Edward Pei-Chuan Huang

Dr. Edward Pei-Chuan Huang


Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 July 2024 in Electronics
Verified
Journal Article
Open access
PDF file
MBMadankumar Balasubramani
Madankumar Balasubramani
CSChih-Wei Sung
Chih-Wei Sung
MHMu-Yang Hsieh
Mu-Yang Hsieh
EHEdward Pei-Chuan Huang
Edward Pei-Chuan Huang
Jiann-Shing ShiehJiann-Shing Shieh
Jiann-Shing Shieh
Maysam AbbodMaysam Abbod
Maysam Abbod
Accurate segmentation of the left ventricle (LV) using echocardiogram (Echo) images is essential for cardiovascular analysis. Conventional techniques are labor-intensive and exhibit inter-observer variability. Deep learning has emerged as a powerful tool for automated medical image segmentation, offering advantages in speed and potentially superior accuracy. This study explores the efficacy of employing a YOLO (You Only Look Once) segmentation model for automated LV segmentation in Echo images. YOLO, a cutting-edge object detection model, achieves exceptional speed–accuracy balance through its well-designed architecture. It utilizes efficient dilated convolutional layers and bottleneck blocks for feature extraction while incorporating innovations like path aggregation and spatial attention mechanisms. These attributes make YOLO a compelling candidate for adaptation to LV segmentation in Echo images. We posit that by fine-tuning a pre-trained YOLO-based model on a well-annotated Echo image dataset, we can leverage the model’s strengths in real-time processing and precise object localization to achieve robust LV segmentation. The proposed approach entails fine-tuning a pre-trained YOLO model on a rigorously labeled Echo image dataset. Model performance has been evaluated using established metrics such as mean Average Precision (mAP) at an Intersection over Union (IoU) threshold of 50% (mAP50) with 98.31% and across a range of IoU thresholds from 50% to 95% (mAP50:95) with 75.27%. Successful implementation of YOLO for LV segmentation has the potential to significantly expedite and standardize Echo image analysis. This advancement could translate to improved clinical decision-making and enhanced patient care.

13 December 2023 in European Journal of Emergency Medicine
Verified
Journal Article
Closed access
YCYu-Chun Chien
Yu-Chun Chien
WCWen-Chu Chiang
Wen-Chu Chiang
CCChi-Hsin Chen
Chi-Hsin Chen
JSJen-Tang Sun
Jen-Tang Sun
SJSabariah Faizah Jamaluddin
Sabariah Faizah Jamaluddin
TNThe PATOS Clinical Research Network
The PATOS Clinical Research Network
Background and importance: This study compared the on-scene Glasgow Coma Scale (GCS) and the GCS-motor (GCS-M) for predictive accuracy of mortality and severe disability using a large, multicenter population of trauma patients in Asian countries. Objective: To compare the ability of the prehospital GCS and GCS-M to predict 30-day mortality and severe disability in trauma patients. Design: We used the Pan-Asia Trauma Outcomes Study registry to enroll all trauma patients >18 years of age who presented to hospitals via emergency medical services from 1 January 2016 to November 30, 2018. Settings and participants: A total of 16,218 patients were included in the analysis of 30-day mortality and 11 653 patients in the analysis of functional outcomes. Outcome measures and analysis: The primary outcome was 30-day mortality after injury, and the secondary outcome was severe disability at discharge defined as a Modified Rankin Scale (MRS) score ≥4. Areas under the receiver operating characteristic curve (AUROCs) were compared between GCS and GCS-M for these outcomes. Patients with and without traumatic brain injury (TBI) were analyzed separately. The predictive discrimination ability of logistic regression models for outcomes (30-day mortality and MRS) between GCS and GCS-M is illustrated using AUROCs. Main results: The primary outcome for 30-day mortality was 1.04% and the AUROCs and 95% confidence intervals for prediction were GCS: 0.917 (0.887–0.946) vs. GCS-M:0.907 (0.875–0.938), P = 0.155. The secondary outcome for poor functional outcome (MRS ≥ 4) was 12.4% and the AUROCs and 95% confidence intervals for prediction were GCS: 0.617 (0.597–0.637) vs. GCS-M: 0.613 (0.593–0.633), P = 0.616. The subgroup analyses of patients with and without TBI demonstrated consistent discrimination ability between the GCS and GCS-M. The AUROC values of the GCS vs. GCS-M models for 30-day mortality and poor functional outcome were 0.92 (0.821–1.0) vs. 0.92 (0.824–1.0) (P = 0.64) and 0.75 (0.72–0.78) vs. 0.74 (0.717–0.758) (P = 0.21), respectively. Conclusion: In the prehospital setting, on-scene GCS-M was comparable to GCS in predicting 30-day mortality and poor functional outcomes among patients with trauma, whether or not there was a TBI.

