Steele Catriona M, Peladeau-Pigeon Melanie, Nagy Ahmed, Waito Ashley A
Swallowing Rehabilitation Research Laboratory, Toronto Rehabilitation Institute, University Health Network, Ontario, Canada.
Department of Speech-Language Pathology, Rehabilitation Sciences Institute, Faculty of Medicine, University of Toronto, Ontario, Canada.
J Speech Lang Hear Res. 2020 May 22;63(5):1404-1415. doi: 10.1044/2020_JSLHR-19-00314. Epub 2020 May 7.
Purpose The field lacks consensus about preferred metrics for capturing pharyngeal residue on videofluoroscopy. We explored four different methods, namely, the visuoperceptual Eisenhuber scale and three pixel-based methods: (a) residue area divided by vallecular or pyriform sinus spatial housing ("%-Full"), (b) the Normalized Residue Ratio Scale, and (c) residue area divided by a cervical spine scalar (%(C2-4)). Method This study involved retrospective analysis of an existing data set of videofluoroscopies performed in 305 adults referred on the basis of suspected dysphagia, who swallowed 15 boluses each (six thin and three each of mildly, moderately, and extremely thick 20% w/v barium). The rest frame at the end of the initial swallow of each bolus was identified. Duplicate measures of pharyngeal residue were made independently by trained raters; interrater reliability was calculated prior to discrepancy resolution. Frequency distributions and descriptive statistics were calculated for all measures. Kendall's τ tests explored associations between Eisenhuber scale scores and pixel-based measures, that is, %-Full and %(C2-4). Cross-tabulations compared Eisenhuber scale scores to 25% increments of the %-Full measure. Spearman rank correlations evaluated relationships between the %-Full and %(C2-4) measures. Results Complete data were available for 3,545 boluses: 37% displayed pharyngeal residue (thin, 36%; mildly thick, 41%; moderately thick, 35%; extremely thick, 34%). Eisenhuber scale scores showed modest positive associations with pixel-based measures but inaccurately estimated residue severity when compared to %-Full measures with errors in 20.6% of vallecular ratings and 14.2% of pyriform sinus ratings. Strong correlations ( < .001) were seen between the %-Full and %(C2-4) measures, but the %-Full measures showed inflation when spatial housing area was small. Conclusions Generally good correspondence was seen across different methods of measuring pharyngeal residue. Pixel-based measurement using an anatomical reference scalar, for example, (C2-4) is recommended for valid, reliable, and precise measurement.
目的 对于在视频荧光透视检查中捕捉咽部残留物的首选指标，该领域尚未达成共识。我们探讨了四种不同的方法，即视觉感知的艾森胡伯量表和三种基于像素的方法：(a) 残留物面积除以梨状窝或梨状窦空间容纳面积（“%-满”），(b) 归一化残留比量表，以及 (c) 残留物面积除以颈椎标量（%(C2 - 4)）。方法 本研究涉及对一个现有视频荧光透视数据集的回顾性分析，该数据集来自305名因疑似吞咽困难而转诊的成年人，他们每人吞咽15个食团（6个稀食团，以及轻度、中度和极度浓稠的20% w/v钡剂食团各3个）。确定每个食团初始吞咽结束时的静止帧。由训练有素的评估者独立进行咽部残留物的重复测量；在解决差异之前计算评估者间信度。计算所有测量的频率分布和描述性统计量。肯德尔τ检验探索艾森胡伯量表评分与基于像素的测量指标（即%-满和%(C2 - 4)）之间的关联。交叉表将艾森胡伯量表评分与%-满测量指标的25%增量进行比较。斯皮尔曼等级相关性评估%-满和%(C2 - 4)测量指标之间的关系。结果 共有3545个食团的完整数据：37%显示有咽部残留物（稀食团，36%；轻度浓稠，41%；中度浓稠，35%；极度浓稠，34%）。艾森胡伯量表评分与基于像素的测量指标显示出适度的正相关，但与%-满测量指标相比，在估计残留物严重程度时不准确，梨状窝评级中有20.6%的误差，梨状窦评级中有14.2%的误差。%-满和%(C2 - 4)测量指标之间存在强相关性（<0.001），但当空间容纳面积较小时，%-满测量指标会出现膨胀。结论 不同的咽部残留物测量方法总体上具有良好的一致性。建议使用基于解剖学参考标量（例如(C2 - 4)）的基于像素的测量方法进行有效、可靠和精确的测量。
