Recurrent laryngeal neuropathy (RLN) is a common upper airway disease in horses resulting in paralysis of the cricoarytenoideus dorsalis muscle and loss of left arytenoid cartilage abduction. The clinical consequence of RLN is inspiratory airflow obstruction resulting in decreased athletic performance.1–3 The prosthetic laryngoplasty procedure can return airflow parameters to normal values by fixing the left arytenoid cartilage in an abducted position using a permanent suture.3–10 Abduction loss, the most common post-operative complication, occurs in 76–95% of horses within the first 6 weeks after surgery, after which the arytenoid position stabilizes.11–13 Proposed mechanisms of abduction loss include poor surgical technique, suture cut-through or migration through the cartilages, failure to apply sutures with even tension, cartilage fracture, suture stretching under cyclic loading, suture breakage, and suture slippage off the caudal edge of the cricoid cartilage.3,11,14–22 Arthrotomy into the cricoarytenoid joint (CAJ) performed as a modification of laryngoplasty increases arytenoid stability by inducing perilaryngeal fibrosis and CAJ ankylosis.23,24 However, abduction loss can still occur before perilaryngeal fibrosis is complete and if not recognized promptly can impair the ability to abduct the arytenoid at repeat surgery.
Dysphagia and coughing are also clinically important post-operative complications.13,25–32 Coughing has been reported in up to 57% of cases during the immediate post-operative period but is reported to persist long-term (> 4 months) in 5–43% of cases.7,8,10,11,33,34 Dysphagia occurs less commonly, in 1.3–16% of cases, but can be life-threatening.15,19,27,35 Mechanisms that cause post-operative coughing and dysphagia are not well understood. Damage to the cranial esophageal sphincter, excessive arytenoid abduction, nerve trauma, nerve irritation, and entrapment of the esophageal adventitia within the suture loop may all play a role in the development of post-operative coughing and/or dysphagia.7,10,11,13,25,28,30,31,33,34 Chronic dysphagia and coughing have been managed successfully in some cases with prosthesis removal and vocal fold augmentation.25,32,36,37 Recent publications have described the proximity of the esophageal adventitia and pathways of several critical vagal nerve branches that may be damaged during the surgical approach for laryngoplasty.18,31,38 A minimally invasive procedure to abduct the arytenoid cartilage may limit complications related to surgical trauma.
Three-dimensional (3D) imaging, such as computed tomography (CT), is used for pre-operative planning, procedure simulation, and intraoperative guidance in human surgery across many fields. Minimally invasive procedures developed using CT guidance have demonstrated improved post-operative outcomes such as reduced incisional complications, decreased post-operative pain, and improved cosmesis due to greater procedural accuracy, decreased dissection, and shorter surgery time.39 In equids, pre-operative CT-generated 3D models have been used for surgical planning and the use of intraoperative CT has improved implant placement for the repair of complicated fractures.40,41
In an attempt to address our overarching goal of achieving rigid fixation of the arytenoid cartilage in abduction while limiting the risk of dysphagia, we propose a novel surgical approach to fix the arytenoid cartilage in abduction using a rigid implant placed across the CAJ in a minimally invasive fashion using CT planning and guidance to facilitate surgical accuracy. The rationale for our proposed approach is as follows. (1) To address surgical error as a cause of abduction loss and/or dysphagia by utilizing CT-generated models intra-operatively to accurately determine the degree of abduction before and after implant placement and to allow planning of the size and orientation of implants to be placed through a minimally invasive stab incision or mini-approach to the muscular process of the arytenoid cartilage. (2) To eliminate some of the proposed causes of abduction loss, many of which implicate the caudal aspect of the cricoid cartilage and suture or knot properties. We have chosen to take a stepwise approach toward the development of the procedure, given the technical nature of the proposed surgical procedure. Because 3D models will be required at every stage of procedural development, the first step is to establish a methodology for generating 3D models to be used in the first study presented here. These models will be used to confirm standard methods of abduction measurement, confirm that a modeled implant can be placed across the CAJ based on the models, and establish if arthrotomy would alter the mobility of the cartilage and the anatomic ability to place an implant across the joint. Future steps in developing this novel surgery will include placing implants across the CAJ in cadaveric larynges and subjecting them to negative airflow pressure conditions, followed by the development of the method for arytenoid abduction and minimally invasive surgical approach for implant placement in situ. The final major step will be the development of equipment for CT-guided drilling of the pilot hole for the implant with depth control to avoid penetration of the laryngeal mucosa. The findings of each progressive step may of course pose new questions to be answered and result in modification of the overall plan before the procedure is performed in research animals to determine in vivo efficacy of the procedure.
The objectives of the study are to develop 3D models of larynges to compare measurements of arytenoid abduction between specimens and models, to investigate the anatomic feasibility of placing a rigid implant across the CAJ, and finally, to determine if incision into the CAJ alters the anatomic relationship of the cartilages. The first hypothesis of this study was that left-to-right quotient (LRQ) angle and cross-sectional area (CSA) measurements will be comparable between photographic images of abducted larynges and 3D laryngeal models. The second hypothesis of this study was that there will be sufficient tissue dimension based on 3D models to allow the positioning of modeled implants within the muscular process and across the contact area of the CAJ when the arytenoid cartilage is abducted. The third hypothesis of the study was that arthrotomy of the CAJ will result in different measures of arytenoid translation and arytenoid contact area compared with intact joints as measured on 3D models.
Materials and Methods
Laryngeal collection and storage
This study was approved by the University of Florida’s Institutional Animal Care and Use Committee. Larynges (n = 9) were harvested from adult horses (age range = 2–15 years; weight range = 400–700 kg) euthanized for reasons unrelated to the study and free of airway disease. Larynges were collected within 3 hours after death. To minimize freezer-induced tissue oxidation and dehydration, the fresh specimens were wrapped in non-permeable plastic wrap and placed in double resealable, zipper storage bags. The specimens were cooled in a refrigerator for 24 hours before being transferred to a −20 °C freezer. Twenty-four hours before experimentation, the specimens were allowed to thaw at room temperature in their original wrapping.
