Abstract
Objective
To evaluate the agreement between radiographic and tracheobronchoscopic assessments of grade IV tracheal collapse (TC) in dogs and characterize the prevalence of the radiographic axial-rotation pattern.
Methods
This retrospective study included dogs diagnosed with grade IV TC via tracheobronchoscopy from 2021 through 2024. Inspiratory and expiratory right-lateral thoracic radiographs and tracheobronchoscopy images were reviewed. Collapse at 4 sites (midcervical, thoracic inlet, midthoracic, and carina) was categorized as mild (≤ 50%) or severe (> 50%). The most severe radiographic observation per site, the dominant radiographic finding, was compared with tracheobronchoscopic grade by weighted κ statistics and percentages of agreement.
Results
78 dogs had inspiratory radiographs, and 68 also had expiratory radiographs. Radiography identified severe collapse in 68 of 78 dogs (87.2%) at the thoracic inlet but less frequently at midcervical (38 of 78 [48.7%]), midthoracic (9 of 78 [11.5%]), and carinal sites (9 of 78 [11.5%]). Agreement was fair for the midthoracic trachea (weighted κ = 0.24) and slight to poor elsewhere. The percentage of agreement for severe versus nonsevere collapse was 87.2% (68 of 78) at the thoracic inlet, followed by midthoracic (76.9% [60 of 78]), carina (66.7% [52 of 78]), and midcervical (47.4% [37 of 78]) sites. Mild or no radiographic collapse was observed in 11 of 78 cases (14.1%) despite grade IV collapse on tracheobronchoscopy. Radiographic pattern, known as TC with axial-rotation (Rad-AR) confined to the thoracic inlet was present in 17 of 78 cases (21.8%).
Conclusions
Thoracic radiographs often underestimate the grade IV TC and cannot exclude disease when collapse is not visible.
Clinical Relevance
Radiography is a useful screening tool for localizing severe TC when collapse is evident, but tracheobronchoscopy remains essential for definitive staging and therapeutic planning.
Tracheal collapse (TC) is a prevalent cause of chronic cough in older small-breed dogs, characterized by a reduction in tracheal lumen diameter and a loss of the normal round-to-oval shape of the trachea.1 The condition is primarily caused by the degeneration and weakening of tracheal and bronchial cartilage, leading to dorsoventral narrowing of the tracheal lumen.1–3 The severity of TC is commonly assessed using the Tangner and Hobson grading system, which classifies the condition into 4 grades based on the extent of tracheal lumen narrowing.4 Accurate grading is critical for determining the location and severity of airway collapse, monitoring disease progression, and selecting appropriate therapeutic interventions.1 Advanced cases, particularly grade IV, can result in severe clinical signs, such as respiratory distress, cyanosis, and collapse, often necessitating surgical intervention.1
Various diagnostic modalities are available for assessing TC, including radiography,5 fluoroscopy,6,7 ultrasonography,8 CT,9 and tracheoscopy.1 Of these, tracheoscopy is considered the gold standard for assessing the grade and extent of collapse,1,4 although it requires general anesthesia, which can be risky in dogs with compromised respiratory function.
Radiography is widely used as a screening tool due to its availability, affordability, and lack of the need for anesthesia. Despite its limitations, such as static imaging and 2-D representation, radiography continues to play a practical role in clinical assessment. Previous studies have consistently shown that radiography underestimates the degree of TC compared to dynamic imaging techniques.7,10 In particular, radiographic assessment has shown the greatest sensitivity and predictive value in the cervical and thoracic inlet regions but tends to miss or underestimate collapse in the thoracic and carinal regions.7 Furthermore, while moderate agreement between radiography and tracheobronchoscopy at the trachea has been documented (κ = 0.49), false positives and discordant findings are common, highlighting its diagnostic limitations.10
Despite this known underestimation and inaccuracy, the current literature has several limitations that support further investigation. Several studies have used only inspiratory radiographs10,11; one evaluated both inspiratory and expiratory radiographs in only a small number of dogs,6 and another assessed both views but reported only combined results without specific data for each phase.7 Consequently, the evidence remains lacking for interpreting dynamic changes. Few studies have described or quantified the radiographic appearance of TC in different anatomical regions of trachea during different respiratory phases, and these comparisons have typically been made to fluoroscopy rather than to tracheobronchoscopy.6,7
In addition, a distinct radiographic pattern, known as TC with axial rotation (Rad-AR), characterized by dorsoventral widening of the tracheal lumen and increased luminal opacity obscuring the tracheal margins, has been reported in some dogs with severe TC.9 Previous studies6,7,10,11 evaluating the diagnostic accuracy of radiography for TC have often excluded cases exhibiting this radiographic finding. Therefore, the clinical relevance and diagnostic value of this finding remain unclear.
