N2 gas egress from patients’ airways during LN2 spray cryotherapy

ARTICLE IN PRESS

JID: JJBE

[m5G;March 17, 2017;11:37]

Medical Engineering and Physics 0 0 0 (2017) 1–10

Contents lists available at ScienceDirect

Medical Engineering and Physics journal homepage: www.elsevier.com/locate/medengphy

N2 gas egress from patients’ airways during LN2 spray cryotherapy John P. O’Connor∗, Brian M. Hanley, Thomas I. Mulcahey, Ellen E. Sheets, Kacey W. Shuey CSA Medical, Inc. 91 Hartwell Avenue, Lexington MA 02421, USA

a r t i c l e

i n f o

Article history: Received 21 March 2016 Revised 14 February 2017 Accepted 26 February 2017 Available online xxx Keywords: LN2 spray cryotherapy Cryoablation N2 egress Passive venting

a b s t r a c t Spray cryotherapy using liquid nitrogen (LN2 ) is a general surgical tool used to ablate benign or malignant lesions. Adequate egress of the gaseous nitrogen (N2 ) generated during this process must be provided for safe use when LN2 is used within the body rather than topically. When delivered to either the gastrointestinal tract (requiring active venting via a suction tube) or body cavities open to room barometric pressure (such as lung airways) allowing for passive venting, the N2 gas generated from the boiling process must be evacuated. This work will examine the egress of N2 during procedures requiring passive venting from human airways undergoing liquid nitrogen spray cryotherapy. Venting characteristics for safe N2 egress will be presented and discussed based on analytical modeling using fluid mechanics simulations and experimental studies of N2 venting with laboratory and porcine models. © 2017 The Author(s). Published by Elsevier Ltd on behalf of IPEM. This is an open access article under the CC BY-NC-ND license. (http://creativecommons.org/licenses/by-nc-nd/4.0/)

List of abbreviations CFD EES ETT N2 ID LN2 MFM OD

computational fluid dynamics engineering equation solver endotracheal tube gaseous nitrogen inside diameter liquid nitrogen mass flow meter outside diameter

1. Introduction Spray cryotherapy using liquid nitrogen (LN2 ) has been used to ablate benign and/or malignant lesions in locations requiring active venting of gas (e.g. gastrointestinal tract) [1,2] or passive venting of gas (e.g. airway) [3–5]. During application of the LN2 , the spray is targeted at desired regions inside the body. Fig. 1 schematically shows the insertion of an endotracheal tube (ETT) into the trachea (left image) and also the insertion of a bronchoscope into the airways (right image). In many situations, the bronchoscope passes through the ETT into the airway system. CSA Medical’s (Lexington, MA) proprietary, commercially available medical device, the truFreezeTM System, employs a computer-driven method and apparatus to transport LN2 from a Dewar inside a control console through a catheter that ∗

Corresponding author. E-mail addresses: [email protected], [email protected] (J.P. O’Connor).

passes through the bronchoscope working channel and into a patient allowing LN2 to be sprayed directly on the targeted tissue [8]. In the LN2 Dewar, the pressure is approximately 20 psig so utilizing the curve in Fig. 2, the saturation temperature at this pressure is approximately −187 °C. When the LN2 exits the tip of the catheter, it is at a pressure of less than 1 psig. At this pressure, the temperature of the saturated nitrogen mixture is approximately −195.2 °C [9]. The cryospray impinges on the walls of the airway where the LN2 spray draws heat away from the tissue to produce a flash freezing of the unwanted material and create a localized kill region. As the LN2 absorbs heat from the tissue, the LN2 passes through a phase change from a liquid state to a gaseous state. The gaseous nitrogen (N2 ) convectively warms on its path out of the lumen. The conversion from liquid to gaseous state and subsequent heating to room temperature is accompanied by an increase in the volume occupied by the N2 of approximately 696 times that of the LN2 (i.e., a 1 cm3 vol of LN2 at −195 °C results in a 696 cm3 vol of N2 at 20 °C) [9,10]. This gaseous nitrogen must egress from the patient to prevent an unsafe pressure increase within the application area to avoid undesired sequelae, such as a pneumothorax. This work assesses the theoretical and experimental egress of gaseous N2 expected from patients undergoing passively vented applications of LN2. The airway will be discussed and modeled as a prototypical passive venting procedure due to several factors needing consideration for adequate egress. This study includes a mathematical analysis of the human airway system and evaluates the resulting model with inputs of typical levels of gaseous N2 generated during LN2 spray cryotherapy within the trachea and generations 1 through 3 of the lungs. The simulation outputs are compared

http://dx.doi.org/10.1016/j.medengphy.2017.02.017 1350-4533/© 2017 The Author(s). Published by Elsevier Ltd on behalf of IPEM. This is an open access article under the CC BY-NC-ND license. (http://creativecommons.org/licenses/by-nc-nd/4.0/)

Please cite this article as: J.P. O’Connor et al., N2 gas egress from patients’ airways during LN2 spray cryotherapy, Medical Engineering and Physics (2017), http://dx.doi.org/10.1016/j.medengphy.2017.02.017

ARTICLE IN PRESS

JID: JJBE 2

[m5G;March 17, 2017;11:37]

J.P. O’Connor et al. / Medical Engineering and Physics 000 (2017) 1–10

Fig. 1. Schematic representation of endotracheal tube insertion (left [6]) and bronchoscope insertion in airway (right [7]).

