ChiCTR2600120927 版本V1.0 版本创建时间2026/03/23 11:13:13 中国临床试验注册中心

审核状态:

Project audit state:

通过审核

Successful

注册号:

Registration number:

ChiCTR2600120927 

最近更新日期:

Date of Last Refreshed on:

2026-03-23 11:13:03 

注册时间:

Date of Registration:

2026-03-23 00:00:00 

注册号状态:

预注册

Registration Status:

Prospective registration

注册题目:

整合眼部生物测量与眼底影像的儿童散瞳屈光度预测模型构建及验证

Public title:

Construction and Validation of a Model for Predicting Cycloplegic Refraction in Children by Integrating Ocular Biometry and Fundus Images.

注册题目简写:

English Acronym:

研究课题的正式科学名称:

整合眼部生物测量与眼底影像的儿童散瞳屈光度预测模型构建及验证

Scientific title:

Construction and Validation of a Model for Predicting Cycloplegic Refraction in Children by Integrating Ocular Biometry and Fundus Images

研究课题代号(代码):

Study subject ID:

在二级注册机构或其它机构的注册号:

The registration number of the Partner Registry or other register:

申请注册联系人:

李丹 

研究负责人:

李丹 

Applicant:

Dan Li 

Study leader:

Li Dan 

申请注册联系人电话:

Applicant telephone:

+86 574 87862255

研究负责人电话:

Study leader's
telephone:

+86 574 87862255

申请注册联系人传真 :

Applicant Fax:

研究负责人传真:

Study leader's fax:

申请注册联系人电子邮件:

Applicant E-mail:

1083098400@qq.com

研究负责人电子邮件:

Study leader's E-mail:

1083098400@qq.com

申请单位网址(自愿提供):

Applicant website(voluntary supply):

研究负责人网址(自愿提供):

Study leader's website(voluntary supply):

申请注册联系人通讯地址:

浙江省宁波市鄞州区北明程路599号

研究负责人通讯地址:

浙江省宁波市鄞州区北明程路599号

Applicant address:

599 North Meicheng Road, Yinzhou District, Ningbo City, Zhejiang Province

Study leader's address:

599 North Meicheng Road, Yinzhou District, Ningbo City, Zhejiang Province

申请注册联系人邮政编码:

Applicant postcode:

研究负责人邮政编码:

Study leader's postcode:

申请人所在单位:

宁波市眼科医院

Applicant's institution:

Ningbo Eye Hospital

研究负责人所在单位:

宁波市眼科医院

Affiliation of the Leader:

Ningbo Eye Hospital

是否获伦理委员会批准:

Approved by ethic committee:

Yes

伦理委员会批件文号:

Approved No. of ethic committee:

甬眼医伦审2026研第001号

伦理委员会批件附件:

Approved file of Ethical Committee:

查看附件View

批准本研究的伦理委员会名称:

宁波市眼科医院伦理委员会

Name of the ethic committee:

Ethics Committee of Ningbo Eye Hospital

伦理委员会批准日期:

Date of approved by ethic committee:

2026-03-21 00:00:00

伦理委员会联系人:

郑双双

Contact Name of the ethic committee:

Zheng shuangshuang

伦理委员会联系地址:

北明程路599号

Contact Address of the ethic committee:

599 beimingcheng Road

伦理委员会联系人电话:

Contact phone of the ethic committee:

+86 574 87862233

伦理委员会联系人邮箱:

Contact email of the ethic committee:

752740695@qq.com

研究实施负责(组长)单位:

宁波市眼科医院

Primary sponsor:

Ningbo Eye Hospital

研究实施负责(组长)单位地址:

北明程路599号

Primary sponsor's address:

599 beimingcheng Road

试验主办单位(项目批准或申办者):

Secondary sponsor:

国家:

中国

省(直辖市):

浙江省

市(区县):

Country:

China

Province:

Zhejiang

City:

单位(医院):

宁波市眼科医院

具体地址:

北明程路599号

Institution
hospital:

Ningbo Eye Hospital

Address:

599 beimingcheng Road

经费或物资来源:

整合眼部生物测量与眼底影像的儿童散瞳屈光度预测模型构建及验证

Source(s) of funding:

Construction and Validation of a Model for Predicting Cycloplegic Refraction in Children by Integrat

研究疾病:

正视;远视储备不足;近视  

Target disease:

Emmetropia; Insufficient hyperopia reserve; Myopia

研究疾病代码:

Target disease code:

研究类型:

诊断试验

Study type:

Diagnostic test

研究所处阶段:

其它 

Study phase:

N/A

研究设计:

诊断试验诊断准确性 

Study design:

Diagnostic test for accuracy 

研究目的:

本研究旨在开发并验证一种基于多模态数据融合的深度学习模型,准确预测儿童散瞳后屈光度,并评估模型在眼球发育过程中的预测稳定性。具体目标包括:构建整合非散瞳验光度数、年龄、性别等临床信息的基线预测模型;基于眼轴长度、角膜曲率、前房深度、晶状体厚度等生物测量参数,特别是利用SRK/T公式计算的晶状体前表面曲率半径等特征,构建生物测量增强模型,量化晶状体调节作用对预测精度的贡献;开发基于卷积神经网络的眼底影像特征提取模块,自动识别豹纹状眼底、视盘椭圆度等形态学特征,构建多模态融合模型,识别屈光度与眼底结构不匹配的高危儿童;在具有1年随访数据的儿童队列中验证模型的纵向预测稳定性;通过可解释性分析揭示各类特征的相对重要性,最终建立适用于临床门诊、学校筛查和流行病学研究的儿童散瞳屈光度预测系统。  

Objectives of Study:

