1.1.

Respiratory Physiology and Clinical Relevance of Pulmonary Volumes

The respiratory system consists of conducting and respiratory airways (Fig. 1). The conducting airways lack alveoli and take no part in gas exchange. Their main function is to conduct, clean, heat up, and humidify the air to 37°C and 100% relative humidity (RH) as it reaches the alveolar sacks.² The respiratory airways make up most of the lung and is the region where gas exchange occurs, and it is composed by millions of alveoli ($\sim 0.3\mathrm{mm}$ in diameter each).³

Fig. 1

Diagram of the respiratory airways. The conducting airways warm up and moisten the inhaled air, the gas exchange occurs in the respiratory zone where the alveoli are located.

The pulmonary gas volume is constantly changing during respiration. The four standard volumes corresponding to different degrees of inhalation or exhalation are classified as: tidal volume (amount of gas that can be inhaled or exhaled during a respiratory cycle: normal breathing); inspiratory reserve volume (the maximum amount of gas that can be inhaled beyond the normal tidal volume: deep breathing); expiratory reserve volume (volume of gas that can be exhaled by force beyond exhalation of the normal tidal volume); and residual volume (volume of gas that remains in the lung after maximal expiration). To assess and diagnose obstructive and restrictive pathologies, the measurement of respiratory volumes is of clinical importance.⁴

1.2.

Gas in Scattering Media Absorption Spectroscopy

Gas in scattering media absorption spectroscopy (GASMAS) is a gas spectroscopy technique that uses tunable diode laser spectroscopy to measure the concentration of gases enclosed by diffuse media in a noninvasive way.⁵ The clinical applications of GASMAS include: assessment of infection in paranasal sinuses,⁶ diagnosis of otitis,⁷ early diagnosis of osteonecrosis,⁸ and monitoring of oxygen (${\mathrm{O}}_2$) and water vapor (${\mathrm{H}}_2\mathrm{O}$) in the lungs of neonates.^9–12 The studies advancing the clinical translation of GASMAS into respiratory healthcare have focused on neonates as the thickness of protective organs surrounding their lungs is within the limits of penetration depth for near-infrared light.^13,14

GASMAS exploits the fact that the absorption signal of gases is unique against the tissue background. The absorption bands of ${\mathrm{O}}_2$ and ${\mathrm{H}}_2\mathrm{O}$ are typically 0.001 nm,¹⁵ which can be distinguished from the absorption imprint of the organs around the lungs (both liquid and solid), which is generally in the 10 nm range.¹⁶ In clinical studies, the thoracic walls are illuminated with a diffuse source, and the scattered light reaching a photodetector is used to identify the absorption signal of molecular ${\mathrm{O}}_2$, by means of the Beer–Lambert law equation, which states that the concentration of the gas $c$ can be calculated from the absorbance (logarithm of the ratio of the source light intensity $I_0$ to detected intensity$I$) when the molar absorption [$\epsilon (\lambda )$] and the gas absorption path length ($l$)¹⁷ are known:

In the case of gas assessment in the lung, the thorax is illuminated with two wavelengths chosen to minimize the absorption by tissue:¹² tunable diode lasers operating at 760 nm are typically selected to match the absorption bands of ${\mathrm{O}}_2$, and 820 or 935 nm to match absorption of ${\mathrm{H}}_2\mathrm{O}$. The last is used as reference gas. Since temperature and RH of the lungs are well known, the concentration of ${\mathrm{H}}_2\mathrm{O}$ is calculated using the ideal gas law, with the partial pressure computed by means of the Arden Buck equation, which states that the partial pressure of vapor saturation in moist air (${\mathit{e}}_{\mathit{w}}^{\prime }$) at ambient temperature $T$ and pressure $P$ is given as¹⁸

The Beer–Lambert law is applied to the ${\mathrm{H}}_2\mathrm{O}$ absorption spectra to estimate $l$ inside the lung. Consequently, the calculated value of $l$ is input into Eq. (1) together with the absorption signal at 760 nm to estimate the ${\mathrm{O}}_2$ concentration. This gas absorption path length approximation is valid for wavelengths that are spectrally close and interacting with identical media.¹⁹

The lung volume assessment is currently done using spirometry, flow volume curves, and/or plethysmography. However, these tests do not provide information of localized volume changes and patients are asked to make inspiratory and expiratory manoeuvres for volume assessment to be performed.²⁰ This represents a limitation in measuring the pulmonary volumes in neonates or patients under anesthesia. GASMAS can potentially solve such limitations given that the light absorbance depends on $l$, which is directly proportional to gas volume.

