Improved Classification of Benign and Malignant Breast Lesions Using Deep Feature Maximum Intensity Projection MRI in Breast Cancer Diagnosis Using Dynamic Contrast-enhanced MRI

Hu, Qiyuan; Whitney, Heather M.; Li, Hui; Ji, Yu; Liu, Peifang; Giger, Maryellen L.

doi:10.1148/ryai.2021200159

Cited by 29 publications

(24 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Images from Database A had been acquired in the axial plane and images from Database B were acquired in the sagittal plane. The details of the imaging protocols are available elsewhere [ 5 , 13 , 14 , 15 ]. Information regarding subtypes of benign lesions and cancers had been collected from pathology and imaging reports.…”

Section: Methodsmentioning

confidence: 99%

Multi-Stage Harmonization for Robust AI across Breast MR Databases

Whitney

et al. 2021

Cancers

Self Cite

View full text Add to dashboard Cite

Radiomic features extracted from medical images may demonstrate a batch effect when cases come from different sources. We investigated classification performance using training and independent test sets drawn from two sources using both pre-harmonization and post-harmonization features. In this retrospective study, a database of thirty-two radiomic features, extracted from DCE-MR images of breast lesions after fuzzy c-means segmentation, was collected. There were 944 unique lesions in Database A (208 benign lesions, 736 cancers) and 1986 unique lesions in Database B (481 benign lesions, 1505 cancers). The lesions from each database were divided by year of image acquisition into training and independent test sets, separately by database and in combination. ComBat batch harmonization was conducted on the combined training set to minimize the batch effect on eligible features by database. The empirical Bayes estimates from the feature harmonization were applied to the eligible features of the combined independent test set. The training sets (A, B, and combined) were then used in training linear discriminant analysis classifiers after stepwise feature selection. The classifiers were then run on the A, B, and combined independent test sets. Classification performance was compared using pre-harmonization features to post-harmonization features, including their corresponding feature selection, evaluated using the area under the receiver operating characteristic curve (AUC) as the figure of merit. Four out of five training and independent test scenarios demonstrated statistically equivalent classification performance when compared pre- and post-harmonization. These results demonstrate that translation of machine learning techniques with batch data harmonization can potentially yield generalizable models that maintain classification performance.

show abstract

Section: Methodsmentioning

confidence: 99%

Multi-Stage Harmonization for Robust AI across Breast MR Databases

Whitney

et al. 2021

Cancers

Self Cite

View full text Add to dashboard Cite

show abstract

“…T2, DWI), 76–78 and (iv) fusing DL feature extraction methods with traditional machine learning classifiers. 74,76,77 Figure 5 depicts a fused architecture integrating DL feature extraction and machine learning classifiers for lesion classification from subtraction images of dynamic contrast-enhanced (DCE) MRI sequences. 74 Several papers show that DL methods for breast MRI outperform traditional machine learning methods, particularly as training data set sizes increase.…”

Section: Magnetic Resonance Imagingmentioning

confidence: 99%

“… 74,76,77 Figure 5 depicts a fused architecture integrating DL feature extraction and machine learning classifiers for lesion classification from subtraction images of dynamic contrast-enhanced (DCE) MRI sequences. 74 Several papers show that DL methods for breast MRI outperform traditional machine learning methods, particularly as training data set sizes increase. 25 Of the many studies published on DL for breast MRI, only a few include reader studies (two showing that DL models performed similarly to humans, 75,76 and one showing that the DL model was inferior 71 ).…”

Section: Magnetic Resonance Imagingmentioning

confidence: 99%

“… An illustration of combined DL feature extraction methods with traditional machine learning classifiers for lesion classification from cropped ROI’s of MIP and 3D RGB of subtraction images of DCE MRI sequences. 74 The top layer illustrates the construction of the cropped ROI’s from the DCE MRI sequences. MIP and 3D RGB features were integrated with max-pooling and then passed to a machine learning classifier.…”

Section: Magnetic Resonance Imagingmentioning

confidence: 99%

“…MIP and 3D RGB features were integrated with max-pooling and then passed to a machine learning classifier. Reprinted by permission from Radiology: Artificial Intelligence, 74 copyright Radiological Society of North America 2021. DCE, dynamic contrast-enhanced; DL, deep learning; MIP, maximum intensity projection; ROI, region of interest.…”

