A computed tomography-based radiomics prediction model for BRAF mutation status in colorectal cancer DOI Creative Commons
Baohua Zhou,

Huaqing Tan,

Yuxuan Wang

et al.

Abdominal Radiology, Journal Year: 2025, Volume and Issue: unknown

Published: May 15, 2025

The aim of this study was to develop and validate CT venous phase image-based radiomics predict BRAF gene mutation status in preoperative colorectal cancer patients. In study, 301 patients with pathologically confirmed were retrospectively enrolled, comprising 225 from Centre I (73 mutant 152 wild-type) 76 II (36 40 wild-type). cohort randomly divided into a training set (n = 158) an internal validation 67) 7:3 ratio, while served as independent external 76). whole tumor region interest segmented, characteristics extracted. To explore whether expansion could improve the performance objectives, contour extended by 3 mm study. Finally, t-test, Pearson correlation, LASSO regression used screen out features strongly associated mutations. Based on these features, six classifiers-Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF), Logistic Regression (LR), K-Nearest Neighbors (KNN), Extreme Gradient Boosting (XGBoost)-were constructed. model clinical utility evaluated using receiver operating characteristic (ROC) curves, decision curve analysis, accuracy, sensitivity, specificity. Gender predictor unexpanded RF model, constructed 11 imaging histologic demonstrated best predictive performance. For cohort, it achieved AUC 0.814 (95% CI 0.732-0.895), accuracy 0.810, sensitivity 0.620. 0.798 0.690-0.907), 0.761, 0.609. 0.737 0.616-0.847), 0.658, 0.667. A machine learning based can effectively mutations cancer. optimal

Language: Английский

Real-World Efficacy and Safety of Anti-PD-1 Antibody Plus Apatinib and Temozolomide for Advanced Acral Melanoma DOI Creative Commons
Jiaran Zhang,

Huichun Tian,

Lili Mao

et al.

Cancer Management and Research, Journal Year: 2025, Volume and Issue: Volume 17, P. 905 - 916

Published: May 1, 2025

The combination of programmed cell death-1 (PD-1) blockade camrelizumab plus apatinib (an antiangiogenic agent) and temozolomide has displayed promising therapeutic effects in patients with advanced acral melanoma (AM) a non-randomized Phase II clinical trial (NCT04397770). aim this retrospective study was to evaluate the efficacy safety triplet regimen for AM real-world setting. data who received anti-PD-1 antibody at Peking University Cancer Hospital Institute between September 2019 December 2023 were analyzed. primary endpoint overall response rate (ORR). secondary endpoints included progression-free survival (PFS), (OS), disease control (DCR), duration (DOR), treatment-related adverse events (TRAEs). Overall, 250 eligible analysis. ORR 38.1% DCR 92.2%. median PFS, OS, DOR 8.5, 18.0, 13.2 months, respectively. When used as first-line treatment, 48.1%, PFS 12.0 OS 24.8 months. number lines therapy (≥2 lines), elevated lactate dehydrogenase, presence brain or liver metastasis negative predictors survival. 92.4% 45.2% experienced any-grade grade 3-4 TRAEs, This provides evidence that support effectiveness combined antibody, treating AM, demonstrating considerable prolonged survival, well acceptable tolerability.

Language: Английский

Citations

0
A computed tomography-based radiomics prediction model for BRAF mutation status in colorectal cancer DOI Creative Commons
Baohua Zhou,

Huaqing Tan,

Yuxuan Wang

et al.

Abdominal Radiology, Journal Year: 2025, Volume and Issue: unknown

Published: May 15, 2025

The aim of this study was to develop and validate CT venous phase image-based radiomics predict BRAF gene mutation status in preoperative colorectal cancer patients. In study, 301 patients with pathologically confirmed were retrospectively enrolled, comprising 225 from Centre I (73 mutant 152 wild-type) 76 II (36 40 wild-type). cohort randomly divided into a training set (n = 158) an internal validation 67) 7:3 ratio, while served as independent external 76). whole tumor region interest segmented, characteristics extracted. To explore whether expansion could improve the performance objectives, contour extended by 3 mm study. Finally, t-test, Pearson correlation, LASSO regression used screen out features strongly associated mutations. Based on these features, six classifiers-Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF), Logistic Regression (LR), K-Nearest Neighbors (KNN), Extreme Gradient Boosting (XGBoost)-were constructed. model clinical utility evaluated using receiver operating characteristic (ROC) curves, decision curve analysis, accuracy, sensitivity, specificity. Gender predictor unexpanded RF model, constructed 11 imaging histologic demonstrated best predictive performance. For cohort, it achieved AUC 0.814 (95% CI 0.732-0.895), accuracy 0.810, sensitivity 0.620. 0.798 0.690-0.907), 0.761, 0.609. 0.737 0.616-0.847), 0.658, 0.667. A machine learning based can effectively mutations cancer. optimal

Language: Английский

Citations

0