Respiratory disease in cattle represents a significant and widespread health concern within the livestock industry, adversely affecting animal welfare and economic productivity.1–3 Among the various respiratory conditions, interstitial pneumonia (IP) in cattle remains particularly challenging to diagnose early and treat successfully due in part to limited comprehensive descriptions available in the literature.4–8 Numerous potential causes, contributing factors, and even the basic pathologic processes underlying the development of IP in feedyard cattle remain poorly understood, and this syndrome is often clinically indistinguishable from other respiratory diseases.4,6,8–10 The 3 most common respiratory diagnoses in feedyard mortalities are bronchopneumonia (BP), BP with IP (BIP), and atypical IP (AIP).6,7,11 Combining AIP and BIP cases, total IP cases are reported to represent upward of 60% of feedyard mortalities.7 A more detailed understanding of IP and development of effective and early diagnostic tools are crucial for enhancing disease management and treatment strategies.9,12–14 The most reliable IP diagnostic method is histopathologic evaluation7,11,15; however, this is not feasible at the time of clinical diagnosis antemortem. Treatment decisions are made chuteside, so the inability to correctly differentiate types of pneumonias antemortem is a significant problem for which improved chuteside diagnostics are needed.
Differentiating IP from other respiratory illnesses is important due to the influence on treatment strategies and management decisions. Antimicrobial therapy is common in BP cases due to their infectious nature, whereas IP typically requires anti-inflammatory or supportive treatments to address the underlying pathogenic mechanisms.4 Therefore, the ability to accurately identify and differentiate these conditions using targeted thoracic point-of-care ultrasound (TT-POCUS) could lead to more targeted and effective therapeutic interventions, ultimately improving animal health outcomes and reducing mortality rates.
Point-of-care ultrasound has emerged as a valuable diagnostic modality in human and veterinary medicine, offering accessible and noninvasive real-time imaging capabilities.16–23 One limitation of point-of-care ultrasound applied to cattle respiratory disease is that the systematic thoracic evaluation method is time consuming, with previous literature reporting evaluations ranging from 7 to 45 minutes.19,24 This study proposes a TT-POCUS method specifically focusing on the right pulmonary caudodorsal region in cattle. The selected area for evaluation is commonly affected in cases of IP of feedyard cattle,6 so targeted evaluation of only 1 region of the pulmonary area limits the overall time necessary to complete the evaluation.
Overall, the objective of this study is to elucidate the potential associations between TT-POCUS imaging parameters observed at respiratory disease treatment and the subsequent presence of IP in cattle mortalities. By enhancing our understanding of IP and refining diagnostic techniques, this research aims to contribute to more effective disease management practices in feedyard operations.
Methods
This study was approved by the IACUC at Kansas State University (protocol #4897).
Experimental design and enrollment criteria
This experiment was designed as a cross-sectional observational study. Feedyard cattle respiratory morbidities were evaluated at the time of treatment using TT-POCUS by a trained veterinarian (LFBBF) from July through December 2023 at 1 feedyard located in the High Plains region of the US. Respiratory morbidity was defined by a treatment event for which an animal was pulled for either BP or AIP based on visual clinical signs (depression, nasal discharge, dyspnea, progressive weight loss) assessed by the feedyard personnel, including first pulls and relapsed treatments. Systematic autopsies were performed by trained technicians (MJ, LC, and LFBBF) supervised by a veterinarian (BJW) to collect pulmonary samples for histopathologic evaluation when mortality occurred in cases previously evaluated by TT-POCUS. Study enrollment required both TT-POCUS evaluation and a histopathologic diagnosis. Animals were not included in the study if mortality did not occur following TT-POCUS or if pulmonary samples could not be collected or were not evaluated histopathologically. Technicians were only present for necropsy on weekdays; therefore, all cattle that died following TT-POCUS may not have been included.
The categories of days on feed (DOF) were determined based on previous reported work that categorizes the temporal epidemiological characteristics of respiratory morbidities in feedyard cattle, where 0 to 42 days are considered early DOF, 43 to 71 mid-DOF, and > 71 late DOF.25 Individual animal body weight (BW) was also recorded at the time of treatment; animal BW was categorized as < 272, 272 to 362, 362 to 453, and ≥ 453 kg.
Targeted thoracic point-of-care ultrasound method
Commercial feedyard cattle were evaluated with TT-POCUS where animals were restrained in a hydraulic squeeze chute, and ultrasound lung scans were no longer than 60 seconds to maintain commercial feedyard work pace. Probe placement for TT-POCUS allowed evaluation of only the right caudodorsal pulmonary lobe assessing between the 8th and 11th rib (Figure 1). The same ultrasound probe unit was utilized for all ultrasonographic evaluations in this study (iQ + Vet; Butterfly Network Inc; 1 MHz to 10 MHz), a point-of-care specialized probe with a 5 X 3-cm footprint set on the optimized preset for lung evaluation. Presets configured the probe array to linear and allowed for up to 30-cm depth, and overall gain was set at 65%. Preparation of the scanning area was performed using 70% isopropyl alcohol without trimming/shaving the hair coat.
Right lateral view of adult bovine anatomy. The polygon traced in yellow represents the area scanned with the targeted thoracic point-of-care ultrasonography, 8th-to-11th intercostal space. Caudal to the lungs is the liver, displaying close proximity with the caudodorsal area of the right lung. Targeted thoracic point-of-care ultrasonography evaluations (n = 62) (ultrasound lung score, A-line count categories, B-line count categories) took place in the demarked area.