Thus, this study aims to assess the agreement between radiographic and tracheobronchoscopic assessments of severe TC, such as grade IV, in dogs using both inspiratory and expiratory views. This study also includes cases with Rad-AR, a subgroup typically excluded from previous analyses. We hypothesize that radiography serves as a valuable screening tool for localizing TC in dogs when collapse is radiographically visible, although it may fail to detect all cases. Consequently, dogs with normal radiographic findings may still require further evaluation via tracheobronchoscopy. Additionally, we hypothesize that Rad-AR is a relatively common finding in dogs with grade IV TC.
Methods
Case selection
A retrospective study was conducted using medical records from AMC Suematsu Animal Hospital, spanning from 2021 through 2024. Dogs diagnosed with grade IV TC via tracheobronchoscopy were included. Inclusion criteria required the presence of an inspiratory radiograph in right lateral recumbency obtained during the same presentation as the tracheobronchoscopy. Expiratory radiographs were also evaluated when available. Cases were excluded if radiographs or video recordings of the tracheobronchoscopy were unavailable, radiographs were of poor quality, or the entire trachea and carina were not included in the radiographic collimation. An American College of Veterinary Radiology board-certified radiologist (MM) and a veterinarian with 15 years of experience in canine respiratory surgery (MS) determined case eligibility. Breed and age data were recorded for each included case.
Tracheobronchoscopic evaluation
All tracheobronchoscopies were performed and recorded by a veterinarian (MS) with 15 years of experience in canine respiratory surgery. All tracheobronchoscopies were performed under general anesthesia with propofol (Propoflo 28). Propofol was administered IV through a catheter at dosages tailored to achieve patient immobilization while preserving spontaneous respiration, with total doses not exceeding 5 mg/kg body weight. This protocol was chosen specifically to minimize airway manipulation and maintain physiological tracheal dynamics during evaluation. The same veterinarian retrospectively reviewed the tracheobronchoscopic images to assess the presence and degree of TC at 4 anatomical levels: the midcervical trachea, thoracic inlet, midthoracic trachea, and carina. Tracheal collapse was graded using the Tangner and Hobson grading system4 when narrowing was present at each site. To facilitate statistical comparison between tracheoscopic findings and radiographic appearance, TC was further categorized as mild (grade I and II, ≤ 50% decrease in tracheal diameter) or severe (grade III and IV, > 50% decrease). This classification was necessary as radiographic evaluation does not reliably differentiate among all 4 grades.
Radiographic evaluation
Two American College of Veterinary Radiology board-certified radiologists (CF and MM) performed consensus radiographic evaluations of the 4 anatomical levels—the midcervical trachea, thoracic inlet, midthoracic trachea, and carina—using DICOM viewing software (Horos Project) while blinded to the bronchoscopy results.
Radiographic assessment cannot reliably distinguish between all 4 grades of the Tangner-Hobson classification4 used in tracheobronchoscopic evaluation. Therefore, we categorized tracheal narrowing as either mild (≤ 50% decrease in dorsoventral tracheal diameter) or severe (> 50% decrease in dorsoventral tracheal diameter) for radiographic evaluation as this distinction is more feasible on radiographs. Cases were classified as severe when there was a dorsoventral widening of the tracheal diameter or effacement of the tracheal margin with increased soft-tissue opacity within the tracheal lumen on the radiographs; these were also recorded as Rad-AR. The presence of a superimposed soft-tissue opaque band on the dorsal aspect of the cervical trachea was not considered TC but was also recorded. The frequency (percentage) of each radiographic finding (mild or severe collapse) and the frequency of Rad-AR and soft-tissue opaque band at each anatomical site were calculated.
Radiographic assessment was conducted in 2 sequential steps. First, inspiratory and expiratory radiographs were independently analyzed to determine the presence and severity of TC at each anatomical site. Second, findings from both respiratory phases were compared to identify discrepancies in the detection of severe TC, and the dominant radiographic finding (DRF) was recorded. To enhance diagnostic accuracy, the presence of TC in at least 1 respiratory phase was considered indicative of disease. Consequently, the most severe manifestation between the 2 phases was designated as the DRF, which was recorded and used for statistical analysis.
Statistical analysis
A formal power or sample size analysis was not performed due to the retrospective nature of the study. The sample size was determined by the availability of cases meeting the inclusion criteria, which is a limitation of this study. We will acknowledge this in the Discussion section.
The agreement between radiographic and tracheobronchoscopic findings in the midcervical, thoracic inlet, midthoracic, and carina regions was evaluated using the Cohen κ and weighted κ coefficients. Each region was assessed using both diagnostic methods, and the degree of collapse was categorized as normal/absent, mild, or severe. Contingency tables were created to display the number of cases in each category for both methods in each region. The Cohen κ coefficient was calculated to determine the agreement between the 2 methods, considering the possibility of chance agreement. Weighted κ coefficients were computed using linear weights to account for the degree of disagreement between categories. All statistical calculations were based on the DRF to ensure consistency in evaluating the most severe radiographic manifestation of TC. Additionally, the percentage of agreement in categorizing TC severity (absent/normal to mild vs severe) between radiography and tracheobronchoscopy was calculated for each anatomical site. Exact 95% CIs for proportions were calculated with the Wilson score method.