In this study, we prove the feasibility of measuring changes in gas volume using GASMAS technology. The absorption imprint of molecular ${\mathrm{O}}_2$ was acquired using a pulmonary phantom model that includes air pockets mimicking the size, temperature, and RH of alveoli, surrounded by liquid phantom matching absorption and scattering of lung tissue.

2.1.

Construction of Lung Tissue Phantom

The phantoms previously created to investigate the clinical application of GASMAS in neonatal respiratory healthcare had a hollow cavity with the outer geometry of the lung without the inner structure mimicking the alveolar anatomy.^21,22 The phantom designed in this study recreated inflated alveoli surrounded by lung tissue.

The simplified pulmonary model required a three-dimensional (3D) rectangular holder ($0.025\times 0.014\times 0.035\mathrm{m}$), printed with clear resin (RS-F2-GPCL-04 Formlabs, Form 2) filled with a set of 229 glass capillaries (maximum amount of capillaries that fit in the holder forming a gridded array) of $0.25\times {10}^{-3}\mathrm{m}$ inner radius and $125\times {10}^{-3}\mathrm{m}$ length (SIGMA, batch Z114952). A fraction of the capillaries was filled with air (mimicking the alveoli) and the remaining capillaries were filled with liquid optical phantom of lung tissue (${\mu }_a=0.50{\mathrm{cm}}^{-1}$ and ${\mu }_s^{\prime }=5.4{\mathrm{cm}}^{-1}$, see Sec. 2.1).¹⁶ The number of capillaries filled with liquid phantom was altered and distributed randomly to reduce or increase the amount of air in the probed volume. To perform the measurements, the capillary array was enclosed in a controlled environment chamber (designed to fit the source and detector of a GASMAS bench top system), keeping the temperature and RH (see Sec. 1.1).

2.2.

Preparation of Liquid Phantom with Optical Properties of Lung Tissue

The optical properties of lung tissue at 760 nm were assigned by filling the capillaries with liquid phantom. To produce 100 ml phantom, 8.5 ml of intralipid (SIGMA, Lot # MKCL8461) was mixed with 1.5 ml of stock ink-water solution (1% Winsor & Newton® black India ink diluted in 99 % distilled water) and 90 ml distilled water. The g-factor was defined by the scattering of intralipid $g=0.6$,²³ and the optical properties were determined using the fiber optic probe system based on diffuse reflectance and diffuse spectroscopy from Ref. 24.

The capillary tubes used to mimic lung absorption and scattering were filled using capillary action. In some cases, air bubbles prevented the rising of the liquid in the tube, which was solved by gently tapping on the external side of the tube. The top end of the tubes was then sealed with starch-based modeling clay (Play-Doh) to prevent leakage during measurements.

The liquid phantom and air-filled capillaries were then randomly placed by hand into the holder to make up the capillary arrays for absorption measurements. The probed air volume was changed by varying the ratio of air-filled capillaries in the array (Sec. 3.2).

2.3.

Chamber for Controlled Environment, Mimicking Lung Relative Humidity, and Temperature

Figure 2 shows the diagram of all the components used in the lung model. An enclosure with temperature and RH control was built using 12-mm thick acrylic sheets (Blarney Trading. Cork, Ireland). The overall dimensions of the enclosure were $0.28\times 0.22\times 0.20\mathrm{m}$ ($L\times H\times W$). The temperature was controlled using a heater (50W, STEGO, Radionics Ireland) and a thermostat (Schneider Electric, Radionics Ireland) set to 37°C. To help maintain the temperature in the enclosure, the outer walls were covered with 0.3-m thick panels of expanded polystyrene foam. The RH was controlled via a generic fish tank water atomizer coupled to a humidity meter (Schneider Electric, Radionics Ireland) set to 90% RH (maximum range in the controller settings). The atomizer was placed inside a 200-ml Pyrex beaker filled with 50 ml of tap water. To prevent large drops of water being ejected from the beaker by the atomizer, an in-house 3D-printed catch lid was installed into the beaker. Both temperature and RH were monitored using a digital sensor (TSP01, ThorLabs). An opening was drilled in one of the walls of the enclosure for the dual laser head of the MicroLab Dual ${\mathrm{O}}_2/{\mathrm{H}}_2\mathrm{O}$ (see Sec. 3.1). The detector was mounted inside the enclosure on adjustable posts (RA90/M and TR75V/M, ThorLabs) and positioned directly in front of the source, to measure the gas content in the capillary array.