Citation: American Journal of Veterinary Research 86, 2; 10.2460/ajvr.24.08.0233
Scanning procedures involved probe placement placed parallel to the ribs, specifically starting at the 11th intercostal space (ICS) and moved from the dorsal aspect of the ICS (identified by visualization of the ventral part of longissimus dorsi), then sliding the probe ventrally until visualizing the liver, diaphragm, and caudal tip of the lung, also called as “curtain sign.” Due to animal variability, the “curtain sign” may not be visible at the 11th ICS, requiring movement of the probe cranially to the next ICS (10th). Each ICS from 11th to 8th was evaluated dorsally to ventrally similarly. Probe “fanning” was also performed, where the angle relative to the skin (side-to-side angle) is dynamically changed to facilitate the visualization of the pulmonary tissue between the bone structures (ribs).
Data collection from TT-POCUS evaluation was carried out in real-time and post hoc analysis by a trained veterinarian. Real-time variables collected included: ultrasound lung score (ULS), B-line count, A-line count, presence of pleural effusion, and presence of pleural abnormalities (moth sign [pleural indentations] and pleural thickening). A 1-to-5 scale was used for ULS, similar to previous research, with ULS scores defined as: 1, presence of < 3 thin B-lines (healthy lung); 2, presence of 3 or more thin B-lines; 3, presence of merged B-lines; 4, presence of several wide/merged B-lines and abnormal pleural (moth sign); and 5, consolidated lung (presence of consolidation with > 1 cm of depth) and might be accompanied with moth sign and effusion based on description of lung injury; consolidated lung is defined by an ultrasound image of parenchymal organs like, the liver, commonly called hepatization.22,23,26 The B-line count was based on a representative count where B-lines were defined as hyperechoic vertical artifacts resembling comet tails. B-line count was categorized (0 to 2, 3 to 5, and > 5) based on the theorized degree of severity, where more B-lines could translate to greater tissue injury. A-lines were defined as hyperechoic horizontal artifacts (reflections of the pleural line) and were categorized in 2 groups: 0 to 2 A-lines and ≥ 3 A-lines. These groups were based on the goal of evenly dividing the population. Both A-line counts and B-line counts were determined based on the lung field with the most abnormalities. The presence/absence of pleural effusion was determined by the presence of a hypoechoic/anechoic layer in pleural space, and pleural thickening/irregularity was recorded as present or not.
Other variables were generated post hoc using the software ImageJ, version 1.54 (US NIH); measurements were taken from a single frame from the 60-second clip, where the frame of choice was the frame with the most tissue abnormalities, allowing a comparison based on tissue injury within the caudodorsal region. The post hoc variables included image brightness (defined as pixel intensity density = average gray value X pixel count), average gray value (the mean intensity of pixels from the image histogram), and B-line area (cm2). Image brightness was measured using a 7-cm linear selection, starting at the pleural line and extending diagonally across the image, from top to bottom. The average gray value was also determined using the same 7-cm selection to evaluate the intensity units within the image. Total B-line area in cm2 was calculated by the use of the polygon selection tool, where all the B-lines in the image would be measured, and the sum of these areas would equate to the total affected B-line area in cm2, with the objective to account for potential differences in B-line width and merged B-lines. It is important to report that A-line evaluation has not been reported in the literature. On the other hand, B-line count has been reported extensively in the literature.27,28 B-line count is often categorized into normal (0 to 2) versus abnormal (≥ 3).23 In the present study, a category was created for B-line counts > 5 based on the theory of higher tissue injury with higher B-line counts.
Histopathology
Pulmonary tissue samples collected for histopathologic diagnosis were collected as a 1-cm2 tissue section from the right cranioventral lobe and right caudodorsal lung lobes. Samples were fixed by immersion in 10% neutral-buffered formalin, paraffin embedded, and then processed routinely by the histology laboratory at the Kansas State Veterinary Diagnostic Laboratory (KSVDL). Four-micrometer-thick tissue sections were cut to prepare slides prior to routine staining with H&E using a Leica ST5010 Autostainer XL (Leica Biosystems) and then digitized using a Leica Aperio AT2 (Leica Biosystems) digital slide scanner and evaluated by board-certified veterinary pathologists (AF and BLP) who were blinded to TT-POCUS evaluation and animal records. This study did not differentiate AIP cases from BIP anatomically; for example, lesions of IP in either section of lung resulted in categorization as IP for any individual case. Key histopathologic lesions indicative of IP included: alveolar fibrin exudation, alveolar hyaline membranes, or septal necrosis supportive of (acute IP); type 2 pneumocyte hyperplasia and hypertrophy supportive of (subacute) IP; and interstitial fibroplasia and bronchiolitis obliterans supportive of (chronic) IP. Without these histopathologic features supportive of IP in either sample, cases were allocated to the non-IP group.
Statistical analysis
Excel (Microsoft Corp) and R Studio (R Studio, version 2023.12.1; R Core Team) software were utilized for data handling. For descriptive purposes, a table comparing the outcome variables was built using the packages “dplyr”29 and “tidyr”30 in R Studio. A generalized linear model was utilized to analyze data using the “glm” with “logit” link function of the native “stats”31 package of R Studio. Histopathological diagnosis was the outcome variable of interest, where the diagnosis was classified binomially as IP (1) for IP or non-IP (0) for BP and all other respiratory syndromes without IP.
