Establishment of a predictive nomogram for breast cancer lympho-vascular invasion based on radiomics obtained from digital breast tomography and clinical imaging features DOI Creative Commons
Gang Liang, Suxin Zhang,

Yiquan Zheng

et al.

BMC Medical Imaging, Journal Year: 2025, Volume and Issue: 25(1)

Published: Feb. 26, 2025

To develop a predictive nomogram for breast cancer lympho-vascular invasion (LVI), based on digital tomography (DBT) data obtained from intra- and peri-tumoral regions. One hundred ninety-two patients were enrolled in this retrospective study 2 institutions, which Institution 1 served as the basis training (n = 113) testing 49) sets, while external validation set 30). Tumor regions of interest (ROI) manually-delineated DBT images, ROI was defined mm around intra-tumoral ROI. Radiomics features extracted, logistic regression used to construct intra-, peri-, + radiomics models. Patient clinical analyzed by both uni- multi-variable analyses identify independent risk factors non-radiomics imaging model, combination most optimal models comprised comprehensive model. The best-performing model out 3 types (radiomics, imaging, comprehensive) identified using receiver operating characteristic (ROC) curve analysis, nomogram. LVI, maximum tumor diameter (odds ratio [OR] 1.486, 95% confidence interval [CI] 1.082–2.041, P 0.014), suspicious malignant calcification (OR 2.898, CI 1.232 ~ 6.815, 0.015), axillary lymph node (ALN) metastasis 3.615, 1.642–7.962, < 0.001) Furthermore, accurate predicting LVI occurrence, with areas under (AUCs) 0.889, 0.916, 0.862, for, respectively, training, compared (0.858, 0.849, 0.844) (0.743, 0.759, 0.732). resulting nomogram, incorporating well diameter, calcification, ALN metastasis, had great correspondence actual diagnoses calibration curve, high utility decision analysis. derived features, highly identifying future occurrence cancer, demonstrating its potential an assistive tool clinicians devise individualized treatment regimes.

Language: Английский

Establishment of a predictive nomogram for breast cancer lympho-vascular invasion based on radiomics obtained from digital breast tomography and clinical imaging features DOI Creative Commons
Gang Liang, Suxin Zhang,

Yiquan Zheng

et al.

BMC Medical Imaging, Journal Year: 2025, Volume and Issue: 25(1)

Published: Feb. 26, 2025

To develop a predictive nomogram for breast cancer lympho-vascular invasion (LVI), based on digital tomography (DBT) data obtained from intra- and peri-tumoral regions. One hundred ninety-two patients were enrolled in this retrospective study 2 institutions, which Institution 1 served as the basis training (n = 113) testing 49) sets, while external validation set 30). Tumor regions of interest (ROI) manually-delineated DBT images, ROI was defined mm around intra-tumoral ROI. Radiomics features extracted, logistic regression used to construct intra-, peri-, + radiomics models. Patient clinical analyzed by both uni- multi-variable analyses identify independent risk factors non-radiomics imaging model, combination most optimal models comprised comprehensive model. The best-performing model out 3 types (radiomics, imaging, comprehensive) identified using receiver operating characteristic (ROC) curve analysis, nomogram. LVI, maximum tumor diameter (odds ratio [OR] 1.486, 95% confidence interval [CI] 1.082–2.041, P 0.014), suspicious malignant calcification (OR 2.898, CI 1.232 ~ 6.815, 0.015), axillary lymph node (ALN) metastasis 3.615, 1.642–7.962, < 0.001) Furthermore, accurate predicting LVI occurrence, with areas under (AUCs) 0.889, 0.916, 0.862, for, respectively, training, compared (0.858, 0.849, 0.844) (0.743, 0.759, 0.732). resulting nomogram, incorporating well diameter, calcification, ALN metastasis, had great correspondence actual diagnoses calibration curve, high utility decision analysis. derived features, highly identifying future occurrence cancer, demonstrating its potential an assistive tool clinicians devise individualized treatment regimes.

Language: Английский

Citations

0