Assessment of Adrenal Computed Tomography Characteristics in Cats with Nonadrenal Disease
ABSTRACT
Adrenal computed tomography characteristics (aCTc) in healthy cats are known, but reference intervals for diseased cats are lacking. aCTc of cats without evidence of adrenal disease (NAD group) were compared to parameters of cats with possible concurrent adrenal disease (PAD group). The PAD group was assessed for adrenal masses or other morphological deviations using the NAD group as reference. Associations of aCTc with patient variables were explored, and all results were compared with published aCTc of healthy cats. No incidental adrenal masses were identified in the PAD group (n = 92), and only few aCTc differed compared to the NAD group (n = 30). The NAD group showed similar associations of patient variables and aCTc as length (right: 11.5 ± 2.2 mm, left: 11.8 ± 1.7 mm), width (right: 6.4 ± 1.2 mm, left: 5.4 ± 0.8 mm), height (right: 4.5 ± 0.9 mm, left: 4.5 ± 0.8 mm), attenuation (right: 33.1 ± 5.0 Hounsfield units, left: 32.5 ± 5.3 Hounsfield units) and position, but markedly more mineralization (right: 10%, left: 13.3%) than reported in healthy cats. This study provides references of aCTc for diseased cats without evidence of adrenal disease. The result suggests that adrenal incidentalomas seem to be rare in cats.
Introduction
Advanced diagnostic imaging, especially computed tomography (CT), is increasingly used to identify morphological changes of abdominal organs including the adrenal glands in cats. Accurate reference intervals and awareness of normal variants of adrenal glands facilitate the discrimination of adrenal disease from normal morphology.
Several ultrasonographic studies have described the adrenal characteristics of healthy and diseased cats with nonadrenal or adrenal disease.1–4 Two recent studies investigated adrenal CT characteristics (aCTc) in healthy cats, but examinations including a group of diseased cats are lacking.5,6 In cats with nonadrenal diseases, reference intervals of aCTc are not yet known, and the prevalence of anatomic variants and pathological changes still needs to be determined. CT studies about adrenal disease in cats are currently limited to case reports about adrenal neoplastic masses.7–9 With an incidence of 0.03%, feline adrenal neoplasia is rare, and, unlike in humans and dogs, the prevalence of feline adrenal incidental masses is not known.10–14
Our objectives were to (1) describe aCTc in cats with no evidence of adrenal disease who underwent CT for nonadrenal disease (NAD group) and (2) determine the prevalence of adrenal masses as well as to evaluate other morphological deviations of aCTc in cats with possible concurrent adrenal disease (PAD group), using the NAD group as a reference. Additionally, we aimed to compare all results with previously published adrenal gland variables in healthy cats.
Materials and Methods
The study was set up as an observational study that includes one prospectively (NAD group) and one retrospectively (PAD group) selected study group.
NAD Group: Prospective Study (Cats with No Evidence of Adrenal Disease)
Data for the NAD group were collected during the period from October 2018 until July 2019. Recruitment of this group was done by asking clients of the Clinical Unit of Diagnostic Imaging of the University of Veterinary Medicine Vienna with cats meeting the inclusion criteria to participate in the study. A written owner informed consent was obtained for all cats participating in the study. Inclusion criteria consisted of: age ≥1 yr and presence of nonadrenal disease that warranted CT examination. A full clinical history, complete physical examination, complete blood count, extensive plasma biochemistry profile including electrolytes, total thyroxine concentration (if ≥6 yr of age), and urinalysis were performed in each cat. A body condition score (BCS) of 1 (thin) to 5 (obese) according to Shoveller and others was applied.15 Systolic blood pressure was determined to exclude patients with severe systolic hypertension (≥179 mm Hg).16
Exclusion criteria included the administration of drugs and diseases that could influence adrenal gland morphology. Specifically, cats treated with corticosteroids, spironolactone, or an angiotensinconverting enzyme inhibitor were excluded. Furthermore, animals that underwent administration of iodinated contrast media within the last 10 days before the study as well as cats with suspected or confirmed diagnosis of the following conditions were excluded: feline idiopathic cystitis, chronic kidney disease, hyperadrenocorticism, hypoadrenocorticism, diabetes mellitus, hyperthyroidism, primary hyperaldosteronism, hypersomatotropism, abnormal sexual hormone production, severe hypertension, and adrenal neoplasia. Presence of nonadrenal diseases that were unlikely to impact adrenal function or adrenal mineralization did not warrant exclusion. The remaining cats were considered to have normal adrenal gland structure and function and were thus included.