The model-building process included using the “scope” parameter to always include the fixed variables (sex and DOF) in the model iterations. The variables sex (heifer vs steer) and DOF at time of chuteside evaluation (DOF) categories (0 to 42, 43 to 71, and > 71) were fixed in the model at the experimental design phase of the study due to previous studies11,25,26 describing the association of sex and late DOF in AIP. The stepwise selection with the backward direction using the “stepAIC” function in the “MASS”32 package of R Studio, where all the variables collected were included (BW, ULS, B-line categories, A-line categories, pleural effusion, pleural abnormality, image brightness, average gray value, and B-line area) and least significant variable is removed at each selection step based on best model fitness assessed by lowest Akaike information criterion coefficient. Multicollinearity amongst variables was assessed using the “car”33 package in R Studio, where calculates the variance inflation factor (VIF), variable was considered to have considerable collinearity VIF coefficient when > 2.5,34 therefore removed from the model. The log-odds ratio from the logistic regression model was then converted to probabilities to facilitate interpretation by using the package “emmeans”35 in R Studio.
Results
A total of 1,200 animals were evaluated using TT-POCUS at the time of respiratory treatment, with 66 animals resulting in mortalities on days technicians were present for necropsy and pulmonary sample collections. Four cases were excluded for incomplete data or missing histopathologic diagnosis, resulting in 62 cases enrolled in the study. The population included 60% (37 of 62) of the cases histopathologically diagnosed with IP and 40% (25 of 62) diagnosed with non-IP. There were 38 of 62 heifers (61%) and 24 of 62 steers (39%) enrolled, with descriptive statistics listed in Table 1.
Descriptive statistics of animals (n = 62) according to their demographics (sex, weight, days on feed categories, and treatment count) and targeted thoracic ultrasonography variables (ultrasound lung score, A-line count categories, and B-line count categories).
| Histopathologic diagnosis | ||
|---|---|---|
| Non-IPa | IPb | |
| Sex | ||
| Heifer | 15 | 23 |
| Steer | 10 | 14 |
| Days on feed categories | ||
| 0–42 | 20 | 15 |
| 43–71 | 3 | 9 |
| ≥ 71 | 2 | 13 |
| Body weight, average (SEM; kg) | 312 (± 11) | 390 (± 14) |
| Body weight, categories (kg) | ||
| 7 | 3 | |
| 272–362 | 16 | 14 |
| 362–453 | 2 | 13 |
| ≥ 453 | 0 | 7 |
| Treatment count (n) | ||
| 1 | 11 | 18 |
| 2 | 8 | 15 |
| 3 | 6 | 4 |
| Real-time measurements | ||
| Ultrasound lung score, headc | ||
| 1 | 0 | 0 |
| 2 | 6 | 3 |
| 3 | 5 | 11 |
| 4 | 9 | 22 |
| 5 | 5 | 1 |
| A-line count categories, headc | ||
| 0–2 | 10 | 22 |
| ≥ 3 | 15 | 15 |
| B-line count categories, headc | ||
| 0–2 | 6 | 11 |
| 3–5 | 16 | 13 |
| 3 | 13 | |
| Post hoc measurementsd | ||
| B-line area, cm2, average (SEM) | 17 (± 2.2) | 20 (± 2) |
| Gray value, average (SEM) | 99 (± 9) | 96 (± 6) |
| Intensity density, average (SEM) | 14 (± 2) | 14 (± 1) |
IP = Interstitial pneumonia.
Cases diagnosed as non-IP by histopathology.
Cases diagnosed with IP by histopathologic evaluation.
Variables collected by real-time targeted thoracic point-of-care ultrasound.
Post hoc measurements were performed using ImageJ, version 1.54. Measurements were taken from a single frame from the 60-second clip, where the frame of choice was the frame with the most tissue abnormalities, allowing a comparison based on tissue injury within the caudodorsal region.
The distribution of cases enrolled by DOF at the time of treatment grouped by histopathologic diagnosis indicates that IP cases occurred dispersed throughout the feeding period (median, 50 DOF; IQR, 67 DOF). In contrast, non-IP case distribution showed a tighter distribution of cases, with a median of 28 DOF and an IQR of 16 DOF (Figure 2).
Raincloud-box plot displays temporal distribution of the 62 cases grouped by histopathological diagnoses by days on feed (DOF) at the time of chuteside evaluation (point-of-care ultrasound [POCUS]) on individuals enrolled in this study (n = 62). Histopathological diagnosis (Histopathology Dx) of interstitial pneumonia (IP). Non-IP indicates other respiratory syndromes without IP. The raincloud plots (colored gray for non-IP and purple for IP) illustrate the density and shape of the data distributions. The box plots overlaid within each raincloud show the IQR, with the thick black line representing the median DOF for each group (non-IP median, 28 days; IQR, 16 days; IP median, 50 days; IQR, 67 days). Individual data points are jittered within each group to show the distribution of DOF at the time of chuteside evaluation, with each dot representing 1 observation.
Citation: American Journal of Veterinary Research 86, 2; 10.2460/ajvr.24.08.0233
In addition to DOF, the interval in days from when the animal was evaluated at chuteside to the day of death is also displayed with the raincloud plot and box-plot combination (Figure 3), where IP cases had a shorter day interval window from chuteside evaluation to death, with a median of 4 days and an IQR of 5 days. Non-IP cases displayed a longer interval window from chuteside evaluation to death, with median of 8 days and an IQR of 6 days.
