Automated surgery planning for an obstructed nose by combining computational fluid dynamics with reinforcement learning

Septoplasty and turbinectomy are among the most common interventions in the field of rhinology. Their constantly debated success rates and the lack of quantitative flow data of the entire nasal airway for planning the surgery necessitate methodological improvement. Thus, physics-based surgery planning is highly desirable. In this work, a novel and accurate method is developed to enhance surgery planning by physical aspects of respiration, i.e., to plan anti-obstructive surgery, for the first time a reinforcement learning algorithm is combined with large-scale computational fluid dynamics simulations. The method is integrated into an automated pipeline based on computed tomography imaging. The proposed surgical intervention is compared to a surgeon's initial plan, or the maximum possible intervention, which allows the quantitative evaluation of the intended surgery. Two criteria are considered: (i) the capability to supply the nasal airway with air expressed by the pressure loss and (ii) the capability to heat incoming air represented by the temperature increase. For a test patient suffering from a deviated septum near the nostrils and a bony spur further downstream, the method recommends surgical interventions exactly at these locations. For equal weights on the two criteria (i) and (ii), the algorithm proposes a slightly weaker correction of the deviated septum at the first location, compared to the surgeon's plan. At the second location, the algorithm proposes to keep the bony spur. For a larger weight on criterion (i), the algorithm tends to widen the nasal passage by removing the bony spur. For a larger weight on criterion (ii), the algorithm's suggestion approaches the pre-surgical state with narrowed channels that favor heat transfer. A second patient is investigated that suffers from enlarged turbinates in the left nasal passage. For equal weights on the two criteria (i) and (ii), the algorithm proposes a nearly complete removal of the inferior turbinate, and a moderate reduction of the middle turbinate. An increased weight on criterion (i) leads to an additional reduction of the middle turbinate, and a larger weight on criterion (ii) yields a solution with only slight reductions of both turbinates, i.e., focusing on a sufficient heat exchange between incoming air and the air-nose interface. The proposed method has the potential to improve the success rates of the aforementioned surgeries and can be extended to further biomedical flows.

Analysis of Upper Airway Flow Dynamics in Robin Sequence Infants Using 4-D Computed Tomography and Computational Fluid Dynamics

Robin Sequence (RS) is a potentially fatal craniofacial condition characterized by undersized jaw, posteriorly displaced tongue, and resultant upper airway obstruction (UAO). Accurate assessment of UAO severity is crucial for management and diagnosis of RS, yet current evaluation modalities have significant limitations and no quantitative measures of airway resistance exist. In this study, we combine 4-dimensional computed tomography and computational fluid dynamics (CFD) to assess, for the first time, UAO severity using fluid dynamic metrics in RS patients. Dramatic intrapopulation differences are found, with the ratio between most and least severe patients in breathing resistance, energy loss, and peak velocity equal to 40:1, 20:1, and 6:1, respectively. Analysis of local airflow dynamics characterized patients as presenting with primary obstructions either at the location of the tongue base, or at the larynx, with tongue base obstructions resulting in a more energetic stenotic jet and greater breathing resistance. Finally, CFD-derived flow metrics are found to correlate with the level of clinical respiratory support. Our results highlight the large intrapopulation variability, both in quantitative metrics of UAO severity (resistance, energy loss, velocity) and in the location and intensity of stenotic jets for RS patients. These results suggest that computed airflow metrics may significantly improve our understanding of UAO and its management in RS.

