AUTHOR=Zhang Jianxing , Tao Xing , Jiang Yanhui , Wu Xiaoxi , Yan Dan , Xue Wen , Zhuang Shulian , Chen Ling , Luo Liangping , Ni Dong 

TITLE=Application of Convolution Neural Network Algorithm Based on Multicenter ABUS Images in Breast Lesion Detection

JOURNAL=Frontiers in Oncology

VOLUME=Volume 12 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2022.938413

DOI=10.3389/fonc.2022.938413

ISSN=2234-943X

ABSTRACT=Objective: This study aimed to evaluate a convolution neural network algorithm for breast lesion detection with multi-center ABUS image data developed based on ABUS image and Yolo v5. 
Methods: A total of 741 cases with 2538 volume data of ABUS examination were analyzed which were recruited from 7 hospitals between October 2016 and December 2020,452 volume data of 413 cases were used as internal validation data, and 2086 volume data of 328 cases were used as external validation data.There were 1178 breast lesions in 413 women (161 malignant and 1017 benign) and 1936 lesions in 328 women (57 malignant and 1879 benign). The efficiency and accuracy of the algorithm were analyzed in detecting lesions with different allowable false positive values and lesion sizes, and the difference was compared and analyzed, which included the various indicators in internal validation and external validation data.
Results: The study found that the algorithm had high sensitivity for all categories of lesions, even whether using internal or external validation data. The overall detection rate of the algorithm achieved as high as 78.1% and71.2% in internal and external validation sets, respectively. The algorithm could detect more lesions with the increasing nodule size (87.4% in ≥10mm lesions, but less than 50% in <10mm). The detection rate of BI-RADS 4/5 lesions was higher than that of BI-RADS 3 or 2 (96.5% vs 79.7%vs 74.7% internal, 95.8% vs 74.7%vs 88.4% external). Furthermore, the detection performance was better for malignant nodules than benign（98.1% vs 74.9% internal, 98.2% vs 70.4% external). 
Conclusions: This algorithm showed good detection efficiency in the internal and external validation set, especially for category 4/5 lesions and malignant lesions. However, there were still some deficiencies in detecting category 2 and 3 lesions and lesions smaller than 10mm.