Raincloud-box plot displays temporal distribution of the day intervals between mortalities and their chuteside evaluations with targeted thoracic point-of-care ultrasonography on individuals enrolled in this study (n = 62). Non-IP indicates other respiratory syndromes without IP. The raincloud plots (colored gray for non-IP and purple for IP) illustrate the density and shape of the data distributions. The box plots overlaid within each raincloud show the IQR, with the thick black line representing the median DOF for each group (non-IP median, 8 days; IQR, 6 days; IP median, 4 days; IQR, 5 days). Individual data points are jittered within each group to show the distribution of days interval from time of chuteside evaluation to death, with each dot representing 1 observation.
Citation: American Journal of Veterinary Research 86, 2; 10.2460/ajvr.24.08.0233
Multivariate logistic regression was utilized to identify potential associations between variables of interest (sex, DOF, BW, ULS, B-line count category, A-line count category, pleural effusion, moth sign, pixel intensity density, average gray value, and B-line area cm2) and an outcome response of a diagnosis of IP using a stepwise backward selection of variables. The fixed variable sex was not associated with the outcome IP in this study (P = .59), but DOF category was significantly associated with IP (P = .01). The final model included several variables significantly associated with the probability of IP: ULS (P = .02), B-line count category (P = .01), and A-line count category (P = .03). The BW variable was removed from the model due to considerably high VIF (4.4).
Tukey-adjusted probabilities using the fit model showed the probability of IP was higher (P < .05) when animals were 43 to 71 DOF (87 ± 11%) and > 71 DOF (81 ± 16%) compared to 0 to 42 DOF (29 ± 16%). An ultrasound lung score of 1 was not present in any of the 62 cases; thus, the probabilities were only calculated for scores 2, 3, 4, and 5. A ULS of 5 was the least likely to be confirmed as IP by histopathology, with 12 ± 16% likelihood of a diagnosis of IP amongst all the scores (P < .05). Ultrasound lung scores of 2, 3, and 4 did not differ in the probabilities of IP (P > .05; 72 ± 18%, 92 ± 6%, and 85 ± 8%, respectively). A B-line count category of > 5 B-lines in an ultrasound image was associated with an 86 ± 11% likelihood to be diagnosed with IP, whereas animals with 3 to 5 B-lines were associated with a significantly lower probability of a confirmed diagnosis of IP (38 ± 14%; P < .05). A B-line count category of 0 to 2 showed no statistical difference amongst the other 2 categories (75 ± 15% likelihood of IP; P > .05). The probabilities of A-line count categories 0 to 2 versus ≥ 3 A-lines were 83 ± 10% and 51 ± 15%, respectively (P < .05).
Discussion
Respiratory disease is a critical factor contributing to feedyard morbidity and mortality, and accurate classification of disease types at the time of treatment is essential to inform effective management and treatment strategies.36 Results indicate that TT-POCUS can be a valuable diagnostic tool for identifying IP in feedyard cattle morbidities, and similar promising results in identifying IP have been shown in human medicine during the COVID-19 pandemic, where bedside point-of-care ultrasound was utilized to evaluate IP.16,37
Lung evaluation using ultrasonography in veterinary medicine has been reported for respiratory disease detection and prognosis.17,23,26,38–43 However, this study proposed a novel evaluation method, where the evaluation time was ≤ 60 seconds, by targeting the caudodorsal lobe of the right lung and was able to decrease evaluation time drastically to find associations of real-time variables (ULS, B-line count, A-line). In contrast to our study that developed the TT-POCUS, the traditional pulmonary ultrasound evaluation is more comprehensive and could take 7 to 45 minutes19,24; this significant time constraint is a limitation to deploying the method within a commercial operation, and the accessibility to the cranioventral acoustic window can also be challenging and time consuming.
Notably, our study demonstrated that real-time collected variables were the only ones significantly associated with histologic diagnosis of IP, suggesting that TT-POCUS can be effectively used chuteside for immediate diagnostic and therapeutic decision making. Although we tested post hoc calculated variables, none of them showed significant associations with IP, further highlighting the practical utility of real-time ultrasound evaluations in the field. In addition, performing TT-POCUS in a 60-second window increases the feasibility for application of this method in commercial operations where the time spent in diagnosis is a major determining factor for adoption. The ability of adopting TT-POCUS in a commercial operation can enhance diagnostic accuracy in real-time, improve treatment outcomes in feedyard cattle, and aid informed management decisions.
Histopathologic analysis revealed that the majority of cases in our study were diagnosed with IP (37 of 62 [60%]), whereas 40% were diagnosed with non-IP. Higher IP prevalence observed in this study may be best explained by our inclusion of both BIP and AIP in contrast to other surveillance studies. Furthermore, another significant factor that could have contributed to higher-than-expected prevalence is the increased sensitivity due to the histologic evaluation of 2 lung samples per animal. A previous epidemiologic publication reported the prevalence of IP (AIP plus BIP) mortalities in feedyard to be approximately 60.4% of the all the respiratory mortalities, whereas 40.4% of the respiratory mortalities were classified as non-IP.7,11
Interstitial pneumonia cases were more dispersed across the feeding period, with a median DOF of 50 days and an IQR of 67 days, indicating a wide distribution of cases. In contrast, the non-IP cases had a narrower distribution, with a median DOF of 28 days and an IQR of 16 days (Figure 2). Additionally, the interval from chuteside evaluation to death was shorter for IP cases (median, 4 days; IQR, 5 days) compared to non-IP cases (median, 8 days; IQR, 6 days) (Figure 3); this shorter time interval from clinical diagnosis to death is the theorized quick development of IP, also described as acute.8,44,45
This variability in interval from TT-POCUS evaluation to death (necropsy) may influence the findings related to pulmonary lesions as longer intervals could allow for progression or resolution of certain lung pathologies, which may obscure the initial TT-POCUS findings. Specifically, the rapid progression of IP cases may mean that TT-POCUS findings at the time of diagnosis align more closely with the acute lesions observed on histopathology, whereas non-IP cases may have had more time for lesion development or alteration.
