Establishment of a predictive nomogram for breast cancer lympho-vascular invasion based on radiomics obtained from digital breast tomography and clinical imaging features

Abstract

Background: To develop a predictive nomogram for breast cancer lympho-vascular invasion (LVI), based on digital breast tomography (DBT) data obtained from intra- and peri-tumoral regions.

Methods: One hundred ninety-two breast cancer patients were enrolled in this retrospective study from 2 institutions, in which Institution 1 served as the basis for training (n = 113) and testing (n = 49) sets, while Institution 2 served as the external validation set (n = 30). Tumor regions of interest (ROI) were manually-delineated on DBT images, in which peri-tumoral ROI was defined as 1 mm around intra-tumoral ROI. Radiomics features were extracted, and logistic regression was used to construct intra-, peri-, and intra- + peri-tumoral radiomics models. Patient clinical data was analyzed by both uni- and multi-variable logistic regression analyses to identify independent risk factors for the non-radiomics clinical imaging model, and the combination of both the most optimal radiomics and clinical imaging models comprised the comprehensive model. The best-performing model out of the 3 types (radiomics, clinical imaging, comprehensive) was identified using receiver operating characteristic (ROC) curve analysis, and used to construct the predictive nomogram.

Results: The most optimal radiomics model was the intra- + peri-tumoral model, and 3 independent risk factors for 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, 95% CI = 1.232 ~ 6.815, P = 0.015), and axillary lymph node (ALN) metastasis (OR = 3.615, 95% CI = 1.642-7.962, P < 0.001) were identified by the clinical imaging model. Furthermore, the comprehensive model was the most accurate in predicting LVI occurrence, with areas under the curve (AUCs) of 0.889, 0.916, and 0.862, for, respectively, the training, testing and external validation sets, compared to radiomics (0.858, 0.849, 0.844) and clinical imaging (0.743, 0.759, 0.732). The resulting nomogram, incorporating radiomics from the intra- + peri-tumoral model, as well as maximum tumor diameter, suspicious malignant calcification, and ALN metastasis, had great correspondence with actual LVI diagnoses under the calibration curve, and was of high clinical utility under decision curve analysis.

Conclusions: The predictive nomogram, derived from both radiomics and clinical imaging features, was highly accurate in identifying future LVI occurrence in breast cancer, demonstrating its potential as an assistive tool for clinicians to devise individualized treatment regimes.

Keywords: Breast cancer; Digital breast tomography; Lympho-vascular invasion; Radiomics.

Fig. 1

Flow chart showing breast cancer patient recruitment from the First Hospital of Shanxi Medical University (Institution 1) and Shanxi Provincial People's Hospital (Institution 2)

Fig. 2

Representative digital breast tomography (DBT) images from a invasive breast ductal carcinoma, for defining intra- and peri-tumoral regions of interest (ROI). A Original DBT, B Intra-tumoral ROI (green), and (C) Peri-tumoral ROI (yellow; 1 mm around the intra-tumoral ROI) from the craniocaudal view. D Original DBT, E Intra-tumoral ROI (green), and (F) Peri-tumoral ROI (yellow; 1 mm around the intra-tumoral ROI) from the mediolateral oblique view

Fig. 3

Receiver operating characteristic (ROC) curves comparing the clinical imaging, radiomics, and comprehensive (predictive nomogram) models, with respect to (A) training, (B) testing, and (C) external validation datasets. ROC curves comparing intra-, peri-, and intra- + peri-tumoral radiomics models, with respect to (D) training and (E) test sets

Fig. 4

Establishment and evaluation of the predictive nomogram, based on the comprehensive model. A Predictive nomogram, incorporating the clinical imaging characteristics of axillary lymph node (ALN) metastases presence, maximum tumor diameter, and suspiciously malignant calcifications, as well as the Radscore, based on the intra- + peri-tumoral radiomics model. B Calibration curve analysis comparing bias-corrected nomogram predictions versus actual (apparent) lympho-vascular invasion (LVI) occurrence. C Decision curve analysis comparing the clinical utility of clinical imaging, radiomics, and predictive nomogram, as well as assuming that all or none of the patients had LVI

Fig. 5

Case examples demonstrating the effectiveness of the nomogram for predicting LVI. A DBT images, from the craniocaudal view, of a invasive ductal breast carcinoma patient. B Predictive nomogram, based on the patient having a lesion size of 3.60 cm, no ALN metastasis, and amorphous calcification, yielding a LVI likelihood of 82%. C Post-operative histopathological analysis confirmed the presence of LVI; breast tissues were stained with hematoxylin & eosin (H&E). D DBT images, from the craniocaudal view, of a invasive ductal breast carcinoma patient. E Predictive nomogram, based on the patient having a lesion size of 2.10 cm, as well as no ALN metastasis and calcification, yielding LVI likelihood of 21%. F Post-operative histopathological analysis confirmed that no LVI was present, under H&E staining