Dual-Intended Deep Learning Model for Breast Cancer Diagnosis in Ultrasound Imaging
Automated medical data analysis demonstrated a significant role in modern medicine, and cancer diagnosis/prognosis to achieve highly reliable and generalizable systems. In this study, an automated breast cancer screening method in ultrasound imaging is proposed. A convolutional deep autoencoder mode...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-05-01
|
| Series: | Cancers |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-6694/14/11/2663 |
| _version_ | 1797493918719279104 |
|---|---|
| author | Nicolle Vigil Madeline Barry Arya Amini Moulay Akhloufi Xavier P. V. Maldague Lan Ma Lei Ren Bardia Yousefi |
| author_facet | Nicolle Vigil Madeline Barry Arya Amini Moulay Akhloufi Xavier P. V. Maldague Lan Ma Lei Ren Bardia Yousefi |
| author_sort | Nicolle Vigil |
| collection | DOAJ |
| description | Automated medical data analysis demonstrated a significant role in modern medicine, and cancer diagnosis/prognosis to achieve highly reliable and generalizable systems. In this study, an automated breast cancer screening method in ultrasound imaging is proposed. A convolutional deep autoencoder model is presented for simultaneous segmentation and radiomic extraction. The model segments the breast lesions while concurrently extracting radiomic features. With our deep model, we perform breast lesion segmentation, which is linked to low-dimensional deep-radiomic extraction (four features). Similarly, we used high dimensional conventional imaging throughputs and applied spectral embedding techniques to reduce its size from 354 to 12 radiomics. A total of 780 ultrasound images—437 benign, 210, malignant, and 133 normal—were used to train and validate the models in this study. To diagnose malignant lesions, we have performed training, hyperparameter tuning, cross-validation, and testing with a random forest model. This resulted in a binary classification accuracy of 78.5% (65.1–84.1%) for the maximal (full multivariate) cross-validated model for a combination of radiomic groups. |
| first_indexed | 2024-03-10T01:26:53Z |
| format | Article |
| id | doaj.art-8416f4d3f82347bbbebdf22380294de0 |
| institution | Directory Open Access Journal |
| issn | 2072-6694 |
| language | English |
| last_indexed | 2024-03-10T01:26:53Z |
| publishDate | 2022-05-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Cancers |
| spelling | doaj.art-8416f4d3f82347bbbebdf22380294de02023-11-23T13:49:02ZengMDPI AGCancers2072-66942022-05-011411266310.3390/cancers14112663Dual-Intended Deep Learning Model for Breast Cancer Diagnosis in Ultrasound ImagingNicolle Vigil0Madeline Barry1Arya Amini2Moulay Akhloufi3Xavier P. V. Maldague4Lan Ma5Lei Ren6Bardia Yousefi7Fischell Department of Bioengineering, University of Maryland, College Park, MD 20742, USAFischell Department of Bioengineering, University of Maryland, College Park, MD 20742, USADepartment of Radiation Oncology, City of Hope Comprehensive Cancer Center, Duarte, CA 91010, USADepartment of Computer Science, Perception Robotics and Intelligent Machines (PRIME) Research Group, University of Moncton, New Brunswick, NB E1A 3E9, CanadaDepartment of Electrical and Computer Engineering, Laval University, Quebec City, QC G1V 0A6, CanadaFischell Department of Bioengineering, University of Maryland, College Park, MD 20742, USADepartment of Radiation Oncology, University of Maryland School of Medicine, Baltimore, MD 21201, USAFischell Department of Bioengineering, University of Maryland, College Park, MD 20742, USAAutomated medical data analysis demonstrated a significant role in modern medicine, and cancer diagnosis/prognosis to achieve highly reliable and generalizable systems. In this study, an automated breast cancer screening method in ultrasound imaging is proposed. A convolutional deep autoencoder model is presented for simultaneous segmentation and radiomic extraction. The model segments the breast lesions while concurrently extracting radiomic features. With our deep model, we perform breast lesion segmentation, which is linked to low-dimensional deep-radiomic extraction (four features). Similarly, we used high dimensional conventional imaging throughputs and applied spectral embedding techniques to reduce its size from 354 to 12 radiomics. A total of 780 ultrasound images—437 benign, 210, malignant, and 133 normal—were used to train and validate the models in this study. To diagnose malignant lesions, we have performed training, hyperparameter tuning, cross-validation, and testing with a random forest model. This resulted in a binary classification accuracy of 78.5% (65.1–84.1%) for the maximal (full multivariate) cross-validated model for a combination of radiomic groups.https://www.mdpi.com/2072-6694/14/11/2663ultrasound imagingbreast cancermedical image analysisdimensionality reductiondeep learningradiomics |
| spellingShingle | Nicolle Vigil Madeline Barry Arya Amini Moulay Akhloufi Xavier P. V. Maldague Lan Ma Lei Ren Bardia Yousefi Dual-Intended Deep Learning Model for Breast Cancer Diagnosis in Ultrasound Imaging Cancers ultrasound imaging breast cancer medical image analysis dimensionality reduction deep learning radiomics |
| title | Dual-Intended Deep Learning Model for Breast Cancer Diagnosis in Ultrasound Imaging |
| title_full | Dual-Intended Deep Learning Model for Breast Cancer Diagnosis in Ultrasound Imaging |
| title_fullStr | Dual-Intended Deep Learning Model for Breast Cancer Diagnosis in Ultrasound Imaging |
| title_full_unstemmed | Dual-Intended Deep Learning Model for Breast Cancer Diagnosis in Ultrasound Imaging |
| title_short | Dual-Intended Deep Learning Model for Breast Cancer Diagnosis in Ultrasound Imaging |
| title_sort | dual intended deep learning model for breast cancer diagnosis in ultrasound imaging |
| topic | ultrasound imaging breast cancer medical image analysis dimensionality reduction deep learning radiomics |
| url | https://www.mdpi.com/2072-6694/14/11/2663 |
| work_keys_str_mv | AT nicollevigil dualintendeddeeplearningmodelforbreastcancerdiagnosisinultrasoundimaging AT madelinebarry dualintendeddeeplearningmodelforbreastcancerdiagnosisinultrasoundimaging AT aryaamini dualintendeddeeplearningmodelforbreastcancerdiagnosisinultrasoundimaging AT moulayakhloufi dualintendeddeeplearningmodelforbreastcancerdiagnosisinultrasoundimaging AT xavierpvmaldague dualintendeddeeplearningmodelforbreastcancerdiagnosisinultrasoundimaging AT lanma dualintendeddeeplearningmodelforbreastcancerdiagnosisinultrasoundimaging AT leiren dualintendeddeeplearningmodelforbreastcancerdiagnosisinultrasoundimaging AT bardiayousefi dualintendeddeeplearningmodelforbreastcancerdiagnosisinultrasoundimaging |