|
10:45 - 11:15
|
Manuscript ID. 0901
Paper No. 2022-SAT-S0603-I001
|
Invited Speaker:
Bi-Chang Chen
|
Probing the protein orientation at its native state
Bi-Chang Chen, Academia Sinica (Taiwan)
3D optical imaging of biological tissue at high spatial resolution over a large scale bridges the observation and understanding of biological systems at cellular and tissue level. Imaging and reconstruction of physically sectioned tissue slices was perhaps the only way to perform such study. Although it is possible to achieve submicron, even super resolution by imaging thin tissue slices, the method wasn’t widely adopted due to the technical limitations and practical difficulties to implement the technique in different researches.
Expansion microscopy (ExM) is an emerging technology that enables biological samples to be imaged with nanoscale precision and resolution on ordinary, diffraction-limited microscopes. ExM generally works by physically magnifying a specimen in a uniform (isotropic) manner in three dimensions, which could easily achieve super-resolved image on the diffraction-limited microscope.
The holy grail in optical microscopy is imaging 3D biological specimen with the resolution, the same as in electron microscopy. Electron microscopy (EM) has a lot better spatial resolution but no color information and presumably limited to 2D. Instead, optical microscopy has chemical information due to fluorescent labeling with 3D fashion. Here, we are performing lightsheet microscopy combined with expansion microscopy to let optical microscopy meet electron microscopy and aim to map the orientation of the protein of interest at the intact tissue without physically sectioning.
 Preview abstract
|
|
|
11:15 - 11:30
|
Manuscript ID. 0062
Paper No. 2022-SAT-S0603-O001
|
| Hsin-Jou Wang
|
Bone mineral density prediction by deep learning with optical bone densitometry
Hsin-Jou Wang, National Yang Ming Chiao Tung University (Taiwan); Wei-Chun Chang, Taipei Municipal Wan fang Hospital (Taiwan); Tsai-Hsueh Leu, Taipei City Hospital Renai Branch (Taiwan); Yi-Min Wang, Gautam Takhellambam, Chia-Wei Sun, National Yang Ming Chiao Tung University (Taiwan)
Osteoporosis is a severe health problem in an aging society, causing patients to suffer a higher risk of bone injury. As a result, early diagnosis is essential. To achieve quick, easy, low-cost, and non-invasive detection of osteoporosis, we develop an optical bone densitometer to predict bone mineral density via deep learning algorithm. The recent results are improved from our previous study, revealing the potential of OBD in BMD prediction.
Preview abstract
|
|
|
11:30 - 11:45
|
Manuscript ID. 0178
Paper No. 2022-SAT-S0603-O002
|
| Chang-Yi Lee
|
Machine learning classification applying the combination of functional near-infrared spectroscopy and APACHE-II scoring on extracorporeal membrane oxygenation patients
Chang-Yi Lee, Ting-Wei Chiang, National Yang Ming Chiao Tung University (Taiwan); Hsiao-Huang Chang, Taipei Veterans General Hospital (Taiwan); Chia-Wei Sun, National Yang Ming Chiao Tung University (Taiwan)
Our study aims to use non-invasive near-infrared spectroscopy (NIRS) to detect the changes in blood oxygen concentration of extracorporeal membrane oxygenation (ECMO) patients when adjusting speed. First, we apply various data processing to the measured blood oxygen value and record the patient's APACHE-II scale score. Patients are divided into two groups by scores of 24. Afterward, we combine the metric blood oxygen information with the APACHE-II scores and do binary classification. In Veno-Arterial (VA) group, we get the train and test accuracies of 83.3% and 81.8%, while in Veno-Venous (VV) group, the corresponding accuracies are 81.8% and 72.7%.
Preview abstract
|
|
|
11:45 - 12:00
|
Manuscript ID. 0177
Paper No. 2022-SAT-S0603-O003
|
| Yung-Chang Chen
|
Application of machine learning for operation outcome from peripheral arterial occlusive disease
Yung-Chang Chen, Pin-Yu Kuo, Biomedical Optical Imaging Lab (Taiwan); Jen-Kuang Lee, Chau-Chung Wu, Division of Cardiology (Taiwan); Chia-Wei Sun, Biomedical Optical Imaging Lab (Taiwan)
Providing an appropriate prognosis for patients with peripheral arterial occlusive disease (PAOD) is essential. In this study, near-infrared spectroscopy (NIRS) was used to monitor the change of blood oxygenation of the lower limbs, combined with support vector machine (SVM) to build a prediction model of surgical effect. The model training and testing accuracies are 89.19 % and 80 %, respectively. In addition, the result of feature selection showed that tissue saturation index (TSI) could reflect the adjustment and stability of blood oxygenation of patients under external influence.
Preview abstract
|
|