PAD Group: Retrospective Study (Cats with Possible Concurrent Adrenal Disease)
The archive of the Clinical Unit of Diagnostic Imaging of the University of Veterinary Medicine Vienna was searched for all CT studies of cats that included the adrenal gland region (June 2009 to September 2018). A slice thickness ≤1 mm, a quality of the CT scan which ensured unrestricted adrenal assessment, and a patient’s age of ≥1 yr was required for inclusion. Unlike the NAD group, all cats were included regardless of underlying disease process, including diseases that may alter adrenal morphology. Patients who had received iodinated contrast media within the last 10 days before the study or with signs of incomplete contrast media excretion were excluded.
Data collected from medical records included age, breed, sex, body weight, BCS, indication for CT, CT acquisition parameters, and body position during CT.
All procedures in this study were approved by the institutional ethics and animal welfare committee of the University of Veterinary Medicine Vienna in accordance with good scientific practice guidelines and national legislation (ETK-09/07/2018 - number of ethical committee for approval).
Image Analysis
Images were analysed by one radiologist (S.G.), using a dedicated workstation and the imaging software JiveX Diagnostic Advanced 5.0.4.4a. Adrenal measurements were conducted using soft-tissue window settings (window level: +350 Hounsfield units [HU], window width: +40 HU). Delineation was graded as very good (0), good (1), moderate (2), or poor (3). All adrenal dimensions were measured with a digital caliper in two planes, and the larger measurement was recorded. Length was measured from the cranial to the caudal borders, width was measured from the medial to the lateral borders, and height was measured from the ventral to the dorsal borders, all at the level of maximal dimension (Figures 1A, B). Adrenal glands were categorized into one of three commonly recognized shape patterns: bipolar (oval with a depression over the minor axis), oval (without depression), or arrowhead (wedge shape with rounded caudal tip) (Figures 2A, B). Adrenal gland attenuation was measured in HU by placing a region of interest outline over the largest cross-sectional area of the adrenal gland in the sagittal or dorsal plane (Figure 3A). The outline of the region of interest was traced by hand. The phrenicoabdominal vessels and the adrenal gland borders were excluded to avoid partial volume effects and exclude periadrenal fat. The adrenal position relative to major vasculature, bordering of vessels and organs, presence of adrenal mineralization (Figure 3B) or masses was noted.
Citation: Journal of the American Animal Hospital Association 58, 3; 10.5326/JAAHA-MS-7140
Citation: Journal of the American Animal Hospital Association 58, 3; 10.5326/JAAHA-MS-7140
Citation: Journal of the American Animal Hospital Association 58, 3; 10.5326/JAAHA-MS-7140
Intra- and Interobserver Agreement
To assess intra-observer agreement, 30 randomly selected anonymized cats from the PAD group were re-evaluated by the same radiologist following a 3 mo interval. A second observer, a board-certified imaging diplomate (E.L.), evaluated the same cats once to permit interobserver agreement assessment.
CT Acquisition
The NAD group received a standardized IV protocol for anaesthesia: midazolamb (0.2 mg/kg), butorphanolc (0.2 mg/kg), ketamined (1–5 mg/kg), propofole (2–4 mg/kg), and sevofluranef (1.5–2.2% in 100% of O2 at 2 L/min). Scans were performed in sternal position with the breath-holding technique. In addition to the region, clinically indicated by the cats nonadrenal disease, abdominal images were acquired from cranial to the diaphragm to caudal to the coxofemoral joint with 130 kVp, 80–120 mAs, a pitch of 1, a scan matrix of 512 × 512, and a slice thickness of 0.75 mm. In both groups, a 16-slice helical CT scannerg was used. Because the PAD group was analyzed retrospectively, the protocol for general anesthesia and scanning was not standardized. The scanned region and body position depended on the clinical indication. Acquisition settings included algorithm-thoracic, abdominal, total-body, spinal, 110–130 kVp, 80–180 mAs, a pitch of 1, a scan matrix of 512 × 512, a slice thickness of 0.75 (n = 82) or 1 mm (n = 10), a field of view of 59–383 mm.
