通过计算机面部分析准确分类困难插管。

Accurate classification of difficult intubation by computerized facial analysis.

机构信息

Department of Anesthesiology, Perioperative and Pain Medicine, Brigham and Women's Hospital, Boston, Massachusetts, USA.

出版信息

Anesth Analg. 2011 Jan;112(1):84-93. doi: 10.1213/ANE.0b013e31820098d6. Epub 2010 Nov 16.

DOI:10.1213/ANE.0b013e31820098d6

PMID:21081769

Abstract

BACKGROUND

Bedside airway evaluation is conduced before anesthesia, but all current methods perform modestly, with low sensitivity and positive predictive value. We hypothesized that subjective features of patients' anatomies improve anesthesiologists' ability to predict difficult intubation, and derived a computer model to do so, based on analysis of photographs of patients' faces.

METHODS

Eighty male patients were divided into 2 equal cohorts for model derivation and validation. Each cohort consisted of 20 easy and 20 challenging intubations, defined as >1 attempt by an operator with at least 12 months of anesthesia experience, grade 3 or 4 laryngoscopic view, need for a second operator, or nonelective use of an alternative airway device. Photographs of each subject's face were analyzed by software that resolves each face into 61 facial proportions derived from an algorithm that models the face as a single point in a 50-dimensional eigenspace. Each parameter was tested for discriminatory ability by logistic regression, and combinations of 11 variables with P ≤ 0.1, plus Mallampati class and thyromental distance, were tested exhaustively by all possible binomial quadratic logistic regression models. Candidate models were cross-validated by maximizing the product of the area under the receiver operating characteristic curves obtained in the derivation and validation cohorts.

RESULTS

The best model included 3 facial parameters and thyromental distance. It correctly classified 70 of 80 subjects (P < 10(-8)). In contrast, the best combination of Mallampati class and thyromental distance correctly classified 47 of 80 (P = 0.073). Sensitivity, specificity, and area under the curve for the computer model were 90%, 85%, and 0.899, respectively.

CONCLUSIONS

Computerized analysis of facial structure and thyromental distance can classify easy versus difficult intubation with accuracy significantly outperforming popular clinical predictive tests.

摘要

背景

在麻醉前进行床边气道评估，但所有当前的方法表现都欠佳，其灵敏度和阳性预测值均较低。我们假设患者解剖结构的主观特征可以提高麻醉医生预测困难插管的能力，并基于对患者面部照片的分析，提出了一种计算机模型来实现这一目标。

方法

将 80 名男性患者分为模型推导和验证两组，每组各有 20 例容易插管和 20 例困难插管。其中，困难插管的定义为：操作者至少有 12 个月的麻醉经验，喉镜检查分级为 3 级或 4 级，需要第二名操作者，或非选择性地使用替代气道设备。通过软件对每位患者的面部照片进行分析，该软件将每个面部解析为 61 个面部比例，这些比例是由一种算法得出的，该算法将面部模拟为 50 维特征空间中的一个单点。通过逻辑回归对每个参数进行判别能力测试，然后对 P≤0.1 的 11 个变量和 Mallampati 分级及甲状软骨-下颌距离进行穷尽的二项式二次逻辑回归模型组合测试。通过在推导和验证队列中最大化接收者操作特征曲线下面积的乘积来对候选模型进行交叉验证。

结果

最好的模型包括 3 个面部参数和甲状软骨-下颌距离。它正确分类了 80 名患者中的 70 名（P<10(-8)）。相比之下，Mallampati 分级和甲状软骨-下颌距离的最佳组合正确分类了 80 名患者中的 47 名（P=0.073）。计算机模型的灵敏度、特异性和曲线下面积分别为 90%、85%和 0.899。