Weiqi Chen

Weiqi Chen

Auto-generated profile, based on publicly available information

Weiqi Chen

Joint Shantou International Eye Center of Shantou University and The Chinese University of Hong Kong, Shantou, China

Research Focus & Metrics

Academic Career Length

13Years (2014-Present)

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

Representative Publications

View All Publications
10 August 2021 in Nature Communications
Verified
Journal Article
Open access
PDF file
LCLing-Ping Cen
Ling-Ping Cen
JJJie Ji
Jie Ji
JLJian-Wei Lin
Jian-Wei Lin
SJSi-Tong Ju
Si-Tong Ju
HLHong-Jie Lin
Hong-Jie Lin
MZMingzhi Zhang
Mingzhi Zhang
Retinal fundus diseases can lead to irreversible visual impairment without timely diagnoses and appropriate treatments. Single disease-based deep learning algorithms had been developed for the detection of diabetic retinopathy, age-related macular degeneration, and glaucoma. Here, we developed a deep learning platform (DLP) capable of detecting multiple common referable fundus diseases and conditions (39 classes) by using 249,620 fundus images marked with 275,543 labels from heterogenous sources. Our DLP achieved a frequency-weighted average F1 score of 0.923, sensitivity of 0.978, specificity of 0.996 and area under the receiver operating characteristic curve (AUC) of 0.9984 for multi-label classification in the primary test dataset and reached the average level of retina specialists. External multihospital test, public data test and tele-reading application also showed high efficiency for multiple retinal diseases and conditions detection. These results indicate that our DLP can be applied for retinal fundus disease triage, especially in remote areas around the world.

28 June 2014 in BMC Ophthalmology
Verified
Journal Article
Open access
PDF file
DLDusheng Lin
Dusheng Lin
WCWeiqi Chen
Weiqi Chen
GZGuihua Zhang
Guihua Zhang
HHHuichun Huang
Huichun Huang
ZZZhaotao Zhou
Zhaotao Zhou
HCHaoyu Chen
Haoyu Chen
Acute Vogt-Koyanagi-Harada (VKH) disease and acute central serous chorioretinopathy (CSCR) are two common disorders with serous retinal detachment caused by dysfunction of choroid. The purpose of this study is to compare the morphological changes of these two diseases with spectral domain optical coherence tomography (SD-OCT).

28 July 2022 in BMJ Open
Verified
Journal Article
Open access
PDF file
GZGuihua Zhang
Guihua Zhang
JLJian-Wei Lin
Jian-Wei Lin
JWJi Wang
Ji Wang
JJJie Ji
Jie Ji
LCLing-Ping Cen
Ling-Ping Cen
MZMingzhi Zhang
Mingzhi Zhang
Objective To develop and validate a real-world screening, guideline-based deep learning (DL) system for referable diabetic retinopathy (DR) detection. Design This is a multicentre platform development study based on retrospective, cross-sectional data sets. Images were labelled by two-level certificated graders as the ground truth. According to the UK DR screening guideline, a DL model based on colour retinal images with five-dimensional classifiers, namely image quality, retinopathy, maculopathy gradability, maculopathy and photocoagulation, was developed. Referable decisions were generated by integrating the output of all classifiers and reported at the image, eye and patient level. The performance of the DL was compared with DR experts. Setting DR screening programmes from three hospitals and the Lifeline Express Diabetic Retinopathy Screening Program in China. Participants 83 465 images of 39 836 eyes from 21 716 patients were annotated, of which 53 211 images were used as the development set and 30 254 images were used as the external validation set, split based on centre and period. Main outcomes Accuracy, F1 score, sensitivity, specificity, area under the receiver operating characteristic curve (AUROC), area under the precision-recall curve (AUPRC), Cohen’s unweighted κ and Gwet’s AC1 were calculated to evaluate the performance of the DL algorithm. Results In the external validation set, the five classifiers achieved an accuracy of 0.915–0.980, F1 score of 0.682–0.966, sensitivity of 0.917–0.978, specificity of 0.907–0.981, AUROC of 0.9639–0.9944 and AUPRC of 0.7504–0.9949. Referable DR at three levels was detected with an accuracy of 0.918–0.967, F1 score of 0.822–0.918, sensitivity of 0.970–0.971, specificity of 0.905–0.967, AUROC of 0.9848–0.9931 and AUPRC of 0.9527–0.9760. With reference to the ground truth, the DL system showed comparable performance (Cohen’s κ: 0.86–0.93; Gwet’s AC1: 0.89–0.94) with three DR experts (Cohen’s κ: 0.89–0.96; Gwet’s AC1: 0.91–0.97) in detecting referable lesions. Conclusions The automatic DL system for detection of referable DR based on the UK guideline could achieve high accuracy in multidimensional classifications. It is suitable for large-scale, real-world DR screening.

