Obstructive sleep apnoea (OSA) results from a combination of anatomical and neuromuscular defects. The mechanical properties of the airway and muscular activity in response to neural drive are poorly characterized.

Objectives:​ To quantify the movement and strain of the upper airway tissues using dynamic MRI techniques and to improve understanding of airway mechanics.

Methods:​ Spatial Modulation of Magnetization (SPAMM) and Magnetic Resonance Elastography (MRE) were used to investigate the mechanics of the tissues surrounding the airway in OSA patients. EMG was performed to quantify any respiratory related geniohyoid activity in quiet breathing and in the genioglossus with 90Hz vibration to ensure the vibration stimulus used during MRE was not activating the muscle as a potential confounder.

SPAMM allows a grid of saturated magnetization to be superimposed on tissues using MRI and tissue motion to be followed with ultra-fast imaging as the grid deforms. SPAMM was used to image the movement of the tongue and lateral walls of the airway associated with quiet breathing in OSA patients compared to controls matched for age, sex and BMI. It was also used to image the movement of the tissues surrounding the airway associated with mandibular advancement across a spectrum of OSA (AHI range 0 – 75) and the effect of increasing airway pressure from 5 to 10cmH2O in subjects with moderate or severe OSA.

MRE is a method of quantifying the shear storage modulus of tissue by imaging the propagation speed of a shear wave produced by vibrating the tissue. Oscillating motion sensitive field gradients are synchronized with the propagating waves in order to image the wave and a quantitative map of tissue mechanical properties is produced by mathematically reconstructing the wave in three dimensions. The shear modulus of the tissue is related to the wavelength of the wave in the tissue. Viscosity can be measured from the attenuation of the amplitude of the wave. MRE was used to measure the shear modulus and viscosity of the tongue and soft palate in OSA patients compared to matched controls and to compare the stiffness of the tissues in patients with OSA with and without 10cmH2O CPAP. Diffusion tensor imaging (DTI) mapped the fibre direction of the tongue so that anisotropic analysis of the MRE data could be performed.

EMG of the geniohyoid was performed in normal subjects to correlate geniohyoid EMG activation with previous imaging studies using SPAMM (Cheng, Butler et al. 2008) which showed the genioglossus sliding over a stationary geniohyoid in normal subjects.

Results:​ EMG of the geniohyoid showed minimal respiratory related activity. EMG with 90Hz vibration showed no change in activity from baseline. The SPAMM study of quiet breathing in OSA patients compared to matched controls showed that all subjects had airway opening in the sagittal plane associated with inspiration. AHI correlated negatively with the amount of movement of the nasopharyngeal lateral walls at inspiration (R = - 0.542, p = 0.006). In the sagittal plane, four movement patterns of the tongue were associated with inspiration:​ ‘en bloc’ movement of the whole posterior tongue;​ movement of only the oropharyngeal tongue;​ rotational movement;​ or minimal movement. The minimal movement pattern was significantly associated with high AHI (p = 0.009).

Shear modulus of the tongue in OSA patients was reduced compared to matched controls (2.68 ± 0.35 kPa compared to 2.98 ± 0.44 kPa, mean ± standard deviation, p < 0.001) and AHI was significantly negatively associated with isotropic tongue shear modulus (R = - 0.496, p = 0.043). Anisotropic analysis revealed that parallel to the muscle fibres, the shear moduli of the tongue and soft palate were significantly less in OSA patients (p = 0.028, 0.041).

During dynamic imaging, the mandible was advanced a mean 5.6 ± 1.8 mm. This produced lateral movement of the lateral walls through a connection from the ramus of the mandible to the pharyngeal lateral walls in all subjects. In the sagittal plane, 3 patterns of posterior tongue deformation were seen with mandibular advancement - A) en bloc anterior movement, B) anterior movement of the oropharyngeal region and C) minimal anterior movement. Subjects with lower AHI were more likely to have en bloc movement (p = 0.04) than minimal movement. Antero-posterior elongation of the tongue increased with AHI (R= 0.461, p = 0.01).

Application of 10cmH20 CPAP increased the shear modulus of the soft palate from mean 1.8 ± 0.3 to 2.3 ± 0.5 kPa (p = 0.02) but there was no significant effect on the stiffness of the tongue. SPAMM sequences with increasing CPAP revealed no movement in 48/111 (43%) sequences, with 27/111 (24%) showing expansion and 36/111 (32%) showing narrowing. In the expansion sequences, the oropharynx moved more on average (2.2 ± 1.3mm) than the nasopharynx (1.3 ±1.5 mm, p = 0.007).

Conclusions:​ In the most severe OSA patients the tongue moves little and is less stiff than matched controls. Increased multiunit EMG activity (Mezzanotte, Tangel et al. 1992) previously found in OSA patients do not seem to translate into movement or increased stiffness of the muscle, which may suggest that neural drive is ineffective. There was no change in tongue stiffness with CPAP application which suggests that decreased drive to the tongue associated with CPAP (Lo, Jordan et al. 2006) did not change the stiffness of the tongue, which is also evidence for ineffective neural drive. In subjects with intermediate severity there is heterogeneous movement which may represent intermediate progression of disease or inter-subject anatomical differences which result in regional regulation of the airway patency. Imaging of mandibular advancement suggested two mechanisms by which mandibular advancement improves airway collapsibility in some OSA patients;​ firstly subjects with lower AHI had movement of the entire tongue forward during mandibular advancement and secondly by increasing lateral airway dimensions via a direct connection from the lateral ramus of the mandible, which is a novel target of future investigation. It was likely the underlying structure in this region was the pterygomandibular raphe.