Analysis of flow field and turbulence predictions in a lung model applying RANS and implications for particle deposition

Airflow and aerosol deposition in the human airways are important aspects for applications such as pulmonary drug delivery and human exposure to aerosol pollutants. Numerical simulations are widely used nowadays to shed light in airflow features and particle deposition patterns inside the airways. For that purpose, the Euler/Lagrange approach is adopted for predicting flow field and particle deposition through point-particle tracking. Steady-state RANS (Reynolds-averaged Navier-Stokes) computations of flow evolution in an extended lung model applying different turbulence models were conducted and compared to measurements as well as high resolution LES (large-eddy simulations) for several flow rates. In addition, various inlet boundary conditions were considered and their influence on the predicted flow field was analysed. The results showed that the mean velocity field was simulated reasonably well, however, turbulence intensity was completely under-predicted by two-equation turbulence models. Only a Reynolds-stress model (RSM) was able predicting a turbulence level comparable to the measurements and the high resolution LES. Remarkable reductions in wall deposition were observed when wall effects were accounted for in the drag and lift force expressions. Naturally, turbulence is an essential contribution to particle deposition and it is well known that two-equation turbulence models considerably over-predict deposition due to the spurious drift effect. A full correction of this error is only possible in connection with a Reynolds-stress turbulence model whereby the predicted deposition in dependence of particle diameter yielded better agreement to the LES predictions. Specifically, with the RSM larger deposition is predicted for smaller particles and lower deposition fraction for larger particles compared to LES. The local deposition fraction along the lung model was numerically predicted with the same trend as found from the measurements, however the values in the middle region of the lung model were found to be somewhat larger.

Lung function assessment in the Pacific walrus (Odobenus rosmarus divergens) while resting on land and submerged in water

In the present study, we examined lung function in healthy resting adult (born in 2003) Pacific walruses (Odobenus rosmarus divergens) by measuring respiratory flow ([Formula: see text]) using a custom-made pneumotachometer. Three female walruses (670-1025 kg) voluntarily participated in spirometry trials while spontaneously breathing on land (sitting and lying down in sternal recumbency) and floating in water. While sitting, two walruses performed active respiratory efforts, and one animal participated in lung compliance measurements. For spontaneous breaths, [Formula: see text] was lower when walruses were lying down (e.g. expiration: 7.1±1.2 l s-1) as compared with in water (9.9±1.4 l s-1), while tidal volume (VT, 11.5±4.6 l), breath duration (4.6±1.4 s) and respiratory frequency (7.6±2.2 breaths min-1) remained the same. The measured VT and specific dynamic lung compliance (0.32±0.07 cmH2O-1) for spontaneous breaths were higher than those estimated for similarly sized terrestrial mammals. VT increased with body mass (allometric mass-exponent=1.29) and ranged from 3% to 43% of the estimated total lung capacity (TLCest) for spontaneous breaths. When normalized for TLCest, the maximal expiratory [Formula: see text] ([Formula: see text]exp) was higher than that estimated in phocids, but lower than that reported in cetaceans and the California sea lion. [Formula: see text]exp was maintained over all lung volumes during spontaneous and active respiratory manoeuvres. We conclude that location (water or land) affects lung function in the walrus and should be considered when studying respiratory physiology in semi-aquatic marine mammals.

A flow-leak correction algorithm for pneumotachographic work-of-breathing measurement during high-flow nasal cannula oxygen therapy

Measuring work of breathing (WOB) is an intricate task during high-flow nasal cannula (HFNC) therapy because the continuous unidirectional flow toward the patient makes pneumotachography technically difficult to use. We implemented a new method for measuring WOB based on a differential pneumotachography (DP) system, equipped with one pneumotachograph inserted in the HFNC circuit and another connected to a monitoring facemask, combined with a leak correction algorithm (LCA) that corrects flow measurement errors arising from leakage around the monitoring facemask. To test this system, we used a mechanical lung model that provided data to compare LCA-corrected respiratory flow, volume and time values with effective values obtained with a third pneumotachograph used instead of the LCA to measure mask flow leaks directly. Effective and corrected volume and time data showed high agreement (Bland-Altman plots) even at the highest leak. Studies on two healthy adult volunteers confirmed that corrected respiratory flow combined with esophageal pressure measurements can accurately determine WOB (relative error < 1%). We conclude that during HFNC therapy, a DP system combined with a facemask and an algorithm that corrects errors due to flow leakages allows pneumotachography to measure reliably the respiratory flow and volume data needed for calculating WOB.

