Case-based Similar Image Retrieval for Weakly Annotated Large Histopathological Images of Malignant Lymphoma Using Deep Metric Learning

8 Jul 2021 · Noriaki Hashimoto, Yusuke Takagi, Hiroki Masuda, Hiroaki Miyoshi, Kei Kohno, Miharu Nagaishi, Kensaku Sato, Mai Takeuchi, Takuya Furuta, Keisuke Kawamoto, Kyohei Yamada, Mayuko Moritsubo, Kanako Inoue, Yasumasa Shimasaki, Yusuke Ogura, Teppei Imamoto, Tatsuzo Mishina, Ken Tanaka, Yoshino Kawaguchi, Shigeo Nakamura, Koichi Ohshima, Hidekata Hontani, Ichiro Takeuchi ·

In the present study, we propose a novel case-based similar image retrieval (SIR) method for hematoxylin and eosin (H&E)-stained histopathological images of malignant lymphoma. When a whole slide image (WSI) is used as an input query, it is desirable to be able to retrieve similar cases by focusing on image patches in pathologically important regions such as tumor cells. To address this problem, we employ attention-based multiple instance learning, which enables us to focus on tumor-specific regions when the similarity between cases is computed. Moreover, we employ contrastive distance metric learning to incorporate immunohistochemical (IHC) staining patterns as useful supervised information for defining appropriate similarity between heterogeneous malignant lymphoma cases. In the experiment with 249 malignant lymphoma patients, we confirmed that the proposed method exhibited higher evaluation measures than the baseline case-based SIR methods. Furthermore, the subjective evaluation by pathologists revealed that our similarity measure using IHC staining patterns is appropriate for representing the similarity of H&E-stained tissue images for malignant lymphoma.

PDF Abstract