Statistical Analysis
Statistical analysis was performed using IBM SPSS version 24 softwareh. The assumption of normal distribution was assessed using the Kolmogorov-Smirnov test. Most data did not show a normal distribution. Differences between groups in age, sex, bodyweight, and BCS were analyzed using Mann-Whitney U tests. The Student t test was applied, when distribution was normal. Spearman’s correlation coefficient (r) was used to analyze the correlation between adrenal gland variables (size, attenuation, delineation grade, and the presence of mineralization or masses) and age, sex, laterality, BCS, and body weight of cats. The Pearson correlation coefficient (r) and paired samples t test were performed to test intra- and interobserver agreements. For all statistical analyses P < .05 was considered significant.
Results
Signalment, mean body weight, and mean BCS of both groups are summarized in Table 1. Thirty and 92 cats were included in the NAD and PAD groups, respectively. There was no significant difference between these groups in terms of age, sex, breed distribution, body weight, and BCS.
The adrenal glands were visualized in all cats. Adrenal gland dimensions, shape, attenuation, and the frequency of mineralization are summarized for both groups in Table 2, along with comparable data of current sonographic and CT publications. Although adrenal disease was not excluded in the PAD group, no cat had adrenal disease as an indication for CT imaging.
The maximum dimensions and ranges were larger for the PAD group; however, only the mean right adrenal height (P = .01) and the mean left adrenal width (P = .01) achieved significance when compared with the NAD group. In the PAD group, dimensions of some adrenal glands were several millimeters above the maximum of the NAD group. No masses, pronounced asymmetries, or abnormal shapes were detectable in any of the adrenal glands. Adrenal gland height was significantly larger on the left side in both groups when compared to the right side (NAD-group: P = .02, PAD group: P = .01). Right adrenal gland width was significantly wider than the left side in the NAD group (P = .01).
Delineation of the left adrenal gland did not diverge significantly (P = .27) between the NAD group (mean 1.2 ± 0.8) and the PAD group (mean 1.4 ± 0.7). In contrast, the grade of delineation of the right adrenal gland was significantly (P = .049) better in the NAD group (mean 1.4 ± 0.6) compared with the PAD group (mean 1.7 ± 0.6). Subjectively, the left adrenal gland was easier to delineate than the right in both groups; however, differences in delineation only achieved statistical significance in the PAD group (P = .04).
Adrenal glands without mineralization exhibited homogenous parenchyma. There was no layering, consistent with corticomedullary differentiation, detectable. Adrenal glands were positioned craniomedial to the ipsilateral kidney. In both groups, the right adrenal gland was typically bordering the liver and the vena cava caudalis (VCcd) and positioned dorsolateral or lateral to the VCcd and ventrolateral to the aorta abdominalis (Aabd). The left adrenal gland was regularly bordering the VCcd, less often the arteria mesenterica cranialis and seldom the vena renalis or Aabd or other surrounding vessels or organs. It was positioned lateral or dorsolateral to the VCcd, frequently ventrolateral, seldom ventral to the Aabd.
Adrenal gland shape did not differ significantly between groups and was best recognized in the dorsal plane. The sagittal plane was misleading when evaluating the three-dimensional adrenal gland shape, especially regarding the arrowhead shape (Figures 2A, B). In both groups, a bipolar or oval shape was most frequently observed. The arrowhead shape was least frequently observed and identified in 7 (23.3%) and 25 (27.2%) right adrenal glands in the NAD group and the PAD group, respectively. In the NAD group, none of the left adrenal glands were of arrowhead shape, but in the PAD group, 7 (7.6%) adrenal glands were.
Mean adrenal gland attenuation did not differ significantly between left and right within groups and between groups. The range of attenuation in the PAD group was wider, but cats in the PAD group only mildly exceeded the range of the NAD group (≤13.7 HU). The frequency of mineralization did not differ significantly between groups or lateralizations (NAD group, right: 3 [10%], left: 4 [13.3%]; PAD group, right: 10 [10.9%], left: 6 [6.5%]).
Intra-observer agreement of adrenal size measurements and attenuation was good to excellent. It was r = 0.92 for length, r = 0.83 for width, r = 0.84 for height, and r = 0.81 for attenuation with P ≤ .01 for all results. Interobserver agreement was moderate to good. It was r = 0.88 for length, r = 0.80 for width, r = 0.72 for height, and r = 0.54 for attenuation with P ≤ .01 for these results. There were no significant differences between means of the adrenal gland dimensions within and between observers.
Few associations were identified between adrenal gland CT parameters and signalment. Right adrenal gland height in the NAD group was the only aCTc that correlated (r = 0.51, P = .01) with cats’ age. Left adrenal gland delineation in the PAD group exhibited mild negative correlation (r = −0.31, P = .01) with BCS (therefore improving with a rising BCS). Delineation grade did not correlate with body positioning. There was no statistically significant association between adrenal gland size, attenuation, delineation grade, presence of mineralization or masses and age, sex, BCS, and body weight.
