EVALUATION OF THE GASTROINTESTINAL TRACT IN DOGS USING COMPUTED TOMOGRAPHY

Abstract

Abdominal computed tomography (CT) studies of 19 dogs with no history or clinical signs of gastrointestinal disease, and two dogs with a histological diagnosis of gastrointestinal neoplasia were examined retrospectively. Gastrointestinal segments were evaluated subjectively for conspicuity, contrast enhancement, and wall layering after contrast medium administration. In dogs without gastrointestinal disease, there were 62.8% of gastrointestinal segments (serosa to serosa) and 77.7% of gastrointestinal walls (serosa to mucosa) visualized. Wall layering on postcontrast images was seen in 21.8% of gastrointestinal segments. There was significant association between gastrointestinal diameter and wall thickness. There was significant association between weight and gastrointestinal wall thickness in the following regions: gastric fundus, gastric body, gastric pylorus, gastric pyloric antrum, duodenal cranial flexure, jejunum and ascending colon, and between patient weight and gastrointestinal diameter in cranial duodenal flexure, descending duodenum, transverse duodenum, ascending duodenum, and jejunum. Measurements acquired from CT studies correlated well with previously published normal reference ranges for radiographic and ultrasonographic studies. Gastrointestinal neoplasia, diagnosed in two dogs, had a gastrointestinal wall thickness greater than the range of the dogs without gastrointestinal disease. Computed tomography offers identification of the gastrointestinal tract segments in dogs, allows for evaluation of gastrointestinal diameter and aids in investigation of gastrointestinal wall thickness.

Introduction

Gastrointestinal diameter and wall thickness are assessed frequently in dogs suspected of having bowel disease. Reference values have been established for normal radiographic gastrointestinal diameter.[1-3] Similarly, normal ultrasonographic gastrointestinal wall thickness has been characterized and jejunal and duodenal thickness related to patient weight.[4-6] Limitations of these imaging modalities include superimposition of abdominal contents in radiographs and limited sonographic assessment of deep structures due to bowel gas.[5]

Computed tomography (CT) has been applied in the dog to evaluate the liver, spleen, adrenal gland, lymphatic system, vascular system, pancreas, and urinary system.[7-13] We found no information related to the use of CT to evaluate the canine gastrointestinal tract. We hypothesized that pre- and postcontrast CT of the gastrointestinal tract will (1) allow identification of all abdominal gastrointestinal segments, (2) depict gastrointestinal tract wall layers, and (3) allow assessment of gastrointestinal diameter and wall thickness, and that these measurements will be dependent on age and weight, and correlate with known ranges determined using radiology and ultrasonography. We further hypothesized that these parameters could be applied to patients with histopathologically diagnosed gastrointestinal neoplasia, to assess changes to gastrointestinal wall thickness and changes to gastrointestinal wall layering.

Materials and Methods

Medical records of 19 dogs that underwent CT examination of the abdomen between November 2007 and July 2011 were reviewed. Dogs were selected for analysis if there was (1) no history or clinical sign of gastrointestinal disease, (2) no abdominal discomfort or palpable abnormalities, (3) no significant hematologic or blood chemistry abnormalities, and (4) a final diagnosis of a nongastrointestinal disease. The medical records of two clinical patients with confirmed gastrointestinal disease were also reviewed.

Of dogs without gastrointestinal disease, breeds were Yorkshire terrier (5), mixed breed (3), Labrador retriever (2), bichon frise (2), and one each of beagle, Chihuahua, Scottish terrier, miniature schnauzer, Drahthaar, American Staffordshire terrier, and Doberman pinscher. Age ranged from 6 months to 12 years (median 21 months), and weight from 1.8 to 36.5 kg (median 6.5 kg). There were 12 males (six neutered, six intact), and seven females (six neutered, one intact). These dogs were examined for suspicion of portosystemic shunt (12), renal infarction (1), subcutaneous mass (2), adrenal disease (1), osteomyelitis (1), and a pulmonary mass (1). These patients were diagnosed with normal abdominal CT studies (15), urinary bladder mass (1), cutaneous mass (2), and pulmonary neoplasia (1).