This study aims to develop and validate a deep learning model based on multimodal data fusion for the accurate prediction of cycloplegic refractive error in children, and to evaluate the predictive stability of the model during ocular development. The specific objectives are as follows: to construct a baseline prediction model integrating clinical information such as non-cycloplegic refractive values, age, and gender; to establish a biometric-enhanced model using ocular biometric parameters including axial length, corneal curvature, anterior chamber depth, and lens thickness, particularly incorporating features such as the anterior lens surface curvature radius calculated via the SRK/T formula, so as to quantify the contribution of lens accommodation to prediction accuracy; to develop a convolutional neural network-based fundus image feature extraction module to automatically identify morphological characteristics such as tessellated fundus and optic disc ellipticity, and construct a multimodal fusion model to screen high-risk children with mismatched refractive status and fundus structure; to validate the longitudinal predictive stability of the model in a pediatric cohort with 1-year follow-up data; to reveal the relative importance of various features through interpretability analysis, and ultimately establish a prediction system for cycloplegic refractive error in children applicable to clinical outpatient settings, school screening, and epidemiological research.

药物成份或治疗方案详述:

 

Description for medicine or protocol of treatment in detail:

 

纳入标准:

1.从EDC系统中筛选符合以下标准的研究对象:(1) 符合原项目纳入标准:年龄6-12岁的小学生,男女不限;睫状肌麻痹后验光证实为双眼正视、轻度远视及低中度近视(等效球镜度数SER <=-6.00 D);散光度数范围:-1.50 D <= 散光 <= 0 D。 2.从EDC系统中筛选符合以下标准的研究对象:(2) 拥有完整的检查数据:包括非睫状肌麻痹验光(小瞳验光)、睫状肌麻痹验光(散瞳验光)、IOL Master 700生物测量数据以及眼底彩色照相图像。 3.从EDC系统中筛选符合以下标准的研究对象:(3) 图像质量符合要求:眼底照片无明显运动伪影或曝光异常,可用于后续分析。

Inclusion criteria

1. Select study subjects from the EDC system who meet the following criteria: (1) Meet the original project inclusion criteria: primary school students aged 6-12, of any gender; after cycloplegia, refraction confirmed as emmetropia, mild hyperopia, or low to moderate myopia in both eyes (spherical equivalent refraction, SER <= -6.00 D); astigmatism range: -1.50 D <= astigmatism <= 0 D. 2. Select study subjects from the EDC system who meet the following criteria: (2) Have complete examination data: including non-cycloplegic refraction (small pupil refraction), cycloplegic refraction (dilated refraction), IOL Master 700 biometry data, as well as fundus color photographs. 3. Select study subjects from the EDC system who meet the following criteria: (3) Image quality meets requirements: fundus photos without obvious motion artifacts or exposure abnormalities, suitable for subsequent analysis.

排除标准:

(1)符合原项目排除标准: 1.屈光参差患者(双眼屈光度差异>=1.00D); 2.光敏感患者; 3.父母有高度近视家族史患者; 4.有斜视、弱视、圆锥角膜、青光眼、白内障、眼睑下垂等,有活动性眼周或眼部炎症(例如:睑缘炎、感染性结膜炎、角膜炎、巩膜炎、葡萄膜炎、眼内炎),有影响视力的眼底病变; 5.可能影响屈光发育的全身状况者(如唐氏综合征、马方综合征); 6.过敏体质,有全身性疾病或衰竭性疾病及免疫力缺陷,合并有严重的心血管、肝、肾和造血系统等原发性疾病、免疫系统疾病、精神病的患者; 7.筛选前4周接受眼部手术者,入组6个月内计划接受眼部手术者; 8.入组6个月内计划采用角膜塑形镜、硬性接触镜、多焦点渐进镜进行视力矫正者; 9.筛选前3个月内参加过其他药物临床研究者; 10.研究者认为不适合参与临床试验者。 (2)角膜塑形镜佩戴者。 (3)关键数据缺失>20%。

Exclusion criteria:

(1) Meeting original project exclusion criteria: 1. Patients with anisometropia (difference in refraction between both eyes >=1.00D); 2. Patients with photosensitivity; 3. Patients with a family history of high myopia in parents; 4. Patients with strabismus, amblyopia, keratoconus, glaucoma, cataract, ptosis, active periocular or ocular inflammation (e.g., blepharitis, infectious conjunctivitis, keratitis, scleritis, uveitis, endophthalmitis), or fundus lesions affecting vision; 5. Patients with systemic conditions that may affect refractive development (e.g., Down syndrome, Marfan syndrome); 6. Patients with allergic constitution, systemic diseases or wasting diseases, immune deficiencies, combined with severe primary cardiovascular, liver, kidney, hematopoietic system diseases, immune system diseases, or psychiatric disorders; 7. Patients who underwent ocular surgery within 4 weeks before screening, or plan to undergo ocular surgery within 6 months after enrollment; 8. Patients planning to use orthokeratology lenses, rigid contact lenses, or multifocal progressive lenses for vision correction within 6 months after enrollment; 9. Patients who participated in other drug clinical trials within 3 months before screening; 10. Patients deemed unsuitable for clinical trial participation by the investigator. (2) Orthokeratology lens wearers. (3) Missing key data >20%.

研究实施时间:

Study execute time:

From 2026-02-11 00:00:00 To 2028-01-31 00:00:00  

征募观察对象时间:

Recruiting time:

From 2026-03-23 00:00:00 To 2026-04-30 00:00:00

诊断试验:

Diagnostic Tests:

金标准或参考标准(即可准确诊断某疾病的单项方法或多项联合方法,在本研究中用于诊断是否有该病的临床参考标准):

研究将“睫状肌麻痹验光(散瞳验光)”作为诊断儿童真实屈光状态的金标准。

Gold Standard or Reference Standard (The clinical reference standards required to establish the presence or absence of the target condition in the tested population in present study):

The study uses "cycloplegic refraction" as the gold standard for diagnosing the true refractive status of children.

