Current Affiliation Organization 【 display / non-display 】

Current Affiliation Organization 【 display / non-display 】

University 【 display / non-display 】

University 【 display / non-display 】

External Career 【 display / non-display 】

External Career 【 display / non-display 】

Published Papers 【 display / non-display 】

Published Papers 【 display / non-display 】

• A Practical Study of Data Requirements for Self-Supervised Learning in Medical Image Analysis

  Mitsuki Hommyo, Takumasa Tsuji, Shinobu Kumagai, Kenshiro Shiraishi, Jun'ichi Kotoku

  Cureus ( Springer Science and Business Media LLC )    2025.11 [ Peer Review Accepted ]

  Type of publication: Research paper (scientific journal)

  • Access this article

    DOI

• Overcoming Clinical Inertia in the Context of Non-Adherence With Guideline-Directed Medical Therapy for Heart Failure

  Takahiro Okumura, Kenya Kusunose, Takumasa Tsuji, Jun'ichi Kotoku, Koji Todaka, Keita Saku

  Cardiology Discovery ( Ovid Technologies (Wolters Kluwer Health) )  5 ( 3 ) 246 - 256   2025.09 [ Peer Review Accepted ]

  Type of publication: Research paper (scientific journal)

  　View Summary

  Abstract Heart failure (HF) remains a leading cause of morbidity and mortality worldwide, despite advancements in guideline-directed medical therapies (GDMTs). A major obstacle to optimal HF management is clinical inertia, defined as the failure of health care providers to initiate or intensify therapy when indicated. This review examined the current state, contributing factors, and strategies for overcoming clinical inertia in HF. Studies have revealed substantial treatment gaps, with sub-optimal prescription rates and dosing of GDMT classes, including angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, angiotensin receptor-neprilysin inhibitors, β-blockers, mineralocorticoid receptor antagonists, and sodium-glucose cotransporter-2 inhibitors. These treatment gaps persist over time and are associated with increased patient mortality and hospitalizations as well as reduced quality of life. Multiple factors contribute to clinical inertia, including patient-related factors (medication adherence and treatment understanding), provider-related factors (guideline familiarity, concerns regarding side effects, complex treatment decision-making), and health care system-related factors (fragmented care models and quality assessment frameworks). Strategies for overcoming clinical inertia involve patient empowerment through education and shared decision-making, provider education and clinical decision support tools, and redesigning HF care delivery. Specialized HF management systems, multidisciplinary collaboration, remote monitoring, and digital tools can promote guideline adherence. Continuous quality improvement by integrating research and practice is also essential. Addressing clinical inertia requires a multifaceted approach targeting patients, providers, and health care systems. By implementing targeted strategies, health care systems can bridge the evidence-practice gap, optimize GDMT utilization, and ultimately improve outcomes for this vulnerable patient population.

  • Access this article

    DOI

• <i>In silico</i> Prediction of Loop-Mediated Isothermal Amplification Using a Generalized Linear Model

  Kenshiro Taguchi, Satoru Michiyuki, Takumasa Tsuji, Jun'ichi Kotoku

  Synthetic Biology     2025.03 [ Peer Review Accepted ]

  Type of publication: Research paper (scientific journal)

  　View Summary

  <jats:title>ABSTRACT</jats:title> <jats:p>Loop-Mediated Isothermal Amplification (LAMP), a DNA amplification technique under isothermal conditions, provides the important benefits of high sensitivity, specificity, rapidity, and simplicity. Maximizing LAMP features necessitates the design of a complex LAMP primer set (LPS) consisting of four primers for six regions of a given target DNA. Furthermore, the LPS of a given target DNA is designed with LPS design support software such as Primer Explorer. However, even if the design is completed, we still must do many in vitro experiments and evaluations. Consequently, designing LPS often fails to achieve high performance, including efficient amplification. For this study, we examined in silico LAMP: a generalized linear model to predict DNA amplification from LPS. Using logistic regression with elastic net regularization, we identified factors that strongly affect LPS design. These factors, combined with domain knowledge for LPS design, led to the creation of LAMP kernel variables that are highly essential for high LAMP reaction. In silico LAMP, constructed using logistic regression with LAMP kernel variables, allows classification and performance prediction of LPS with an AUC of 0.86. These results suggest that a high LAMP reaction can be predicted using LAMP kernel variables and GLM. Moreover, an LPS with high performance can be constructed without in vitro experimentation.</jats:p>

  • Access this article

    DOI

• Cone Beam Computed Tomography Image-Quality Improvement Using "One-Shot" Super-resolution.