Prosthetic laryngoplasty and CT
CT scans of the larynges were acquired sequentially in 3 positions: neutral, bilateral arytenoid abduction using the traditional laryngoplasty procedure (ABD), and bilateral arytenoid abduction using the traditional laryngoplasty procedure after incision into the left CAJ (ARTH). Standardized photographic images (Canon EOS Rebel T6) were obtained of each larynx in neutral, ABD, and ARTH positions. For each CT exam, larynges were positioned ventrodorsally in a small animal comfort pad (“V” trough). Transverse plane images of the larynges were obtained in a soft tissue algorithm (window width: 350 Hounsfield units (HU), window level: 35 HU), using a 160-slice multidetector-row CT scanner (Toshiba Aquilion Prime, Cannon Medical Systems) with the following settings: 80 kVp, 200 mA, 2 mm slice thickness, 0.75 second tube rotation time, and a pitch of 0.638. Images were stored in Digital Imaging and Communications in Medicine (DICOM) format.
After the first scan of the larynges in a neutral position, prosthetic laryngoplasty was performed as previously described with minor modifications.2 The right arytenoid cartilage was first secured in maximal abduction. The procedure was then repeated on the left arytenoid cartilage to achieve between 80–100% of maximal abduction on the left side. Briefly, the procedure was as follows: 2 strands of 5 braided polyester suture (Ethibond, Ethicon) on a swaged trocar point needle were placed between the cricoid cartilage and the muscular process of the arytenoid cartilage. The first suture penetrated the cricoid cartilage caudally, 2–3 mm lateral to the midline, and then exited 2 cm cranial and 1 cm lateral to the sagittal ridge on the dorsal midline. Care was taken to avoid penetrating the laryngeal mucosa. The suture was then passed through the muscular process in a caudomedial to the craniolateral direction at a 30–40 degree angle from the frontal plane. The second suture was placed through the cricoid cartilage, approximately 1 cm lateral to the first, and then through the muscular process, taking care not to impale the first suture. Each suture was tied with 6 throws to secure the arytenoid cartilage in abduction. The suture ends were left long on the left side for later use. For abduction standardization, LRQ quotient angles were determined before the second CT scan as previously described.24,35 The abducted larynges were secured on a foam mat and the epiglottis was pinned ventrally and still photographic images were obtained for each larynx. The images were uploaded into Image J (National Institute of Health) and the LRQ angles were calculated for each larynx. The scale was set on each image by correlating the distance in pixels with the measurements on a measuring tape that was placed in a standard position in every image. The height of the larynx, as measured from dorsal to ventral, was determined for each larynx. The dorsal landmark was set on the midline at the junction of the corniculate processes. The ventral landmark was located on the floor of the larynx, also on the midline, and between the vocal folds. The length between the dorsal and ventral points was then increased by one-third from the dorsal midline. Tangential lines were drawn from the extended dorsal midline to the lateral aspect of the right and left arytenoid cartilages. The right and left angles were determined and recorded. The left angle was divided by the right to calculate the LRQ angle for each larynx. The minimum LRQ deemed acceptable was 0.8 or 80% of maximum abduction. If the LRQ angle was not at least 80%, the sutures were released, re-tied, and then the LRQ angle was calculated again until a value above the accepted cut-off was achieved. The CT scan was then repeated as described previously.
The sutures that held the left arytenoid cartilage in abduction during the first treatment were untied. A modified arthrotomy was performed for the ARTH treatment group. The CAJ was approached laterally, directly under the attachment of the tendon of insertion of the cricoarytenoideus dorsalis muscle on the muscular process. Curved Metzenbaum scissors were used to incise the CAJ capsule while leaving the cricoarytenoideus dorsalis muscle tendon of insertion intact. Mosquito hemostats were opened within the incision to further enlarge the incision bluntly. The sutures were retied to secure the left arytenoid cartilage in abduction. The larynx was again secured to a foam mat and the epiglottis was pinned ventrally. Still, photographic images were obtained, and the position of the left arytenoid was adjusted until the LRQ matched that obtained for the ABD treatment for each larynx as closely as possible and the CT scan was repeated.
Image segmentation and 3D modeling
CT volumes were processed using 3D reconstruction software, Materialise Mimics™ (Materialize Medical Imaging Software Suite, Materialise). DICOM images were segmented into 3D models that included laryngeal soft tissue models, individual laryngeal cartilage models, and airway luminal volume models (Figure 1). Segmentation was performed by creating a range of thresholds for each 3D object known as masks. Thresholds were based on HU determined by the pixel data within the DICOM images. Threshold ranges for the soft tissue, laryngeal cartilage, and luminal volume models of the larynx were −100 to 400 HU, 80 to 600 HU, and −1,024 to −900 HU, respectively. The laryngeal cartilages were separated further using the split mask function and manual identification of each cartilage on axial plane CT images. When there was contact between the arytenoid and cricoid cartilage that prevented the computer from separating the structures, the author’s (H.A.R.) best judgment was used to divide the medial articular surface of the arytenoid cartilage from the articular facet on the cricoid lamina to separately define the arytenoid cartilage and cricoid cartilage. The final modeled arytenoid cartilages did not include the thin elastic cartilage of the corniculate processes that could not be discerned from surrounding soft tissues based on HUs (Figure 1). The 3D objects were saved in the standard stereolithography file (STL) format. The STL files were imported into 3-Matic™ biomedical software (Materialize Medical Imaging Software Suite, Materialise) and the 3D objects were processed according to a standard protocol to create the models. All 3D objects were trimmed to eliminate disconnected tissues included in the mask by the computer program during segmentation. The 3D objects were then wrapped with a standard gap closing distance of 2.0 mm and the smallest detail was set at 0.5 mm. Standardization of the processing procedure eliminated the potential for measurement artifacts and/or object distortion created by methodology inconsistencies.