指标试验(即本研究的待评估诊断试验,无论为方法、生物标志物或设备,均请列出名称):

非睫状肌麻痹验光(小瞳验光)和睫状肌麻痹验光(散瞳验光)度数、眼轴、前房深度、晶体厚度等。

Index test:

Non-cycloplegic refraction (small-pupil refraction) and cycloplegic refraction (dilated refraction) measurements, axial length (AL), anterior chamber depth (ACD), lens thickness (LT), etc.

目标人群(可以是某种疾病患者或正常人群,详细描述其疾病特征,注意应纳入符合分布特点的全序列病例,具有良好的代表性)

6-12岁双眼正视、轻度远视及低中度近视的儿童

例数:

Sample size:

1200

Target condition (The target condition is a particular disease or disease stage that the index test will be intended to identify. Please specify the characteristics in detail; the population should has a complete spectrum and good representative):

Children aged 6-12 years with bilateral emmetropia, mild hyperopia, and low-to-moderate myopia.

容易混淆的疾病人群(即与目标疾病不易区分的一种或多种不同疾病,应避免采用正常人群对照的病例-对照设计):

例数:

Sample size:

0

Population with condition difficult to distinguish from the target condition, the normal population in a case-control study design should be avoid:

None

研究实施地点:

Countries of recruitment and research settings:

国家:

中国

省(直辖市):

浙江省 

市(区县):

 

Country:

China

Province:

Zhejiang

City:

单位(医院):

宁波市眼科医院 

单位级别:

三级甲等 

Institution
hospital:

Ningbo Eye Hospital

Level of the institution:

Tertiary A

测量指标:

Outcomes:

指标中文名:

一年眼轴增长量

指标类型:

次要指标

Outcome:

One-year axial length elongation

Type:

Secondary indicator

测量时间点:

随访时(第1年)

测量方法:

数据基础来源于EDC系统。计算公式:一年眼轴增长量(mm)= 随访时眼轴长度(mm)? 基线时眼轴长度(mm);年化增长率(mm/年)= 眼轴增长量 / 随访时间间隔(年)。保留2位小数。

Measure time point of outcome:

Follow-up (Year 1)

Measure method:

Derived variable calculated from EDC data. Axial elongation (mm) = follow-up AL ? baseline AL. Annualised rate (mm/year) = axial elongation / follow-up interval (years). Precision to 0.01 mm.

指标中文名:

视盘椭圆度

指标类型:

次要指标

Outcome:

Optic disc ovality index

Type:

Secondary indicator

测量时间点:

基线时(第0年)

测量方法:

基于EDC系统中的眼底彩色照相图像,采用计算机辅助测量。对视盘进行最佳拟合椭圆,分别测量椭圆长径和短径。计算公式:视盘椭圆度 = 短径 / 长径。无量纲比值,取值范围0–1,保留2位小数。

Measure time point of outcome:

Baseline (Year 0)

Measure method:

Computer-assisted measurement from fundus photographs in EDC. Best-fit ellipse fitted to the optic disc; short and long diameters measured. Ovality index = short diameter / long diameter. Dimensionless ratio (range 0–1), precision to 0.01.

指标中文名:

睫状肌麻痹验光球镜等效度(散瞳后球镜等效度)

指标类型:

主要指标

Outcome:

Cycloplegic spherical equivalent refraction (post-cycloplegic SER)

Type:

Primary indicator

测量时间点:

基线时(第0年);随访时(第1年,仅纵向验证集)

测量方法:

Data extracted from EDC system. Cycloplegia achieved with compound tropicamide eye drops (1 drop every 5 minutes × 3 doses); autorefraction performed 30 minutes after the last dose using Topcon CT-1 a

Measure time point of outcome:

Baseline (Year 0); Follow-up (Year 1, longitudinal validation cohort only)

Measure method:

Data extracted from EDC system. Cycloplegia achieved with compound tropicamide eye drops (1 drop every 5 minutes × 3 doses); autorefraction performed 30 minutes after the last dose using Topcon CT-1 autorefractor. Three measurements per eye, averaged. SER = sphere + 0.5 × cylinder (diopters, D).

指标中文名:

分层预测性能指标

指标类型:

次要指标

Outcome:

Subgroup prediction performance metrics

Type:

Secondary indicator

测量时间点:

模型训练完成后,在验证集和测试集上计算

测量方法:

将样本按三个维度分层,分别计算各亚组的MAE、RMSE和R2:①年龄分层(6–8岁、8–10岁、10–12岁);②基线屈光状态分层(远视≥+1.00D、正视?0.50D至+1.00D、近视≤?0.50D);③眼轴增长速度分层(慢速<0.2 mm/年、中速0.2–0.3 mm/年、快速≥0.3 mm/年,仅纵向验证集)。采用方差分析评估亚组间差异。

Measure time point of outcome:

After model training, evaluated on validation and test sets

Measure method:

MAE, RMSE, and R2 computed for each subgroup defined by: ① age (6–8, 8–10, 10–12 years); ② baseline refractive status (hyperopia ≥ +1.00D, emmetropia ?0.50 to +1.00D, myopia ≤ ?0.50D); ③ axial elongation rate (slow < 0.2, moderate 0.2–0.3, fast ≥ 0.3 mm/year; longitudinal cohort only). ANOVA used to assess between-subgroup differences.

