News | Artificial Intelligence | April 06, 2021

Artificial Intelligence Improves Accuracy of Breast Ultrasound Diagnoses

Scientists train deep learning algorithms to classify breast lesions from ultrasound images in a large multi-center study

Ultrasound is an invaluable diagnostic tool for the early detection of breast cancer, but the classification of lesions is sometimes challenging and time consuming. Could artificial intelligence hold the answer to solving these problems? Graphic courtesy of Chinese Medical Journal

Ultrasound is an invaluable diagnostic tool for the early detection of breast cancer, but the classification of lesions is sometimes challenging and time consuming. Could artificial intelligence hold the answer to solving these problems? Graphic courtesy of Chinese Medical Journal

April 6, 2021 — In 2020, the International Agency for Research on Cancer of the World Health Organization stated that breast cancer accounts for most cancer morbidities and mortalities in women worldwide. This alarming statistic not only necessitates newer methods for the early diagnosis of breast cancer, but also brings to light the importance of risk prediction of the occurrence and development of this disease. Ultrasound is an effective and noninvasive diagnostic procedure that truly saves lives; however, it is sometimes difficult for ultrasonologists to distinguish between malignant tumors and other types of benign growths. In particular, in China, breast masses are classified into four categories: benign tumors, malignant tumors, inflammatory masses, and adenosis (enlargement of milk-producing glands). When a benign breast mass is misdiagnosed as a malignant tumor, a biopsy usually follows, which puts the patient at unnecessary risk. The correct interpretation of ultrasound images is made even harder when factoring in the large workload of medical specialists.

Could deep learning algorithms be the solution to this conundrum? Professor Wen He, M.D., (Beijing Tian Tan Hospital, Capital Medical University, China) thinks so. "Artificial intelligence is good at identifying complex patterns in images and quantifying information that humans have difficulty detecting, thereby complementing clinical decision making," he states. Although much progress has been made in the integration of deep learning algorithms into medical image analysis, most studies in breast ultrasound deal exclusively with the differentiation of malignant and benign diagnoses. In other words, existing approaches do not try to categorize breast masses into the four abovementioned categories.

To tackle this limitation, He, in collaboration with scientists from 13 hospitals in China, conducted the largest multicenter study on breast ultrasound yet in an attempt to train convolutional neural networks (CNNs) to classify ultrasound images. As detailed in their paper published in Chinese Medical Journal, the scientists collected 15,648 images from 3,623 patients and used half of them to train and the other half to test three different CNN models. The first model only used 2D ultrasound intensity images as input, whereas the second model also included color flow Doppler images, which provide information on blood flow surrounding breast lesions. The third model further added pulsed wave Doppler images, which provide spectral information over a specific area within the lesions.

Each CNN consisted of two modules. The first one, the detection module, contained two main submodules whose overall task was to determine the position and size of the breast lesion in the original 2D ultrasound image. The second module, the classification module, received only the extracted portion from the ultrasound images containing the detected lesion. The output layer contained four categories corresponding to each of the four classifications of breast masses commonly used in China.

First, the scientists checked which of the three models performed better. The accuracies were similar and around 88%, but the second model including 2D images and color flow Doppler data performed slightly better than the other two. The reason the pulsed wave Doppler data did not contribute positively to performance may be that few pulsed wave images were available in the overall dataset. Then, researchers checked if differences in tumor size caused differences in performance. While larger lesions resulted in increased accuracy in benign tumors, size did not appear to have an effect on accuracy when detecting malignancies. Finally, the scientists put one of their CNN models to the test by comparing its performance to that of 37 experienced ultrasonologists using a set of 50 randomly selected images. The results were vastly in favor of the CNN in all regards, as He remarked "The accuracy of the CNN model was 89.2%, with a processing time of less than two seconds. In contrast, the average accuracy of the ultrasonologists was 30%, with an average time of 314 seconds."

This study clearly showcases the capabilities of deep learning algorithms as complementary tools for the diagnosis of breast lesions through ultrasound. Moreover, unlike previous studies, the researchers included data obtained using ultrasound equipment from different manufacturers, which hints at the remarkable applicability of the trained CNN models regardless of the ultrasound devices present at each hospital. In the future, the integration of artificial intelligence into diagnostic procedures with ultrasound could speed up the early detection of cancer. It would also bring about other benefits, as Dr. He explains: "Because CNN models do not require any type of special equipment, their diagnostic recommendations could reduce predetermined biopsies, simplify the workload of ultrasonologists, and enable targeted and refined treatment."

Let us hope artificial intelligence soon finds a home in ultrasound image diagnostics so doctors can work smarter, not harder.

For more information: www.cactusglobal.com/

Related Content

The impact of deploying artificial intelligence (AI) for radiation cancer therapy in a real-world clinical setting has been tested by Princess Margaret researchers in a unique study involving physicians and their patients.

