Simukayi Mutasa1, Peter Chang2, John Nemer1, Eduardo Pascual Van Sant1, Mary Sun1, Alison McIlvride1, Maham Siddique1, Richard Ha3. 1. Department of Radiology, Columbia University Medical Center, New York, NY. 2. Center for Artificial Intelligence in Diagnostic Medicine (CAIDM), Division of Neuroradiology, UCI Health, Department of Radiological Sciences, Orange, CA. 3. Department of Radiology, Columbia University Medical Center, New York, NY; Breast Imaging Section, Columbia University Medical Center, New York, NY. Electronic address: rh2616@cumc.columbia.edu.
Abstract
INTRODUCTION: We previously developed a convolutional neural networks (CNN)-based algorithm to distinguish atypical ductal hyperplasia (ADH) from ductal carcinoma in situ (DCIS) using a mammographic dataset. The purpose of this study is to further validate our CNN algorithm by prospectively analyzing an unseen new dataset to evaluate the diagnostic performance of our algorithm. MATERIALS AND METHODS: In this institutional review board-approved study, a new dataset composed of 280 unique mammographic images from 140 patients was used to test our CNN algorithm. All patients underwent stereotactic-guided biopsy of calcifications and underwent surgical excision with available final pathology. The ADH group consisted of 122 images from 61 patients with the highest pathology diagnosis of ADH. The DCIS group consisted of 158 images from 79 patients with the highest pathology diagnosis of DCIS. Two standard mammographic magnification views (craniocaudal and mediolateral/lateromedial) of the calcifications were used for analysis. Calcifications were segmented using an open source software platform 3D slicer and resized to fit a 128 × 128 pixel bounding box. Our previously developed CNN algorithm was used. Briefly, a 15 hidden layer topology was used. The network architecture contained 5 residual layers and dropout of 0.25 after each convolution. Diagnostic performance metrics were analyzed including sensitivity, specificity, accuracy, and area under the receiver operating characteristic curve. The "positive class" was defined as the pure ADH group in this study and thus specificity represents minimizing the amount of falsely labeled pure ADH cases. RESULTS: Area under the receiver operating characteristic curve was 0.90 (95% confidence interval, ± 0.04). Diagnostic accuracy, sensitivity, and specificity was 80.7%, 63.9%, and 93.7%, respectively. CONCLUSION: Prospectively tested on new unseen data, our CNN algorithm distinguished pure ADH from DCIS using mammographic images with high specificity.
INTRODUCTION: We previously developed a convolutional neural networks (CNN)-based algorithm to distinguish atypical ductal hyperplasia (ADH) from ductal carcinoma in situ (DCIS) using a mammographic dataset. The purpose of this study is to further validate our CNN algorithm by prospectively analyzing an unseen new dataset to evaluate the diagnostic performance of our algorithm. MATERIALS AND METHODS: In this institutional review board-approved study, a new dataset composed of 280 unique mammographic images from 140 patients was used to test our CNN algorithm. All patients underwent stereotactic-guided biopsy of calcifications and underwent surgical excision with available final pathology. The ADH group consisted of 122 images from 61 patients with the highest pathology diagnosis of ADH. The DCIS group consisted of 158 images from 79 patients with the highest pathology diagnosis of DCIS. Two standard mammographic magnification views (craniocaudal and mediolateral/lateromedial) of the calcifications were used for analysis. Calcifications were segmented using an open source software platform 3D slicer and resized to fit a 128 × 128 pixel bounding box. Our previously developed CNN algorithm was used. Briefly, a 15 hidden layer topology was used. The network architecture contained 5 residual layers and dropout of 0.25 after each convolution. Diagnostic performance metrics were analyzed including sensitivity, specificity, accuracy, and area under the receiver operating characteristic curve. The "positive class" was defined as the pure ADH group in this study and thus specificity represents minimizing the amount of falsely labeled pure ADH cases. RESULTS: Area under the receiver operating characteristic curve was 0.90 (95% confidence interval, ± 0.04). Diagnostic accuracy, sensitivity, and specificity was 80.7%, 63.9%, and 93.7%, respectively. CONCLUSION: Prospectively tested on new unseen data, our CNN algorithm distinguished pure ADH from DCIS using mammographic images with high specificity.
Authors: Christopher V Nguyen; Constance T Albarracin; Gary J Whitman; Adriana Lopez; Nour Sneige Journal: Ann Surg Oncol Date: 2010-10-23 Impact factor: 5.344
Authors: Eunyoung Ko; Wonshik Han; Jong Won Lee; Jihyoung Cho; Eun-Kyu Kim; So-Youn Jung; Mee Joo Kang; Woo Kyung Moon; In Ae Park; Sung-Won Kim; Ku Sang Kim; Eun Sook Lee; Kyu Hong Min; Seok Won Kim; Dong-Young Noh Journal: Breast Cancer Res Treat Date: 2007-12-04 Impact factor: 4.872
Authors: Lin-Ying Chen; Jintao Hu; Julia Y S Tsang; Michelle A Lee; Yun-Bi Ni; Siu-Ki Chan; Gary M K Tse Journal: Histopathology Date: 2019-07-16 Impact factor: 5.087
Authors: S Bendifallah; S Defert; N Chabbert-Buffet; N Maurin; J Chopier; M Antoine; C Bezu; D Touche; S Uzan; O Graesslin; R Rouzier Journal: Eur J Cancer Date: 2011-11-17 Impact factor: 9.162