指标中文名:

白到白距离

指标类型:

次要指标

Outcome:

White-to-white distance (WTW)

Type:

Secondary indicator

测量时间点:

基线时(第0年);随访时(第1年,仅纵向验证集)

测量方法:

数据从EDC系统提取。使用IOL Master 700测量水平方向角膜缘之间的距离。每眼测量5次取平均值。单位为mm,精确到0.01 mm。

Measure time point of outcome:

Baseline (Year 0); Follow-up (Year 1, longitudinal validation cohort only)

Measure method:

Data extracted from EDC system. IOL Master 700 measures horizontal limbus-to-limbus distance. Five measurements per eye, averaged. Unit: mm, precision to 0.01 mm.

指标中文名:

睫状肌麻痹验光球镜度数

指标类型:

主要指标

Outcome:

Cycloplegic sphere power

Type:

Primary indicator

测量时间点:

基线时(第0年);随访时(第1年,仅纵向验证集)

测量方法:

数据从EDC系统提取。受试儿童使用复方托吡卡胺滴眼液进行睫状肌麻痹,每5分钟滴注1滴,共3次,末次给药后30分钟使用电脑验光机(拓普康CT-1)测量,每只眼睛测量3次取平均值。球镜等效度计算公式:SER = S + 0.5×C,单位为屈光度(D)。

Measure time point of outcome:

Baseline (Year 0); Follow-up (Year 1, longitudinal validation cohort only)

Measure method:

Data extracted from EDC system. Cycloplegia achieved with compound tropicamide eye drops (1 drop every 5 minutes × 3 doses); autorefraction performed 30 minutes after the last dose using Topcon CT-1 autorefractor. Three measurements per eye, averaged. SER = sphere + 0.5 × cylinder (diopters, D).

指标中文名:

睫状肌麻痹验光柱镜度数

指标类型:

主要指标

Outcome:

Cycloplegic cylinder power

Type:

Primary indicator

测量时间点:

基线时(第0年);随访时(第1年,仅纵向验证集)

测量方法:

数据从EDC系统提取。测量程序与睫状肌麻痹验光球镜度数相同,柱镜度数和散光轴位同时记录,采用负柱镜形式,单位为D,精确到0.25D,每只眼睛测量3次取平均值。

Measure time point of outcome:

Baseline (Year 0); Follow-up (Year 1, longitudinal validation cohort only)

Measure method:

Data extracted from EDC system. Same measurement procedure as cycloplegic sphere power; cylinder power and axis recorded simultaneously. Negative cylinder convention; unit: D, precision to 0.25D. Three measurements per eye, averaged.

指标中文名:

眼轴/角膜曲率半径比值(AL/CR)

指标类型:

次要指标

Outcome:

Axial length to corneal radius ratio (AL/CR)

Type:

Secondary indicator

测量时间点:

基线时(第0年);随访时(第1年,仅纵向验证集)

测量方法:

衍生指标,数据基础来源于EDC系统。计算公式:AL/CR = 眼轴长度(mm)/ 角膜曲率半径(mm),其中角膜曲率半径(mm)= 337.5 / 平均角膜曲率(D)。无量纲,保留3位小数。

Measure time point of outcome:

Baseline (Year 0); Follow-up (Year 1, longitudinal validation cohort only)

Measure method:

Derived variable calculated from EDC data. AL/CR = axial length (mm) / corneal radius (mm), where corneal radius (mm) = 337.5 / mean K (D). Dimensionless, reported to 3 decimal places.

指标中文名:

前房深度

指标类型:

次要指标

Outcome:

Anterior chamber depth (ACD)

Type:

Secondary indicator

测量时间点:

基线时(第0年);随访时(第1年,仅纵向验证集)

测量方法:

数据从EDC系统提取。使用IOL Master 700测量从角膜内皮至晶状体前表面的距离,在非睫状肌麻痹状态下测量。每眼测量5次取平均值,质量控制标准:信噪比≥2.0、标准差<0.02 mm。单位为mm,精确到0.01 mm。

Measure time point of outcome:

Baseline (Year 0); Follow-up (Year 1, longitudinal validation cohort only)

Measure method:

Data extracted from EDC system. IOL Master 700 measures distance from corneal endothelium to anterior lens surface under non-cycloplegic conditions. Five measurements per eye, averaged. QC criteria: SNR ≥ 2.0, SD < 0.02 mm. Unit: mm, precision to 0.01 mm.

指标中文名:

临床可接受预测比例

指标类型:

次要指标

Outcome:

Proportion of clinically acceptable predictions

Type:

Secondary indicator

测量时间点:

模型训练完成后,在验证集和测试集上计算

测量方法:

计算预测误差绝对值分别在≤0.50D、≤0.75D和≤1.00D范围内的样本比例。计算公式:可接受比例 = 符合标准的样本数 / 总样本数 × 100%。以百分比表示,保留1位小数。

Measure time point of outcome:

After model training, evaluated on validation and test sets

Measure method:

Proportion of predictions with absolute error ≤ 0.50D, ≤ 0.75D, and ≤ 1.00D computed separately. Formula: acceptable proportion = n(|error| ≤ threshold) / N × 100%. Expressed as percentage (%), precision to 0.1%.

指标中文名:

预测误差漂移量(仅纵向验证)

指标类型:

主要指标

Outcome:

Prediction error drift (longitudinal validation only)

Type:

Primary indicator

测量时间点:

纵向验证时计算

测量方法:

计算公式:预测误差漂移量 = 随访时预测误差 ? 基线时预测误差,其中预测误差 = 预测值 ? 真实值。单位为屈光度(D),保留2位小数。计算全体纵向验证样本的平均漂移量及漂移量的标准差。

Measure time point of outcome:

Longitudinal validation

Measure method:

Drift = (predicted ? true) at follow-up ? (predicted ? true) at baseline. Unit: D, precision to 0.01D. Mean drift and SD of drift computed across all subjects in the longitudinal validation cohort.

