https://scholars.lib.ntu.edu.tw/handle/123456789/112326
標題: | 雷射原位角膜成型術後近視回退之預測模型 Predictive Model for Myopic Regression after Laser in Situ Keratomileusis (LASIK) |
作者: | 陳韻怡 Chen, Yun-i |
關鍵字: | 近視回退;雷射原位角膜成型術;myopic regression;laser in situ keratomileusis (LASIK) | 公開日期: | 2005 | 摘要: | 目的: 在學童近視盛行率日益升高的台灣,成年後接受雷射原位角膜成型術 ( LASIK ) 來矯正近視之患者亦有日益增多之趨勢。然而醫師們於術後追蹤期間發現某些患者有部分近視度數回退的情形,回退度數深者甚至可能大幅影響患者裸眼視物的清晰程度而需要再次手術。因此本研究為評估何種術前因子可能影響方法: 本研究回溯蒐集311位接受LASIK手術之近視患者(共615 隻眼睛)的資料,包括患者之年齡及性別,術前屈光狀態,角膜厚度及弧度,眼壓值,基礎淚水分泌量( Schirmer test),雷射削切深度及光學區大小等相關因子,並紀錄術後患者之屈光狀態變化情形以判別其有無近視回退的發生。本研究提出三種預測模型: 邏輯迴歸模型、Cox氏比例危害回歸迴歸模型、以及interval censored 三種模型來預估近視患者在LASIK術結果: 根據邏輯迴歸模型,可能影響LASIK 術後之度數回退且具統計學上顯著意義之預測因子包括: 術前屈光度數,術前角膜弧度,光學區之大小及術後度數欠矯。年紀則具邊緣性顯著意義。Cox氏比例危害迴歸模型及interval censored 模型亦顯示上述的預測因子 (除了年紀),會影響LASIK 術後較早發生度數回退。我們利用接受操作特徵曲線分析( ROC curve analysis )作模型的準確度預測,三種模式之準確度皆相當接近。 結論: 此研究以回溯性世代 (retrospective cohort) 及交互驗證(cross-validated) 的研究設計來評估可能導致LASIK術後之近視度數回退之預測因子。利用三種預測模型: 邏輯迴歸模式、Cox氏比例危害迴歸模型、以及interval censored 模式來預估近視患者在LASIK 術後追蹤期間發生近視度數回退的可能性。 Objectives: To evaluate the effects of preoperative characteristics on refractive outcomes (myopic regression) after laser in situ keratomileusis (LASIK) for myopia and myopic astigmatism and to develop a predictive model to predict myopic regression after LASIK. Setting: Universal Eye Center, an ophthalmology clinic franchise, Taipei, Taiwan Methods: The medical records of 615 eyes of 311 patients undergoing LASIK from June 2003 to Dec 2003 were retrospectively reviewed. The surgeries were performed by one experienced surgeon using the Nidek's Excimer Laser EC-5000. Preoperative variables such as refraction errors, corneal thickness, keratometry reading, intraocular pressure (IOP), basic tear secretion (Schirmer test), ablation depth of transition zone and optic zone, diameter of optic zone and transition zone were collected from chart reviews. Age, sex were also recorded. Postoperative examinations of refraction errors were done 1 day, 1 week, 1, 3, 6, 12 months or even more if it is feasible. Three types of predictive models including logistic regression mode, Cox proportional hazards regression model, and interval-censored model were adopted to develop predictive model for estimating the probability of myopic regression after LASIK given specific time of follow-up. Results: According to the logistic analysis, greater manifest refraction errors, lower preoperative keratometry reading, smaller size of optic zone, greater undercorrection and older age (borderline significance) lead to higher probability of myopic regression after LASIK. After taking into account time to myopic regression, the results of Cox proportional hazards regression model revealed that the variables identified in the logistic regression model except age, led to higher probability of earlier myopic regression. Similar findings were noted while using interval-censored model. According to ROC curve analysis, the predictive validity of the three models were rather identical. Conclusions: A retrospective cohort with a cross-validated design was employed to identify significant preoperative variables associated with myopic regression covering all eyes with mild and severe regression after LASIK. Three predictive models, logistic regression, Cox proportional hazards regression model, and interval-censored model, were successfully developed on the basis of these associations to predict the probability of myopic regression by follow-up time. |
URI: | http://ntur.lib.ntu.edu.tw//handle/246246/59212 | 其他識別: | en-US |
顯示於: | 流行病學與預防醫學研究所 |
檔案 | 描述 | 大小 | 格式 | |
---|---|---|---|---|
ntu-94-P92846002-1.pdf | 23.31 kB | Adobe PDF | 檢視/開啟 |
在 IR 系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。