Computed tomography examinations were performed with the dogs under general anesthesia maintained with isoflourane. A single detector row helical CT unit1 was used in helical acquisition mode. A bolus of 100% contrast medium (300 mg I/kg IV) was injected through a peripheral intravenous catheter after inducing short-term apnea with manual hyperventilation. Dogs without gastrointestinal disease were in sternal 12 of 19 (63.1%), dorsal 6 of 19 (31.6%), or lateral 1 of 19 (5.3%) recumbency. A 2–5-mm slice thickness and a helical reconstruction interval were used. Clinically affected patients were in dorsal recumbency, with a slice thickness of 3–5 mm. A medium frequency reconstruction algorithm was employed for all dogs.

For the purpose of gastrointestinal segment identification, a window width of 400 and a window level of 40 were used to standardize viewing conditions. Images acquired both pre- and postintravenous contrast administrations were evaluated. The gastrointestinal tract was divided into 13 anatomic segments: gastric fundus, gastric body, gastric pylorus, gastric pyloric antrum, cranial duodenal flexure, descending duodenum, transverse duodenum, ascending duodenum, jejunum, ileum, ascending colon, transverse colon, and descending colon. Each segment was stated to be identified if the portion of the gastrointestinal tract could be identified from serosa to serosa, such that the entire diameter of a gastrointestinal tract segment could be identified.

For the purposes of gastrointestinal wall identification, each segment was examined for wall conspicuity; using images acquired both pre and postintravenous contrast administration. A gastrointestinal segment wall was stated to be identified if the wall could be distinguished from serosa to mucosa. Contrast enhancement of the individual segment walls was evaluated subjectively, as well as the conspicuity of individual wall layering. Each segment wall was examined for contrast enhancement, with contrast enhancement stated to be either present or absent. Contrast enhancement of a segmental wall was stated to be present if there was subjectively increased attenuation within the wall on postcontrast images that was not present on pre-contrast images.

Measurements of the diameter and wall thickness of each segment of the gastrointestinal tract were made using electronic calipers, using postcontrast studies. Individual gastrointestinal tract segment diameters were measured from serosa to serosa, and wall thickness measured from serosa to mucosa (Fig. 1). Measurements of the individual segments were acquired in a plane as close as possible to short axis. Diameter and wall thickness of each segment were acquired at up to eight points along the length of each segment, and a mean value obtained for each segment.

Figure 1. Postcontrast transverse CT image illustrating the method used for the measurement of wall thickness (A, serosa to mucosa) and diameter (B, serosa to serosa) of a jejunal loop.

The small intestinal diameter was compared subjectively to the height of the mid-body of the fifth lumbar vertebra, and the colon diameter was compared subjectively to the length of the seventh vertebral body.[1-3] The wall thickness was compared subjectively to previously published reference values for gastrointestinal wall thickness measurements obtained ultrasonographically.[4-6] The obtained measurement values of diameter and wall thickness were compared to patient age and patient weight using a one-way analysis of variance. Patients were divided into previously published weight-based reference ranges. The Wilcoxon Rank Sum test was used to determine whether gastrointestinal wall thickness and diameter measurements were related to body weight. Significance level was set at P < 0.05.

The medical records of two clinically affected patients were reviewed. Gastrointestinal segments were examined as previously described, and obtained values were compared to those of the nonaffected population.

Results

There was subjective contrast enhancement of the intestinal wall, facilitating increased conspicuity of the wall. Identification of gastrointestinal segments was possible in 155 of a total of 247 segments (62.8%), by identifying serosa to serosa. In several gastrointestinal segments identifying serosa to serosa was difficult, due to serosa of one segment being continuous with the serosa of an adjacent segment, or intestinal loops being collapsed. However, one aspect of the gastrointestinal wall (mucosa to serosa) could be identified, allowing for an overall gastrointestinal wall identification of 192 of 247 (77.7%). The 55 of 247 (22.3%) segments whose walls were unidentified included: pylorus 1 of 19 (5%), pyloric antrum 5 of 19 (26.3%), cranial duodenal flexure 4/19 (21%), descending duodenum 4 of 19 (21%), transverse duodenum 7 of 19 (36.8%), ascending duodenum 7 of 19 (36.8%), jejunum 8 of 19 (42.1%), ileum 14 of 19 (73.6%), descending colon 4 of 19 (21%), and transverse colon 1 of 19 (5%).

