Weakly supervised learning for classification of lung cytological images using attention-based multiple instance learning

• Published in: Scientific reports Vol. 11; no. 1; pp. 20317 - 9
• Main Authors: Teramoto, Atsushi, Kiriyama, Yuka, Tsukamoto, Tetsuya, Sakurai, Eiko, Michiba, Ayano, Imaizumi, Kazuyoshi, Saito, Kuniaki, Fujita, Hiroshi
• Format: Journal Article
• Language: English
• Published: London Nature Publishing Group UK 13.10.2021; Nature Publishing Group; Nature Portfolio
• Subjects: 631/114/1305; 639/705/258; 692/699/67/1612; Adenocarcinoma - diagnostic imaging; Carcinoma, Squamous Cell - diagnostic imaging; Chromatin - chemistry; Classification; Deep Learning; Humanities and Social Sciences; Humans; Image Interpretation, Computer-Assisted - methods; Image Processing, Computer-Assisted - methods; Learning; Lung - diagnostic imaging; Lung cancer; Lung Neoplasms - diagnostic imaging; Malignancy; Medical Informatics - methods; multidisciplinary; Neural networks; Neural Networks, Computer; Pattern Recognition, Automated; Reproducibility of Results; Retrospective Studies; Science; Science (multidisciplinary); Small Cell Lung Carcinoma - diagnostic imaging; Supervised Machine Learning; Thorax
• ISSN: 2045-2322; 2045-2322
• DOI: 10.1038/s41598-021-99246-4

In cytological examination, suspicious cells are evaluated regarding malignancy and cancer type. To assist this, we previously proposed an automated method based on supervised learning that classifies cells in lung cytological images as benign or malignant. However, it is often difficult to label all cells. In this study, we developed a weakly supervised method for the classification of benign and malignant lung cells in cytological images using attention-based deep multiple instance learning (AD MIL). Images of lung cytological specimens were divided into small patch images and stored in bags. Each bag was then labeled as benign or malignant, and classification was conducted using AD MIL. The distribution of attention weights was also calculated as a color map to confirm the presence of malignant cells in the image. AD MIL using the AlexNet-like convolutional neural network model showed the best classification performance, with an accuracy of 0.916, which was better than that of supervised learning. In addition, an attention map of the entire image based on the attention weight allowed AD MIL to focus on most malignant cells. Our weakly supervised method automatically classifies cytological images with acceptable accuracy based on supervised learning without complex annotations.