Automated identification of abnormal respiratory ciliary motion in nasal biopsies

Abstract

Motile cilia lining the nasal and bronchial passages beat synchronously to clear mucus and foreign matter from the respiratory tract. This mucociliary defense mechanism is essential for pulmonary health, because respiratory ciliary motion defects, such as those in patients with primary ciliary dyskinesia (PCD) or congenital heart disease, can cause severe sinopulmonary disease necessitating organ transplant. The visual examination of nasal or bronchial biopsies is critical for the diagnosis of ciliary motion defects, but these analyses are highly subjective and error-prone. Although ciliary beat frequency can be computed, this metric cannot sensitively characterize ciliary motion defects. Furthermore, PCD can present without any ultrastructural defects, limiting the use of other detection methods, such as electron microscopy. Therefore, an unbiased, computational method for analyzing ciliary motion is clinically compelling. We present a computational pipeline using algorithms from computer vision and machine learning to decompose ciliary motion into quantitative elemental components. Using this framework, we constructed digital signatures for ciliary motion recognition and quantified specific properties of the ciliary motion that allowed high-throughput classification of ciliary motion as normal or abnormal. We achieved >90% classification accuracy in two independent data cohorts composed of patients with congenital heart disease, PCD, or heterotaxy, as well as healthy controls. Clinicians without specialized knowledge in machine learning or computer vision can operate this pipeline as a “black box” toolkit to evaluate ciliary motion.

Assessment of ciliary phenotype in primary ciliary dyskinesia by micro-optical coherence tomography

“…It should be noted that, while μOCT provides in situ measurements that are key to understanding the severity and context of the defect, HSVM remains highly useful, especially for visualizing the individual cilium due to its superior resolution at present. Although the current μOCT system described here provides a number of metrics surrounding ciliary motion and stroke that provide a comprehensive view of the PCD defect, more simple imaging systems, such as conventional OCT (33) or even HSVM, can also provide metrics surrounding CBF (14) and waveforms (21), which could be informative even as a single measure and may be simple to implement in vivo. We acknowledge that the current μOCT imaging system allows only for visualization of planar motion.…”

“…Cilia ultrastructural analysis has limited detection sensitivity, since not all PCD mutations cause cilia defects visualized by electron microscopy and the method is subject to artifacts. Functional analysis represents another important approach toward understanding ciliary motility (14)(15)(16)(17), but this is technically challenging, and the relationship between ciliary motion and mucus transport has not been well characterized, since HSVM measurements are not conducted in situ (13,(18)(19)(20)(21). For example, absence of motile cilia, reduced ciliary beat frequency (CBF), or abnormal ciliary stroke can each confer reduced mucus transport, but their relative importance to MCT is not fully understood and cannot be captured with HSVM alone (12).…”

Primary Ciliary Dyskinesia

“…There have been multiple approaches to try and improve accuracy of HSVM, but these modifications do not eliminate all of the limitations (43-45). A recent publication touted the use of “automated” identification of abnormal ciliary motion (46), but this approach has not been independently validated, and should not be used to diagnose PCD. It is also now clear through genetic testing that many forms of PCD have no, or only subtle, abnormalities of ciliary ultrastructure and/or beat frequency and waveform.…”

The use of biophysical approaches to understand ciliary beating

“…In clinical studies, it is common to diagnose cilia function by optical microscopy, having scraped cells from the nose or upper airways. Traditionally a clinician with experience on what constitutes a range of healthy or problematic waveforms would classify the dynamics, but recently there have been reports showing that this approach could be automated [10] [41].…”