指标中文名:

特征重要性得分

指标类型:

次要指标

Outcome:

Feature importance score

Type:

Secondary indicator

测量时间点:

模型训练完成后计算

测量方法:

使用SHAP(SHapley Additive exPlanations)值计算各输入特征对预测结果的贡献度,量化每个特征对最终预测的平均边际贡献。计算所有样本各特征SHAP绝对值的均值作为特征重要性得分,归一化至0–1范围。对于眼底图像模块,同时采用梯度加权类激活映射(Grad-CAM)生成热力图,可视化模型关注的关键区域。

Measure time point of outcome:

After model training

Measure method:

Feature importance quantified using SHAP (SHapley Additive exPlanations) values based on Shapley value theory from cooperative game theory. Mean absolute SHAP value per feature computed across all samples and normalised to 0–1 range. For the fundus image module, Gradient-weighted Class Activation Mapping (Grad-CAM) additionally applied to generate heatmaps visualising key regions attended by the model.

指标中文名:

Bland-Altman 95%一致性界限

指标类型:

次要指标

Outcome:

Bland-Altman 95% limits of agreement

Type:

Secondary indicator

测量时间点:

模型训练完成后,在验证集和测试集上计算

测量方法:

计算预测值与真实值差异的均值±1.96倍标准差。单位为屈光度(D),保留2位小数。绘制Bland-Altman图,横轴为真实值,纵轴为预测误差(预测值?真实值)。

Measure time point of outcome:

After model training, evaluated on validation and test sets

Measure method:

Computed as mean(predicted ? true) ± 1.96 × SD(predicted ? true). Unit: D, precision to 0.01D. Bland-Altman plot generated with true values on the x-axis and prediction error (predicted ? true) on the y-axis.

指标中文名:

平均绝对误差(MAE)

指标类型:

次要指标

Outcome:

Mean absolute error (MAE)

Type:

Secondary indicator

测量时间点:

模型训练完成后,在验证集和测试集上计算

测量方法:

MAE = (1/n) × Σ|预测值 ? 真实值|,其中n为样本数量。单位为屈光度(D),保留2位小数。分别计算球镜等效度、球镜度数和柱镜度数的MAE。

Measure time point of outcome:

After model training, evaluated on validation and test sets

Measure method:

MAE = (1/n) × Σ|predicted ? true|, where n = sample size. Unit: D, precision to 0.01D. Computed separately for SER, sphere power, and cylinder power.

指标中文名:

皮尔逊相关系数(r)

指标类型:

次要指标

Outcome:

Pearson correlation coefficient (r)

Type:

Secondary indicator

测量时间点:

模型训练完成后,在验证集和测试集上计算

测量方法:

计算预测值与真实值之间的皮尔逊相关系数。无量纲,取值范围?1至1,保留3位小数。分别计算球镜等效度、球镜度数和柱镜度数的相关系数。

Measure time point of outcome:

After model training, evaluated on validation and test sets

Measure method:

Pearson r computed between predicted and true values. Dimensionless (range ?1 to 1), precision to 0.001. Computed separately for SER, sphere power, and cylinder power.

指标中文名:

晶状体厚度

指标类型:

次要指标

Outcome:

Lens thickness (LT)

Type:

Secondary indicator

测量时间点:

基线时(第0年);随访时(第1年,仅纵向验证集)

测量方法:

数据从EDC系统提取。使用IOL Master 700测量晶状体前表面至后表面的距离,在非睫状肌麻痹状态下进行。每眼测量5次取平均值,质量控制标准:信噪比≥2.0、标准差<0.02 mm。单位为mm,精确到0.01 mm。

Measure time point of outcome:

Baseline (Year 0); Follow-up (Year 1, longitudinal validation cohort only)

Measure method:

Data extracted from EDC system. IOL Master 700 measures distance from anterior to posterior lens surface under non-cycloplegic conditions. Five measurements per eye, averaged. QC criteria: SNR ≥ 2.0, SD < 0.02 mm. Unit: mm, precision to 0.01 mm.

指标中文名:

决定系数(R2)

指标类型:

次要指标

Outcome:

Coefficient of determination (R2)

Type:

Secondary indicator

测量时间点:

模型训练完成后,在验证集和测试集上计算

测量方法:

计算公式:R2 = 1 ? [Σ(真实值 ? 预测值)2 / Σ(真实值 ? 真实值均值)2]。无量纲,取值范围0–1,保留3位小数。分别计算球镜等效度、球镜度数和柱镜度数的R2。

Measure time point of outcome:

After model training, evaluated on validation and test sets

Measure method:

R2 = 1 ? [Σ(true ? predicted)2 / Σ(true ? mean_true)2]. Dimensionless (range 0–1), precision to 0.001. Computed separately for SER, sphere power, and cylinder power.

指标中文名:

晶状体前表面曲率半径

指标类型:

次要指标

Outcome:

Anterior lens surface radius of curvature

Type:

Secondary indicator

测量时间点:

基线时(第0年);随访时(第1年,仅纵向验证集)

测量方法:

数据基础来源于EDC系统。借鉴Su等的方法,利用SRK/T人工晶状体度数计算公式的逆向推导计算,通过输入非睫状肌麻痹验光度数、眼轴长度、角膜曲率、前房深度等参数,反向推算晶状体前表面曲率半径。单位为mm,保留2位小数。

Measure time point of outcome:

Baseline (Year 0); Follow-up (Year 1, longitudinal validation cohort only)

Measure method:

Derived variable calculated from EDC data following Su et al. Reverse derivation of the SRK/T IOL power formula using non-cycloplegic refraction, axial length, keratometry, and ACD as inputs to back-calculate anterior lens surface radius of curvature. Unit: mm, precision to 0.01 mm.