Getty Images

News | Artificial Intelligence | June 15, 2021
June 15, 2021 — The impact of deploying ...
The prevalence of genetic mutations associated with breast cancer in Black and white women is the same, according to a new JAMA Oncology study of nearly 30,000 patients led by researchers in the Basser Center for BRCA at the Abramson Cancer Center.

Getty Images

News | Women's Health | June 15, 2021
June 15, 2021 — The prevalence of genetic mutations associated with breast cancer in Black and white women is the sam
A cardiac MRI of athletes who had COVID-19 is seven times more effective in detecting inflammation of the heart than symptom-based testing, according to a study led by researchers at The Ohio State University Wexner Medical Center and College of Medicine with 12 other Big Ten programs.

Cardiac Magnetic Resonance Imaging in Athletes With Clinical and Subclinical Myocarditis A-D, Athlete A with subclinical possible myocarditis was asymptomatic with normal electrocardiogram (ECG), echocardiogram, and high-sensitivity troponin findings. A, T2 mapping showing elevated T2 in basal-mid inferolateral wall in short axis view. B, late gadolinium enhancement (LGE) in the basal inferolateral wall in short axis view. C, Postcontrast steady state-free precession (SSFP) images showing contrast uptake in the basal-mid inferolateral wall in short axis view. D, LGE in the inferolateral wall in 3-chamber view. E-H, Athlete B with subclinical probable myocarditis was asymptomatic with normal ECG, normal echocardiogram, and elevated high-sensitivity troponin findings. E, T2 mapping showing elevated T2 in the anteroseptal wall in short axis view. F, LGE in the anteroseptal wall in 3-chamber view. G, T2 mapping showing elevated T2 in the anteroseptal wall in 3-chamber view. F, Postcontrast SSFP image showing pericardial effusion in short axis view. I-K, Athlete C with clinical myocarditis and chest pain, dyspnea, abnormal ECG, normal echocardiogram, and normal troponin findings. I, T2 mapping showing elevated T2 in the lateral wall short axis view. J, Postcontrast SSFP images showing contrast uptake in midlateral wall in short axis view. K, LGE in the epicardial midlateral wall in short axis view. L-N, Athlete D with clinical myocarditis, chest pain, abnormal ECG, echocardiogram, and troponin findings. L, T1 mapping showing elevated native T1 in midlateral wall in short axis view. M, T2 mapping showing elevated T2 in the midlateral wall in short axis view. N, LGE in the epicardial midlateral wall in short axis view. IR indicates inferior right view; IRP, inferior, right, posterior view; PLI, posterior, left, inferior view; SL, superior left view; SLA, superior, left, anterior view. Image courtesy of JAMA Cardiol. Published online May 27, 2021. doi:10.1001/jamacardio.2021.2065

News | Cardiac Imaging | June 15, 2021
June 15, 2021 — A...
Rensselaer algorithm can identify risk of cardiovascular disease using lung cancer scan #CT
News | Computed Tomography (CT) | June 14, 2021
June 14, 2021 — Heart disease and cancer are the ...
A new imaging technique has the potential to detect neurological disorders — such as Alzheimer's disease — at their earliest stages, enabling physicians to diagnose and treat patients more quickly. Termed super-resolution, the imaging methodology combines position emission tomography (PET) with an external motion tracking device to create highly detailed images of the brain.

Result of the Hoffman brain phantom study. Top row: same PET slice reconstructed with A) 2mm static OSEM, B) 1mm static OSEM, C) proposed SR method and D) corresponding CT slice (note that the CT image can be treated as a high-resolution reference). Middle row: zoom on region of interest for corresponding images. Bottom row: Line profiles for corresponding data. Image created by Y Chemli, et al., Gordon Center for Medical Imaging: Department of Radiology Massachusetts General Hospital, Harvard Medical School, Boston, MA.

News | PET Imaging | June 14, 2021
June 14, 2021 — A new imaging technique has the potential to detect neurological disorders — such as...
Prediction performance of DL compared to quantitative measures and Kaplan-Meier curves for quartiles of DL. Image created by Singh et al., Cedars-Sinai Medical Center, Los Angeles, CA.

Prediction performance of DL compared to quantitative measures and Kaplan-Meier curves for quartiles of DL. Image created by Singh et al., Cedars-Sinai Medical Center, Los Angeles, CA.

News | SPECT Imaging | June 14, 2021
June 14, 2021 — An advanced artificial i...
Accuray Incorporated announced the company has received CE Mark certification for its ClearRT helical fan-beam kVCT imaging capability.
News | Radiation Therapy | June 11, 2021
June 11, 2021 — Accuray Incorporated announced the company has received CE Mark certification for its...