-182 -184

Vapor

TSat (°C)

-186 -188 -190 -192

Liquid

-194 -196 0

5

10

15

20

25

30

35

Pressure (psig) Fig. 2. Nitrogen liquid–vapor saturation curve [9].

to experimental results acquired from pressure and flow measurements of anatomical lung models inputted with gaseous N2 . Lastly, results from spray cryotherapy performed in porcine models will be discussed in light of the analytical and experimental measurements. These results provide theoretical and experimental foundations to understand the volumes of gaseous N2 inputted into a patient and its egress during spray cryotherapy procedures requiring passive venting of the resulting N2 gas. Requirements for adequate N2 egress conditions are also presented. 2. Analytical modeling methods and results 2.1. Modeling methods The morphology [11–13] and hydrodynamics [14,15] of human airway flows have been extensively studied and modeled. This information is important to understanding patterns resulting from flow area reductions, expansions, turns, and entrance and exit conditions which are all variables that must be characterized in order to model the airflow in the lungs. Weibel [16] previously reported quantity, geometry, and impedance data for the various generations of the bronchial tree, which are reproduced in Table 1. In Weibel’s model, each generation bifurcates from an existing generation to the next with reductions in diameters and lengths with increasing generation (exception of generation 3 that exhibits

a shorter length (0.76 cm) than anticipated). It is to be noted that a number of the values articulated in Table 1 will vary during respiration as the muscles of the diaphragm expand and contract to bring air into or expel air from the lung and airways. To model the airflow in the lungs, one can use fluid mechanics simulations to determine which flow passages exhibit turbulent, transitional and laminar flow with details presented in the following paragraphs. By evaluating the Reynolds number, Re, it can be determined that the trachea and generations 1 to 5 all exhibit primarily turbulent flow [17] during maximal forced expiration and spray cryotherapy. Generation 6 exhibits primarily transitional flow and the remaining generations (7 to 23) exhibit laminar flow [17]. The following describes the physics associated with spray cryotherapy during passive venting. Subsequent to the cryogen exit from the catheter to the point of expiration, myriad complex thermal and hydrodynamic processes are experienced. The cryospray is a mixture of saturated liquid and gas in a dispersed multiphase flow regime (misty flow) at the saturation temperature corresponding to the cavity pressure of the lung. As the flow travels toward the tissue, droplet breakup occurs concurrent with boiling due to ambient heating and the spray spreads conically. Interacting with moist tissue, the impinging droplets wet the mucosa due to the momentum of the impinging droplets and the miscibility of nitrogen and water-based fluids. The wetted LN2 absorbs heat from the tissue during the boiling process, and transitions to a saturated vapor. As the pressure-driven gas traverses the bronchial network, it convectively absorbs sensible heat and evaporates moisture from the mucosa, increasing the gas temperature and humidity. Simultaneously, the expanding gas flow dissipates energy due to aerodynamic drag, turbulent mixing, imperfect expansion at bifurcations between lung generations, as well as flow constrictions and turns as it passes through the annulus between the bronchoscope and endotracheal tube and finally of the patient. Assessment of all of the aforementioned contributions would be impractical; therefore only primary flow drivers and losses are considered. The complicated 3-dimensional geometry of the bronchial tree, if modeled explicitly, requires significant computational power, but could be solved using computational fluid dynamics (CFD) simulation software. In order to provide a tractable scoping model for assessing the impact of surgical equipment geometry on egress pressures, the flow path can be modeled to first order as an isothermal network of straight tubes and annuli with minor loss in fidelity. Geometry, thermal effects, and multiphase phenomena are secondary and are therefore not used in the simple scoping model. The following describes the semi-empirical modeling method used to simulate gas egress as a function of injection

Please cite this article as: J.P. O’Connor et al., N2 gas egress from patients’ airways during LN2 spray cryotherapy, Medical Engineering and Physics (2017), http://dx.doi.org/10.1016/j.medengphy.2017.02.017

ARTICLE IN PRESS

JID: JJBE

[m5G;March 17, 2017;11:37]

J.P. O’Connor et al. / Medical Engineering and Physics 000 (2017) 1–10

3

Table 1 Schematic representation of airway branching in the human lung with approximate dimensions. (Adapted from [16]).

location within the lung, average gas temperature, and pressure differential. In order to simulate gas egress from the bronchial tree via an endotracheal tube, the flow path is simulated as a pipe network with viscous effects and transition losses. The steady-state energy conservation equation evaluated between two points in an incompressible flow network with head losses is given by Eq. (1) [18].