All statistical analyses were performed using R software (version 4.4.3; R Foundation for Statistical Computing).
Results
Study population
Seventy-eight dogs were included in the study, with body weights ranging from 1.2 to 11.5 kg (mean ± SD, 4.2 ± 2.2 kg) and ages ranging from 1.0 to 16 years (8.3 ± 3.6 years). The sample included various breeds: Pomeranian (n = 19), Chihuahua (n = 18), Yorkshire Terrier (n = 15), Toy Poodle (n = 10), Shiba (n = 8), mixed-breed (n = 5), Shih Tzu (n = 2), and Jack Russell Terrier (n = 1).
Tracheobronchoscopic evaluation
Among the 78 dogs diagnosed with grade IV TC by tracheobronchoscopy, 74 dogs had a characteristic “goose-honk” cough, and 23 dogs were experiencing respiratory distress. Although tracheobronchoscopy identified varying degrees of TC at all sites (the midcervical trachea, thoracic inlet, midthoracic trachea, and carina), grade IV TC was observed at the thoracic inlet in all dogs. The number of cases for each grade of TC identified by tracheobronchoscopy at each site is listed in Table 1. The number of cases tracheobronchoscopically diagnosed as severe TC including grade III or IV at each location was as follows: 61 of 78 (midcervical [78.2%]), 78 of 78 (thoracic inlet [100%]), 22 of 78 (midthoracic [26.9%]), and 22 of 78 (carina [28.2%]).
Tracheobronchoscopic grading of tracheal collapse (TC) at different anatomical sites in 78 dogs diagnosed with grade IV TC described in Figure 1 (2021 through 2024).
| Grade | Midcervical trachea (%; 95% CI) | Thoracic inlet trachea (%; 95% CI) | Midthoracic trachea (%; 95% CI) | Carina (%; 95% CI) |
|---|---|---|---|---|
| Absent/normal | 0 (0%; 0–4.7%) | 0 (0%; 0–4.7%) | 12 (15.4%; 9.0–25.0%) | 20 (25.6%; 17.3–36.3%) |
| I | 1 (1.3%; 0.2–6.9%) | 0 (0%; 0–4.7%) | 18 (23.1%; 15.1–33.6%) | 22 (28.2%; 19.4–39.0%) |
| II | 16 (20.5%; 13.0–30.8%) | 0 (0%; 0–4.7%) | 27 (34.6%; 25.0–45.7%) | 14 (17.9%; 11.0–27.9%) |
| III | 47 (60.3%; 49.2–70.4%) | 0 (0%; 0–4.7%) | 18 (23.1%; 15.1–33.6%) | 17 (21.8%; 14.1–32.2%) |
| IV | 14 (17.9%; 11.0–27.9%) | 78 (100%; 95.3–100%) | 3 (3.8%; 1.3–10.7%) | 5 (6.4%; 2.8–14.1%) |
The number of cases and corresponding percentages are reported for each grade of TC (Tangner-Hobson grades I–IV) as well as for dogs with a normal trachea at each site. Percentages are followed by exact 95% CIs (Wilson score method).
Radiographic evaluation
All 78 dogs underwent inspiratory radiography, whereas expiratory radiographs were available for 68 dogs. Independent evaluations of inspiratory and expiratory radiographs were conducted separately. On inspiratory radiographs in 78 dogs, severe TC was detected at each location, with the number of cases diagnosed as follows (Figure 1; Table 2): 32 of 78 (midcervical [41.0%]), 67 of 78 (thoracic inlet [85.9%]), 6 of 78 (midthoracic [7.7%]), and 4 of 78 (carina [5.1%]). On expiratory radiographs in 68 dogs, severe TC was detected at each location, with the number of cases diagnosed as follows: 19 of 68 (27.9%) at the midcervical trachea, 53 of 68 (77.9%) at the thoracic inlet, 6 of 68 (8.8%) at the midthoracic region, and 7 of 68 (10.3%) at the carina.
Radiographic findings of grade IV tracheal collapse (TC) in dogs. Right-lateral thoracic radiographs were obtained from 2021 through 2024 from 78 client-owned dogs with grade IV TC confirmed by tracheobronchoscopy. A—Severe TC involving the caudal cervical region and thoracic inlet in a dog. B—Severe TC with axial rotation at the thoracic inlet, characterized by dorsoventral widening of the trachea and increased radiopacity of the tracheal lumen. C—A soft-tissue opaque band superimposed on the dorsal half of the caudal cervical trachea, representing a redundant dorsal tracheal membrane in a dog.