No epidemiologic association was identified in this study between IP and sex. The existing literature often reports a higher susceptibility of heifers to AIP and a higher proportion of steers presenting with BIP.4,6,46 Our data did not show a statistically significant difference between heifers and steers in terms of IP probability, and the absence of a significant difference in our study may be attributed to classifying BIP and AIP cases as IP, specific management practices, sample size, or environmental factors unique to our study population. Although this study did not differentiate between AIP and BIP, which could have affected the differences between heifers and steers, the decision to group these conditions together was based on their overlapping pathology presentations at the caudodorsal level. Future studies with larger sample sizes and clearer differentiation of these subtypes are warranted to explore potential sex-related differences in IP prevalence.
The distribution of cases based on DOF showed that most animals diagnosed with non-IP were in the 0 to 42 DOF category, which is considered the early-feeding phase. This temporal pattern was observed by previous studies,25,47 where non-IP including BP had higher morbidity early in the feeding phase (< 42 DOF). In contrast, animals diagnosed with IP were distributed across the 3 time categories: 40% (15 of 37) were observed in the 0 to 42 DOF category, while 60% (22 of 37) were found in the combined categories of 43 to 71 DOF and greater than 71 DOF. These DOF patterns for non-IP (BP and other respiratory diseases not IP related) and IP (Figure 3) are similar to a recent epidemiologic study11 showing that IP occurred throughout the feeding phase. Although IP distribution was dispersed throughout the feeding period in the current study, the model-adjusted probabilities indicate that IP is more likely to happen mid and late (> 42 DOF). These probabilities are corroborated by field data indicating that IP is often described as a late-day disease.4,45,46
The inclusion of epidemiological data in this study enhances the understanding of how TT-POCUS findings could be interpreted alongside other demographics to improve IP diagnosis. While TT-POCUS provides valuable information on lung pathology, integrating this data with demographic factors, such as DOF or BW, could improve the diagnostic accuracy. Hence, DOF and sex were kept in the model.
Ultrasound lung score indicated that animals with a score of 5 had higher prevalence (5 of 6 of all ULS 5 [83%]) in cattle in the non-IP group. Hence, the probability of these animals with confirmed IP was very low among those animals with ULS 5 from TT-POCUS (12 ± 16% likelihood), reinforcing that pulmonary tissue consolidation identified during ultrasound evaluation is not likely to be associated with a confirmed diagnosis of IP37,48 (Figure 4). The B-line count category was also associated with IP outcome, where animals with more than 5 B-lines identified at TT-POCUS displayed the greatest likelihood to be diagnosed with IP (86 ± 11%). The B-line is the artifact most evaluated in pulmonary ultrasonography, and its relevance in the context of IP in our study is that more B-lines appear in the image when there are areas of decreased air content.23,27,49 The use of these real-time variables in differentiating respiratory diseases have not been reported in other studies. In contrast, thoracic ultrasonography has been used extensively in critical care for respiratory disease detection and assessment of disease severity.23 One limitation to B-line evaluation count is that as disease develops and pulmonary injury worsen, B-lines have the potential to merge (Figure 4), making absolute quantification difficult.28 In an attempt to evaluate the potential association of merged B-lines, this study analyzed the post hoc variable B-line area to account for B-line merging. However, the statistical analysis revealed no evidence of an association of B-line area with an ability to differentiate of IP from non-IP. We theorized that the merging of B-lines was most likely accounted for in the statistical model using ULS.
Histologic photomicrographs from an animal (ID #453) diagnosed with IP. A—H&E stain; scale bar = 200 μm: Diffuse alveolar filling by proteinaceous fluid, fibrin, and inflammatory cells. B—H&E stain; scale bar = 125 μm: Presence of alveolar hyaline membranes (arrowhead) and pneumocyte type II hypertrophy (*). C—H&E stain; scale bar = 150 μm: Chronic interstitial fibroplasia (f) and bronchiolitis obliterans (bo). D—Transthoracic right lung point of care ultrasonography image (ID #453): Absence of A-lines, merged B-lines, diffuse hyperechoic lung, ultrasound lung score (ULS) 3. E–H—Histologic photomicrographs and ultrasound from an animal (ID #767) diagnosed with bronchopneumonia. E—H&E stain; scale bar = 325 μm: Bronchial, bronchiolar, and alveolar filling by hemorrhage, neutrophils intermixed with necrotic debris (n), and fibrin (f). F—H&E stain; scale bar = 65 μm: Fibrinonecrotizing and neutrophilic bronchiolitis and bronchiolectasis. G—H&E stain; scale bar = 65 μm: Neutrophilic bronchiolitis. H—(ID #767) Transthoracic right lung point-of-care ultrasonography image; lung consolidation, absence of A-lines, and ULS 5.