Respiratory flow and vital signs associated with the intensity of functional electrical stimulation delivered to human abdominal muscles during quiet breathing

[Purpose] The purpose of this study was to examine the effects of increasing the intensity of functional electrical stimulation delivered to abdominal muscles during quiet breathing on respiratory flow, vital signs and pain in healthy subjects. [Subjects and Methods] Electrical stimulation was delivered bilaterally using one pair of high-conductivity gel-skin plate electrodes, placed on both sides of the abdomen, to nine healthy males. Subjects were required to breathe normally through a face mask for 2 minutes while in a supine position. The stimulation intensity was incrementally increased by 10 mA until reaching 100 mA. Respiratory parameters, vital signs and pain based on the visual analog scale were measured for each intensity of electrical stimulation. [Results] Transcutaneous oxygen saturation showed a slight upward trend in association with increasing stimulation intensity, but there were no significant changes in pulse or blood pressure. Respiratory flow, tidal volume, and minute ventilation increased significantly as the stimulation intensity rose. [Conclusion] This study revealed that functional electrical stimulation can be safely delivered to human abdominal muscles without causing vital sign abnormalities. It was also found that the appropriate intensity level of electrical stimulation for achieving effects on respiratory flow while also minimizing pain is 60–80 mA.

Correlation between acoustic rhinometry, computed rhinomanometry and cone-beam computed tomography in mouth breathers with transverse maxillary deficiency

INTRODUCTION

To provide clinical information and diagnosis in mouth breathers with transverse maxillary deficiency with posterior crossbite, numerous exams can be performed; however, the correlation among these exams remains unclear.

OBJECTIVE

To evaluate the correlation between acoustic rhinometry, computed rhinomanometry, and cone-beam computed tomography in mouth breathers with transverse maxillary deficiency.

METHODS

A cross-sectional study was conducted in 30 mouth breathers with transverse maxillary deficiency (7-13 y.o.) patients with posterior crossbite. The examinations assessed: (i) acoustic rhinometry: nasal volumes (0-5cm and 2-5cm) and minimum cross-sectional areas 1 and 2 of nasal cavity; (ii) computed rhinomanometry: flow and average inspiratory and expiratory resistance; (iii) cone-beam computed tomography: coronal section on the head of inferior turbinate (Widths 1 and 2), middle turbinate (Widths 3 and 4) and maxilla levels (Width 5). Acoustic rhinometry and computed rhinomanometry were evaluated before and after administration of vasoconstrictor. Results were compared by Spearman's correlation and Mann-Whitney tests (α=0.05).

RESULTS

Positive correlations were observed between: (i) flow evaluated before administration of vasoconstrictor and Width 4 (Rho=0.380) and Width 5 (Rho=0.371); (ii) Width 2 and minimum cross-sectional areas 1 evaluated before administration of vasoconstrictor (Rho=0.380); (iii) flow evaluated before administration of vasoconstrictor and nasal volumes of 0-5cm (Rho=0.421), nasal volumes of 2-5cm (Rho=0.393) and minimum cross-sectional areas 1 (Rho=0.375); (iv) Width 4 and nasal volumes of 0-5cm evaluated before administration of vasoconstrictor (Rho=0.376), nasal volumes of 2-5cm evaluated before administration of vasoconstrictor (Rho=0.376), minimum cross-sectional areas 1 evaluated before administration of vasoconstrictor (Rho=0.410) and minimum cross-sectional areas 1 after administration of vasoconstrictor (Rho=0.426); (v) Width 5 and Width 1 (Rho=0.542), Width 2 (Rho=0.411), and Width 4 (Rho=0.429). Negative correlations were observed between: (i) Width 4 and average inspiratory resistance (Rho=-0.385); (ii) average inspiratory resistance evaluated before administration of vasoconstrictor and nasal volumes of 0-5cm (Rho=-0.382), and average expiratory resistance evaluated before administration of vasoconstrictor and minimum cross-sectional areas 1 (Rho=-0.362).

CONCLUSION

There were correlations between acoustic rhinometry, computed rhinomanometry, and cone-beam computed tomography in mouth breathers with transverse maxillary deficiency.