Kuan Chen

Mr. Kuan Chen


Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

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

Representative Publications

View All Publications
27 March 2021 in Tomography
Verified
Journal Article
Open access
PDF file
TLTzu-Wei Lee
Tzu-Wei Lee
CCCheng-Yu Chen
Cheng-Yu Chen
Kuan ChenKuan Chen
Kuan Chen
CTChao-Wei Tso
Chao-Wei Tso
HLHui-Hsien Lin
Hui-Hsien Lin
HLHua-Shan Liu
Hua-Shan Liu
The presence of a swallow-tail sign in the nigrosome-1 with hyperintense signal shown on the susceptibility-weighted imaging (SWI) has been shown to be sensitive in detecting the abnormal iron deposits in this area. A systematic evaluation in healthy subjects is required before this tool can be recommended in a widespread application. We evaluated a simple and practical SWI approach using a multiecho gradient-echo sequence with an improved contrast-to-noise ratio (CNR). We also evaluated the association of the neuromelanin imaging contrast behavior with the susceptibility and relaxation measurements. Twenty-five older and 23 young healthy adults were evaluated. The CNRs of the nigrosome-1 were compared along with method and group. Correlations of the nigrosome-1 neuromelanin signal in the neuromelanin-sensitive imaging with CNRs in the susceptibility, T1 and T2 mappings were examined. Two different coils were used to confirm the reproducibility. Compared with the single-echo, multiecho SWI can improve the CNR of the swallow-tail sign. We found significant correlations between neuromelanin signal and CNRs in the susceptibility and T2 mappings, and T1 value. The older subjects exhibited increased CNRs compared with the young adults. No significant difference was observed in the measurements between 20 and 64 channels. The multiecho technique allows the high-quality nigrosome-1 images in SWI and allows for a joint analysis of T2* and quantitative-susceptibility mapping at high spatial resolution. The correlations of neuromelanin-sensitive imaging with susceptibility and T2 imply that the iron content in the nigrosome-1 may have significant influences on the hyperintensity of neuromelanin in the magnetization transfer-related contrast.

3 July 2025 in Healthcare
Verified
Journal Article
Open access
PDF file
Wen-rui HaoWen-rui Hao
Wen-rui Hao
Chun Chao ChenChun Chao Chen
Chun Chao Chen
Kuan ChenKuan Chen
Kuan Chen
LLLong-Chen Li
Long-Chen Li
CCChun-Chih Chiu
Chun-Chih Chiu
YLYu-Chuan (Jack) Li
Yu-Chuan (Jack) Li
Background: Large language models (LLMs) like ChatGPT are increasingly being explored for medical applications. However, their reliability in providing medication advice for patients with complex clinical situations, particularly those with multiple comorbidities, remains uncertain and under-investigated. This study aimed to systematically evaluate the performance, consistency, and safety of ChatGPT in generating medication recommendations for complex cardiovascular disease (CVD) scenarios. Methods: In this simulation-based study (21 January–1 February 2024), ChatGPT 3.5 and 4.0 were prompted 10 times for each of 25 scenarios, representing five common CVDs paired with five major comorbidities. A panel of five cardiologists independently classified each unique drug recommendation as “high priority” or “low priority”. Key metrics included physician approval rates, the proportion of high-priority recommendations, response consistency (Jaccard similarity index), and error pattern analysis. Statistical comparisons were made using Z-tests, chi-square tests, and Wilcoxon Signed-Rank tests. Results: The overall physician approval rate for GPT-4 (86.90%) was modestly but significantly higher than that for GPT-3.5 (85.06%; p = 0.0476) based on aggregated data. However, a more rigorous paired-scenario analysis of high-priority recommendations revealed no statistically significant difference between the models (p = 0.407), indicating the advantage is not systematic. A chi-square test confirmed significant differences in error patterns (p < 0.001); notably, GPT-4 more frequently recommended contraindicated drugs in high-risk scenarios. Inter-model consistency was low (mean Jaccard index = 0.42), showing the models often provide different advice. Conclusions: While demonstrating high overall physician approval rates, current LLMs exhibit inconsistent performance and pose significant safety risks when providing medication advice for complex CVD cases. Their reliability does not yet meet the standards for autonomous clinical application. Future work must focus on leveraging real-world data for validation and developing domain-specific, fine-tuned models to enhance safety and accuracy. Until then, vigilant professional oversight is indispensable.