Citation: American Journal of Veterinary Research 2025; 10.2460/ajvr.25.04.0150
Dominant radiographic findings of grade IV TC confirmed by tracheobronchoscopy and associated findings in 78 dogs described in Figure 1.
| Grade | Midcervical trachea (%; 95% CI) | Thoracic inlet trachea (%; 95% CI) | Midthoracic trachea (%; 95% CI) | Carina (%; 95% CI) |
|---|---|---|---|---|
| Absent/normal | 14 (17.9%; 11.0–27.9%) | 4 (5.1%; 2.0–12.5%) | 30 (38.5%; 28.4–49.6%) | 58 (74.4%; 63.7–82.7%) |
| Mild | 26 (33.3%; 23.9–44.4%) | 6 (7.7%; 3.6–15.8%) | 39 (50.0%; 39.2–60.8%) | 11 (14.1%; 8.1–23.5%) |
| Severe | 38 (48.7%; 37.9–59.6%) | 51 (65.4%; 54.3–75.0%) | 9 (11.5%; 6.2–20.5%) | 9 (11.5%; 6.2–20.5%) |
Tracheal collapse was classified as absent/normal, mild, or severe based on radiographic assessment. Percentages are followed by exact 95% CIs (Wilson score method).
Radiographic pattern, known as TC with axial rotation was observed only at the thoracic inlet in 12 of 78 dogs (15.4%) on inspiratory radiographs and in 13 of 68 dogs (19.1%) on expiratory radiographs (Figure 1). A superimposed soft-tissue opaque band at the cervical trachea was noted in 33 of 78 dogs (42.3%) in inspiratory radiographs and in 27 of 68 dogs (39.7%) on expiratory radiographs.
Comparison between inspiratory and expiratory radiographs revealed inconsistencies in the detection of severe TC in 19 of 68 cases (27.9%) at the midcervical trachea, 8 of 68 cases (11.8%) at the thoracic inlet, 6 of 68 cases (8.8%) at the midthoracic trachea, and 6 of 68 cases (8.8%) at the carina. Expiratory radiographs identified severe collapse that was not evident on inspiratory radiographs in 6 of 68 cases (8.8%) at the midcervical trachea, 1 of 68 cases (1.5%) at the thoracic inlet, 3 of 68 cases (4.4%) at the midthoracic trachea, and 5 of 68 cases (7.4%) at the carina. Conversely, inspiratory radiographs detected severe collapse that was absent on expiratory radiographs in 13 of 68 cases (19.1%) at the midcervical trachea, 7 of 68 cases (10.3%) at the thoracic inlet, 3 of 68 cases (4.4%) at the midthoracic trachea, and 1 of 68 cases (1.5%) at the carina.
All further results are based on the DRF, which incorporates the most severe collapse observed on either inspiratory or expiratory radiographs. Using DRF, severe TC was identified at the midcervical trachea in 38 of 78 cases (48.7%), at the thoracic inlet in 68 of 78 cases (87.2%), at the midthoracic region in 9 of 78 cases (11.5%), and at the carina in 9 of 78 cases (11.5%). Radiographic pattern, known as TC with axial rotation was present in 17 of 78 cases (21.8%) at the thoracic inlet. A superimposed soft-tissue opaque band at the cervical trachea was noted in 38 of 78 cases (48.7%).
Agreement between radiography and tracheobronchoscopy
Contingency tables displaying the number of cases in each category for each region are presented in Supplementary Table S1. Cohen κ coefficients for the midcervical, thoracic inlet, midthoracic, and carina regions were −0.03, 0, 0.17, and 0.03, respectively, indicating negligible to slight agreement between radiography and tracheobronchoscopy in all regions except the midthoracic region, which approached fair agreement (κ = 0.17; P = .039). Weighted κ coefficients for the midcervical, thoracic inlet, midthoracic, and carina regions were −0.03, 0, 0.24, and 0.08, respectively. Applying severity weighting improved agreement only in the midthoracic region, resulting in a statistically significant fair agreement (weighted κ = 0.24; P = .002). Due to the uniform severity observed in the thoracic inlet region during tracheobronchoscopy, statistical assessment using weighted κ was deemed inappropriate for this region. Instead, percentages of agreement (Table 3) were considered more suitable for evaluating radiographic capability in detecting severe TC at the thoracic inlet.
The percentage of agreement in TC severity between radiography and tracheobronchoscopy at 4 anatomical locations in dogs with grade IV TC confirmed by tracheobronchoscopy described in Figure 1.
| Midcervical trachea (%; 95% CI) | Thoracic inlet trachea (%; 95% CI) | Midthoracic trachea (%; 95% CI) | Carina (%; 95% CI) | |
|---|---|---|---|---|
| Agreement | 37 (47.4%; 36.7–58.4%) | 68 (87.2%; 78.0–92.9%) | 60 (76.9%; 66.4–84.9%) | 52 (66.7%; 55.6–76.1%) |
| Both severe | 29 (37.2%; 27.3–48.3%) | 68 (87.2%; 78.0–92.9%) | 6 (7.7%; 3.6–15.8%) | 1 (1.3%; 0.2–6.9%) |
| Both absent/normal to mild | 8 (10.3%; 5.3–19.0%) | 0 (0%; 0–4.7%) | 54 (69.2%; 58.3–78.4%) | 51 (65.4%; 54.3–75.0%) |
| Radiography (severe), tracheobronchoscopy (absent/normal to mild) | 9 (11.5%; 6.2–20.5%) | 0 (0%; 0–4.7%) | 3 (3.8%; 1.3–10.7%) | 5 (6.4%; 2.8–14.1%) |
| Radiography (absent/normal to mild), tracheobronchoscopy (severe) | 32 (41.0%; 30.8–52.1%) | 10 (12.8%; 7.1–22.0%) | 15 (19.2%; 12.0–29.3%) | 21 (26.9%; 18.3–37.7%) |
On tracheobronchoscopy, grades I and II were classified as mild, and grades III and IV were classified as severe. On radiography, the severe category included radiographic pattern, known as TC with axial rotation. Agreement and discrepancies are presented for the midcervical trachea, thoracic inlet, midthoracic trachea, and carina. Percentages are followed by exact 95% CIs (Wilson score method).
