Cancer3.AI › Latest Research

Latest Research

A cross-section of the Cancer3.AI database: the newest publication from each body region first, then the next one from each — so the review spans cancer types instead of running in blocks. Summaries are generated by Claude Sonnet (Anthropic) and link to the original publications.

ICD: C50 WHO Vol. 2 Breast
2026-09-10

Mass-Forming Ductal Carcinoma In Situ >10 mm on Ultrasonography: Sonographic-Pathological Correlation Underlying Mass Formation and Implications for Invasion Assessment.

Michishita Y, et al

A retrospective study of 52 surgically confirmed breast DCIS cases presenting as masses larger than 10 mm on ultrasound identified three distinct sonographic-pathological patterns — intracystic (9/52, 17.3%), clustered outline (34/52, 65.4%), and sclerosing stromal (7/52, 13.5%) — and found that background sclerosis, especially the sclerosing stromal pattern, is the chief driver of false-positive invasion suspicion on imaging. Preoperative ultrasound incorrectly suggested invasive carcinoma in 26 of 52 cases (50.0%), with the sclerosing stromal pattern carrying the highest rate of suspected invasion (6/7, 85.7%), followed by the clustered outline pattern (17/34, 50.0%) and the intracystic pattern (2/9, 22.2%). Histological contributors to invasion-mimicking appearances included background sclerosing adenosis, radial sclerosing lesion, inflammatory changes, and fibrosis. Awareness of these three patterns may help radiologists avoid overestimating tumour stage and improve preoperative surgical planning for patients with large-mass DCIS.

Ultrasound in medicine & biology

Source →
2026-09-10 • AI

Interpretable four-class machine learning prediction of initial I-131 therapy responses in differentiated thyroid cancer.

Lei Y, et al

A four-class Random Forest machine learning framework achieved a micro-average AUC of 0.894 (95% CI: 0.842–0.901) in the test cohort and 0.885 (95% CI: 0.832–0.912) in temporal validation when predicting four distinct response categories — excellent response, indeterminate response, biochemical incomplete response, and structural incomplete response — to initial post-operative I-131 therapy in differentiated thyroid cancer. Developed and validated in 948 patients, SHAP analysis showed that pre-treatment stimulated thyroglobulin (pre-sTg) was the dominant predictor of biochemical responses (peak weight 31.9% for biochemical incomplete response), while diagnostic whole-body scintigraphy drove structural incomplete response predictions (45.1% weight), with distant metastasis on scintigraphy raising structural incomplete response risk dramatically (OR: 171.89) and pre-sTg ≥ 10 ng/mL significantly increasing both structural (OR: 7.18) and biochemical incomplete response (OR: 5.48) risks. By separating transitional states obscured in conventional binary models, this interpretable tool offers clinicians a practical pre-therapeutic risk-stratification framework for personalizing radioiodine management in thyroid cancer patients.

Annals of medicine

Source →
ICD: C61 WHO Vol. 8 Male Reproductive System
2026-09-09

Magnetic Resonance Imaging-Guided Transurethral Ultrasound Ablation for Localized Prostate Cancer: A GRADE-Assessed Systematic Review and Meta-Analysis.

Nadeem A, et al

A systematic review and meta-analysis of 13 studies encompassing 650 patients found that MRI-guided transurethral ultrasound ablation (TULSA) for localized prostate cancer achieved 100% technical success with major complications (grade ≥ 3) in only 2% of cases. Urinary continence was preserved in 97% of treatment-naive patients compared to just 47% in the salvage setting (p < 0.0001), and PSA levels fell by a mean of 5.02 ng/mL — representing an 85% reduction from baseline. Biochemical recurrence was observed in 13% of patients, clinically significant cancer on follow-up biopsy in 22%, and 14% required salvage treatment progression, though oncologic certainty remains very low due to high heterogeneity and a maximum follow-up of only 36 months. The authors conclude that current evidence supports TULSA for carefully selected treatment-naive prostate cancer patients, pending long-term comparative data.

Cardiovascular and interventional radiology

Source →
ICD: C50 WHO Vol. 2 Male Breast
2026-09-09

Stereotactic radiosurgery for brain metastases from male breast cancer: a multicenter retrospective study.

Duzkalir AH, et al

A multicenter retrospective study of 19 male breast cancer patients with 185 brain metastases treated with stereotactic radiosurgery (SRS) achieved an exceptionally high crude local control rate of 97.3% and a median overall survival of 28.0 months. Patients with controlled systemic disease at the time of SRS survived significantly longer than those without (40.0 vs. 22.0 months; p = 0.032), identifying systemic disease status as the primary determinant of survival. Neurological symptom burden fell markedly after SRS (from 63.2% to 26.3%; p = 0.016), and target volume — with an optimal threshold of 0.98 cc on ROC analysis — was the principal dosimetric predictor of both local failure and radiation toxicity. These results support applying established SRS practice parameters to this exceptionally rare patient population.

Journal of neuro-oncology

Source →
2026-09-09

Circulating tumor HPV DNA reappearance as an early indicator of recurrence after radiotherapy-based treatment for HPV-associated oropharyngeal cancer: analysis of more than 2000 assessments in 232 patients.

Jabłońska IA, et al

A prospective study of 232 patients with HPV-positive oropharyngeal squamous cell carcinoma or carcinoma of unknown primary demonstrated that serial blood-based ctHPV DNA monitoring via quantitative PCR detects disease recurrence with 90.3% sensitivity and 97.9% specificity, and preceded clinical confirmation of relapse in 58.1% of recurrent cases across 2,928 total assessments. ctHPV DNA positivity measured one month after completing radiotherapy-based treatment was a powerful independent predictor of outcomes, associated with a hazard ratio of 19.1 (95% CI 7.5–48.8) for overall survival and 15.4 (95% CI 6.7–35.4) for disease-free survival. The high negative predictive value of 98.4% means a negative test result reliably identifies patients at very low short-term risk of recurrence, supporting the potential integration of ctHPV DNA surveillance into routine post-treatment follow-up.

Oral oncology

Source →