Discussion
To the authors’ knowledge, this is the first study that investigated the presence of adrenal gland structural changes in a group of cats using CT. Adrenal masses were not identified in any cat under investigation. This is of interest because the prevalence of incidentally identified adrenal masses, so-called incidentalomas, was 3–7% in human studies and 4–9% in canine studies, respectively.11,12,14,17,18 The fact that no adrenal mass was found in 92 cats is a significant contrast to human and canine studies but in line with the reported incidence of 0.03% for feline adrenal neoplasia.10
Radiological findings indicating the presence of adrenal gland neoplasia include increased size, heterogeneity, irregular borders, and vascular invasion whether identified in ultrasonography, CT, or MRI.10 Although the mean right adrenal gland height and left adrenal gland width were higher and the range of dimensions was larger in the cats with possible adrenal disease compared with cats with likely normal adrenal glands, adrenal gland masses were not identified.
It is important to note that, as the measurements and characteristics of normal and diseased adrenal glands overlap, neoplasia or bilateral hyperplasia can remain occult on ultrasound and CT.19,20 Different body positioning can lead to differing adrenal size measurements, and this could have contributed to the wider size range in the PAD group.21 The absence of adrenal masses in this selected group of cats suggests that, in contrast to other species, adrenal masses rarely occur in cats. Their identification should prompt the clinician to perform further investigations to clarify the clinical importance.
Both groups consisted of cats undergoing CT for reasons unrelated to adrenal disease. The NAD group was recruited prospectively and screened for clinical signs and laboratory changes associated with adrenal disease. Thus, it is rather unlikely that adrenal disease was present within this group. Furthermore, mean adrenal gland size and standard deviation of the NAD group were consistent with a recent study in healthy cats.5 The fact that slightly lower dimensions were reported by Phoomvuthisarn and others could be the consequence of including juvenile cats in that study.6 Consistent with a recent healthy cat study, the right adrenal in the NAD group was significantly wider than the left.5 We hypothesize this is because of the arrowhead shape that predominantly occurred on the right side. Ultrasonographic studies do not report this difference in widths, probably due to the differing manner of measurement.4
Not unexpected, adrenal lesions were very rare in the PAD group and comparable to changes found in the NAD group. This can be explained by overall low incidence of feline adrenal disease including adrenal masses in cats and is consistent with recent ultrasonographic studies, in which aCTc of healthy cats were very similar to that of animals with nonendocrine diseases.4,10,22 Nevertheless, the likelihood to identify adrenal lesions was higher in the PAD group because this group was recruited retrospectively without prior screening for adrenal disease. In the NAD group, cats with diseases that could influence adrenal function or mineralization were excluded. Nevertheless, it is possible that nonadrenal diseases could affect adrenal size or other aCTc. This might have influenced the differences found between both groups.
Generally, the reported ultrasonographic dimensions are smaller than our findings.1,2,4,22 This is not surprising because measurements of abdominal organs in CT are regularly higher in direct comparison to ultrasound.23 We hypothesize that the reason for this is that it is easier to recognize the ideal plane for maximal dimension in multiplanar reconstructions of CT, unlike ultrasound, in which animal conformation and bowel gas can hinder evaluation.
