers add substantially to the smoking-related burden.
EClinicalMedicine
Source →Lipid Metabolism and Oxidative Stress in Uterine Corpus Endometrial Carcinoma: Prognostic and Immune Landscape Analysis.
Li H, et al
Researchers investigated the role of genes involved in lipid metabolism and oxidative stress in uterine corpus endometrial carcinoma (UCEC), one of the most common gynecologic cancers, using large-scale transcriptomic and clinical datasets. By applying advanced clustering techniques, the team identified two distinct molecular subtypes of UCEC: one characterized by higher immune cell infiltration and better survival outcomes, and another marked by immune suppression and poorer prognosis. A 10-gene prognostic signature was developed and validated, demonstrating strong ability to predict patient outcomes, with four genes — EPHX2, SLC8A1, ORMDL2, and MYLIP — further confirmed in laboratory experiments to actively drive tumor growth, migration, and invasion. A diagnostic model built using 12 machine learning algorithms showed superior accuracy in distinguishing high-risk patients. These findings highlight lipid metabolism and oxidative stress genes as promising biomarkers that could guide more personalized treatment strategies for endometrial cancer patients.
Asia-Pacific journal of clinical oncology
Source →Primary Synovial Sarcoma of the Breast Initially Mimicking Dermatofibrosarcoma Protuberans: A Case Report.
Mkhinini I, et al
Researchers report a rare case of primary synovial sarcoma of the breast in a 60-year-old postmenopausal woman, a malignancy accounting for less than 1% of all breast tumors. The patient presented with a 10 cm firm mass that was initially misidentified as dermatofibrosarcoma protuberans (DFSP) based on a core needle biopsy showing strong CD34 positivity. Following simple mastectomy, advanced immunohistochemical staining and molecular analysis revealed key markers including TLE1 and BCL2 positivity alongside a confirmed SYT-SSX1 gene fusion, establishing the correct diagnosis of monophasic synovial sarcoma. Surgical margins were clear, and the patient remained disease-free at one-year follow-up. This case underscores the critical importance of comprehensive molecular and immunohistochemical testing when diagnosing rare spindle-cell tumors of the breast, as initial biopsy findings can be misleading and lead to incorrect treatment planning.
European journal of breast health
Source →Metastatic Prostate Cancer Masquerading as Locally Advanced Breast Cancer.
Howard T, et al
Researchers report a rare case of a 72-year-old man whose prostate cancer spread to breast tissue affected by gynecomastia, along with involvement of nearby lymph nodes under the arm. The case is unusual because the spread occurred without widespread metastatic disease and without prolonged hormone therapy, which are the conditions most commonly associated with such metastases. Initial imaging, including a specialized PSMA PET scan ordered due to rising PSA levels, strongly suggested a primary breast cancer rather than a metastatic deposit from the prostate. Only tissue biopsy and fine-needle aspiration confirmed the true diagnosis of metastatic prostate adenocarcinoma, which was later verified through surgical removal of the lesion. This case highlights the critical importance of tissue confirmation before assuming that a breast lesion represents a primary breast cancer, particularly in men with a history of prostate cancer.
Clinical case reports
Source →