Apart from the thoracic inlet, the most pronounced discrepancies were observed in the midcervical region, where κ values were negative, indicating poor agreement. These findings suggest that radiographic evaluation of TC is particularly unreliable in the midcervical and carina regions, whereas reliability is comparatively better in the midthoracic region. Due to statistical limitations, the reliability of radiographic assessment for the thoracic inlet region could not be determined using κ statistics.
The percentage of agreement of TC severity (none/mild vs severe) between radiography and tracheobronchoscopy (Figure 2) was 47.4% (37 of 78) in the midcervical region, 87.2% (68 of 78) in the thoracic inlet region, 76.9% (60 of 78) in the midthoracic region, and 66.7% (52/78) in the carina region (Table 3). Eight dogs had grade IV TC on tracheobronchoscopy but no radiographic evidence of collapse (absent/normal) at any location. Additionally, 3 dogs had grade IV TC on tracheobronchoscopy but were considered mild on radiographs.
Comparison of tracheoscopy and radiography images in dogs with grade IV TC described in Figure 1. Images were obtained from dogs with grade IV confirmed by tracheobronchoscopy. Tracheoscopy images of the thoracic inlet (A and C) are paired with right-laterolateral thoracic radiographs (B and D). A and B—Agreement in one dog, where both tracheoscopic (A) and radiographic (B) evaluations confirm severe TC at the thoracic inlet. C and D—Discrepancy in another dog, where tracheoscopy (C) shows grade IV TC, but the corresponding radiograph (D) shows no evidence of collapse.
Citation: American Journal of Veterinary Research 2025; 10.2460/ajvr.25.04.0150
Discussion
We hypothesized that (1) thoracic radiographs would accurately localize severe TC when collapse is visible yet fail to detect all grade IV cases and (2) the Rad-AR would be a relatively common radiographic finding in dogs with grade IV TC. Our data support both hypotheses. Radiography correctly identified severe collapse in most dogs at the thoracic inlet (percentage of agreement = 87.2% [68 of 78]) but underestimated disease at the midcervical, midthoracic, and carinal sites, confirming its value as a screening tool rather than a definitive test. In addition, Rad-AR was documented at the thoracic inlet in 17 of 78 cases (21.8%), demonstrating that this pattern is relatively common and should be recognized as part of the radiographic spectrum of severe TC. Consequently, normal or equivocal thoracic radiographs cannot exclude grade IV TC, and tracheobronchoscopy remains necessary for definitive staging and treatment planning.
Radiography is a widely used imaging modality for evaluating TC in dogs; however, its diagnostic accuracy is limited, particularly for detecting dynamic airway collapse and assessing the severity of tracheal abnormalities. Radiography often underestimates both the frequency and severity of collapse compared to more advanced imaging modalities, such as fluoroscopy, CT, and bronchoscopy, with reported sensitivities ranging from 60% to over 90%.4,6,7,10 This discrepancy is particularly evident in the carina, where radiography frequently fails to identify collapse.6,7,10 Furthermore, static radiographic images may not capture the dynamic nature of TC, leading to false negative results.12,13 Radiographic assessment is also less reliable for diagnosing bronchial collapse, which often coexists with TC and contributes to clinical signs.11 Despite these limitations, radiography remains a useful screening tool for identifying dogs with potential TC.6,7 Its high agreement (68 of 78 at the thoracic inlet [87.2%]) in identifying severe collapse, as observed in the present study, highlights its utility in detecting grade IV TC. However, the considerable rates of underdiagnosis at different tracheal regions (32 of 78 at the midcervical region [41.0%], 10 of 78 at the thoracic inlet [12.8%], 15 of 78 at the midthoracic region [19.2%], and 21 of 78 at the carina [26.9%]) indicate its inability to reliably exclude TC, even in severe cases. In the present study, a proportion of dogs with grade IV TC confirmed on tracheobronchoscopy exhibited absent/normal (8 of 78 [10.3%]) or only mild collapse (3 of 78 [3.8%]) on radiographs, emphasizing the necessity of tracheobronchoscopy for a definitive diagnosis. These findings are consistent with previous literature highlighting the limitations of radiography in detecting milder forms of collapse and its tendency to underestimate tracheal diameter compared to CT, potentially impacting clinical decisions, such as stent sizing. Furthermore, thoracic radiographs remain valuable for evaluating concurrent conditions, such as cardiac disease or pulmonary abnormalities, which may mimic or exacerbate the clinical signs of TC.14 Although other imaging modalities, such as fluoroscopy, offer superior diagnostic accuracy by capturing dynamic airway collapse in real time, radiography continues to play an important role as an accessible and noninvasive screening modality, particularly in general practice.