  Takumasa Tsuji, Soichiro Yoshida, Mitsuki Hommyo, Asuka Oyama, Shinobu Kumagai, Kenshiro Shiraishi, Jun'ichi Kotoku

  Journal of Imaging Informatics in Medicine   38 ( 4 ) 2120 - 2133   2024.12 [ Peer Review Accepted ]

  Type of publication: Research paper (scientific journal)

  　View Summary

  Cone beam computed tomography (CBCT) images are convenient representations for obtaining information about patients' internal organs, but their lower image quality than those of treatment planning CT images constitutes an important shortcoming. Several proposed CBCT image-quality improvement methods based on deep learning require large amounts of training data. Our newly developed model using a super-resolution method, "one-shot" super-resolution (OSSR) based on the "zero-shot" super-resolution method, requires only small amounts of training data to improve CBCT image quality using only the target CBCT image and the paired treatment planning CT image. For this study, pelvic CBCT images and treatment planning CT images of 30 prostate cancer patients were used. We calculated the root mean squared error (RMSE), peak signal-to-noise ratio (PSNR), and structural similarity (SSIM) to evaluate image-quality improvement and normalized mutual information (NMI) as a quantitative evaluation of positional accuracy. Our proposed method can improve CBCT image quality without requiring large amounts of training data. After applying our proposed method, the resulting RMSE, PSNR, SSIM, and NMI between the CBCT images and the treatment planning CT images were as much as 0.86, 1.05, 1.03, and 1.31 times better than those obtained without using our proposed method. By comparison, CycleGAN exhibited values of 0.91, 1.03, 1.02, and 1.16. The proposed method achieved performance equivalent to that of CycleGAN, which requires images from approximately 30 patients for training. Findings demonstrated improvement of CBCT image quality using only the target CBCT images and the paired treatment planning CT images.

  • Access this article

    DOI
  • Search related information
    PubMed

• Echocardiographic artificial intelligence for pulmonary hypertension classification

  Hirata, Y; Tsuji, T; Kotoku, J; Sata, M; Kusunose, K

  HEART ( Heart )  110 ( 8 ) 586 - 593   2024.04 [ Peer Review Accepted ]

  Type of publication: Research paper (scientific journal)

  　View Summary

  OBJECTIVE: The classification of pulmonary hypertension (PH) is crucial for determining the appropriate therapeutic strategy. We investigated whether machine learning (ML) algorithms may assist in echocardiographic PH prediction, where current guidelines recommend integrating several different parameters. METHODS: We obtained physical and echocardiographic data from 885 patients who underwent right heart catheterisation (RHC). Patients were classified into three groups: non-PH, precapillary PH and postcapillary PH, based on values obtained from RHC. Using 24 parameters, we created predictive models employing four different classifiers and selected the one with the highest area under the curve. We then calculated the macro-average classification accuracy for PH on the derivation cohort (n=720) and prospective validation data set (n=165), comparing the results with guideline-based echocardiographic assessment obtained from each cohort. RESULTS: Logistic regression with elastic net regularisation had the highest classification accuracy, with areas under the curves of 0.789, 0.766 and 0.742 for normal, precapillary PH and postcapillary PH, respectively. The ML model demonstrated significantly better predictive accuracy than the guideline-based echocardiographic assessment in the derivation cohort (59.4% vs 51.6%, p<0.01). In the independent validation data set, the ML model's accuracy was comparable to the guideline-based PH classification (59.4% vs 57.8%, p=0.638). CONCLUSIONS: This preliminary study suggests promising potential for our ML model in predicting echocardiographic PH. Further research and validation are needed to fully assess its clinical utility in PH diagnosis and treatment decision-making.

  • Access this article

    DOI
  • Search related information
    PubMed

display all >>

Presentations 【 display / non-display 】

Presentations 【 display / non-display 】

• Optimizing Data Augmentation for Medical Image Classification

  Kazuki Wakao, Takumasa Tsuji, Mitsuki Hommyo, Atsuhiro Tauchi, Yuta Hori, Jun'ichi Kotoku

  The 10th Japan-Korea Joint Meeting on Medical Physics  1900.01  -  1900.01 

• Memory Consumption Reduction in 3D U-Net for Medical Image Segmentation Using Patch-Based Method

  Masaya Mtatsuki, Takumasa Tsuji, Jun'ichi Kotoku

  The 10th Japan-Korea Joint Meeting on Medical Physics  1900.01  -  1900.01 

• Machine learning to classify pulmonary hypertension using echocardiographic measurements

  Takumasa Tsuji, Yukina Hirata, Kenya Kusunose, Masataka Sata, Jun'ichi Kotoku

  The 125th Scientific Meeting of the Japan Society of Medical Physics  1900.01  -  1900.01 

• Detecting pulmonary hypertension using chest X-ray images by convolutional neural network

  Takumasa Tsuji, Yukina Hirata, Kenya Kusunose, Masataka Sata, Jun'ichi Kotoku

  The 121th Scientific Meeting of the Japan Society of Medical Physics  1900.01  -  1900.01 

• Detecting of elevated pulmonary arterial wedge pressure using chest X-ray image by convolutioanl neural network

  Takumasa Tsuji, Yukina Hirata, Kenya Kusunose, Masataka Sata, Jun'ichi Kotoku

  The 123th Scientific Meeting of the Japan Society of Medical Physics  1900.01  -  1900.01 

display all >>

Media Coverage 【 display / non-display 】

Media Coverage 【 display / non-display 】

• Enhancing Cone Beam CT Imaging with "One-Shot" Super-Resolution  Internet

  The Healthcare Management and Leadership Portal  The Healthcare Management and Leadership Portal  https://healthmanagement.org/c/imaging/News/enhancing-cone-beam-ct-imaging-with-one-shot-super-resolution  2024.12

  Author： Other