P V 2  + z + = hL,i i γ 2g

(1)

where γ is the specific weight of the fluid z is the height relative to a gravitational datum (m), V is the mean velocity of the fluid (m/s) and hL, i are the major and minor head losses (m). Viscous losses (major loss) can be characterized by the Darcy– Weisbach equation [18], which may be written as: (N/m3 ),

hL = f

L V2 D 2g

(2)

where D is the diameter of the lumen (m), L is the length of the lumen (m), and f is the friction factor. The friction factor depends on the relative roughness of the lumen, ε /D, where ε is the characteristic roughness dimension (m), and the Reynolds number Re given by Eq. (3).

ρV D Re = μ

(3)

where ρ is the density (kg/m3 ) and μ is the viscosity (Pa-s). The friction factor is dependent on whether the flow is laminar (Re < 2100) or turbulent (Re > 40 0 0), and the magnitude is computed using Eqs. (4a) and (4b), which are valid for fully-developed pipe flow [18]. For transitional flow, the friction factor is interpolated between Eqs. (4a) and (4b).

Laminar : f =

64 Re

(4a)

 1

Turbulent : 

f

= −2.0 log

ε /D 3.7

+

2.51 Re



 (4b)

f

For annular flows, the laminar friction factor is calculated based on an empirically-derived correction based on the ratio of the outer tube diameter to the inner tube diameter. The expression for laminar friction factor in an annulus is given in Eq. (4c) [18] for the diameter ratios present in the current work.

95 Laminar annulus (1.0 > Do /Di > 0.5) : f ∼ = Reh

(4c)

where Reh is the Reynolds number based on the hydraulic diameter of the annulus, Dh = Do −Di , where the subscripts o and i represent outer and inner dimensions, respectively. The turbulent friction factor for annular flow is computed using Eq. (4b), substituting Reh . Minor head losses due to flow area reductions, expansions, turns, and entrance and exit conditions can be accounted for using empirical loss coefficients denoted by KL , which are incorporated into Eq. (1) using the relation:

hL,minor = KL

V2 2g

(5)

The empirical coefficients KL are evaluated from standard curves present in most introductory fluid mechanics texts [18] and are computed based on geometry. The computed coefficients for our equivalent pipe network are given in Table 2. The net flowrate can be computed between the point at which the flow is introduced and point where the flow is exhausted based on a given differential pressure by evaluating all head losses in the flow path. The volumetric flow rate φ (m3 /s) may be written as φ = AV, where A is the cross-sectional area (m2 ). The conservation of mass is evaluated at each node in a branching pipe network based on Eq. (6):



m˙ in =



m˙ out

(6)

Please cite this article as: J.P. O’Connor et al., N2 gas egress from patients’ airways during LN2 spray cryotherapy, Medical Engineering and Physics (2017), http://dx.doi.org/10.1016/j.medengphy.2017.02.017

ARTICLE IN PRESS

JID: JJBE 4

[m5G;March 17, 2017;11:37]

J.P. O’Connor et al. / Medical Engineering and Physics 000 (2017) 1–10 Table 2. Analytical model dimensions and minor loss coefficients. Symbol

Magnitude

Units

Description

DETT Dscope Dtrachea DB 1 DB 2 DB 3 Dexit LETT Ltrachea LB 1 LB 2 LB 3

8.5 variable 18.0 12.2 8.3 5.6 9.5 355.6 120 47.6 19.0 7.6 4.6 × 10– 4 7.5 × 10– 3 1.5 × 10– 4 0.8 1.1 1.0 0.28 0.28 0.28

mm mm mm mm mm mm mm mm mm mm mm mm m m m – – – – – –

Endotracheal tube inner diameter (ID) Bronchoscope outer diameter (OD) Trachea ID Bronchus generation 1 ID Bronchus generation 2 ID Bronchus generation 3 ID Endotracheal tube exit plenum ID Endotracheal tube length Trachea length Bronchus generation 1 length Bronchus generation 2 length Bronchus generation 3 length Equivalent roughness of endoscope Equivalent roughness of trachea Equivalent roughness of bronchus Loss coefficient, contraction from trachea to ETT/scope annulus, modeled as reentrant tube Loss coefficient, mitered 90° elbow branching flow out of ETT Loss coefficient, exit expansion efficiency to atmospheric pressure Loss coefficient, expansion from generation 1 to trachea Loss coefficient, expansion from generation 2 to generation 1 Loss coefficient, expansion from generation 3 to generation 2

ε scope ε trachea εB

K1 K2 K3 KB 1 _trachea KB 2 _B 1 KB 3 _B 2

Fig. 3. Pipe network representation/schematic indicating losses and coefficients in each section.