Section: Magnetic Resonance Imagingmentioning

confidence: 99%

See 2 more Smart Citations

Deep learning in breast imaging

Bhowmik

Eskreis‐Winkler

2022

BJR|Open

View full text Add to dashboard Cite

Millions of breast imaging exams are performed each year in an effort to reduce morbidity and mortality rates of breast cancer. Breast imaging exams are mainly performed for cancer screening, diagnostic workup of suspicious findings, evaluating extent of disease in recently diagnosed breast cancer patients, and determining treatment response. Yet, the interpretation of breast imaging can be subjective, tedious, time-consuming, and prone to human error. Retrospective and small reader studies suggest that deep learning (DL) has great potential to perform medical imaging tasks at or above human-level performance, and may be used to automate aspects of the breast cancer screening process, improve cancer detection rates, decrease unnecessary callbacks and biopsies, optimize patient risk assessment, and open up new possibilities for disease prognostication. Prospective trials are urgently needed to validate these proposed tools, and to evaluate whether they are ready for real-world clinical use. Additionally, new regulatory frameworks are needed to address the unique ethical, medicolegal, and quality control issues that DL algorithms present. In this article, we review the basics of DL, describe recent DL breast imaging applications including cancer detection and risk prediction, and finally conclude by discussing the challenges and future directions of artificial intelligence-based systems in the field of breast cancer.

show abstract

Four‐Dimensional Machine Learning Radiomics for the Pretreatment Assessment of Breast Cancer Pathologic Complete Response to Neoadjuvant Chemotherapy in Dynamic Contrast‐Enhanced MRI

Caballo

Sanderink

Han

et al. 2022

Magnetic Resonance Imaging

View full text Add to dashboard Cite

Background: Breast cancer response to neoadjuvant chemotherapy (NAC) is typically evaluated through the assessment of tumor size reduction after a few cycles of NAC. In case of treatment ineffectiveness, this results in the patient suffering potentially severe secondary effects without achieving any actual benefit. Purpose: To identify patients achieving pathologic complete response (pCR) after NAC by spatio-temporal radiomic analysis of dynamic contrast-enhanced (DCE) MRI images acquired before treatment. Study type: Single-center, retrospective. Population: A total of 251 DCE-MRI pretreatment images of breast cancer patients. Field strength/sequence: 1.5 T/3 T, T1-weighted DCE-MRI. Assessment: Tumor and peritumoral regions were segmented, and 348 radiomic features that quantify texture temporal variation, enhancement kinetics heterogeneity, and morphology were extracted. Based on subsets of features identified through forward selection, machine learning (ML) logistic regression models were trained separately with all images and stratifying on cancer molecular subtype and validated with leave-one-out cross-validation. Statistical tests: Feature significance was assessed using the Mann-Whitney U-test. Significance of the area under the receiver operating characteristics (ROC) curve (AUC) of the ML models was assessed using the associated 95% confidence interval (CI). Significance threshold was set to 0.05, adjusted with Bonferroni correction. Results: Nine features related to texture temporal variation and enhancement kinetics heterogeneity were significant in the discrimination of cases achieving pCR vs. non-pCR. The ML models achieved significant AUC of 0.707 (all cancers, n = 251, 59 pCR), 0.824 (luminal A, n = 107, 14 pCR), 0.823 (luminal B, n = 47, 15 pCR), 0.844 (HER2 enriched, n = 25, 11 pCR), 0.803 (triple negative, n = 72, 19 pCR). Data Conclusions: Differences in imaging phenotypes were found between complete and noncomplete responders. Furthermore, ML models trained per cancer subtype achieved high performance in classifying pCR vs. non-pCR cases. They may, therefore, have potential to help stratify patients according to the level of response predicted before treatment, pending further validation with larger prospective cohorts.

show abstract

Improved Classification of Benign and Malignant Breast Lesions Using Deep Feature Maximum Intensity Projection MRI in Breast Cancer Diagnosis Using Dynamic Contrast-enhanced MRI

Cited by 29 publications

References 34 publications

Multi-Stage Harmonization for Robust AI across Breast MR Databases

Multi-Stage Harmonization for Robust AI across Breast MR Databases

Deep learning in breast imaging

Four‐Dimensional Machine Learning Radiomics for the Pretreatment Assessment of Breast Cancer Pathologic Complete Response to Neoadjuvant Chemotherapy in Dynamic Contrast‐Enhanced MRI

Contact Info

Product

Resources

About