指标中文名:

均方根误差(RMSE)

指标类型:

次要指标

Outcome:

Root mean square error (RMSE)

Type:

Secondary indicator

测量时间点:

模型训练完成后,在验证集和测试集上计算

测量方法:

计算公式:RMSE = √[(1/n) × Σ(预测值 ? 真实值)2]。单位为屈光度(D),保留2位小数。分别计算球镜等效度、球镜度数和柱镜度数的RMSE。

Measure time point of outcome:

After model training, evaluated on validation and test sets

Measure method:

RMSE = √[(1/n) × Σ(predicted ? true)2]. Unit: D, precision to 0.01D. Computed separately for SER, sphere power, and cylinder power.

指标中文名:

模型校准度(Calibration)

指标类型:

次要指标

Outcome:

Model calibration

Type:

Secondary indicator

测量时间点:

模型训练完成后,在验证集和测试集上计算

测量方法:

将真实值按每0.5D一组进行区间分组,计算每组的真实值均值与预测值均值。绘制校准曲线,横轴为真实值均值,纵轴为预测值均值。计算校准曲线斜率和截距,理想情况下斜率=1、截距=0。

Measure time point of outcome:

After model training, evaluated on validation and test sets

Measure method:

True values grouped into 0.5D intervals; mean true value and mean predicted value computed per interval. Calibration curve plotted (x-axis = mean true value, y-axis = mean predicted value). Slope and intercept of calibration line computed; ideal calibration: slope = 1, intercept = 0.

指标中文名:

一年随访时间间隔

指标类型:

次要指标

Outcome:

Follow-up interval

Type:

Secondary indicator

测量时间点:

随访时(第1年)

测量方法:

数据从EDC系统提取。根据基线检查日期和随访检查日期计算,以天为单位精确计算后转换为年,保留2位小数。

Measure time point of outcome:

Follow-up (Year 1)

Measure method:

Data extracted from EDC system. Calculated as follow-up examination date minus baseline examination date (in days), then converted to years; reported to 2 decimal places.

指标中文名:

平均角膜曲率

指标类型:

次要指标

Outcome:

Mean keratometry (mean K)

Type:

Secondary indicator

测量时间点:

基线时(第0年);随访时(第1年,仅纵向验证集)

测量方法:

数据从EDC系统提取。使用IOL Master 700测量角膜前表面在3 mm光学区内陡峭轴和平坦轴的曲率值,计算平均角膜曲率。每眼测量5次取平均值。单位为D,精确到0.01D。换算公式:K(D)= 337.5 / R(mm)。

Measure time point of outcome:

Baseline (Year 0); Follow-up (Year 1, longitudinal validation cohort only)

Measure method:

Data extracted from EDC system. IOL Master 700 measures anterior corneal surface curvature in the 3 mm optical zone (steep and flat meridians); mean K = (K_steep + K_flat) / 2. Five measurements per eye, averaged. Unit: D, precision to 0.01D. Conversion: K (D) = 337.5 / R (mm).

指标中文名:

非睫状肌麻痹验光球镜等效度(小瞳球镜等效度)

指标类型:

次要指标

Outcome:

Non-cycloplegic spherical equivalent refraction (small-pupil SER)

Type:

Secondary indicator

测量时间点:

基线时(第0年);随访时(第1年,仅纵向验证集)

测量方法:

数据从EDC系统提取。在标准验光室照明环境下(光照度300–500 lux),使用电脑验光机(拓普康CT-1)测量,未使用睫状肌麻痹药物,每只眼睛测量3次取平均值。SER = S + 0.5×C,单位为D。

Measure time point of outcome:

Baseline (Year 0); Follow-up (Year 1, longitudinal validation cohort only)

Measure method:

Data extracted from EDC system. Autorefraction performed under standard illumination (300–500 lux) using Topcon CT-1 without cycloplegic agents. Three measurements per eye, averaged. SER = sphere + 0.5 × cylinder (D).

指标中文名:

性别

指标类型:

次要指标

Outcome:

Sex

Type:

Secondary indicator

测量时间点:

基线时(第0年)

测量方法:

数据从EDC系统提取。性别为二分类变量,记录为"男"或"女",数据分析时编码为0(女)或1(男)。

Measure time point of outcome:

Baseline (Year 0)

Measure method:

Data extracted from EDC system. Binary variable recorded as male/female; encoded as 0 (female) or 1 (male) for model input.

指标中文名:

眼轴长度

指标类型:

次要指标

Outcome:

Axial length (AL)

Type:

Secondary indicator

测量时间点:

基线时(第0年);随访时(第1年,仅纵向验证集)

测量方法:

数据从EDC系统提取。使用IOL Master 700扫频光学生物测量仪测量从角膜顶点到视网膜色素上皮层的距离。每眼测量5次取平均值,质量控制标准:信噪比≥2.0、标准差<0.02 mm。单位为mm,精确到0.01 mm。

Measure time point of outcome:

Baseline (Year 0); Follow-up (Year 1, longitudinal validation cohort only)

Measure method:

Data extracted from EDC system. Measured using IOL Master 700 swept-source optical biometer (corneal vertex to retinal pigment epithelium). Five measurements per eye, averaged. QC criteria: SNR ≥ 2.0, SD < 0.02 mm. Unit: mm, precision to 0.01 mm.

