Quantification of intratumoral heterogeneity using habitat-based MRI radiomics to identify HER2-positive, -low and -zero breast cancers: a multicenter study

Breast Cancer Research, Journal Year: 2024, Volume and Issue: 26(1)

Published: Nov. 22, 2024

Human epidermal growth factor receptor 2-targeted (HER2) therapy with antibody-drug conjugates has proven effective for patients HER2-low breast cancer. However, intratumoral heterogeneity (ITH) poses a great challenge in identifying tumors. ITH signatures were developed by quantifying to differentiate HER2-positive, -low and -zero cancers. This retrospective study included 614 from two institutions. The was structured into primary tasks: task 1 between HER2-positive -negative tumors, followed 2 Whole-tumor radiomics features habitat extracted MRI construct the signatures. Multivariable logistic regression analysis used determine significant independent predictors. A combined model integrating clinicopathologic variables, signature, signature (1) Subsequently, better-performing established using same approach (2) area under receiver operating characteristic curve (AUC) assess performance of each model. Task comprised (training, n = 348; validation, 149; test cohorts, 117). encompassed 501 283; 122; 96). For task1, showed outstanding performance, achieving AUCs 0.81, 0.81 training, validation respectively. achieved improved 0.83, 0.84 0.83 across three task2, maintained superior 0.94, 0.93 indicated that none characteristics retained as predictors associated odds Our quantified habitat-based radiomics, differentiating HER2-postive further

Language: Английский

---

Multiparametric MRI Radiomics With Machine Learning for Differentiating HER2-Zero, -Low, and -Positive Breast Cancer: Model Development, Testing, and Interpretability Analysis

American Journal of Roentgenology, Journal Year: 2024, Volume and Issue: 224(1)

Published: Oct. 16, 2024

BACKGROUND. MRI radiomics has been explored for three-tiered classification of HER2 expression levels (i.e., HER2-zero, HER2-low, or HER2-positive) in patients with breast cancer, although an understanding how such models reach their predictions is lacking. OBJECTIVE. The purpose this study was to develop and test multiparametric machine learning differentiating as well explain the contributions model features through local global interpretations use Shapley additive explanation (SHAP) analysis. METHODS. This retrospective included 737 (mean age, 54.1 ± 10.6 [SD] years) cancer from two centers (center 1 [n = 578] center 2 159]), all whom underwent had determined after excisional biopsy. Analysis entailed tasks: HER2-negative HER2-zero HER2-low) tumors HER2-positive (task 1) HER2-low 2). For each task, were randomly assigned a 7:3 ratio training set 1: n 405; task 2: 284) internal 173; 122); formed external 159; 105). Radiomic extracted early phase dynamic contrast-enhanced (DCE) imaging, T2-weighted DWI. support vector (SVM) used feature selection, score (radscore) computed using weights SVM correlation coefficients, conventional combined constructed, performances evaluated. SHAP analysis provide outputs. RESULTS. In set, 1, AUCs model, radscore, 0.624, 0.757, 0.762, respectively; 2, AUC radscore 0.754, no could be constructed. identified DCE imaging having strongest influence both tasks; also prominent role 2. CONCLUSION. findings indicate suboptimal performance noninvasive characterization expression. CLINICAL IMPACT. provides example interpretation better understand imaging-based models.

Language: Английский

---

Preoperative diagnosis of meningioma sinus invasion based on MRI radiomics and deep learning: a multicenter study

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

Published: Feb. 28, 2025

Abstract Objective Exploring the construction of a fusion model that combines radiomics and deep learning (DL) features is great significance for precise preoperative diagnosis meningioma sinus invasion. Materials methods This study retrospectively collected data from 601 patients with confirmed by surgical pathology. For each patient, 3948 features, 12,288 VGG 6144 ResNet 3072 DenseNet were extracted MRI images. Thus, univariate logistic regression, correlation analysis, Boruta algorithm applied further feature dimension reduction, selecting DL highly associated Finally, models constructed using random forest (RF) algorithm. Additionally, diagnostic performance different was evaluated receiver operating characteristic (ROC) curves, AUC values compared DeLong test. Results Ultimately, 21 invasion selected, including 6 2 7 features. Based on these five constructed: model, DL-radiomics (DLR) model. demonstrated superior performance, 0.818, 0.814, 0.769 in training set, internal validation independent external respectively. Furthermore, results test indicated there significant differences between both ( p < 0.05). Conclusions The combining exhibits It expected to become powerful tool clinical plan selection patient prognosis assessment.

