Visual and quantitative evaluation of microcalcifications in mammograms with deep learning-based super-resolution. Issue 154 (September 2022)

• Record Type: Journal Article
• Authors: Honjo, Takashi; Ueda, Daiju; Katayama, Yutaka; Shimazaki, Akitoshi; Jogo, Atsushi; Kageyama, Ken; Murai, Kazuki; Tatekawa, Hiroyuki; Fukumoto, Shinya; Yamamoto, Akira; Miki, Yukio
• Is Part Of: European journal of radiology. Issue 154(2022)
• Publication Date: 2022-09
• Subjects: MLO medio-lateral oblique -- DL deep learning -- SR super-resolution -- CNN convolutional neural network -- FSRCNN Fast Super-Resolution Convolutional Neural Network -- SRCNN Super-Resolution Convolutional Neural Network -- SSIM structural similarity -- PSNR peak signal-to-noise ratio -- BI-RADS® Breast Imaging Reporting and Data System -- PIQE perception-based image-quality evaluator; Breast Cancer -- Mammography -- Microcalcification -- Deep Learning -- Artificial Intelligence -- Super Resolution; Medical radiology -- Periodicals; Radiology -- Periodicals; Radiologie médicale -- Périodiques; Medical radiology; Periodicals; 616.075705
• DOI: 10.1016/j.ejrad.2022.110433
• Languages: English
• ISSNs: 0720-048X

Abstract

Highlights: The first study to visually evaluate microcalcifications of mammograms using a deep learning-based super-resolution (SR). Our SR model can improve microcalcification visibility in mammography. The mean PIQE (perception-based image-quality evaluator) of SR mammograms was significantly rated better than the originals. Abstract: Purpose: To evaluate visually and quantitatively the performance of a deep-learning-based super-resolution (SR) model for microcalcifications in digital mammography. Method: Mammograms were consecutively collected from 5080 patients who underwent breast cancer screening from January 2015 to March 2017. Of these, 93 patients (136 breasts, mean age, 50 ± 7 years) had microcalcifications in their breasts on mammograms. We applied an artificial intelligence model known as a fast SR convolutional neural network to the mammograms. SR and original mammograms were visually evaluated by four breast radiologists using a 5-point scale (1: original mammograms are strongly preferred, 5: SR mammograms are strongly preferred) for the detection, diagnostic quality, contrast, sharpness, and noise of microcalcifications. Mammograms were quantitatively evaluated using a perception-based image-quality evaluator (PIQE). Results: All radiologists rated the SR mammograms better than the original ones in terms of detection, diagnostic quality, contrast, and sharpness of microcalcifications. These ratings were significantly different according to the Wilcoxon signed-rank test (p <.001), while the noise score of the three radiologists was significantly lower (p <.001). According to PIQE, SR mammograms were rated better than the original mammograms, showing a significant difference by paired t -test (p <.001). Conclusion: An SR model based on deep learning can improve the visibility of microcalcifications in mammography and help detect and diagnose them in mammograms.