Tracheal collapse in dogs most commonly and severely affects the caudal cervical trachea and the thoracic inlet region,1,4,6,15–18 where the tracheal cartilage is anatomically thinner and more prone to collapse due to external compressive forces.16 These regions are consistently identified as the most affected sites in numerous studies, with structural vulnerability and external pressures from surrounding tissues contributing to their predisposition. The thoracic inlet is especially susceptible to collapse during inspiration, the phase typically captured in static radiographic images, making it appear as the most frequently affected site in radiographic evaluations.1,6,19 However, it is suggested that the intrathoracic trachea, particularly the carina, may also be a significant site of collapse as dynamic imaging modalities, such as fluoroscopy, better capture the expiratory and coughing maneuvers that exacerbate collapse in this region.1,6,19–21 In the current study, severe collapse at the thoracic inlet was universally observed in all dogs with grade IV TC on tracheobronchoscopy, and radiographic detection was most accurate at this site. The agreement between radiography and tracheobronchoscopy was strongest at the thoracic inlet, reflected by high Cohen κ and weighted κ coefficients. Conversely, discrepancies were more pronounced in the midcervical and midthoracic trachea, where radiography frequently underestimated the severity of collapse. At the carina, radiographic assessment also underestimated the degree of collapse compared to tracheobronchoscopy, consistent with previous findings highlighting the limitations of static imaging in this region.6,7,10 These results emphasize the importance of utilizing dynamic imaging techniques, such as fluoroscopy, to accurately identify and localize TC in dogs, particularly in regions less reliably assessed by radiographs.
Anecdotally, cervical TC in dogs has been considered more prominent during inspiration, whereas collapse of the intrathoracic trachea and carina occurs predominantly during expiration. However, limited scientific data specifically address this concept.6,10 One study6 evaluated the consistency of inspiratory and expiratory radiographs in detecting TC at different anatomical locations in dogs. This study compared radiographic detection of collapse in the cervical, thoracic inlet, thoracic, and carinal regions between inspiratory and expiratory views. The authors found that reviewing both phases of respiration improved accuracy only marginally, and agreement between inspiratory and expiratory radiographs in identifying the same collapse location was relatively low (62.5%).6 However, their findings were based on a small sample size of only 8 dogs with paired radiographs. Another study10 evaluated both inspiratory and expiratory radiographs but did not specifically compare tracheal segments between these phases. Instead, it focused on comparing radiography with fluoroscopy and bronchoscopy. The present study of 68 dogs supports anecdotal observations by confirming that severe collapse occurs more frequently in the cervical trachea during inspiration and in the intrathoracic segments during expiration. While our results indicate that the diagnosis of TC is generally consistent between inspiratory and expiratory radiographs, there was a substantial discrepancy, particularly in the middle cervical trachea (19 of 68 [27.9%]). This finding supports the clinical importance of evaluating both respiratory phases to increase diagnostic confidence. It also suggests that the most severe collapse, whether identified on inspiratory or expiratory views, should be considered as the clinical and analytical focus for decision making.
Axial rotation of the trachea can occur in dogs with TC, creating diagnostic challenges and influencing treatment decisions.9 In this condition, the trachea rotates along its longitudinal axis, resulting in changes in tracheal morphology and radiographic appearance. On radiographs, axial rotation can manifest as an apparent increase in the dorsoventral luminal dimension, which may be misinterpreted as a normal finding or mistaken for intraluminal abnormalities, such as soft-tissue masses or foreign bodies.9 Interestingly, physiologic axial displacement has been reported in nearly half of asymptomatic dogs on CT imaging, suggesting that axial rotation is not always pathologic and may represent a natural variation in tracheal alignment in some cases.22 In addition, Rad-AR can be affected by patient positioning and may not be reproducible in different recumbent positions. Therefore, Rad-AR should be considered a radiographic observation rather than a definitive confirmation of tracheal rotation. In the present retrospective study, true axial rotation could not be fully assessed by tracheobronchoscopy. However, we report Rad-AR because it has often been excluded in previous studies, but in the present study, Rad-AR was observed exclusively at the thoracic inlet, with a prevalence of 21.8% (17 of 78) among dogs with grade IV TC. This finding supports the importance of reporting this finding and including these findings as TC in dogs to accurately evaluate radiographic diagnosis of TC in dogs. Although this was beyond the focus of the current study, further prospective research utilizing CT or tracheoscopy is necessary to clarify whether Rad-AR reflects a consistent anatomic variant, an intermittent artifact, or a clinically significant form of tracheal rotation.