where m˙ = ρφ is the mass flow rate. In an LN2 cryospray application, one must consider the dependence of the flow path on the location of the spray within the network and calculate the egress flow based on that specific location. Also, flow resistances for each of the lumens must be evaluated as well as resistances due to the inclusion of medical devices that are used for clinical applications such as bronchoscopes and endotracheal tubes. An idealized straight and annular pipe network representing the steady-state gas egress path is schematically shown in Fig. 3 for a case where flow is injected in a 3rd generation bronchus. The model inputs are given in Table 2. Relative roughness values were constrained to physically reasonable values and used as empirical fitting parameters to best correlate the model to experimental data, which will be presented later in this article. Parallel branches are not included in this model as during steady-state, they do not contribute to the net outflow. The full system of equations for this model is not repeated here for brevity. 2.2. Analytical simulation results In the following analysis, LN2 sprays originating in lung generations 0 to 3 as illustrated in Fig. 3 have been simulated by solving systems of equations based on Eqs. (1–6) and the input parameters given in Table 2 using Engineering Equation Solver (EES) [19]. EES contains physical property routines which compute the temperature- and pressure-dependent viscosity and density of the fluid based on the NIST equations of state for nitrogen. An iterative or nonlinear solver is required due to the implicit nature of the viscous head loss terms, where the friction factor depends on the flow, and vice versa. Friction factors described in Eq. (4a–c) are im-

plemented in EES through internal functions called MoodyChart() for pipe flows and AnnularFlow() for annular segments. In a simulated bronchial procedure, an 8.5 mm ID ETT is placed in the trachea with the cuff inflated (representing a worst case scenario for gas egress) as shown in Fig. 1 (left image). Calculations were completed for both room temperature N2 gas (20 °C) and for cold (not liquid) N2 gas (−90 °C). Simulations incorporated a bronchoscope routed through the ETT (representing simulated use) with distal end located at the following junctions (see Fig. 4): a. b. c. d.

The The The The

exit of the ETT for the trachea measurement trachea for the main stem measurement main stem for the lobar bronchus measurement lobar bronchus for the segmental bronchus measurement.

Simulation results are shown in Table 3 for a pressure differential of 40 cm H2 O, representing a safe clinical condition consistent with the operating pressure of common ventilator circuits. Simulations for bronchoscope diameters of 4.4 mm, 5.0 mm, and 6.0 mm illustrate the decrease in the maximum safe flow rate as the bronchoscope occupies more of the venting area. An 8.0 mm ID ETT was also simulated for comparison to clinical results obtained using a 6.0 mm bronchoscope during cryospray. Note that the propagated uncertainty for simulated flow rates is approximately +12%. Reported uncertainty is attributed to the following: (1) the use of the Colebrook Equation (Eq. 4b) which approximates the Moody diagram (±10%), (2) the uncertainty in the roughness of the lumen walls and bronchoscope (±50%), and (3) the use of fully-developed flow correlations which ignore the effects in the entrance length of each segment. The entrance length

Please cite this article as: J.P. O’Connor et al., N2 gas egress from patients’ airways during LN2 spray cryotherapy, Medical Engineering and Physics (2017), http://dx.doi.org/10.1016/j.medengphy.2017.02.017

ARTICLE IN PRESS

JID: JJBE

[m5G;March 17, 2017;11:37]

J.P. O’Connor et al. / Medical Engineering and Physics 000 (2017) 1–10

5

Fig. 4. Schematic image of airway [20] model test locations. Table 3. Simulation results of N2 egress from airway generations 0–3 given 40 cm H2 O differential pressure, as a function of bronchoscope OD, ETT ID, and mean gas temperature. Gas egress flow rate (L/s) 8.5 mm ETT 4.4 mm scope

8.0 mm ETT 5.0 mm scope

6.0 mm scope

6.0 mm scope

Spray location \ N2 Temp

20 °C

−90 °C

20 °C

−90 °C

20 °C

−90 °C

20 °C

−90 °C

Trachea Main bronchi (1) Lobar bronchi (2) Segmental bronchi (3)

0.99 0.99 0.99 1.03

0.78 0.78 0.79 0.81

0.85 0.85 0.86 0.87

0.68 0.68 0.68 0.69

0.71 0.71 0.71 0.70

0.47 0.47 0.47 0.46

0.55 0.55 0.55 0.53

0.38 0.38 0.38 0.38

Le = 4.4Dh Re^(1/6) was computed for all segments and it was found that the fully developed flow only exists in 79% of the annulus between the bronchoscope and ETT, therefore an error is present in the friction factors for the remaining segments of undeveloped flow. Nearly 75% of the total head loss in the egress path occurs within the ETT annulus, therefore the losses in the actual bronchial tree and their associated uncertainties are small compared to the net system losses. The net resistance of the egress flow path is dominated by the flow through and exit from the endotracheal tube assembly, while the airways themselves only contribute a small percentage of the head losses. As such, the addition of short sections of distal airways in a turbulent flow field yields negligible changes to the overall flowrate. Planned future work includes transient second order modeling of saturated N2 expansion in the bronchial tree using computational fluid dynamics, which will explicitly incorporate local temperature-dependence of the fluid properties rather than using bulk averaged properties, as well as heat transfer with the bronchial lumens. 3. Experimental methods and results 3.1. Measurements of gaseous N2 input into an airway during LN2 spray cryotherapy An apparatus and test method were developed to create a repeatable method to evaluate LN2 spray characteristics and N2 generation. The test configuration is shown schematically in Fig. 5. LN2

