透過您的圖書館登入
IP:3.137.185.180
  • 期刊
  • OpenAccess

Diagnostic Performance in Differentiating Colorectalpolyps: Narrow Band Image and Conventional White Light Image

比較窄頻影像及傳統白光影像在區別大腸息肉之診斷表現

摘要


大部分的大腸癌多經由良性的腺瘤演變而形成,因為有癌化的風險,腺瘤必須切除。如果我們能正確的分辨腺瘤與非腺瘤病灶,就可以避免非腺瘤的過度治療。窄頻影像系統是新近發展的光學染色內視鏡,可以使得黏膜表淺的微細血管突顯,而沒有傳統染色內視鏡在觀察上需要人為噴灑染劑的不便。為了提高診斷正確率,不論傳統染色內視鏡或窄頻影像系統,最好能搭配放大內視鏡,但是在臨床上,放大內視鏡的使用並不普遍。我們分析了36個息肉病灶在白光及窄頻影像下,其紅、綠、藍三種顏色的數位資料,除了窄頻影像紅光的SD及Kurt外,不論是否使用窄頻影像,兩者病灶在SD、Kurt、Entropy與Energy的特徵分析均呈現統計上的差異,可能是因為非腺瘤病灶呈現較均質黏膜表面。在窄頻影像中綠光的平均值在兩者病灶間亦呈明顯的差異,可能是腺瘤病灶具備較大的凹陷型態吸收較少的綠光而造成。在電腦輔助診斷的表現上,窄頻影像優於傳統白光,這與以往臨床上的研究結果相符。

並列摘要


A majority of colon and rectal cancers arise from previously benign adenomas through the adenoma–carcinoma sequence. Discrimination between neoplastic and nonneoplastic lesions is crucial to avoid the overtreatment of nonneoplastic lesions. Narrow-band imaging (NBI) is a newly developed optical chromoendoscopy technique that highlights the microvascular structure of the mucosal surface without the inconvenience of dye spraying that is used in conventional chromoendoscopy. Both conventional chromoendoscopy and NBI utilize magnifying colonoscopy technology to achieve high diagnostic accuracy; however, magnifying colonoscopies are not routinely used in clinical practice. We analyzed the numerical data of the red, green, and blue components under white light (WL) and NBI in 36 polyps. Significant differences between two lesions were compared on the basis of various characteristics, including standard deviation (SD), Kurt, Entropy, and Energy for each component under WL and NBI; however, SD and Kurt were omitted for the red component under NBI. A more homogeneous mucosal surface on a hyperplastic polyp contributes to the presentation. In addition, the mean of the green component under NBI achieved a significant difference that may have resulted from less absorption over the larger pit area in an adenomatous polyp. The accuracy of the computer-aided diagnostic performance was higher for NBI than for WL; this was compatible with previous results from other clinical studies that did not use computer-aided analysis.

延伸閱讀