Segmentation. Segmentation used to create 3D equine laryngeal models from computed tomographic (CT) soft tissue volumes. (A) Axial CT image and (B) the same axial slice with color coding that highlights the different structures used to create the (C) soft tissue, (D) laryngeal cartilage, and (E) luminal volume models of the larynx. (F) Superimposed cartilage and soft tissue models. Note the mucosal outline of the corniculate process represented in the soft tissue model (black arrow) and the absence of the elastic cartilage of the corniculate process from the cartilage model.
Citation: American Journal of Veterinary Research 84, 6; 10.2460/ajvr.22.11.0202
Computer-generated numerical values obtained from the 3D models included the CSA of the rima glottidis based on the lumen volume model and translation of the left arytenoid cartilage from the neutral to the ABD and ARTH positions. Measurements of the models for surgical planning included the contact area between the arytenoid cartilage and cricoid cartilage at the CAJ and the distance between the luminal surface of the cricoid cartilage and the spine of the arytenoid cartilage that was used to determine the length of modeled implants.
CSA and LRQ angle measurements
Soft tissue models were rotated within 3D space to reproduce the perspective of the previously acquired photographs of the matched laryngeal specimens and treatments. The screenshot images were saved as JPEG image files for analysis. All images were imported into ImageJ (National Institute of Health) to determine rima glottidis CSA and LRQ angle. CSA was determined by manually outlining the rima glottidis on images as previously described.42 The rima glottidis CSA and LRQ angle calculations were repeated in triplicate by the same author and the average was used for statistical analysis.
Computer-generated measurements of CSA of the rima glottidis were obtained from the laryngeal luminal volume models of the ABD and ARTH treatment groups (Figure 2). The luminal volume models of each specimen were superimposed with their matched soft tissue laryngeal model. The volume models were cropped in a standardized manner such that the rostral border represented the rima glottidis. The models were first viewed in the sagittal plane. Marker points were placed dorsally at the junction of the corniculate processes of the arytenoid cartilages and ventrally at the junction of the vocal folds on the floor of the larynx. The models were then rotated to ensure the correct placement of markers in the frontal and coronal planes and the luminal volume model was cropped at this location. The rostral face of the cropped model was highlighted excluding the laryngeal saccules that do not contribute to the lumen of the airway but contain air and are connected to the laryngeal lumen in the 3D models. The computer-generated surface area corresponding to the CSA was recorded.
Computer-generated measurement of equine rima glottidis cross-sectional area (CSA). (A) Lateral view of a cropped lumen volume model (green model) superimposed with its matched soft tissue model in 3D space. (B) The rima glottidis CSA was determined based on the computer-generated surface area (cm2) of the rostral face of the cropped lumen volume model highlighted in orange.
Citation: American Journal of Veterinary Research 84, 6; 10.2460/ajvr.22.11.0202
Translation of the arytenoid cartilage
The translation of the arytenoid cartilage from the neutral to the ABD and ARTH positions was determined (Figure 3). Four standardized landmarks on the arytenoid cartilage were used for translation analysis. Landmarks included the vocal process, the center of the muscular process, the axial edge of the arytenoid cartilage, and the rostral edge of the spine of the arytenoid cartilage. Segmentation of the arytenoid cartilage removed the majority of the elastic cartilage; thus, these landmarks were based on the remaining hyaline cartilage. A transformation algorithm was used to determine movement in 3D space of the four landmarks from the neutral to the ABD and ARTH positions. In the neutral position, the landmark of interest in the muscular process was set on the global coordinate system to the (x, y, z) positions of (0,0,0). The location within the global coordinate system of the landmark of interest in the ABD and ARTH positions was then determined based on the movement of the landmark in 3D space from the neutral position. This was repeated for each of the 4 landmarks on the arytenoid cartilage for every larynx in each position.
Arytenoid cartilage translation and results. (A) Translation of the left arytenoid cartilage from the neutral (beige) to the abducted position (blue) in relation to the cricoid cartilage (pink) as viewed from the frontal plane. The arytenoid cartilage is viewed from the (B) rostral, (C) lateral, and (D) dorsal views. Standardized landmarks included the center of the muscular process (black dot), the rostral aspect of the spine of the arytenoid cartilage (red dot), the axial aspect of the arytenoid cartilage (green dot), and the vocal process (purple dot). Panels E–H show the mean ± standard error (n = 9) translation of the 4 standardized landmarks of the left arytenoid cartilage of equine 3D laryngeal models. There was no significant difference in the translation of these landmarks between the abducted and arthrotomy treatment groups (E–H). Paired t test; P < .05.
Citation: American Journal of Veterinary Research 84, 6; 10.2460/ajvr.22.11.0202
CAJ contact area and modeled implant lengths
Measurements representing the potential region across which surgical implant(s) could be inserted were determined by manually measuring the contact area of the CAJ (Figure 4). The model was viewed obliquely at an approximately 45-degree angle in a rostrolateral to caudomedial direction. The arytenoid cartilage was made partially transparent so that the outline of the arytenoid cartilage as it overlaid the cricoid cartilage could be visualized. The superimposition of the arytenoid and cricoid cartilages was outlined manually by the author and the surface area was recorded.
Cricoarytenoid joint contact area. The equine cricoarytenoid joint contact area is defined as the overlap between the arytenoid cartilage (blue) and the cricoid cartilage (pink), which is highlighted in yellow.