exits the truFreezeۛ control console at its typical operational parameters (LN2 Dewar head pressure of 22 psi and temperature of −196 °C) and enters the Catheter. Cryogen then enters the Test Chamber kept at ambient temperature through the Test Chamber Input Port. The LN2 is sprayed into the test chamber where it impacts the walls of the chamber. At this point, in the same process as observed in human airways, the LN2 changes state to gaseous N2 , expands as it approaches room temperature, and exits through the mass flow meter (MFM). The MFM measurement is used to determine the amount of N2 gas that was inputted into the Test Chamber. (It is noted that the modeling described above follows the same pattern utilized in the test apparatus of LN2 cryospray impacting the walls of the airway, conversion to gaseous N2 , and expansion upon warming to body temperatures.) Using this assembly, the gaseous N2 flow rate and total gaseous N2 vol from the CSA Medical spray cryotherapy system to be used in airway applications were determined. The results are shown in Figs. 6 and 7, respectively. As may be observed, during a typical airway cryospray application, the steady-state gaseous N2 flow rate into the simulated airway model is 22 standard liters per minute (SLPM), or 0.37 standard liters per second (SLPS). The total N2 vol generated in a typical application is approximately 4.2 L. The N2 egress of the ETT/bronchoscope/catheter system must be capable of ensuring that a flow rate above 0.37 SLPS is able to efficiently discharge from the patient with a maximum pressure of 40 cm H2 O within the lung cavity to avoid potential adverse events [21].

Please cite this article as: J.P. O’Connor et al., N2 gas egress from patients’ airways during LN2 spray cryotherapy, Medical Engineering and Physics (2017), http://dx.doi.org/10.1016/j.medengphy.2017.02.017

ARTICLE IN PRESS

JID: JJBE 6

[m5G;March 17, 2017;11:37]

J.P. O’Connor et al. / Medical Engineering and Physics 000 (2017) 1–10

Fig. 5. Schematic of N2 gas flow rate and total N2 gas input test apparatus.

Flow Rate (SLPM)

25 20 15 10 5

0 0

2

4

6

8

10

12

14

16

Time (s)

Total volume (liters)

Fig. 6. N2 gaseous flow rate during LN2 application.

4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 0

2

4

6

8

10

12

14

16

Time (s) Fig. 7. Cumulative N2 gaseous volume during LN2 application.

4. Measurements of gaseous N2 egress from airway models In order to assess the effectiveness of N2 egress from proximal regions of the lung, both laboratory and porcine animal studies were completed. 4.1. Laboratory N2 egress test measurements 4.1.1. Passive venting parameter analysis For the truFreeze commercial cryospray system, an analysis of cavity pressures versus passive vent areas was completed. A schematic of the test apparatus is shown in Fig. 8. LN2 exits

the truFreezeۛ control console at its typical operational parameters (LN2 Dewar head pressure of 22 psi) and enters a catheter. The LN2 is sprayed into a test chamber from the catheter that passes through an endoscope. A specified annular area is utilized to permit gaseous N2 venting around the endoscope. The pressure is monitored in the test chamber by a pressure sensor during the cryospray. Results of the maximum cavity pressure versus annular area for esophageal applications of LN2 are shown in Fig. 9. In measurements similar to those noted above (Figs. 6 and 7), the maximum gaseous N2 input into a patient with typical truFreeze LN2 sprays (termed “normal flow” with truFreeze end spray catheters) is 0.65 L/s. As can be observed in Fig. 9, the maximum pressure under these spray conditions with an annular passive venting area of >20 mm2 remains below 40 cm H2 O. This result further evidences the safety of passive venting during LN2 sprays. 4.1.2. N2 egress measurements in a computed tomography Data-generated partial lung model Flow measurements were conducted utilizing a computed tomography (CT) data-generated, partial lung model that is shown schematically in Fig. 10. The bronchial tree was fabricated based on a model created by Advanced Vascular Models of Seaside, CA derived from a CT scan of a 69-year old female patient. The airway model terminated at the 3rd generation. These measurements were assessed under controlled-input pressure conditions for input N2 . Gaseous N2 inputs were connected at three segmental bronchi locations. At the inputs, the pressure was monitored by a pressure sensor. (N2 input at the 1st location is shown in Fig. 10.) An ETT was inserted into the trachea and the cuff was inflated to ensure N2 gas only flowed through the ETT. N2 flow values were determined using a mass flow meter at the exit of the endotracheal tube with input airway pressures ranging from 10 cm H2 O to 50 cm H2 O in 10 cm H2 O increments. (Although each individual measurement was performed with a static pressure input, varying the input pressure from 10 to 50 cm H2 O in 10 cm H2 O increments should present a reasonable approach to approximate a dynamic pressure input over the range.) Note that efficient discharge of the N2 gas from the patient with a maximum pressure of 40 cm H2 O within the lung cavity is required to avoid potential adverse events [21]. The first set of measurements utilized only the endotracheal tube inserted into the trachea of the model with the cuff inflated. For this setup, the test results are shown in Fig. 11 for a simulated spray (N2 input) in the segmental bronchus.