指标中文名:

豹纹状眼底分级

指标类型:

次要指标

Outcome:

Tessellated fundus grading

Type:

Secondary indicator

测量时间点:

基线时(第0年)

测量方法:

基于EDC系统中的眼底彩色照相图像,由两名经过培训的眼科医生或研究人员独立分级评估,一致性系数要求≥0.8。根据豹纹状眼底累及的视盘周围象限数量分级:0级=无豹纹状眼底;1级=累及1个象限;2级=累及2个象限;3级=累及3个象限;4级=累及4个象限。

Measure time point of outcome:

Baseline (Year 0)

Measure method:

Graded from colour fundus photographs in EDC by two independently trained ophthalmologists/researchers; required agreement coefficient ≥ 0.8. Scale: Grade 0 = absent; Grade 1 = 1 peripapillary quadrant; Grade 2 = 2 quadrants; Grade 3 = 3 quadrants; Grade 4 = 4 quadrants involved.

指标中文名:

平均偏差(Mean Bias)

指标类型:

次要指标

Outcome:

Mean prediction bias

Type:

Secondary indicator

测量时间点:

模型训练完成后,在验证集和测试集上计算

测量方法:

计算公式:Mean Bias = (1/n) × Σ(预测值 ? 真实值)。单位为屈光度(D),保留2位小数。正值表示模型系统性高估,负值表示系统性低估。

Measure time point of outcome:

After model training, evaluated on validation and test sets

Measure method:

Mean Bias = (1/n) × Σ(predicted ? true). Unit: D, precision to 0.01D. Positive value = systematic overestimation; negative value = systematic underestimation.

指标中文名:

年龄

指标类型:

次要指标

Outcome:

Age

Type:

Secondary indicator

测量时间点:

基线时(第0年);随访时(第1年,仅纵向验证集)

测量方法:

数据从EDC系统提取。根据检查日期与出生日期计算,以周岁为单位,保留1位小数。

Measure time point of outcome:

Baseline (Year 0); Follow-up (Year 1, longitudinal validation cohort only)

Measure method:

Data extracted from EDC system. Calculated as examination date minus date of birth; expressed in years to 1 decimal place.

指标中文名:

非睫状肌麻痹验光球镜度数

指标类型:

次要指标

Outcome:

Non-cycloplegic sphere power

Type:

Secondary indicator

测量时间点:

基线时(第0年);随访时(第1年,仅纵向验证集)

测量方法:

数据从EDC系统提取。测量程序与非睫状肌麻痹验光球镜等效度相同,每只眼睛测量3次取平均值。单位为D,精确到0.25D。

Measure time point of outcome:

Baseline (Year 0); Follow-up (Year 1, longitudinal validation cohort only)

Measure method:

Data extracted from EDC system. Same procedure as non-cycloplegic SER. Three measurements per eye, averaged. Unit: D, precision to 0.25D.

指标中文名:

角膜散光值

指标类型:

次要指标

Outcome:

Corneal astigmatism

Type:

Secondary indicator

测量时间点:

基线时(第0年);随访时(第1年,仅纵向验证集)

测量方法:

数据从EDC系统提取。使用IOL Master 700测量角膜在陡峭轴与平坦轴的曲率差异。计算公式:角膜散光值(D)= 陡峭轴曲率(D)? 平坦轴曲率(D)。每眼测量5次取平均值。单位为D,精确到0.01D。

Measure time point of outcome:

Baseline (Year 0); Follow-up (Year 1, longitudinal validation cohort only)

Measure method:

Data extracted from EDC system. IOL Master 700 measures curvature difference between steep and flat meridians. Corneal astigmatism (D) = K_steep ? K_flat. Five measurements per eye, averaged. Unit: D, precision to 0.01D.

指标中文名:

非睫状肌麻痹验光柱镜度数

指标类型:

次要指标

Outcome:

Non-cycloplegic cylinder power

Type:

Secondary indicator

测量时间点:

基线时(第0年);随访时(第1年,仅纵向验证集)

测量方法:

数据从EDC系统提取。测量程序与非睫状肌麻痹验光球镜度数相同,采用负柱镜形式,单位为D,精确到0.25D,每只眼睛测量3次取平均值。

Measure time point of outcome:

Baseline (Year 0); Follow-up (Year 1, longitudinal validation cohort only)

Measure method:

Data extracted from EDC system. Same procedure as non-cycloplegic sphere power. Negative cylinder convention; unit: D, precision to 0.25D. Three measurements per eye, averaged.

指标中文名:

一年屈光度变化量

指标类型:

次要指标

Outcome:

One-year change in cycloplegic spherical equivalent refraction

Type:

Secondary indicator

测量时间点:

随访时(第1年)

测量方法:

数据基础来源于EDC系统。计算公式:一年屈光度变化量(D)= 随访时散瞳球镜等效度(D)? 基线时散瞳球镜等效度(D);年化变化率(D/年)= 屈光度变化量 / 随访时间间隔(年)。保留2位小数。正值表示向远视方向变化,负值表示向近视方向变化。

Measure time point of outcome:

Follow-up (Year 1)

Measure method:

Derived variable calculated from EDC data. ΔSER (D) = cycloplegic SER at follow-up ? cycloplegic SER at baseline. Annualised rate (D/year) = ΔSER / follow-up interval (years). Precision to 0.01D. Positive value = hyperopic shift; negative value = myopic shift.

指标中文名:

体重指数(BMI)

指标类型:

次要指标

Outcome:

Body mass index (BMI)

Type:

Secondary indicator

测量时间点:

基线时(第0年);随访时(第1年,仅纵向验证集)

测量方法:

数据从EDC系统提取。计算公式:BMI = 体重(kg)/ 身高2(m2)。保留1位小数,单位为kg/m2。

Measure time point of outcome:

Baseline (Year 0); Follow-up (Year 1, longitudinal validation cohort only)

Measure method:

Data extracted from EDC system. BMI = weight (kg) / height2 (m2). Unit: kg/m2, precision to 1 decimal place.