Citation: American Journal of Veterinary Research 84, 6; 10.2460/ajvr.22.11.0202
Measurements representing the length of the potential surgical implant(s) placed across the arytenoid and cricoid cartilage to affix the left arytenoid cartilage in permanent abduction were determined. Modeled implants were positioned approximately perpendicular to the CAJ coursing from the spine of the muscular process of the arytenoid cartilage to the luminal surface of the cricoid cartilage (Figure 5). Modeled implants were created using cylinders that were 1.3 mm in diameter to represent the dimensions of Arthrex Chondral Darts™ (Arthrex, Inc) positioned along the planned path of the implants. The angle and position of the cylinders were adjusted as needed to allow 2 implants with at least 2–3 mm between them and 3–4 mm between the first implant and the rostral edge of the cricoid cartilage. The length of each cylinder represented the length of potential implants placed in this location and is referred to as the length of modeled implants.
Modeled implants. Modeled implants (black) within an equine laryngeal cartilage model from the (A) rostral, (B) dorsal, and (C) lateral views. (D) The final view is a sectioned cricoid cartilage viewed from the luminal surface.
Citation: American Journal of Veterinary Research 84, 6; 10.2460/ajvr.22.11.0202
Statistical Methods
Descriptive statistics were generated for all continuous variables and the distribution of the data was determined using the Shapiro-Wilk test. Normally distributed data are presented as mean ± standard error (SE) and non-parametric data are presented as median and 95% confidence interval (CI). For statistical analysis of CSA between treatment groups, LRQ between treatment groups and between image analysis methods, translation between treatment groups, and contact area across the CAJ the paired t test was used for parametric data and the Wilcoxon signed rank test was used for non-parametric data. Statistical analysis of CSA across methods of measurement and pin lengths across treatments and locations were performed using ANOVA and Tukey post hoc test. Significance was set at P < .05 throughout.
Results
Larynges were harvested from 5 mares and 4 geldings (n = 9) with a median age of 12 years (range = 2–15) and a median weight of 496 kg (range = 449–669). Breeds included 5 Thoroughbreds, 2 American Quarter Horses, 1 Warmblood, and 1 Florida Cracker Horse.
When comparing the ABD to the ARTH treatment groups, there was no significant difference in CSA measurements within the specimens or the models. Within the ARTH group, the CSA was significantly larger when measured on photographs of the laryngeal specimens than when measured on screen captures of the soft tissue model or the computer-generated CSA from the cropped lumen volume model (P = .0096; Figure 6). Within the ABD group, there was no difference in rima glottidis CSA between the 3 methods of measurement.
Cross-sectional area (CSA) of the rima glottidis results. Representative images of each method of CSA measurement: (A) Photographic image of an equine laryngeal specimen, (B) screen capture of the soft tissue model of the same larynx, and (C) computer-generated lumen volume model cropped at fixed landmarks superimposed with the soft tissue model. The outlines of the CSA in red are manually drawn in (A) and (B), and computer-generated in (C). The orange arrows highlight the wider area measured by the researcher due to the inability to accurately outline structures in the same plane on a 2D image of a 3D object compared with the computer-generated CSA measurement. Also, note the difference in an epiglottic position that affects the area outlined in (A) and (B) but not (C). (D) The mean ± standard error (n = 9) CSA measured on specimens in the laryngoplasty with cricoarytenoid joint arthrotomy group was significantly greater than when the other 2 methods of measurement were used. Different letters indicated significant differences between groups; ANOVA with Tukey post hoc test; P < .05.
Citation: American Journal of Veterinary Research 84, 6; 10.2460/ajvr.22.11.0202
Within the ABD treatment groups, the LRQ angles were 0.97, 95% CI [0.95–0.99] for the specimens, and 0.96, 95% CI [0.93–0.99] for the soft tissue models. Within the ARTH treatment groups, the LRQ angles were 0.95, 95% CI [0.93–1.05] for the specimens, and 0.96, 95% CI [0.93–1.05] for the soft tissue models. The LRQ was not different when measured on images of laryngeal specimens or soft tissue models within the ABD or ARTH groups. There was also no difference in LRQ between the ABD and ARTH treatment groups within the laryngeal specimens or the soft tissue models.
The translation of the left arytenoid cartilage in the x (medial/lateral), y (dorsal/ventral), and z (rostral/caudal) planes were not significantly different between the ABD and the ARTH groups based on the 4 standardized landmarks (Figure 3). In both the ABD and ARTH groups, all landmarks on the arytenoid cartilage moved variable distances laterally, dorsally, and caudally, with the exception of the vocal process that moved laterally, dorsally, and rostrally.
There was no difference in the contact area between the arytenoid and cricoid cartilage across the CAJ between the ABD (101.79 ± 7.19 mm2) and ARTH (109.69 ± 2.78 mm2) groups. Loss of cartilage resulted in the incongruity of the CAJ in 7/9 ARTH specimens. The modeled implant lengths within the ABD treatment group were 20.14 ± 0.51 mm for the rostrally positioned modeled implant and 20.02 ± 0.53 mm for the caudally positioned modeled implant. Within the ARTH group, the mean implant length at the rostral position was 20.06 ± 0.64 mm, and at the caudal position was 19.86 ± 0.56 mm. There was no difference in modeled implant length across treatment groups or locations.
Discussion
Prosthetic laryngoplasty was developed over 50 years ago and is the gold-standard treatment for horses with RLN.3 Despite tremendous research efforts, modest improvements have been made in post-operative outcomes.23 The success of the prosthetic laryngoplasty is limited by post-operative complications that include abduction loss, coughing, and dysphagia among others.11,13,20,22,43 To improve post-operative prognosis, surgical modifications to mitigate or eliminate factors that may contribute to the development of key post-operative complications need to be developed.