Fig. 2

(a) Diagram of the experimental set up designed to study the variations in GASMAS signal for different gas volumes in the lung tissue phantom. (b) Zoom in of source and detector placement in transmitance geometry. (c) The capillary array placed in the 3D printed holder, the sampling volume was defined by the size of light source and detector and the length of the capillary array.

The capillary array was placed between the source and detector in transmission geometry. The long walls of the capillary array were masked with a black film, and the contact surfaces between the capillaries and the system source and detector were daubed with ultrasound gel, to prevent light absorption by the gas outside the capillary tubes.

3.1.

MicroLab Dual O₂/H₂O

The GASMAS bench top system used in this study was MicroLab Dual ${\mathrm{O}}_2/{\mathrm{H}}_2\mathrm{O}$ (GASPOROX AB). The system was equipped with a dual diode laser source operating at 760 and 820 nm as well as a photodetector with $1\times {10}^{-4}{\mathrm{m}}^2$ area. The values of temperature (taken with the Thorlabs probe) and room pressure (1.007 Atm) were input to the system prior to data acquisition.

One wall of the capillary array was illuminated using the 760- and 820-nm lasers. During the illumination, each laser scanned across the absorption lines of ${\mathrm{O}}_2$ (760 nm) and ${\mathrm{H}}_2\mathrm{O}$ (820 nm). The scattered light was collected by the detector placed on the opposite side of capillary array. The system employed wave modulation spectroscopy (WMS) to increase the detection sensitivity of ${\mathrm{O}}_2$ and ${\mathrm{H}}_2\mathrm{O}$. To use this method, the detection band of each gas was modulated to higher frequencies (in the kilohertz range). Simultaneously, the laser wavelength was scanned across the gas absorption line, generating high harmonics of the absorption signal and suppressing the background noise. The characteristic output of this instrument was the amplitude signal obtained by WMS, which was used to detect the proportional light intensity dip for each wavelength.

3.2.

Air Volumes to Test in the Pulmonary Phantom Model

The sampling air volume was defined by the dimensions of the diffuse light source and detector ($0.01\times 0.005\mathrm{m}$ and $0.01\times 0.01\mathrm{m}$, respectively) and the length of the capillary array (0.025 m) as shown in Fig. 2(c). The air volume between the capillaries ($6.89\times {10}^{-7}{\mathrm{m}}^3$), constant to all measurements, was calculated by subtracting the total volume occupied by 229 capillaries ($2.81\times {10}^{-6}{\mathrm{m}}^3$) from a cube with the dimensions of the analyzed volume ($3.5\times {10}^{-6}{\mathrm{m}}^3$).

An alveolar unit holds $\sim 1.41\times {10}^{-8}{\mathrm{m}}^3$ of air, assuming a spherical geometry. The air content of a single capillary in our model contained $1.96\times {10}^{-9}{\mathrm{m}}^3$ of air. The model could hold a maximum of $3.5\times {10}^{-6}{\mathrm{m}}^3$ of air within the sampling volume, analogous to a group of alveolar sacks with at least 250  inflated alveoli. The radius of the capillaries used in this study was in agreement with the physiological radius of an alveolar unit. Despite the cylindrical geometry, the air volumes were comparable with those of pulmonary tissue and the lung model made possible to recreate progressive changes in volume during respiration in a pulmonary section.

The test set 1 started with all the capillaries empty, and the absorption signal was measured 10 times. For each acquisition, the MicroLab Dual ${\mathrm{O}}_2/{\mathrm{H}}_2\mathrm{O}$ system averaged the absorption imprints from ${\mathrm{O}}_2$ and ${\mathrm{H}}_2\mathrm{O}$ over 10 s. The air content between source and detector was reduced by replacing (and allocating randomly by hand) 20 empty capillaries with those filled with liquid phantom, followed by the respective data acquisition. This process was repeated 10 times until a total of 200 capillaries were filled with lung tissue phantom and remaining 29 filled with air, as shown in Table 1. In the test set 2, the inverse process was done. It started with all capillaries filled with optical properties and a progressive increment in air volume of 20 capillaries at a time followed by the data acquisition. The process was repeated until a total of 200 empty and 29 capillaries filled with liquid phantom were placed in the holder.

Table 1

Air volumes in the capillary lung model conformed by 229 capillaries in total and a constant air space of 6.89×10−7  m3.