Please cite this article as: J.P. O’Connor et al., N2 gas egress from patients’ airways during LN2 spray cryotherapy, Medical Engineering and Physics (2017), http://dx.doi.org/10.1016/j.medengphy.2017.02.017

ARTICLE IN PRESS

JID: JJBE

[m5G;March 17, 2017;11:37]

J.P. O’Connor et al. / Medical Engineering and Physics 000 (2017) 1–10

7

Fig. 8. Schematic of test apparatus to measure maximum cavity pressure during cryospray.

Fig. 9. Maximum cavity pressure as function of annular vent area using truFreeze spray cryotherapy.

The N2 flow measurements were repeated with identical setup conditions except that a 4.4 mm OD bronchoscope was inserted into the endotracheal tube with its distal end in the lobar bronchus for the segmental bronchus N2 gaseous input. The test results are shown in Fig. 12. The N2 flow measurements were repeated with identical setup conditions except that the input nitrogen gas was cooled to −14 °C to represent cold N2 in the airways (the 4.4 mm OD bronchoscope was inserted into the endotracheal tube with its distal end in the lobar bronchus for the segmental bronchus N2 gaseous input). The test results are shown in Fig. 13. 4.2. Animal study N2 egress measurements Studies of N2 egress have been conducted in a porcine model utilizing the commercial truFreeze spray cryotherapy systems Three physicians trained on the device rendered cryogen spray in

the airway in one animal each for a total of 3 animals. Passive venting was used to evacuate the N2 gas. (Note: All animal work was carried out in accordance with CSA Medical Inc. guidelines and follow NIH Publications No. 8023, revised 1978.) LN2 spray administration was conducted using a standard 6.0 mm OD flexible therapeutic bronchoscope inserted through a standard 8.0 mm ID ETT. The vent area was the annulus between the bronchoscope and ETT (area = 22 mm2 ). The cuff on the ETT was inflated to create a closed system and to ensure the N2 gas evacuated only through the annulus area between the scope and the ETT. A pressure sensor was placed via the ETT and positioned within bronchoscopic view near the designated cryospray impact area in the airway. Lastly, the scope with the LN2 spray catheter was positioned at the opening to the left or right main stem to create an additional obstruction for N2 gas evacuation. In each animal, the first site was sprayed with typical LN2 output for four cycles of 5 s each for a total of 20 s. The scope was

Please cite this article as: J.P. O’Connor et al., N2 gas egress from patients’ airways during LN2 spray cryotherapy, Medical Engineering and Physics (2017), http://dx.doi.org/10.1016/j.medengphy.2017.02.017

ARTICLE IN PRESS

JID: JJBE 8

[m5G;March 17, 2017;11:37]

J.P. O’Connor et al. / Medical Engineering and Physics 000 (2017) 1–10

Fig. 10. Schematic view of gaseous N2 flow measurement apparatus without bronchoscope inserted.

1.60

N2 Flow Rate (liters/s)

N2 Flow Rate (liters/s)

1.60 1.40 1.20 1.00 0.80 0.60 0.40 0.20

1.20 1.00 0.80 0.60 0.40 0.20 0.00 0

0.00 0

10

20

30

40

50

60

Input Pressure (cm H2O) Fig. 11. N2 egress from lung model, N2 (20 °C) input in segmental bronchus, 8.5 mm ID ETT, no bronchoscope.

1.60

N2 Flow Rate (liters/s)

1.40

1.40

1.20 1.00 0.80 0.60 0.40 0.20 0.00 0

10

20

30

40

50

60

Input Pressure (cm H2O) Fig. 12. N2 egress from lung model, N2 (20 °C) input in segmental bronchus, 8.5 mm ID ETT, 4.4 mm bronchoscope in lobar bronchus.

10

20

30

40

50

60

Input Pressure (cm H2O) Fig. 13. N2 egress from lung model, N2 (−14 °C) input in segmental bronchus, 8.5 mm ID ETT, 4.4 mm bronchoscope in lobar bronchus.

then re-positioned over the opposite mainstem and the flow of the device was reduced to 50% of the initial sprays. Four cycles of 10 s each for a total of 40 s were delivered. As noted above, the gaseous N2 inputted into the animals using the device with typical spray output was 0.65 L/s; for the 50% reduced output, the N2 generation was 0.35 L/s. All three animals were safely sprayed in the pulmonary space into the openings of the right and left mainstems (Generation 1) using the 22 mm2 annulus area for N2 egress without adverse event. The physicians monitored the in situ pressure while also monitoring the animal vital signs (blood pressure, heart rate, oxygen saturation levels, etc.) to ensure the animal safely tolerated the LN2 spray process. For the typical LN2 cryosprays (‘normal flow”) where approximately 0.65 L/s of N2 gas was input into the animal’s airway, the pressure varied from 32 to 70 cm H2 O with an average of 53 cm H2 O. For the LN2 sprays at 50% of typical output (approximately 0.32 L/s of N2 gas input), the pressure varied from 17 to 58 cm H2 O with an average of 37 cm H2 O. All animals

Please cite this article as: J.P. O’Connor et al., N2 gas egress from patients’ airways during LN2 spray cryotherapy, Medical Engineering and Physics (2017), http://dx.doi.org/10.1016/j.medengphy.2017.02.017