We accept our first hypothesis, that the 3D models can be used for measuring arytenoid abduction. We accept our second hypothesis that the anatomic relationships and dimension of the arytenoid cartilage and CAJ will allow an implant to be placed into the muscular process and across the CAJ into the cricoid cartilage. We reject our third hypothesis as the results of translation and contact area did not differ between the ABD and ARTH treatment groups; however, further investigation of the effect of cartilage debridement on joint congruity is warranted.
The only difference identified when evaluating CSA and LRQ on models compared with traditional measures on photographs of laryngeal specimens was a larger CSA measured on specimen photographs compared with either method of measurement using the models in the ARTH group. There are several factors that contribute to variability and inaccuracy in the measurement of CSA outlined by hand on either photographic images of laryngeal specimens or screen capture of 3D models. The epiglottis obscures the most ventral aspect of the rima glottidis in the photographic or screen capture images, which leads to an underestimation of the true CSA (Figure 6). In addition, measurements of rima glottidis CSA are subject to error from the inability to accurately outline a cross-section of a 3D object viewed as a 2D image (Figure 6). For example, the author’s outline of the CSA on the 2D images was wider at the junction of the aryepiglottic folds and the corniculate process bilaterally when compared with the cross-section created using fixed points to crop the lumen volume model. Finally, minor, unavoidable changes in camera position or 3D model rotation during 2D image acquisition can contribute to variability. The computer-generated CSA from the lumen volume model cropped at fixed landmarks is likely the most accurate measure as it is not susceptible to the sources of error listed for free-hand CSA determination based on 2D images. Overall, the underestimation of CSA due to epiglottic position and overestimation at the more rostral attachment of the aryepiglottic fold on the 2D images yielded CSA results comparable with computer-generated CSA based on the lumen volume model and likely represent reasonable estimates. The LRQ angle is less susceptible to the errors listed above as previously described.44 For this reason, the LRQ angle was used to standardize arytenoid abduction to allow a comparison of subsequent measures between the ABD and ARTH groups. There were no differences between the ABD and ARTH groups for CSA or LRQ angle indicating that the degree of arytenoid abduction was successfully standardized for all subsequent measurements. The CSA and LRQ angle measurements were comparable with published analysis methods for arytenoid abduction and can be used to assess arytenoid abduction and for surgical planning.19,24,35,44–47
We determined the area of the CAJ where there was contact between the arytenoid and cricoid cartilages on the models across which implants could be placed and we created computerized models of current commercially available implants to evaluate the ability to position the implants across the CAJ and measured the length of the modeled implant once in position. The contact area dimensions can be used to guide future implant development if needed. The overall range of CAJ contact area across treatment groups was 70.34 to 141.14 mm2. Appropriate implants used for this procedure should be chosen to fit within this area. Two modeled implants with a diameter of 1.3 mm were placed from the arytenoid cartilage to the cricoid cartilage through the CAJ. A high degree of surgical precision would be needed to avoid placing the implants too close to one another, which could weaken the cartilage. We were able to position 2 modeled implants across the CAJ while remaining within the substance of the muscular process and crossing the CAJ within the contact area. Measurements of the length of the modeled implants fell within a narrow range of 17.59 mm to 23.87 mm across groups. This procedure is intended to be performed under CT guidance and intra-operative modeling that would allow real-time measurements to be performed before implant selection. Nevertheless, the small range of measured implant lengths between larynges in this study means that a narrow range of pre-fabricated implants could be readily available for surgical placement in most cases. This would alleviate the necessity for patient-specific implants and decrease associated expenses and time needed to produce such implants. The CT scan and surgical procedure could also be performed under one general anesthetic event compared with the requirement of a second anesthetic event if a custom implant was needed. Future biomechanical studies are required to test potential implants, including the material, size, number, and orientation of implants to determine the ideal implant to maintain rigid arytenoid abduction.
Previous research has documented variability in the caudal prominences on the cricoid cartilage with smaller or absent prominences proposed to increase the risk of suture slippage and loosening on the cricoid.43,48 We also observed similar variability in the cricoid prominences in our models. Prosthesis or implant placement through the rostral lamina of the cricoid cartilage circumvents complications from suture slippage and/or pull-through from the caudal aspect of the cricoid cartilage. Additionally, the rostral lamina of the cricoid cartilage is thicker which should allow increased purchase and security compared with caudally placed prostheses.49 The use of a pin or screw to secure the arytenoid would alleviate the potential for suture lengthening or knot slippage.
Facilitated ankylosis and perilaryngeal fibrosis achieved after arthrotomy and curettage of the CAJ have demonstrated improved outcomes in racehorses compared with the traditional laryngoplasty alone.23,24 The final goal of this study was to investigate the effect of the arthrotomy of the CAJ by comparing surgical planning measurements obtained on the 3D models between the ABD and ARTH treatment groups. In the current study, arthrotomy into the CAJ did not alter any of the measurements assessed compared with leaving the joint intact but defects were identified in the articular surfaces of the cricoid cartilage that would not be desirable when placing an implant across this joint. The defects in the joint surfaces created by the insertion of Halstead Mosquito forceps into the joints were not expected but the alterations in joint congruity are an important consideration. Future investigation of the effect of more aggressive and previously described joint debridement techniques on joint incongruity will be required if arthrotomy is ultimately included in the final surgical procedure.24 While implants may avoid some of the potential causes of abduction loss attributed to suture material, cyclic fatigue of the implant or cartilage fracture would still be possible. If future biomechanical studies demonstrated that these implants were not able to withstand long-term cyclic fatigue, our results suggest arthrotomy of the CAJ in combination with the rigid implant may be a viable modification to prevent arytenoid abduction loss once fibrosis stabilizes the joint.11,13 However, further investigation of the effect of cartilage debridement as previously described on joint congruity would be required.24 The principle effect of adding arthrotomy to the surgical procedure would be the potential need for additional dissection and iatrogenic injury to surrounding structures.31
Measurement of the translation of the arytenoid cartilage was performed to assess the effect of the arthrotomy modification to the laryngoplasty on arytenoid movement when abducted. The arytenoid cartilage has an irregular shape that deforms during the abduction.50 Therefore, 4 standardized landmarks were used to assess the movement of key parts of the arytenoid using the 3D cartilage model (Figure 3). Perkins et al reported the roll, pitch, and yaw of the arytenoid cartilage sequentially during incremental forces that mimicked the contraction of the cricoarytenoideus dorsalis during the abduction.50 In that study, radiopaque markers were placed in standardized locations on the arytenoid and cricoid cartilages similar to the soft tissue model in our study.50 In addition to measuring overall arytenoid movement, we were particularly interested in the translation of the muscular process at the location where implants would be placed. There was no difference between the ABD and ARTH groups for translation of any of the 4 points and despite the lack of the corniculate process on our cartilage model, the direction of our translation results was comparable to the previous publication.50
We used CT scans of cadaver larynges to produce three 3D models. At a minimum, the soft tissue model and the cartilage model will be required for surgical planning and the cartilage model will be required for surgical guidance. The soft tissue model will be required for the determination of the LRQ, which is impossible to measure on any of the other models. The cartilage model will be required for planning the orientation and length of implants as it is the only model on which the relationship between the muscular process and the CAJ can be viewed. The volume model will be required only if accurate representations of the CSA are desired; however, this could be evaluated post-operatively for research purposes without requiring intra-operative production of the model.
