Month: May 2025

Joonho Lee1, Joonyoung Cho1, DoKyung Kim1, Yoon-La Choi2, Kyungsoo Jung2, Tae-Yeong Kwak1, Sun Woo Kim1, Hyeyoon Chang1

1 Deep Bio Inc., 2 Samsung Medical Center

Disclosure: The authors of this abstract have indicated the following conflicts of interest that relate to the content of this abstract: Joonho Lee, Joonyoung Cho, DoKyung Kim,
Tae-Yeong Kwak, and Hyeyoon Chang are employees of Deep Bio Inc., and Sun Woo Kim is CEO of Deep Bio Inc.

This study aims to develop a deep-learning model for cancer diagnosis in frozen section sentinel lymph nodes with limited annotations. This study proposes a method to reduce stain and scanner variations using a multi-institutional dataset with multiple instance learning (MIL) approaches combined with a classifier-isolate training method. The proposed method outperforms the fine-tuned strategy.

Hyeon Seok Yang1, Yunseob Hwang1, DoKyung Kim1, Yoon-La Choi2, Kyungsoo Jung2, Minjung Sung2, Young Kee Shin3, Ji-Hye Nam4, Tae-Yeong Kwak1, Sun Woo Kim1, Hyeyoon Chang1

1 Deep Bio Inc., 2 Samsung Medical Center, 3 Seoul National University, 4 LOGONE Bio-Convergence Research Foundation

Disclosure: The authors of this abstract have indicated the following conflicts of interest that relate to the content of this abstract: Hyeon Seok Yang, Yunseob Hwang, DoKyung Kim, Tae-Yeong Kwak, and Hyeyoon Chang are employees of Deep Bio Inc., and Sun Woo Kim is CEO of Deep Bio Inc.

This study confirmed the correlation between AI-estimated H-scores through cell analysis based on deep learning and image analysis and pathologist-estimated H-scores for MET IHC stained WSI, and compared the distribution of AI-estimated H-scores by tumor subtype.

Yunseob Hwang1, Gui Young Kwon2, Jeongwon Kim3, Jiyoon Jung3, DoKyung Kim1, Tae-Yeong Kwak1, Sun Woo Kim1, Hyeyoon Chang1

1 Deep Bio Inc., 2 Seoul Clinical Laboratories, 3 Hallym University Sacred Heart Hospital

Disclosure: The authors of this abstract have indicated the following conflicts of interest that relate to the content of this abstract: Yunseob Hwang, DoKyung Kim, Tae-Yeong Kwak,
and Hyeyoon Chang are employees of Deep Bio Inc., and Sun Woo Kim is CEO of Deep Bio Inc..

This research explores deep learning-based image analysis to quantify PD-L1 staining intensity in non-small cell lung cancer, demonstrating a strong correlation between AI-driven PD-L1 intensity and clinically assessed TPS values.