Citation: American Journal of Veterinary Research 86, 2; 10.2460/ajvr.24.08.0233
The A-line count category is a novel variable not described as useful for respiratory disease differentiation. We found evidence of statistical association, where A-line count category lower than 3 displayed a higher probability of IP diagnosis (< 3 vs ≥ 3 A-lines were 83 ± 10% and 51 ± 15% probability of IP). Normal lungs are expected to exhibit several A-lines due to the reflection of the ultrasound waves by the pleural line, which is caused by the air-filled lung. A-lines appear as hyperechoic horizontal artifacts, and their presence typically indicates a healthy lung with normal air content.23 In contrast, lower A-line counts in IP cases can be attributed to the pathologic changes occurring in the lung tissue. Interstitial pneumonia involves the infiltration of inflammatory cells into the interstitium and accumulation of fluid within the alveolar spaces, which disrupts the normal air-tissue interface. This disruption reduces the reflective surface area available to produce A-lines. As a result, the number of detectable A-lines decreases in association with decreased lung aeration.
Collectively, several features, including a ULS of 2, 3, and 4; higher B-line counts; and lower A-line counts, occur significantly more often in IP cases compared to non-IP cases, suggesting that these markers could serve as useful diagnostic indicators in veterinary medicine for identifying severe interstitial lung disease. By incorporating ULS, B-line, and A-line counts into TT-POCUS evaluations, field veterinarians can improve their ability to differentiate IP from other respiratory diseases, leading to more targeted and early effective treatment strategies. However, the study was limited in scope (only mortalities); TT-POCUS shows promise as an adjunct diagnostic tool in commercial settings when used in conjunction with routine health monitoring.
The TT-POCUS requires some level of training; the simplicity and speed of the caudodorsal lung evaluation make it a promising tool for adoption by trained operators in commercial feedyards. In 2013, Buczinski et al24 highlighted that with adequate training, even less experienced operators can perform thoracic ultrasonography with consistency and accuracy in detecting lung pathologies in cattle. Variables, such as B-line and A-line counts, are relatively straightforward to learn with adequate training and could provide valuable diagnostic insights when integrated into routine health assessments.
Furthermore, this study did not assess the impact of specific therapeutic interventions, but it is plausible that treatment outcomes might affect lesion progression or resolution over time. Future research should explore these factors further to clarify the relationships between TT-POCUS findings, treatment effects, and histopathologic outcomes in feedyard cattle.
The findings of this study highlight the substantial potential of TT-POCUS as a diagnostic modality for the identification of IP in feedyard cattle. Significant associations were observed between the histopathologic diagnosis of IP and imaging parameters, including B-line count, ULS, and A-line count. These associations show the potential diagnostic efficacy of TT-POCUS in differentiating interstitial lung disease from other respiratory diseases in a timely manner. This diagnostic method can potentially enhance management decisions in cattle with findings compatible with IP, leading to potential improved health outcomes for feedyard cattle morbidities. Integrating TT-POCUS into routine sick animal diagnostics can streamline health monitoring and disease management processes.
The promising results from this study warrant further research and refinement of this diagnostic technique. Future studies should focus on developing this TT-POCUS protocol and validating these findings across diverse cattle populations and settings to fully realize the clinical benefits and ensure the widespread adoption of TT-POCUS in the field.
Acknowledgments
Access to feedyard cattle for this study was provided by Innovative Livestock Services. The ultrasound unit (Butterfly iQ + Vet) was loaned by Butterfly Network. The illustration utilized for Figure 1 was courtesy of Dr. Eduarda Bortoluzzi (Department of Anatomy and Physiology, Kansas State University), made specifically for this publication.
Disclosures
Dr. White is a member of the AJVR Scientific Review Board, but was not involved in the editorial evaluation of or decision to accept this article for publication.
ResearchRabbit.ai was utilized to assist in creating the bibliography collection.
Funding
Funded by the Beef Cattle Institute and by an intramural grant at the Kansas State University College of Veterinary Medicine (mentored clinical, applied, or translational research).
ORCID
B. White https://orcid.org/0000-0002-4293-6128
L. F. B. B. Feitoza https://orcid.org/0000-0001-9995-8929
R. L. Larson https://orcid.org/0000-0003-3633-403X
B. L. Plattner https://orcid.org/0000-0002-5088-7143
References
- 1.↑
Panciera RJ, Confer AW. Pathogenesis and pathology of bovine pneumonia. Vet Clin North Am Food Anim Pract. 2010;26(2):191–214. doi:10.1016/j.cvfa.2010.04.001
- 2.
Smith RA, Step DL, Woolums AR. Bovine respiratory disease: looking back and looking forward, what do we see? Vet Clin North Am Food Anim Pract. 2020;36(2):239–251. doi:10.1016/j.cvfa.2020.03.009
- 3.↑
Johnson KK, Pendell DL. Market impacts of reducing the prevalence of bovine respiratory disease in United States beef cattle feedlots. Front Vet Sci. 2017;4:189. doi:10.3389/fvets.2017.00189
- 4.↑
Woolums AR. Feedlot acute interstitial pneumonia. Vet Clin North Am Food Anim Pract. 2015;31(3):381–389. doi:10.1016/j.cvfa.2015.05.010
- 5.
Doster AR. Bovine atypical interstitial pneumonia. Vet Clin North Am Food Anim Pract. 2010;26(2):395–407. doi:10.1016/j.cvfa.2010.03.002
- 6.↑
Haydock LAJ, Fenton RK, Smerek D, Renaud DL, Caswell JL. Bronchopneumonia with interstitial pneumonia in feedlot cattle: epidemiologic characteristics of affected animals. Vet Pathol. 2023;60(2):226–234. doi:10.1177/03009858221146096
- 7.↑
Schmidt PH, White BJ, Finley A, et al. Determining frequency of common pulmonary gross and histopathological findings in feedyard fatalities. Vet Sci. 2023;10(3):228.
