Oestreich Marc-Alexander, Doswald Isabelle, Salem Yasmin, Künstle Noëmi, Wyler Florian, Frauchiger Bettina S, Kentgens Anne-Christianne, Latzin Philipp, Yammine Sophie
Division of Paediatric Respiratory Medicine and Allergology, Department of Paediatrics, Inselspital, Bern University Hospital, University of Bern, Bern, Switzerland.
University Children's Hospital Basel (UKBB), University of Basel, Basel, Switzerland.
Front Pediatr. 2024 Jun 6;12:1393291. doi: 10.3389/fped.2024.1393291. eCollection 2024.
Multiple-breath washout (MBW) is a sensitive method for assessing lung volumes and ventilation inhomogeneity in infants, but remains prone to artefacts (e.g., sighs). There is a lack of tools for systematic retrospective analysis of existing datasets, and unlike N-MBW in older children, there are few specific quality control (QC) criteria for artefacts in infant SF-MBW.
We aimed to develop a computer-based tool for systematic evaluation of visual QC criteria of SF-MBW measurements and to investigate interrater agreement and effects on MBW outcomes among three independent examiners.
We developed a software package for visualization of raw Spiroware (Eco Medics AG, Switzerland) and signal processed WBreath (ndd Medizintechnik AG, Switzerland) SF-MBW signal traces. Interrater agreement among three independent examiners (two experienced, one novice) who systematically reviewed 400 MBW trials for visual artefacts and the decision to accept/reject the washin and washout were assessed.
Our tool visualizes MBW signals and provides the user with (i) display options (e.g., zoom), (ii) options for a systematic QC assessment [e.g., decision to accept or reject, identification of artefacts (leak, sigh, irregular breathing pattern, breath hold), and comments], and (iii) additional information (e.g., automatic identification of sighs). Reviewer agreement was good using pre-defined QC criteria (κ 0.637-0.725). Differences in the decision to accept/reject had no substantial effect on MBW outcomes.
Our visual quality control tool supports a systematic retrospective analysis of existing data sets. Based on predefined QC criteria, even inexperienced users can achieve comparable MBW results.
多次呼吸洗脱（MBW）是评估婴儿肺容积和通气不均匀性的一种敏感方法，但仍容易出现伪影（如叹息）。目前缺乏对现有数据集进行系统回顾性分析的工具，而且与大龄儿童的氮冲洗法多次呼吸洗脱（N-MBW）不同，婴儿单次呼吸多次呼吸洗脱（SF-MBW）中针对伪影的特定质量控制（QC）标准很少。
我们旨在开发一种基于计算机的工具，用于系统评估SF-MBW测量的视觉QC标准，并调查三位独立检查者之间的检查者间一致性以及对MBW结果的影响。
我们开发了一个软件包，用于可视化原始的Spiroware（瑞士Eco Medics AG公司）和经过信号处理的WBreath（瑞士ndd Medizintechnik AG公司）SF-MBW信号轨迹。评估了三位独立检查者（两位经验丰富，一位新手）之间的检查者间一致性，他们系统地审查了400次MBW试验，以检查视觉伪影以及接受/拒绝冲洗和洗脱的决定。
我们的工具可以可视化MBW信号，并为用户提供（i）显示选项（如缩放），（ii）系统QC评估选项[如接受或拒绝的决定、伪影识别（泄漏、叹息、不规则呼吸模式、屏气）和注释]，以及（iii）附加信息（如叹息的自动识别）。使用预定义的QC标准时，检查者之间的一致性良好（κ值为0.637 - 0.725）。接受/拒绝决定的差异对MBW结果没有实质性影响。
我们的视觉质量控制工具支持对现有数据集进行系统的回顾性分析。基于预定义的QC标准，即使是没有经验的用户也能获得可比的MBW结果。