A redundant dorsal tracheal membrane, or trachealis muscle, is radiographically observed as a soft-tissue opaque band superimposed on the dorsal aspect of the tracheal lumen on lateral radiographs.23 This radiographic appearance can occur in dogs without evidence of TC, although redundancy of the dorsal tracheal membrane is frequently a consequence of TC. As the tracheal rings flatten during TC, the dorsal tracheal membrane stretches and protrudes into the tracheal lumen.4,23 Additionally, pathological thickening of the dorsal tracheal membrane or trachealis muscle, such as from hematoma or abscess, can independently narrow the tracheal lumen.24,25 Interpretation of this finding is further complicated by factors such as respiratory phase (inspiratory vs expiratory) and the anatomical proximity of the esophagus, which can overlap the trachea on radiographs.23 Understanding the interplay between these conditions is essential for accurate diagnosis as distinguishing among them can significantly influence clinical interpretation and treatment strategies. In this study, a soft-tissue opaque band was identified on radiographs of the dorsal cervical trachea in 38 of 77 cases (48.7%), likely representing either a redundant dorsal tracheal membrane or superimposed esophagus. While this finding is often discussed in the context of TC, its precise clinical significance remains poorly understood. Its potential relationship to disease severity and progression, as well as its impact on airflow limitation, warrants further investigation. These findings underscore the importance of careful radiographic interpretation; the consideration of anatomical factors, such as esophageal superimposition; and the use of complementary imaging modalities, like fluoroscopy, to enhance diagnostic accuracy.
The primary limitations of this study include its retrospective design and the relatively small sample size, which may limit the generalizability of the findings. The exclusive focus on grade IV TC may have skewed the agreement metrics as severe cases are inherently easier to identify radiographically. Future studies should include a broader range of collapse severities and larger populations to confirm and extend these findings. Another limitation was the inability to perform a direct comparison between fluoroscopy and tracheobronchoscopy. Fluoroscopy, as a dynamic imaging modality, provides superior real-time visualization of airway dynamics compared to static radiographs. However, our study’s primary objective was to assess the frequency of radiographic features to evaluate diagnostic accuracy due to its wide availability rather than to determine the optimal imaging modality for TC. Additionally, institutional protocols typically favored direct progression from radiography to tracheobronchoscopy; thus, fluoroscopic data were unavailable for comparison. Future prospective studies should include fluoroscopy to further clarify its diagnostic role relative to radiography and tracheobronchoscopy. We also acknowledge the potential influence of anesthetic agents on tracheal membrane laxity during tracheobronchoscopy. In this study, tracheobronchoscopy was performed under sedation with propofol, administered IV via catheter. The dosage was adjusted to maintain patient immobilization while preserving spontaneous respiration. This protocol was specifically chosen to minimize airway manipulation and maintain physiological tracheal dynamics.
This study reaffirms that radiography is a valuable initial screening tool for detecting severe TC but has notable limitations in assessing the condition’s severity and extent, particularly in the midcervical and midthoracic regions. Tracheobronchoscopy remains the gold standard for diagnosis, providing direct visualization and enabling accurate grading. Clinicians should interpret radiographic findings cautiously and consider a multimodal diagnostic approach, combining radiography with tracheobronchoscopy, to ensure precise diagnosis and optimal treatment planning for dogs with TC.
Supplementary Materials
Supplementary materials are posted online at the journal website: avmajournals.avma.org.
Acknowledgments
The authors thank the clinicians and veterinary technicians of AMC Suematsu Animal Hospital for their assistance with image acquisition and data retrieval.
Disclosures
Generative AI, ChatGPT, was used solely for language refinement and writing assistance. No generative AI was used for data analysis, interpretation, or any aspect that could influence the results or conclusions of this research.
Funding
The authors have nothing to disclose.
References
- 1.↑
Johnson L. Tracheal collapse: diagnosis and medical and surgical treatment. Vet Clin North Am Small Anim Pract. 2000;30(6):1253–1266. doi:10.1016/s0195-5616(00)06005-8
- 2.
Dallman MJ, McClure RC, Brown EM. Normal and collapsed trachea in the dog: scanning electron microscopy study. Am J Vet Res. 1985;46(10):2110–2115.
- 3.↑
Dallman MJ, McClure RC, Brown EM. Histochemical study of normal and collapsed tracheas in dogs. Am J Vet Res. 1988;49(12):2117–2125. doi:10.2460/ajvr.1988.49.12.2117
- 4.↑
Tangner CH, Hobson HP. A retrospective study of 20 surgically managed cases of collapsed trachea. Vet Surg. 1982;11(4):146–149. doi:10.1111/j.1532-950X.1982.tb00691.x
- 5.↑
Alexander K. Canine and feline larynx and trachea. In: Thrall DE, ed. Textbook of Veterinary Diagnostic Radiology. 7th ed. WB Saunders; 2018:583–595.