Language: Английский

---

Dual-Modality Virtual Biopsy System Integrating MRI and MG for Noninvasive Predicting HER2 Status in Breast Cancer

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

Published: March 1, 2025

Language: Английский

---

Multi-parametric MRI Radiomics Predicts Different HER2 Expression in Breast Cancer

Research Square (Research Square), Journal Year: 2025, Volume and Issue: unknown

Published: May 19, 2025

Purpose To develop and validate radiomic models using multi-parametric dynamic contrast-enhanced MRI (DCE-MRI) intravoxel incoherent movement (IVIM)-based features for the preoperative differentiation of HER2 expressions levels in breast cancer.Materials Methods This retrospectively study analyzed 227 female cancer patients who underwent 3.0TMRI examination at our institution from December 2019 to 2023. The least absolute shrinkage selection operator (LASSO) ten-fold cross-validation method was used identify positive negative cancer(task 1), further low zero 2). Then were selected combined with clinical characteristics construct predicting logistic regression analysis. area under receiver operating characteristic curve(AUC), sensitivity, specificity evaluate performance models.Results For task 1, AUCs model (histological grade peritumoral edema), DCE IVIM(D + D*+f) clinic 0.785 (95%CI:0.713,0.846), 0.866 (95%CI:0.803,0.915) 0.903 (95%CI:0.846,0.944) respectively. In validation cohort, 0.751 (95%CI:0.633,0.848), (95%CI:0.633,0.848) 0.830 (95%CI:0.720,0.910) 2, IVIM training cohort 0.951 (95%CI:0.888,0.984) 0.853 (95%CI:0.712,0.942) respectively, radiomics score independent predictors cancer.Conclusion signature derived MRI, together edema histological grade, demonstrated strong expression preoperatively cancer, which may support individualized treatment strategies.

Language: Английский

---

MRI-based habitat analysis for Intratumoral heterogeneity quantification combined with deep learning for HER2 status prediction in breast cancer

Magnetic Resonance Imaging, Journal Year: 2025, Volume and Issue: unknown, P. 110429 - 110429

Published: May 1, 2025

Language: Английский

---

Development and Validation of a Deep Learning System to Differentiate HER2‐Zero, HER2‐Low, and HER2‐Positive Breast Cancer Based on Dynamic Contrast‐Enhanced MRI

Journal of Magnetic Resonance Imaging, Journal Year: 2024, Volume and Issue: unknown

Published: Dec. 6, 2024

Background Previous studies explored MRI‐based radiomic features for differentiating between human epidermal growth factor receptor 2 (HER2)‐zero, HER2‐low, and HER2‐positive breast cancer, but deep learning's effectiveness is uncertain. Purpose This study aims to develop validate a learning system using dynamic contrast‐enhanced MRI (DCE‐MRI) automated tumor segmentation classification of HER2‐zero, statuses. Study Type Retrospective. Population One thousand two hundred ninety‐four cancer patients from three centers who underwent DCE‐MRI before surgery were included in the (52 ± 11 years, 811/204/279 training/internal testing/external testing). Field Strength/Sequence 3 T scanners, T1‐weighted 3D fast spoiled gradient‐echo sequence, enhanced sequence turbo field echo sequence. Assessment An model segmented tumors utilizing data, followed by models (ResNetGN) trained classify HER2 Three developed distinguish their respective non‐HER2 categories. Statistical Tests Dice similarity coefficient (DSC) was used evaluate performance model. Evaluation performances statuses involved receiver operating characteristic (ROC) curve analysis area under (AUC), accuracy, sensitivity, specificity. The P ‐values <0.05 considered statistically significant. Results automatic network achieved DSC values 0.85 0.90 compared manual across different sets. ResNetGN AUCs 0.782, 0.776, 0.768 HER2‐zero others training, internal test, external test sets, respectively. Similarly, 0.820, 0.813, 0.787 HER2‐low vs. others, 0.792, 0.745, 0.781 Data Conclusion proposed DCE‐MRI‐based may have potential preoperatively distinct expressions cancers with therapeutic implications. Evidence Level 4 Technical Efficacy Stage

Language: Английский

---

A Comprehensive Model Outperformed the Single Radiomics Model in Noninvasively Predicting the HER2 Status in Patients with Breast Cancer

Academic Radiology, Journal Year: 2024, Volume and Issue: unknown

Published: Aug. 1, 2024

Language: Английский

---

Editor's Notebook: April 2024

American Journal of Roentgenology, Journal Year: 2024, Volume and Issue: 222(4)

Published: April 1, 2024

---

Editorial for “Preoperative Differentiation of HER2‐Zero and HER2‐Low from HER2‐Positive Invasive Ductal Breast Cancers Using BI‐RADS MRI Features and Machine Learning Modeling”

Journal of Magnetic Resonance Imaging, Journal Year: 2024, Volume and Issue: unknown

Published: May 16, 2024

Language: Английский

---