26 April 2023 in International Journal of Cardiology
Verified
Journal Article
Closed access
YHYi-Ju Ho
Yi-Ju Ho
CCChi-Hsin Chen
Chi-Hsin Chen
CSChih-Wei Sung
Chih-Wei Sung
CFCheng-Yi Fan
Cheng-Yi Fan
SLShao-Yung Lin
Shao-Yung Lin
EHEdward Pei-Chuan Huang
Edward Pei-Chuan Huang

21 May 2024 in Preprints.org
Verified
Preprint
Open access
PDF file
MBMadankumar Balasubramani
Madankumar Balasubramani
CSChih-Wei Sung
Chih-Wei Sung
MHMu-Yang Hsieh
Mu-Yang Hsieh
EHEdward Pei-Chuan Huang
Edward Pei-Chuan Huang
Jiann-Shing ShiehJiann-Shing Shieh
Jiann-Shing Shieh
Maysam AbbodMaysam Abbod
Maysam Abbod
Accurate segmentation of the Left Ventricle (LV) in Echocardiogram (Echo) images is essential for cardiovascular analysis. Conventional techniques are labor-intensive and exhibit inter-observer variability. Deep learning has emerged as a powerful tool for automated medical image segmentation, offering advantages in speed and potentially superior accuracy. This study explores the efficacy of employing the YOLO (You Only Look Once) segmentation model for automated LV segmentation in Echo images. YOLO, a cutting-edge object detection model, achieves exceptional speed-accuracy balance through its well-designed architecture. It utilizes efficient dilated convolutional layers and bottleneck blocks for feature extraction, while incorporating innovations like path aggregation and spatial attention mechanisms. These attributes make YOLO a compelling candidate for adaptation to LV segmentation in Echo images. We posit that by fine-tuning a pre-trained YOLO based model on a well-annotated Echo image dataset, we can leverage the model's strengths in real-time processing and precise object localization to achieve robust LV segmentation. The proposed approach entails fine-tuning a pre-trained YOLO model on a rigorously labeled Echo image dataset. Model performance has been evaluated using established metrics such as mean Average Precision (mAP) at an Intersection over Union (IoU) threshold of 50% (mAP50) with 98.31%and across a range of IoU thresholds from 50% to 95% (mAP50:95) with 75.27%. Successful implementation of YOLO for LV segmentation has the potential to significantly expedite and standardize Echo image analysis. This advancement could translate to improved clinical decision-making and enhanced patient care.

16 November 2024 in Resuscitation Plus
Verified
Journal Article
Open access
CFCheng-Yi Fan
Cheng-Yi Fan
SHSih-Shiang Huang
Sih-Shiang Huang
CCChi-Hsin Chen
Chi-Hsin Chen
CSChih-Wei Sung
Chih-Wei Sung
CCChin-Hao Chang
Chin-Hao Chang
EHEdward Pei-Chuan Huang
Edward Pei-Chuan Huang
The PIVOT trial evaluates the clinical outcomes and ventilatory quality of an automatic pneumatic ventilation method compared to a bag-valve-mask ventilation method in patients who have experienced out-of-hospital cardiac arrest and have had an advanced airway placed. The PIVOT trial is a pragmatic, open-label, multicenter randomized controlled trial. It aims to recruit 514 patients in Hsinchu County, Taiwan. Adult, non-trauma patients who experience out-of-hospital cardiac arrest, are treated by emergency medical services, and have an advanced airway in place will be randomized. Biweekly cluster randomization will assign EMS teams to either the automatic pneumatic ventilation group or the bag-valve-mask group. Informed consent is waived. The primary outcome is the return of spontaneous circulation, either prehospital or in-hospital. Secondary outcomes include survival to discharge, neurological outcomes, prehospital ventilatory quality, and the content of prehospital resuscitation. Participants will be followed until they pass away or are discharged from the hospital. The PIVOT trial will provide new insight on the clinical effectiveness of automatic pneumatic ventilation in patients experienced out-of-hospital cardiac arrest. Trial number: NCT06067204 in clinicaltrial.gov