Limitations of this study include the use of cadavers that lack muscle tone, blood supply, and innervation, and are subject to rigor mortis, all of which may have influenced the morphometric findings. The ABD and ARTH procedures had to be performed sequentially because the ARTH procedure could not be undone which precluded randomization. The epiglottic position was not standardized between specimen photographs and CT scanning which may have affected measurements. The epiglottis was secured ventrally to a mat using 14-gauge needles to stabilize the larynx and minimize epiglottic obstruction of the ventral aspect of the rima glottidis when tracing the CSA on photographs. However, the epiglottis was left unsecured during CT scans to avoid imaging artifacts from the needle. The results of this study may not be valid for horses falling outside of the 400 to 700 kg range used in this study, however, with the exception of draft horses, the majority of horses with RLN would fall within this range. We did not measure contact area over a variety of LRQ; therefore, the results of this study are limited to the LRQ reported. Potential implant sizes and positioning were determined on 3D models, but translation in vivo remains theoretical. Future studies should investigate the in vivo translation of the arytenoid cartilage to assess potential differences and/or significance of surgical versus in vivo abduction.
In conclusion, we were able to generate the necessary models for surgical planning and measurement using a standardized protocol. The results of this study support the anatomic feasibility of placement of a surgical implant across the CAJ to affix the left arytenoid cartilage in an abducted position for the treatment of RLN in horses. Measurements obtained on the 3D models were comparable to traditional measurements of arytenoid abduction on 2D images of cadaver laryngeal specimens. There is sufficient cartilage thickness and contact area across the CAJ to accommodate implants placed across the joint. Ex vivo biomechanical and flow chamber studies are required to test the biomechanical stability of potential implants, followed by the development of a minimally invasive surgical approach for in vivo testing.
Acknowledgments
Funding for this study was provided by the UF College of Veterinary Medicine Intramural Grant Program.
The authors declare that there were no conflicts of interest.
The authors would like to thank Christine Fitzgerald for performing the CT scans, Elizabeth Wyman and Brett Rice for technical support, and Hongjia He for assistance with biomedical software and model creation.
References
- 1.↑
Parente EJ. Fifty years of recurring struggles with recurrent laryngeal neuropathy. Equine Vet J. 2018;50(2):155–158. doi:10.1111/evj.12763
- 2.↑
Ducharme NG, Rossignol F. Larynx. In: Auer JA, Stick JA, Kummerle JM, Prange T. Equine Surgery. 5th ed. Elsevier Inc; 2019;734–769.
- 3.↑
Marks D, Mackay-Smith MP, Gushing LS, Leslie JA. Use of a prosthetic device for surgical correction of laryngeal hemiplegia in horses. J Am Vet Med Assoc. 1970;157(2):157–163.
- 4.
Tetens J, Derksen FJ, Stick JA, Lloyd JW, Robinson NE. Efficacy of prosthetic laryngoplasty with and without bilateral ventriculocordectomy as treatments for laryngeal hemiplegia in horses. Am J Vet Res. 1996;57(11):1668–1673.
- 5.
Witte TH, Mohammed HO, Radcliffe CH, Hackett RP, Ducharme NG. Racing performance after combined prosthetic laryngoplasty and ipsilateral ventriculocordectomy or partial arytenoidectomy: 135 Thoroughbred racehorses competing at less than 2400 m (1997–2007). Equine Vet J. 2009;41(1):70–75. doi:10.2746/042516408X343163
- 6.
Dixon PM, Hahn CN, Barakzai SZ. Recurrent laryngeal neuropathy (RLN) research: where are we and to where are we heading? Equine Vet J. 2009;41(4):324–327. doi:10.2746/042516409X423082
- 7.↑
Hawkins JF, Tulleners EP, Ross MW, Evans LH, Raker CW. Laryngoplasty with or without ventriculectomy for treatment of left laryngeal hemiplegia in 230 racehorses. Vet Surg. 1997;26(6):484–491. doi:10.1111/j.1532-950X.1997.tb00521.x
- 8.↑
Strand E, Martin GS, Haynes PF, McClure JR, Vice JD. Career racing performance in Thoroughbreds treated with prosthetic laryngoplasty for laryngeal neuropathy: 52 cases (1981–1989). J Am Vet Med Assoc. 2000;217(11):1689–1696. doi:10.2460/javma.2000.217.1689
- 9.