指标中文名:

视盘旁萎缩弧(PPA)

指标类型:

次要指标

Outcome:

Peripapillary atrophy (PPA)

Type:

Secondary indicator

测量时间点:

基线时(第0年)

测量方法:

基于EDC系统中的眼底彩色照相图像,由两名独立评估者进行,一致性系数要求≥0.8。PPA定义为视盘边缘周围的视网膜色素上皮和脉络膜萎缩区域。①目测法:记录PPA的存在(有/无)和累及象限数量(0–4个象限);②定量法:测量PPA面积占视盘面积的比例(%)或PPA最大宽度(以视盘直径百分比表示)。

Measure time point of outcome:

Baseline (Year 0)

Measure method:

Assessed from fundus photographs in EDC by two independent graders; required agreement coefficient ≥ 0.8. PPA defined as zone of RPE/choroid atrophy surrounding the disc margin. Qualitative: presence (yes/no) and quadrant involvement (0–4 quadrants). Quantitative: PPA area as percentage of disc area (%), or maximum PPA width as percentage of disc diameter (%).

指标中文名:

眼底彩色照相图像

指标类型:

次要指标

Outcome:

Colour fundus photograph

Type:

Secondary indicator

测量时间点:

基线时(第0年)

测量方法:

数据从EDC系统提取。在睫状肌麻痹后使用全自动眼底照相机(SK-680A)拍摄以黄斑为中心、45°视野范围的眼底图像。图像分辨率≥2000×2000像素,以JPEG格式存储。质量控制标准:视盘边缘和视网膜血管清晰可辨,整体亮度适中,黄斑中心凹光反射清晰,无明显运动伪影或睫毛遮挡。

Measure time point of outcome:

Baseline (Year 0)

Measure method:

Data extracted from EDC system. Macula-centred 45° colour fundus photographs acquired post-cycloplegia using SK-680A auto fundus camera; resolution ≥ 2000×2000 pixels; stored as JPEG. QC criteria: clear disc margins and retinal vessels, adequate brightness, visible foveal reflex, no motion artefact or eyelash obstruction.

指标中文名:

晶状体厚度/前房深度比值(LT/ACD)

指标类型:

次要指标

Outcome:

Lens thickness to anterior chamber depth ratio (LT/ACD)

Type:

Secondary indicator

测量时间点:

基线时(第0年);随访时(第1年,仅纵向验证集)

测量方法:

数据基础来源于EDC系统。计算公式:LT/ACD = 晶状体厚度(mm)/ 前房深度(mm)。无量纲,保留3位小数。

Measure time point of outcome:

Baseline (Year 0); Follow-up (Year 1, longitudinal validation cohort only)

Measure method:

Derived variable calculated from EDC data. LT/ACD = lens thickness (mm) / anterior chamber depth (mm). Dimensionless, reported to 3 decimal places.

采集人体标本:

Collecting sample(s)
from participants:

标本中文名:

组织:

Sample Name:

NA

Tissue:

人体标本去向

其它  

说明

Fate of sample:

0thers  

Note:

征募研究对象情况:

Recruiting status:

正在进行

Recruiting

年龄范围:

Participant age:

最小 Min age 6 years
最大 Max age 12 years

性别:

男女均可

Gender:

Both

随机方法(请说明由何人用什么方法产生随机序列):

Randomization Procedure (please state who generates the random number sequence and by what method):

None

是否公开试验完成后的统计结果:

Calculated Results after the Study Completed public access:

不公开/Private

盲法:

Blinding:

None

是否共享原始数据:

IPD sharing

是Yes

共享原始数据的方式(说明:请填入公开原始数据日期和方式,如采用网络平台,需填该网络平台名称和网址):

1.共享时间:临床研究完成并结果发表后6个月;2.共享数据范围:共享去标识化个体水平原始数据、元数据(数据字典、SAP)、研究方案、伦理批件、知情同意书模板;不共享受试者隐私信息、未发表中间数据;3.共享平台:国家医学研究登记备案信息系统及中国临床试验注册中心;4.共享方式:主要研究者发送邮件申请;4.主要研究者联系方式:邮箱:ldan@cdutcm.edu.cn,电话:18215654072。

The way of sharing IPD”(include metadata and protocol, If use web-based public database, please provide the url):

1. Timing of data sharing: Six months after the completion of the clinical trial and the publication of the results; 2. Scope of data to be shared: De-identified raw data at the individual level, metadata (data dictionary, SAP), the study protocol, ethics approval documents, and informed consent form templates; private information of trial participants and unpublished interim data will not be shared; 3. Sharing platform: National Medical Research Registration and Filing Information System; 4. Method of sharing: The principal investigator must submit a request via email; 4. Principal investigator’s contact details: Email: ldan@cdutcm.edu.cn, Tel: 18215654072.

数据采集和管理(说明:数据采集和管理由两部分组成,一为病例记录表(Case Record Form, CRF),二为电子采集和管理系统(Electronic Data Capture, EDC),如ResMan即为一种基于互联网的EDC:

根据纳入标准,采用“csv”格式从科研数据管理平台提取数据,并以“csv”格式的电子文件由专人管理。

Data collection and Management (A standard data collection and management system include a CRF and an electronic data capture:

In accordance with the inclusion criteria, data is extracted from the research data management platform in CSV format and managed by a designated person in the form of electronic files in CSV format.

数据与安全监察委员会:

Data and Safety Monitoring Committee:

有/Yes

注册人:

Name of Registration:

 2026-03-23 11:13:03