ARTICLE IN PRESS

JID: JJBE

[m5G;March 17, 2017;11:37]

J.P. O’Connor et al. / Medical Engineering and Physics 000 (2017) 1–10

Volume Flow Rate (liters/min)

80 Experimental: T=23.5 C 70

Experimental: T=-15 C Simulaon: T=23.5 C

60

Simulaon: T=-15 C

50 40 30 20 10 5

15

25

35

45

Differenal Pressure (cm H 2O) Fig. 14. Comparison of analytical model and benchtop experiment results. Both include a 4.4 mm diameter bronchoscope placed within an 8.5 mm I.D. ETT, with the scope located in the 2nd generation bronchus and gas input at the 3rd generation.

recovered uneventfully from anesthesia and returned to normal diet within 24–48 h of receiving cryospray. The animals were closely monitored for one week post cryospray. At this point, the animals were euthanized and the lungs resected for pathological inspection of the cryo-induced injury. This examination showed cryo-injury to the mucosal and sub-mucosal layers followed by disrupted smooth muscle layer with a depth of injury of approximately 2.0 mm. It is illustrative to compare the fluid dynamics simulations to the laboratory and animal measurements of the N2 egress from airways during LN2 spray cryotherapy. For the purposes of model validation, the previously-described analytical model was given geometry inputs matching the CT bronchial tree setup shown in Fig. 10, including plumbing used to connect the gas source to the inlet and the mass flow meter to the outlet. The differential pressure and gas temperature were then varied in the model to match the experimental permutations obtained with the scope placed in the 2nd generation bronchus, corresponding to a spray in the 3rd generation. The results of this model validation are compared to the corresponding experimental results in Fig. 14.

9

tem averaged 53 cm H2 O (maximum of 70 cm H2 O) with 22 mm2 of passive N2 egress through an endotracheal tube. This result is higher than measured in the laboratory model shown in Figs. 12 and 13 that demonstrated a pressure of <30 cm H2 O under similar egress conditions, with the notable exception that the bench model incorporated an 8.5 mm endotracheal tube compared to the 8.0 mm model used in the animal laboratory. As there is no directly comparable benchtop data, we rely on the analytical model for comparison. When the scope and endotracheal tube diameters are adjusted to match the animal lab configuration, the correlated analytical model predicts a flow rate of 0.63 L/s when the average pressure differential of 53 cm H2 O measured in the animal study is inputted, yielding only 3.1% error relative to the flow rate of 0.65 L/s inputted into the animal. It is to be noted that for LN2 cryospray conditions for airway applications, the measured gaseous N2 input into a patient would be approximately 0.37 L/s (Figs. 6 and 7). It is reasonable to surmise that the measured pressure within the animal airways would be reduced proportionally to the input N2 gas amount. Thus, from the data obtained during the animal study, one would expect that under the airway cryospray conditions that the average pressure would remain below 30 cm H2 O with maximum pressure below 40 cm H2 O. These conditions should prevent undesired sequelae, such as a pneumothorax within the airways [21]. The analytical modeling, laboratory, and animal study results reported here demonstrate that, with appropriate passive venting considerations, LN2 cryospray applications in the airway may be implemented safely and without potential patient issues such as pneumothorax. It is also to be noted that these analyses and experimental studies were conducted under worst case operational conditions when the cuff of the endotracheal tube had been kept inflated and the only gaseous N2 egress was through the annular opening between the inner diameter of the ETT and the outer diameter of the bronchoscope. For LN2 spray cryotherapy in the airway, the cuff on the ETT will be deflated (and the circuit for the ventilator system will be removed). These additional considerations will further enhance the safety factor for the patient with respect to gaseous N2 egress from the lungs during LN2 spray applications of the airways. Conflict of Interest

5. Discussion Fig. 14 directly compares analytical simulations to benchtop experimentation with a CT-data generated partial lung model. The dimensional inputs used to correlate the analytical (pipe and annular flow) model were adjusted to match the features present in the experimental setup, including adjusting the nominal 4.4 mm bronchoscope to 4.6 mm to match the experimental scope, whereas the dimensions from Zarei et al. [17] were used to generate the results in Tables 2 and 3. Fig. 14 shows strong correlation for pressures 30 cm H2 O and greater, with an average error of 1.9% and maximum error of 3.7% relative to experiments. The model shows poor agreement for flows driven by pressures less than 25 cm H2 O, with average error of 22.8% and maximum error of 40.5%. The error at intermediate pressures is due to the transitional flow within the annulus formed between the endotracheal tube and the bronchoscope, for which the interpolated friction factor evidently underestimates viscous losses. The slope increases for pressures less than 10 cm H2 O, corresponding to fully laminar flow within the ETT annulus, which suggests that the model may provide acceptable accuracy at very low flows; however, this range is outside the scope of interest for the present study and no experimental data was collected. In the animal study, the observed pressure within the airway during LN2 cryospray in “normal mode” with the truFreeze sys-