Kidd JA, Slone DE. Treatment of laryngeal hemiplegia in horses by prosthetic laryngoplasty, ventriculectomy and vocal cordectomy. Vet Record. 2002;150(15):481–484. doi:10.1136/vr.150.15.481
- 10.↑
Davenport CLM, Tulleners EP, Parente EJ. The effect of recurrent laryngeal neurectomy in conjunction with laryngoplasty and unilateral ventriculocordectomy in thoroughbred racehorses. Vet Surg. 2001;30(5):417–421. doi:10.1053/jvet.2001.25865
- 11.↑
Dixon PM, McGorum BC, Railton DI, et al. Long-term survey of laryngoplasty and ventriculocordectomy in an older, mixed-breed population of 200 horses. Part 1: maintenance of surgical arytenoid abduction and complications of surgery. Equine Vet J. 2003;35(4):389–396. doi:10.2746/042516403776014172
- 12.
Barakzai SZ, Boden LA, Dixon PM. Postoperative race performance is not correlated with degree of surgical abduction after laryngoplasty in National Hunt Thoroughbred racehorses. Vet Surg. 2009;38(8):934–940. doi:10.1111/j.1532-950X.2009.00605.x
- 13.↑
Barnett TP, O’Leary JM, Parkin TDH, Dixon PM, Barakzai SZ. Long-term maintenance of arytenoid cartilage abduction and stability during exercise after laryngoplasty in 33 horses. Vet Surg. 2013;42(3):291–295. doi:10.1111/j.1532-950X.2013.01109.x
- 14.↑
Froydenlund TJ, Dixon PM. A review of equine laryngoplasty complications. Equine Vet Ed. 2014;26(2):98–106. doi:10.1111/eve.12131
- 15.↑
Dean PW, Nelson JK, Schumacher J. Effects of age and prosthesis material on in vitro cartilage retention of laryngoplasty prostheses in horses. Am J Vet Res. 1990;51(1):114–117.
- 16.
Herde I, Boening KJ, Sasse HHL. Arytenoid cartilage retention of laryngoplasty in horses–in vitro assessment of effect of age, placement site, and implantation technique. Proceedings of the Annual Convention of the AAEP; 2001. San Diego, CA.
- 17.
Witte TH, Cheetham J, Soderholm LV, Mitchell LM, Ducharme NG. Equine laryngoplasty sutures undergo increased loading during coughing and swallowing. Vet Surg. 2010;39(8):949–956. doi:10.1111/j.1532-950X.2010.00742.x
- 18.↑
Brandenberger O, Martens A, Robert C, et al. Anatomy of the vestibulum esophagi and surgical implications during prosthetic laryngoplasty in horses. Vet Surg. 2018;47(7):942–950. doi:10.1111/vsu.12944
- 19.↑
Hawkins JF, Couetil L, Miller MA. Maintenance of arytenoid abduction following carbon dioxide laser debridement of the articular cartilage and joint capsule of the cricoarytenoid joint combined with prosthetic laryngoplasty in horses: an in vivo and in vitro study. Vet J. 2014;199(2):275–280. doi:10.1016/j.tvjl.2013.11.027
- 20.↑
Hardcastle MR, Pauwels FET, Collett MG. Clinicopathologic observations on laryngoplasty failure in a horse. Vet Surg. 2012;41(5):649–653. doi:10.1111/j.1532-950X.2012.00975.x
- 21.
Ahern BJ, Parente EJ. Surgical complications of the equine upper respiratory tract. Vet Clin North Am: Equine Practice. 2008;24(3):465–484.
- 22.↑
Gray SM, Gutierrez-Nibeyro SD, Couëtil LL, et al. Evaluation of the airway mechanics of modified toggle laryngoplasty constructs using a vacuum chamber airflow model. Vet Surg. 2021;50(7):1409–1417. doi:10.1111/vsu.13690
- 23.↑
Aceto H, Parente EJ. Using quarterly earnings to assess racing performance in 70 Thoroughbreds after modified laryngoplasty for treatment of recurrent laryngeal neuropathy. Vet Surg. 2012;41(6):689–695. doi:10.1111/j.1532-950X.2012.01017.x
- 24.↑
Parente EJ, Birks EK, Habecker P. A modified laryngoplasty approach promoting ankylosis of the cricoarytenoid joint. Vet Surg. 2011;40(2):204–210. doi:10.1111/j.1532-950X.2010.00773.x
- 25.↑
Greet TRC, Baker GJ, Lee R. The effect of laryngoplasty on pharyngeal function in the horse. Equine Vet J. 1979;11(3):153–158. doi:10.1111/j.2042-3306.1979.tb01328.x
- 26.
Ahern TJ. Laryngo-pharyngeal desensitization following a prosthetic laryngoplasty. J Equine Vet Sci. 1996;16(3):120–122. doi:10.1016/S0737-0806(96)80047-5
- 27.↑
Ahern TJ. Oral palatopharyngoplasty: a survey of one hundred post operative raced horses. J Equine Vet Sci. 1993;13(12):670–672. doi:10.1016/S0737-0806(06)81554-6
- 28.↑
Barakzai SZ, Dixon PM, Hawkes CS, Cox A, Barnett TP. Upper esophageal incompetence in five horses after prosthetic laryngoplasty. Vet Surg. 2015;44(2):150–155. doi:10.1111/j.1532-950X.2014.12101.x
- 29.
