Program        Invited speakers

Invited speakers

IFMIA Invited Lecture 1 (Wednesday, January 11, 2023, 9:00-9:40)
"PET tracing of microplastic distribution"
Dr. Jin Su Kim,
Korea Institute of Radiological & Medical Sciences, Korea
Microplastics, an environmental threat and human health risk, are widely detected in food, and consequently ingested. Therefore, real-time monitoring of orally administered microplastic is tremendously important to trace them in the body. First part of this talk, to visualize their absorption path, we labeled polystyrene with [64Cu]Cu-DOTA. The absorption path and distribution of [64Cu]Cu-DOTA-polystyrene were determined using positron emission tomography (PET). This study provided evidence of microplastic accumulation and existence in tissue by using PET imaging. I will also talk pre/post-natal exposure to microplastic as a potential risk factor for autism spectrum disorder. Second part of this talk, I will talk the development of combination immunotherapy. Radioimmunotherapy (RIT) using monoclonal antibodies (mAbs) labeled with radionuclides is an attractive approach for cancer treatment. However, due to various limitations, mAbs cannot reach solid tumors, consequently reducing RIT efficacy. Thus, we developed the combination immunotherapy for treatment of incurable cancer such as pancreatic cancer. The combination of drugs or moieties with RIT would be helpful to overcome the barriers that RIT faces for solid tumors.
Jin Su Kim is an active researcher in the filed of PET quantification and targeted immunotherapy. He developed radio/gene immunotherapy for the treatment of pancreatic cancer. Recently, He traced bio-distribution of microplastic using PET. He is the Lab Head, Molecular Imaging and Targeted therapy in Korea Institute of Radiological & Medical Sciences, Korea. He is a Professor in University of Science and Technology, Korea, where he also served as the chair.
IFMIA Invited Lecture 2 (Wednesday, January 11, 2023, 9:40-10:20)
"Toward the Construction of Quantitative Evaluation Criteria for the Diagnosis of Malignant Lymphoma"
Prof. Hidekata Hontani,
Nagoya Institute of Technology, Japan
This talk discusses our attempt to construct an explainable classifier of malignant lymphoma subtypes based on pathological images and the difficulties we face in this attempt. In the diagnosis of malignant lymphoma, the tissue patterns and the morphologies of the cell nuclei in H&E-stained pathological images are first observed to identify the candidate subtypes. Here, this identification is not made based on quantitative evaluation of features of tissue patterns or cell nucleus morphologies but based on criteria acquired by the pathologists through experience. Although pathologists can identify the candidates of subtypes, they are often unable to accurately describe the criteria for their identification. Identifying the diagnostic rationale and allowing its quantitative evaluation would be largely useful in the diagnosis of malignant lymphoma.
For the construction of the evaluation criteria that can be used to quantitatively identify malignant lymphoma subtypes, we employed an approach that first constructs a classifier of the subtypes and second approximates the classification function realized by the constructed classifier with a decision tree based on features that can be interpreted by the pathologists. This talk shows some difficulties we have faced in this approaches that include label-noise found in semi-supervised learning and in outliers in counterfactual image generation.
Hidekata Hontani is a researcher in computer vision. He had been involved in the Research and Development Center of Toshiba Corporation. He is a Professor in Department of Computer Science in Nagoya Institute of Technology in Japan.