All authors are employees of CSA Medical, Inc. 91 Hartwell Avenue, Lexington, MA 02421, USA and report no conflict of interest with this work. Acknowledgments The authors wish to thank the great team at CSA Medical, Inc. for their efforts and support to make this work possible. References [1] Greenwald BD, Dumot JA, Abrams JA, Lightdale CJ, David DS, Nishioka NS, et al. Endoscopic spray cryotherapy for esophageal cancer; safety and efficacy. GIE 2010;71:686–93. [2] Shaheen NJ, Greenwald BD, Peery AF, Dumot JA, Nishioka NS, Wolfsen HC, et al. Safety and efficacy of endoscopic spray cryotherapy for Barrett’s esophagus with high-grade dysplasia. GIE 2010;71:680–5. [3] Browning R, Parrish S, Sarkar S, Turner JF Jr. First report of a novel liquid nitrogen adjustable flow spray cryotherapy (SCT) device in the bronchoscopic treatment of disease of the central trachea-bronchial airways. J Thorac Dis 2013;5(3):E103–6. [4] Fernando HC, Dekeratry D, Downie G, Finley D, Sullivan V, Sarkar S, et al. Feasibility of spray cryotherapy and balloon dilation for non-malignant strictures of the airway. Eur J Cardiothorac Surg 2011;40:1177–80. [5] Krimsky WS, Roidrigues MP, Malayaman N, Sarkar S. Spray cryotherapy for the treatment of glottic and subglottic stenosis. Laryngoscope 2010;120:473–7. [6] Cuffed Endotracheal Tube Medical Dictionary. http://medical-dictionary. thefreedictionary.com/endotracheal+tube; 2016 (retrieved October 4, 2016).

Please cite this article as: J.P. O’Connor et al., N2 gas egress from patients’ airways during LN2 spray cryotherapy, Medical Engineering and Physics (2017), http://dx.doi.org/10.1016/j.medengphy.2017.02.017

JID: JJBE 10

ARTICLE IN PRESS

[m5G;March 17, 2017;11:37]

J.P. O’Connor et al. / Medical Engineering and Physics 000 (2017) 1–10

[7] Bronchoscopy MDGuidelines. www.mdguidelines.com/bronchoscopy; 2016 (retrieved October 4, 2016). [8] Sachdeva A, Pickering EM, Lee HJ. From electrocautery, balloon dilation, neodymium-doped:yttrium-aluminum-garnet (Nd:YAG) laser to argon plasma coagulation and cryotherapy. J Thorac Dis 2015;7(S4):363–79. [9] Lemmon EW, McLinden MO, Friend DG. Thermophysical properties of fluid systems. NIST chemistry WebBook, NIST standard reference database number 69. Linstrom PJ, Mallard WG, editors. Gaithersburg MD 20899: National Institute of Standards and Technology; 2016 http://webbook.nist.gov retrieved January 28. [10] Unit Conversion Data for Nitrogen Universal Industrial Gases, Inc. http://www. uigi.com/n2_conv.html; 2016 (retrieved July 15, 2016). [11] Parker H, Horsfield K, Cumming G. Morphology of a distal airways in the human lung. J Appl Physiol 1971;31(3):386–91. [12] Paré PD, Wiggs BR, James A. The comparative mechanics and morphology of airways in asthma and in chronic obstructive pulmonary disease. Am Rev Respir Dis 1991;143(5):1189–93. [13] Weibel ER. Morphometry of the human lung. Berlin: Springer-Verlag; 1963.

[14] Mauroy B. 3D Hydrodynamics in the upper human bronchial tree: interplay between geometry and flow distribution. In: Losa GA, Merlini D, Nonnenmacher TF, Weibel ER, editors. Fractals in biology and medicine, 4. Basel: Birkhäuser; 2005. p. 43–53. [15] Bauer K, Brücker C. The influence of airway tree geometry and ventilation frequency on airflow distribution. J Biomech Eng 2015;137(8):081001. [16] Weibel ER. Design of airways and blood vessels considered as branching trees. The Lung: scientific foundations. 2nd ed. New York: Lippincott-Raven; 1997. p. 1061–71. [17] Zarei S, Mirtar A, Andresen B, Salamon P. Modeling the airflow in a lung with cystic fibrosis. J Non-equilib Thermodyn 2013;38:119–40. [18] Munson BR, Young DF, Okiishi TH, Huebsch WW. Fundamentals of fluid mechanics. Hoboken: NJ: J. Wiley & Sons; 2006. [19] Engineering Equation Solver. F-Chart software, LLC. Madison, Wisconsin; 2015. [20] http://www.ivyroses.com/HumanBody/Respiratory/Respiratory_ Tracheobronchial_Tree.php (retrieved 21 Jul 2016). [21] Barash PG, Cullen BF, Stoelting RK, Cahalan MK, Stock MC, editors. Clinical anesthesia. 6th ed Philadelphia: Lippincott: Williams & Wilkins; 2009. p. 251.

Please cite this article as: J.P. O’Connor et al., N2 gas egress from patients’ airways during LN2 spray cryotherapy, Medical Engineering and Physics (2017), http://dx.doi.org/10.1016/j.medengphy.2017.02.017