- 6.↑
Macready DM, Johnson LR, Pollard RE. Fluoroscopic and radiographic evaluation of tracheal collapse in dogs: 62 cases (2001-2006). J Am Vet Med Assoc. 2007;230(12):1870–1876. doi:10.2460/javma.230.12.1870
- 7.↑
Yangwanitset W, Huaijantug S, Tansakul M, Sakcamduang W. The agreement between radiography and fluoroscopy as diagnostic tools for tracheal collapse in dogs. Animals (Basel). 2023;13(9):1434. doi:10.3390/ani13091434
- 8.↑
Rudorf H, Herrtage ME, White RA. Use of ultrasonography in the diagnosis of tracheal collapse. J Small Anim Pract. 1997;38(11):513–518. doi:10.1111/j.1748-5827.1997.tb03308.x
- 9.↑
Heng HG, Lim CK, Gutierrez-Crespo B, Guptill LF. Radiographic and computed tomographic appearance of tracheal collapse with axial rotation in four dogs. J Small Anim Pract. 2018;59(1):53–58. doi:10.1111/jsap.12679
- 10.↑
Johnson LR, Singh MK, Pollard RE. Agreement among radiographs, fluoroscopy and bronchoscopy in documentation of airway collapse in dogs. J Vet Intern Med. 2015;29(6):1619–1626. doi:10.1111/jvim.13612
- 11.↑
Johnson LR, Pollard RE. Tracheal collapse and bronchomalacia in dogs: 58 cases (7/2001-1/2008). J Vet Intern Med. 2010;24(2):298–305. doi:10.1111/j.1939-1676.2009.0451.x
- 12.↑
Amis TC. Tracheal collapse in the dog. Aust Vet J. 1974;50(7):285–289. doi:10.1111/j.1751-0813.1974.tb05312.x
- 13.↑
White RAS, Williams JM. Tracheal collapse in the dog - is there really a role for surgery: a survey of 100 cases. J Small Anim Pract. 1994;35(4):191–196. doi:10.1111/j.1748-5827.1994.tb01685.x
- 14.↑
Tappin SW. Canine tracheal collapse. J Small Anim Pract. 2016;57(1):9–17. doi:10.1111/jsap.12436
- 15.↑
Done SH, Drew RA. Observations on the pathology of tracheal collapse in dogs. J Small Anim Pract. 1976;17(12):783–791. doi:10.1111/j.1748-5827.1976.tb06944.x
- 16.↑
Dabanoglu I, Ocal MK, Kara ME. A quantitative study on the trachea of the dog. Anat Histol Embryol. 2001;30(1):57–59. doi:10.1046/j.1439-0264.2001.00301.x
- 17.
Weisse C, Berent A, Violette N, McDougall R, Lamb K. Short-, intermediate-, and long-term results for endoluminal stent placement in dogs with tracheal collapse. J Am Vet Med Assoc. 2019;254(3):380–392. doi:10.2460/javma.254.3.380
- 18.↑
Kirby BM, Bjorling DE, Rankin JH, Phernetton TM. The effects of surgical isolation and application of polypropylene spiral prostheses on tracheal blood flow. Vet Surg. 1991;20(1):49–54. doi:10.1111/j.1532-950x.1991.tb00305.x
- 19.↑
Della Maggiore A. An update on tracheal and airway collapse in dogs. Vet Clin North Am Small Anim Pract. 2020;50(2):419–430. doi:10.1016/j.cvsm.2019.11.003
- 20.
Maggiore AD. Tracheal and airway collapse in dogs. Vet Clin North Am Small Anim Pract. 2014;44(1):117–127. doi:10.1016/j.cvsm.2013.09.004
- 21.↑
Mostafa AA, Berry CR. Radiographic determination of the tracheal indices at caudal cervical, thoracic-inlet, and intra-thoracic trachea in non-bulldog brachycephalic breeds without evidence of cardiorespiratory disease. BMC Vet Res. 2023;19(1):184. doi:10.1186/s12917-023-03730-0
- 22.↑
Lim S, Jeong J, Heng H, et al. Computed tomographic features of tracheal shapes and dimensions in awake dogs. Vet Med (Praha). 2018;63(3):131–136. doi:10.17221/124/2017-vetmed
- 23.↑
Lindl Bylicki BJ, Johnson LR, Pollard RE. Comparison of the radiographic and tracheoscopic appearance of the dorsal tracheal membrane in large and small breed dogs. Vet Radiol Ultrasound. 2015;56(6):602–608. doi:10.1111/vru.12276
- 24.↑
Lynch AM, O’Toole TE, Rush JE, Rondeau D, O’Neill K, Kudej RK. Acute tracheal compression in a large breed dog due to a dorsal tracheal membrane abscess. J Vet Emerg Crit Care. 2015;25(6):795–800. doi:10.1111/vec.12379
- 25.↑
Pink JJ. Intramural tracheal haematoma causing acute respiratory obstruction in a dog. J Small Anim Pract. 2006;47(3):161–164. doi:10.1111/j.1748-5827.2006.00011.x