4 July 2022 in BMC Ophthalmology
Verified
Journal Article
Open access
PDF file
ZHZijing Huang
Zijing Huang
KQKunliang Qiu
Kunliang Qiu
JYJingsheng Yi
Jingsheng Yi
HLHongjie Lin
Hongjie Lin
DZDezhi Zheng
Dezhi Zheng
WCWeiqi Chen
Weiqi Chen
Background: Capillary non-perfusion is an important characteristic for diabetic retinopathy (DR) indicating microvascular damage and ischemia. Data on the description and treatment outcomes of DR with large area of non-perfusion are lacking to date. We aim to describe the characteristics and treatment outcomes in a series of patients with DR who presented extensively large area of capillary non-perfusion (LACNP). Methods: Fundus fluorescein angiograms from medical charts in patients diagnosed with DR between Jan 2017 and Dec 2019 were retrospectively reviewed. Clinical data in eyes with LACNP including imaging and laboratory findings at the first presentation were analyzed. The LACNP was defined as over 70% area of capillary non-perfusion throughout the whole image retina. The mean follow-up duration was 12.4 ± 16.7 months. Follow-up data including extensive pan-retinal photocoagulation and surgical intervention and treatment outcomes were evaluated. Results: A total of 43 eyes in 24 patients with LACNP were included, accounting for 3.3% of DR populations in the same period. The overall percentage of non-perfusion area was 79.1 ± 8.1%. All patients received proper control of diabetes and hypertension, and extensive pan-retinal laser photocoagulation. During the follow-up periods, 20 eyes (46.5%) developed severe neovascular complications, of which 15 eyes (34.9%) underwent vitrectomy and/or anti-glaucoma surgeries. Conservative therapies including glycemic control and supplemental laser photocoagulation were conducted in 23 eyes (53.5%) without neovascular complications. In the final follow-up, best corrected visual acuity improved or maintained stable in 19 eyes (44.2%) while deteriorated in 24 eyes (55.8%). Conclusions: The presence of LACNP is the hallmark of advanced DR and often indicates a poor visual outcome, although aggressive treatments may slow DR progression and maintain central vision for some time.

18 November 2022 in BMC Ophthalmology
Verified
Journal Article
Open access
PDF file
DZDezhi Zheng
Dezhi Zheng
ZHZijing Huang
Zijing Huang
QZQin Zeng
Qin Zeng
YWYifan Wang
Yifan Wang
SCShirong Chen
Shirong Chen
WCWeiqi Chen
Weiqi Chen
Background: Vitrectomy and peeling of the internal limiting membrane (ILM) was an effective therapeutic approach for myopic foveoschisis with progressive visual loss. This study investigated the anatomical and visual outcomes of fovea-sparing ILM peeling with or without the inverted flap technique for patients with symptomatic myopic foveoschisis (MF).Methods: We retrospectively reviewed the clinical data of patients with MF. Vitrectomy with fovea-sparing ILM peeling and air tamponade was performed in all patients. The primary outcome measures included best-corrected visual acuity (BCVA), mean macular thickness (MMT), and central foveal thickness (CFT). Depending on whether an inverted ILM flap technique was utilized, further subgroup comparisons between the inverted flap group and the non-inverted flap group were conducted.Results: Twenty-six eyes of 22 patients were included. Fifteen eyes were underwent fovea-sparing ILM peeling without inverted ILM flap and 11 of the 26 eyes were treated with fovea-sparing ILM peeling and an inverted ILM flap technique. In the mean follow-up period of 10.74 ± 4.58 months, a significant improvement in BCVA was observed from 0.97 ± 0.45 logMAR to 0.58 ± 0.51 logMAR (P < 0.01), during which the BCVA of 20 eyes (76.92%) improved and remained stable in 5 eyes (19.23%). Moreover, a positive correlation was also found between the preoperative BCVA and the postoperative BCVA (r = 0.50,P = 0.01). At the last visit, the final MMT decreased from 492.69 ± 209.62 μm to 234.73 ± 86.09 μm, and the CFT reduced from 296.08 ± 209.22 μm to 138.31 ± 73.92 μm (allP < 0.01). A subgroup analysis found no significant differences in BCVA, MMT, or CFT between the inverted and non-inverted flap groups (allP > 0.05).Conclusion: Fovea-sparing ILM peeling with or without inverted flap technique resulted in favorable visual and anatomical outcomes for the treatment of MF. An important factor affecting the postoperative visual outcome was the preoperative visual acuity. Our study found no significant difference between the presence and absence of the inverted ILM flap.

20 June 2019 in Eye
Verified
Journal Article
Open access
PDF file
DZDezhi Zheng
Dezhi Zheng
ZHZijing Huang
Zijing Huang
GZGuihua Zhang
Guihua Zhang
DHDingguo Huang
Dingguo Huang
GLGuoqiao Lin
Guoqiao Lin
WCWeiqi Chen
Weiqi Chen
Purpose To investigate the incidence and impact factors of intraoperative loss of light perception (LP) under sub-Tenon’s anesthesia in patients with macular diseases. Methods Eighty-five consecutive patients received standard phacoemulsification combined pars plana vitrectomy (PPV) under sub-Tenon’s anesthesia. At several checkpoints during the surgery (the end of phacoemulsification, the end of vitrectomy, and the end of surgery), participants were interviewed about whether they had LP or not after removing the influence of contralateral eye and the photo-bleaching effect. In patients treated with retinal photocoagulation, visual experience on laser flashes was evaluated. Results Under routine draping, no patients reported loss of LP at all the checkpoints. When the contralateral eye was tightly covered, the rates of LP loss were 84.7%, 97.6%, and 87.1% at the end of phacoemulsification, the end of vitrectomy, and the end of surgery, respectively. When the photo-bleaching effect was also removed, the rates of LP loss were 61.2%, 82.4%, and 56.5% at each checkpoint, respectively, and there were 87.1% (74/85) of patients reporting visual loss in at least one checkpoint. In addition, 76.9% (50/65) of patients could not feel laser flashes during retinal photocoagulation. Conclusion Intraoperative loss of LP under sub-Tenon’s anesthesia was a relatively common and reversible event. The conduction block of optic nerve by anesthetic mainly contributed to the visual loss during surgery. Photo-bleaching effect also has some effect on the LP evaluation. Surgeons need to inform and counsel the patients about the intraoperative loss of LP, to prevent any sudden panic attacks in them.