- 8.↑
Valles JA, Apley MD, Reinhardt CD, Bartle SJ, Thomson DU. Pathologies of acute interstitial pneumonia in feedlot cattle. Am J Anim Vet Sci. 2016;11(1):1–7. doi:10.3844/ajavsp.2016.1.7
- 9.↑
Buczinski S, Pardon B. Bovine respiratory disease diagnosis: what progress has been made in clinical diagnosis? Vet Clin North Am Food Anim Pract. 2020;36(2):399–423. doi:10.1016/j.cvfa.2020.03.004
- 10.↑
Pardon B, Buczinski S. Bovine respiratory disease diagnosis: what progress has been made in infectious diagnosis? Vet Clin North Am Food Anim Pract. 2020;36(2):425–444. doi:10.1016/j.cvfa.2020.03.005
- 11.↑
Bortoluzzi EM, White BJ, Schmidt PH, et al. Epidemiological factors associated with gross diagnosis of pulmonary pathology in feedyard mortalities. Vet Sci. 2023;10(8):522.
- 12.↑
Cooper VL, Brodersen BW. Respiratory disease diagnostics of cattle. Vet Clin North Am Food Anim Pract. 2010;26(2):409–416. doi:10.1016/j.cvfa.2010.04.009
- 13.
Abutarbush SM, Pollock CM, Wildman BK, et al. Evaluation of the diagnostic and prognostic utility of ultrasonography at first diagnosis of presumptive bovine respiratory disease. Can J Vet Res. 2012;76(1):23–32.
- 14.↑
Wolfger B, Timsit E, White BJ, Orsel K. A systematic review of bovine respiratory disease diagnosis focused on diagnostic confirmation, early detection, and prediction of unfavorable outcomes in feedlot cattle. Vet Clin North Am Food Anim Pract. 2015;31(3):351–365. doi:10.1016/j.cvfa.2015.05.005
- 15.↑
Hille MM, Sillman SJ, Brodersen BW. The role of histopathology in ruminant diagnostics. Vet Clin North Am Food Anim Pract. 2023;39(1):73–91. doi:10.1016/j.cvfa.2022.10.005
- 16.↑
Allinovi M, Parise A, Giacalone M, et al. Lung ultrasound may support diagnosis and monitoring of COVID-19 pneumonia. Ultrasound Med Biol. 2020;46(11):2908–2917. doi:10.1016/j.ultrasmedbio.2020.07.018
- 17.↑
Reef VB, Whittier M, Allam LG. Thoracic ultrasonography. Clin Technique Equine Pract. 2004;3(3):284–293. doi:10.1053/j.ctep.2005.02.006
- 18.
Masset N, Assié S, Herman N, Jozan T, Herry V. Ultrasonography of the cranial part of the thorax is a quick and sensitive technique to detect lung consolidation in veal calves. Vet Med Sci. 2022;8(3):1229–1239. doi:10.1002/vms3.774
- 19.↑
Rademacher RD, Buczinski S, Tripp HM, Edmonds MD, Johnson EG. Systematic thoracic ultrasonography in acute bovine respiratory disease of feedlot steers: impact of lung consolidation on diagnosis and prognosis in a case-control study. Bovine Practitioner. 2014;48(1):1–10.
- 20.
Smallwood N, Dachsel M. Point-of-care ultrasound (POCUS): unnecessary gadgetry or evidence-based medicine? Clin Med (Lond). 2018;18(3):219–224.
- 21.
Mojoli F, Bouhemad B, Mongodi S, Lichtenstein D. Lung ultrasound for critically ill patients. Am J Respir Crit Care Med. 2019;199(6):701–714. doi:10.1164/rccm.201802-0236CI
- 22.↑
Lichtenstein DA, Mezière GA. Relevance of lung ultrasound in the diagnosis of acute respiratory failure the BLUE protocol. Chest. 2008;134(1):117–125. doi:10.1378/chest.07-2800
- 23.↑
Bøysen S, Gommeren K, Chalhoub S. The Essentials of Veterinary Point-of-Care Ultrasound: Pleural Space and Lung. Edra; 2022.
