Erqian Wang

Erqian Wang

Auto-generated profile, based on publicly available information

Erqian Wang

Department of Ophthalmology, Peking Union Medical College Hospital, Beijing, China

Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

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

Representative Publications

View All Publications
23 November 2019 in Journal of Ophthalmology
Verified
Journal Article
Open access
PDF file
JYJingyuan Yang
Jingyuan Yang
MYMingzhen Yuan
Mingzhen Yuan
EWErqian Wang
Erqian Wang
YCYouxin Chen
Youxin Chen
The aim of this study was to compare the repeatability of optical coherence tomography angiography (OCT-A) measurements of macular vessel density using four OCT-A systems, including Heidelberg Spectralis HRA, Optovue RTVue XR, Zeiss Cirrus HD-OCT 5000, and Topcon DRI OCT Triton. A cross-sectional design was used for this study. The vascular density and vascular length density of the superficial and deep retinal capillary plexuses were imaged with OCT-A using 3 mm and 6 mm scan patterns and were calculated using ImageJ. Comparisons of intraclass correlation coefficients (ICC) were conducted. We found that the OCT-A systems had various levels of repeatability. Zeiss had better repeatability for vessel density than the other systems (overall ICC = 0.936). Optovue had better repeatability for vessel length density when the 6 mm scan pattern was used (ICC = 0.680 and 0.700 for retinal superficial and deep capillary plexus, respectively). We concluded that repeatability varied when different scan patterns of various OCT-A systems were used for imaging the superficial retinal and deep capillary plexuses. Results should be seen as valid only for a given method. The repeatability of various OCT-A systems should be considered in clinical practice and in clinical trials that use OCT-A metrics as outcome measures.

1 June 2025 in HepatoBiliary Surgery and Nutrition
Verified
Journal Article
Open access
PDF file
EWErqian Wang
Erqian Wang
ZYZuyi Yang
Zuyi Yang
DTDianzhe Tian
Dianzhe Tian
LZLei Zhang
Lei Zhang
Navigating the artificial intelligence (AI) era: a paradigm shift in clinical surgical education

13 November 2026 in Journal of Medical Internet Research
Verified
Journal Article
Open access
ZYZuyi Yang
Zuyi Yang
DTDianzhe Tian
Dianzhe Tian
MHMeiqian He
Meiqian He
EWErqian Wang
Erqian Wang
LZLvyuxing Zhao
Lvyuxing Zhao
YCYouxin Chen
Youxin Chen
Background: DeepSeek, an open-source large language model (LLM) extensively adopted in hospitals across China, presents a locally deployable alternative to proprietary models such as GPT. However, its real-world clinical performance and safeguards remain underexplored. Objective: The aim of this study is to evaluate the performance of the general-purpose and reasoning versions of DeepSeek (DeepSeek AI) and GPT models in diagnosis, medical record analysis, and safeguards against adversarial vulnerabilities. Methods: This cross-sectional study included 50 real-world medical records from 10 hospitals spanning 14 specialties. A total of 7 LLMs were evaluated, including DeepSeek-V3, DeepSeek-R1, GPT-4o (via ChatGPT; OpenAI), o3-mini (via ChatGPT), o3-mini-high (via ChatGPT), GPT-4o (via Copilot; Microsoft), and GPT-4o-reasoning (via Copilot). Models were tasked with generating diagnoses based on admission notes, answering 12 questions related to extraction, summarization, and inference using discharge summaries, and undergoing 4 rounds of simulated adversarial prompt attacks designed to elicit misinformation. The diagnostic evaluation was performed on 47 records after the exclusion of 3 flawed admission notes, while the other tasks used the full set of 50 records. Diagnostic performance was evaluated based on accuracy, with the causes of misdiagnosis further categorized. Medical record analysis was assessed by scores in extraction, summarization, and inference tasks. Safeguard evaluation included identifying the earliest round of successful prompt attack and verifying reference authenticity and accuracy. Results: Among the 7 models included in this study, DeepSeek-R1 and GPT-4o-reasoning (via Copilot) achieved the highest diagnostic accuracy at 64% (30/47). In medical record analysis, both models also excelled across the 3 subdimensions. GPT-4o-reasoning (via Copilot) achieved accuracies of 97.7% (293/300) in extraction, 87.5% (175/200) in summarization, and 87% in inference, while DeepSeek-R1 reached 97.7% (293/300), 83.5% (167/200), and 85% (85/100), respectively. In simulated adversarial safeguard evaluation involving 4 rounds of prompt attacks, all evaluated models showed high vulnerability. DeepSeek-R1 generated medical disinformation in 98% (49/50) of cases, with only a 2% refusal rate. GPT-4o-reasoning (via Copilot) showed the highest refusal rate at 30%, but it still generated disinformation in 70% (35/50) of cases. True hallucination rates ranged from 38.6% (27/70; GPT-4o-reasoning via Copilot) to 73% (73/100; DeepSeek-V3). Conclusions: These findings underscore the clinical application potential of DeepSeek models, while also emphasizing the critical need to strengthen safeguards against adversarial vulnerabilities prior to their clinical deployment.