Yukihiro IMAKIIRE Research Fellow


Research Fellow

  • Chiba University Institute for Advanced Academic Research / Graduate School of Medicine

  • Keywords

    DKD, Diabetic kidney disease, Pathological progression, Renal function, Hidden Markov model

Research Theme

Development of a progression prediction model for diabetic kidney disease that takes into account individual characteristics in disease progression


Today, the number of patients with diabetic nephropathy (DKD), a kidney disease caused by diabetes, is increasing. It is estimated that approximately 30% of diabetic patients suffer from this disease. DKD is progressive; in the worst case, it can progress to renal failure, requiring dialysis or kidney transplantation. Prevention of the onset and worsening of symptoms is therefore essential. On the other hand, the progression of the disease varies greatly depending on individual characteristics, so a large number of related factors and combinations of related factors at each time point need to be considered to achieve a highly accurate prediction of the progression. Therefore, we aim to establish a mathematical foundation to apply to predictive and personalized medicine by introducing AI and using the large amount of big data accumulated through advances in measurement technology.

This investigation aims to derive and characterize latent states based on a hidden Markov model (CNN-HMM) utilizing dimensionality compression with convolutional neural networks (CNN), considering individual characteristics in pathological progression, especially in the speed of pathological progression. Ultimately, we aim to establish a highly accurate prediction model of DKD progression.