The main differences between the present and earlier studies include the adrenal gland shape and the prevalence of mineralization. One recent healthy cat study reported no oval shaped right adrenal glands; conversely, another study reported all adrenal glands to be oval in shape.5,6 Consistent with the first study involving healthy cats, the arrowhead shape was almost entirely limited to the right adrenal gland.5 This contrasts with ultrasound findings, in which a similar shape distribution (elongated, oval, bilobed) between both adrenal glands was reported.1,2 In ultrasonographic studies, the arrowhead shape was not reported; consequently, this is another reason why the frequency of shapes/conformation differed to our groups.1,2,4 Multiplanar reconstruction may result in a differing assessment of shape than ultrasonographic evaluation (Figures 2A, B). Sonographic assessment of shape in the sagittal plane can be influenced by obliqueness.3 In contrast to CT studies involving healthy cats, mineralization was markedly more frequent in our study, consistent with ultrasound in which hyperechoic foci (suspicious for mineralization) occurred in 20% of diseased cats but only in 9% of healthy cats.4 The adrenal glands of the NAD group had mineralization, but there was no evidence of adrenal gland disease. This is consistent with histopathological studies that identified adrenal mineralization as mainly dystrophic, seldom related to adrenalitis, and not indicative of adrenal neoplasia.13,24 When adrenals glands with mineralization and as a consequence higher attenuation were excluded, adrenal gland attenuation of our groups was consistent with studies involving healthy cats.5,6
In this study, the associations of aCTc with signalment and adrenal gland position largely mirrored previous publications. In both groups, the position of the adrenals relative to the kidneys, VCcd, and Aabd was consistent with the previously published literature.6,25 In our study the left adrenal gland was bordering the VCcd with a higher frequency than the Aabd, which contrasts anatomical literature about dogs and cats and, to our knowledge, has not been reported for normal feline adrenal glands in CT studies.26,27 In agreement with several studies, body weight, BCS, and sex did not influence adrenal gland size and attenuation.2,4,5,22 Unlike an earlier CT study including healthy cats, wider adrenal glands were not identified in male cats in our populations.5 Regarding adrenal gland size, the sole moderate positive correlation was right adrenal gland height and age. This contrasts with ultrasound and CT studies in which a mild reduction of adrenal gland length with increasing age or no association with adrenal gland size was reported.3,5 Because of the scarcity of intact male and female cats, the impact of neutering was impossible to assess.
This study has several limitations. The main drawback is the lack of a suitable gold standard like histopathological analysis to prove normal adrenal structure. It is thus possible that we unintentionally included cats with various adrenal diseases. The NAD group was screened prospectively to the best of our ability with clinical and laboratory examinations. It is also possible that subtle structural changes within the adrenal parenchyma were not identified in CT in the PAD group. Contrast-enhanced CT may have increased the diagnostic value of this study. Given its retrospective nature, it was not possible to standardize CT studies for the PAD group. The variable body positioning during the CT scan and the few cases with slightly thicker slices (1 mm instead of 0.75 mm) may have influenced the measurements of the aCTc. However, as partial averaging decreases with thinner slice thickness, we believe the small difference of 0.25mm had minimal effect.28 Although the tube settings varied in the PAD group (80–180 mAs), the studies were of diagnostic quality, and thus difference in current was unlikely to account for any differences identified. The number of cats in the PAD group (n = 92) was considerable, but because of the low prevalence of feline adrenal masses and the applied selection criteria, higher numbers and a representative population are needed to assess the general prevalence of feline adrenal masses in CT reliably. The increasing usage of CT in veterinary medicine will likely facilitate the adequate study number for comparable studies in the future.
Conclusion
Adrenal gland size of cats with nonadrenal disease but no evidence of adrenal disease is similar to that of healthy cats when measured in CT. In contrast, adrenal gland shapes and their frequency differed partially, and adrenal gland mineralization was markedly more frequent than previously reported in healthy cats. Otherwise characteristics and correlations were largely similar. The fact that no adrenal gland mass was detected in both study groups might indicate that adrenal masses overall are rare in cats with no symptoms indicating adrenal disease. The identification of a so-called adrenal incidentaloma should prompt further investigation to clarify the clinical importance. The authors acknowledge the small number in the study population and that a larger study is needed to confirm the frequency and the character of structural adrenal changes in a larger feline population. Ideally, these should be supplemented by histopathological analyses.
Dorsal (A) and transverse (B) CT image of an oval right adrenal gland of a cat displayed in soft-tissue window. All distances for adrenal gland size were measured with a digital caliper in both planes. The maximal dimension was recorded for length (dotted white line), width (solid black line), and height (discontinuous white line). CT, computed tomography.
Dorsal (A) and sagittal (B) CT images of a right arrowhead shaped adrenal gland displayed in soft-tissue window. The arrowhead shape can be clearly seen in the dorsal plane in contrast to the sagittal plane where the cross-sectional area is oval and does not correspond to the three-dimensional shape of the adrenal gland. CT, computed tomography.
Dorsal CT images of the left adrenal gland (A) and both adrenal glands (B) of cats from the NAD group displayed in soft-tissue window. (A) The ROI outline is traversed by hand over the cross-sectional area of the adrenal, excluding vessels and gland borders to avoid partial volume effects and exclude periadrenal fat. The software calculates automatically attenuation values of the ROI in HU. (B) Bilateral mineralization of the adrenal glands. CT, computed tomography; HU, Hounsfield units; NAD, nonadrenal disease; ROI, region of interest.
Contributor Notes