Barakzai SZ, Dixon PM, Hawkes CS, Cox A, Barnett TP. Upper esophageal incompetence in five horses after prosthetic laryngoplasty. Vet Surg. 2015;44(2):150–155. doi:10.1111/j.1532-950X.2014.12101.x
- 30.↑
Barnett TP, O’Leary JM, Dixon PM, Barakzai SZ. Characterisation of palatal dysfunction after laryngoplasty. Equine Vet J. 2014;46(1):60–63. doi:10.1111/evj.12081
- 31.↑
Pisano SRR, Stoffel MH, Bodó G. Ex vivo study of vagal branches at risk for iatrogenic injury during laryngoplasty in horses. Vet Surg. 2021;50(2):425–434. doi:10.1111/vsu.13548
- 32.↑
Luedke LK, Cheetham J, Mohammed HO, Ducharme NG. Management of postoperative dysphagia after prosthetic laryngoplasty or arytenoidectomy. Vet Surg. 2020;49(3):529–539. doi:10.1111/vsu.13389
- 33.↑
Russell AP, Slone DE. Performance analysis after prosthetic laryngoplasty and bilateral ventriculectomy for laryngeal hemiplegia in horses: 70 cases (1986-1991). J Am Vet Med Assoc. 1994;204(8):1235–1241.
- 34.↑
Speirs VC, Bourke JM, Anderson GA. Assessment of the efficacy of an abductor muscle prosthesis for treatment of laryngeal hemiplegia in horses. Aust Vet J. 1983;60(10):294–299. doi:10.1111/j.1751-0813.1983.tb02812.x
- 35.↑
Ahern BJ, Lim YW, van Eps A, Franklin S. In vitro evaluation of the effect of a prototype dynamic laryngoplasty system on arytenoid abduction. Vet Surg. 2018;47(6):837–842. doi:10.1111/vsu.12933
- 36.↑
Fitzharris LE, Lane JG, Allen KJ. Outcomes of horses treated with removal of a laryngoplasty prosthesis. Vet Surg. 2019;48(4):465–472. doi:10.1111/vsu.13150
- 37.↑
Vidovic A, Delling U. Aryepiglottic fold augmentation as treatment for late-onset dysphagia following surgical treatment of recurrent laryngeal neuropathy. Tierärztliche Praxis Ausgabe G: Großtiere/Nutztiere. 2017;45(04):219–225. doi:10.15653/TPG-160712
- 38.↑
Biasutti S, Dart AJ, Jeffcott LB. A review of recent developments in the clinical application of prosthetic laryngoplasty for recurrent laryngeal neuropathy: indications, complications and outcome. Equine Vet Ed. 2017;29(6):337–345. doi:10.1111/eve.12553
- 39.↑
Marescaux J, Diana M. Next step in minimally invasive surgery: hybrid image-guided surgery. J Ped Surg. 2015;50(1):30–36. doi:10.1016/j.jpedsurg.2014.10.022
- 40.↑
Heer C, Fürst AE, Del Chicca F, Jackson MA. Comparison of 3D-assisted surgery and conservative methods for treatment of type III fractures of the distal phalanx in horses. Equine Vet Ed. 2020;32(S10):42–51. doi:10.1111/eve.13232
- 41.↑
de Preux M, Klopfenstein Bregger MD, Brünisholz HP, Van der Vekens E, Schweizer-Gorgas D, Koch C. Clinical use of computer-assisted orthopedic surgery in horses. Vet Surg. 2020;49(6):1075–1087. doi:10.1111/vsu.13486
- 42.↑
Perkins JD, Meighan H, Windley Z, Troester S, Piercy R, Schumacher J. In vitro effect of ventriculocordectomy before laryngoplasty on abduction of the equine arytenoid cartilage. Vet Surg. 2011;40(3):305–310. doi:10.1111/j.1532-950X.2011.00796.x
- 43.↑
Dahlberg JA, Valdes-Martinez A, Boston RC, Parente EJ. Analysis of conformational variations of the cricoid cartilages in Thoroughbred horses using computed tomography. Equine Vet J. 2011;43(2):229–234. doi:10.1111/j.2042-3306.2010.00070.x
- 44.↑
Ahern BJ, Lukas E, Lam K, et al. Evaluation of a prototype dynamic laryngoplasty system in vitro with an equine vacuum airflow system. Vet Surg. 2019;48(2):173–179. doi:10.1111/vsu.13137
- 45.
Tucker ML, Sumner D, Reinink SK, Wilson DG, Carmalt JL. Ex vivo evaluation of arytenoid corniculectomy, compared with three other airway interventions, performed on cadaveric equine larynges with simulated recurrent laryngeal neuropathy. Am J Vet Res. 2019;80(12):1136–1143. doi:10.2460/ajvr.80.12.1136
- 46.
Bischofberger AS, Wereszka MM, Hadidane I, Perkins NR, Jeffcott LB, Dart AJ. Optimal tension, position, and number of prostheses required for maximum rima glottidis area after laryngoplasty. Vet Surg. 2013;42(3):280–285. doi:10.1111/j.1532-950X.2013.01103.x
- 47.↑
Secor EJ, Gutierrez-Nibeyro SD, Horn GP. Biomechanical evaluation of modified laryngoplasty by use of a toggle technique for stabilization of arytenoid cartilage in specimens obtained from equine cadavers. Am J Vet Res. 2018;79(2):226–232. doi:10.2460/ajvr.79.2.226
- 48.↑
Brandenberger O, Rossignol F, Perkins JD, et al. Ex vivo biomechanical stability of 5 cricoid-suture constructs for equine laryngoplasty. Vet Surg. 2017;46(5):705–713. doi:10.1111/vsu.12671
- 49.↑
Passman SN, Cheetham J, Bonassar LJ, Ducharme NG, Rawlinson JJ. Biomechanical characterisation of equine laryngeal cartilage. Equine Vet J. 2011;43(5):592–598. doi:10.1111/j.2042-3306.2010.00315.x
- 50.↑
Perkins JD, Raffetto J, Thompson C, Weller R, Piercy RJ, Pfau T. Three-dimensional biomechanics of simulated laryngeal abduction in horses. Am J Vet Res. 2010;71(9):1003–1010. doi:10.2460/ajvr.71.9.1003