- 24.↑
Buczinski S, Forté G, Bélanger AM. Short communication: ultrasonographic assessment of the thorax as a fast technique to assess pulmonary lesions in dairy calves with bovine respiratory disease. J Dairy Sci. 2013;96(7):4523–4528. doi:10.3168/jds.2013-6577
- 25.↑
Smith KJ, Amrine DE, Larson RL, Theurer ME, White BJ. Determining relevant risk factors associated with mid- and late-feeding-stage bovine respiratory disease morbidity in cohorts of beef feedlot cattle. Appl Anim Sci. 2022;38(4):373–379. doi:10.15232/aas.2022-02312
- 26.↑
Fiore E, Lisuzzo A, Beltrame A, et al. Lung ultrasonography and clinical follow-up evaluations in fattening bulls affected by bovine respiratory disease (BRD) during the restocking period and after tulathromycin and ketoprofen treatment. Animals. 2022;12(8):994. doi:10.3390/ani12080994
- 27.↑
Parri N, Allinovi M, Giacalone M, Corsini I. To B or not to B. The rationale for quantifying B-lines in pediatric lung diseases. Pediatr Pulmonol. 2023;58(1):9–15. doi:10.1002/ppul.26185
- 28.↑
Fischer EA, Minami T, Ma IWY, Yasukawa K. Lung ultrasound for pleural line abnormalities, confluent B-lines, and consolidation: expert reproducibility and a method of standardization. J Ultrasound Med. 2022;41(8):2097–2107. doi:10.1002/jum.15894
- 29.↑
Wickham H, François R, Henry L, Müller K, Vaughan D. dplyr: A Grammar of Data Manipulation. CRAN: Contributed Packages. Published online January 16, 2014. doi:10.32614/CRAN.package.dplyr
- 30.↑
Wickham H, Vaughan D, Girlich M. tidyr: Tidy Messy Data. CRAN: Contributed Packages. Published online July 21, 2014. doi:10.32614/CRAN.package.tidyr
- 31.↑
R: a language and environment for statistical computing. Version 2023.12.1. Accessed October 27, 2024. https://www.R-project.org/
- 32.↑
Venables WN, Ripley BD. Modern Applied Statistics with S. 4th ed. Springer; 2002. Accessed October 27, 2024. https://www.stats.ox.ac.uk/pub/MASS4/
- 33.↑
Fox J, Weisberg S. An {R} Companion to Applied Regression. 3rd ed. Sage; 2019. Accessed October 27, 2024. https://socialsciences.mcmaster.ca/jfox/Books/Companion/
- 34.↑
Johnston R, Jones K, Manley D. Confounding and collinearity in regression analysis: a cautionary tale and an alternative procedure, illustrated by studies of British voting behaviour. Qual Quant. 2018;52(4):1957–1976. doi:10.1007/s11135-017-0584-6
- 35.↑
Lenth RV. Emmeans: estimated marginal means, AKA least-squares means. Accessed October 27, 2024. https://CRAN.R-project.org/package=emmeans
- 36.↑
White BJ, Larson BL. Impact of bovine respiratory disease in U.S. beef cattle. Anim Health Res Rev. 2020;21(2):132–134. doi:10.1017/S1466252320000079
- 37.↑
Volpicelli G, Frascisco MF. Sonographic detection of radio-occult interstitial lung involvement in measles pneumonitis. Am J Emerg Med. 2009;27(1):128.e1–128.e3. doi:10.1016/j.ajem.2008.03.052
- 38.↑
Braun U, Pusterla N, Flückiger M. Ultrasonographic findings in cattle with pleuropneumonia. Vet Rec. 1997;141(1):12–17. doi:10.1136/vr.141.1.12
- 39.
Scott PR, Gessert ME. Ultrasonographic examination of the ovine thorax. Vet J. 1998;155(3):305–310. doi:10.1016/S1090-0233(05)80027-9
- 40.
Rabeling B, Rehage J, Döpfer D, Scholz H. Ultrasonographic findings in calves with respiratory disease. Vet Rec. 1998;143(17):468–471. doi:10.1136/vr.143.17.468
- 41.
Tharwat M, Oikawa S. Ultrasonographic evaluation of cattle and buffaloes with respiratory disorders. Trop Anim Health Prod. 2011;43(4):803–810. doi:10.1007/s11250-010-9766-0
- 42.
Babkine M, Blond L. Ultrasonography of the bovine respiratory system and its practical application. Vet Clin North Am Food Anim Pract. 2009;25(3):633–649. doi:10.1016/j.cvfa.2009.07.001
- 43.↑
Timsit E, Tison N, Booker CW, Buczinski S. Association of lung lesions measured by thoracic ultrasonography at first diagnosis of bronchopneumonia with relapse rate and growth performance in feedlot cattle. J Vet Intern Med. 2019;33(3):1540–1546. doi:10.1111/jvim.15483
- 44.↑
Haydock LAJ, Fenton RK, Sergejewich L, Squires EJ, Caswell JL. Acute interstitial pneumonia and the biology of 3-methylindole in feedlot cattle. Anim Health Res Rev. 2022;23(1):72–81. doi:10.1017/S1466252322000020
- 45.↑
Ayroud M, Popp JD, Vanderkop MA, et al. Characterization of acute interstitial pneumonia in cattle in southern Alberta feedyards. Can Vet J. 2000;41(7):547–554.
- 46.↑
Loneragan GH, Gould DH, Mason GL, et al. Involvement of microbial respiratory pathogens in acute interstitial pneumonia in feedlot cattle. Am J Vet Res. 2001;62(10):1519–1524. doi:10.2460/ajvr.2001.62.1519
- 47.↑
Babcock AH, Renter DG, White BJ, Dubnicka SR, Scott HM. Temporal distributions of respiratory disease events within cohorts of feedlot cattle and associations with cattle health and performance indices. Prev Vet Med. 2010;97(3–4):198–219. doi:10.1016/j.prevetmed.2010.09.003
- 48.↑
Volpicelli G, Mussa A, Garofalo G, et al. Bedside lung ultrasound in the assessment of alveolar-interstitial syndrome. Ame J Emerg Med. 2006;24(6):689–696. doi:10.1016/j.ajem.2006.02.013
- 49.↑
Omer T, Cousins C, Lynch T, Le NN, Sajed D, Mailhot T. Lung ultrasound findings in COVID-19: a descriptive retrospective study. Cureus. 2022;14(3):e23375. doi:10.7759/cureus.23375