Distinct wall layers were conspicuous subjectively in postcontrast images in 54 of 247 (21.8%) segments. Distinct wall layering was seen most often in the stomach in 7 of 19 patients, followed by jejunum 6 of 19 patients and less commonly in the ileum in 2 of 19 patients and duodenum in 2 of 19 patients, with wall layering seen in two regions in four patients (Figs. 2-5). Individual wall layers, such as serosa, muscularis, submucosa, and mucosa were indistinguishable. Patient age was not associated with gastrointestinal wall thickness or gastrointestinal diameter.

Figure 2. Postcontrast transverse CT image wherein distinct contrast enhancement of the wall layering of the gastric body can be seen.

Figure 3. Postcontrast transverse CT image wherein jejunal mucosal (arrowheads) and serosal layering (arrows) are discernible.

Figure 4. Postcontrast transverse CT image. Note the distinct contrast enhancement of duodenal wall layers (arrow).

Figure 5. Postcontrast transverse CT image. Note the homogeneous contrast enhancement of colon wall without discernible wall layers.

Initially, patients were divided into two weight groups using an arbitrary cut-off value of 9 kg to obtain an equivalent number of dogs in each group. The small number of patients prevented division into groups similar to previously published weight groups.[4]

Significant association was present between patient weight and diameter as well as patient weight and wall thickness of individual gastrointestinal segments. Diameter of the cranial duodenal flexure, descending duodenum, transverse duodenum, ascending duodenum and jejunum was increased in patients weighing greater than 9 kg, compared to patients weighing less than 9 kg. Wall thickness of the gastric fundus, gastric body, gastric pylorus, gastric pyloric antrum, cranial duodenal flexure, jejunum and ascending colon was significantly increased in patients weighing greater than 9 kg, compared to patients weighing less than 9 kg (Tables 1 and 2).

Table 1. Regions where Wall Thickness (mm) Was Related to Weight (Median and Interquartile Range)

	0–9 kg	9 kg	P value
Gastric fundus	1.17 (0.98–1.30)	2.63 (2.0–3.08)	0.001
Gastric body	1.45 (1.27–1.59)	3.85 (2.75–4.10)	<0.001
Gastric pylorus	1.75 (1.43–2.13)	3.31 (2.92–3.55)	<0.001
Pyloric antrum	1.86 (1.69–2.74)	3.90 (2.95–4.40)	0.019
Ascending duodenum	3.22 (2.59–3.30)	4.84 (4.61–5.26)	0.026
Jejunum	2.58 (2.00–2.93)	3.62 (3.31–3.87)	0.024
Ascending colon	1.30 (1.05–1.45)	2.08 (1.78–2.35)	0.004

Table 2. Regions where Segment Diameter (mm) was Related to Weight (Median and Interquartile Range)

	0–9 kg	9 kg	P value
Pyloric antrum	10.15 (9.69–12.00)	15.24 (13.54–16.16)	0.045
Duodenal cranial flexure	9.23 (8.40–9.83)	13.20 (11.58–16.47)	0.006
Descending duodenum	10.04 (8.79–10.46)	11.75 (11.60–16.33)	0.001
Transverse duodenum	8.57 (7.35–9.31)	11.10 (10.70–13.28)	0.007
Ascending duodenum	9.10 (8.43–9.98)	14.00 (12.00–15.37)	0.001
Jejunum	7.60 (6.95–8.21)	10.10 (9.00–12.53)	0.004
Transverse colon	10.11 (9.00–11.93)	17.29 (12.84–21.49)	0.009

The ratio of gastrointestinal diameter to vertebral length determined using CT fell within previously published radiographic reference ranges.[1-3] There was partial agreement between CT derived values of wall thickness compared to previously published sonographic reference ranges. All values fell within reference ranges except for two small intestinal segment thickness values and a colonic wall thickness value which were outside the previously published reference ranges, but were within the published standard deviation of the results.[1-3]

Clinically affected patients were one male Labrador retriever dog weighing 34 kg, and one male Bulldog crossbreed dog weighing 26 kg. The Labrador retriever (patient 1) had a heterogeneously contrast enhancing mural mass located within the dorsal aspect of the fundic wall. Fundic wall thickness at the level of the mass was 13.7 mm with disruption, but there was not complete loss of wall layering on postcontrast images (Fig. 6). The histopathologic diagnosis was gastrointestinal stromal tumor, suggestive of leiomyosarcoma. Thus, fundic wall thickness was greater than the range of 2.00–3.08 mm. The bulldog cross-breed (patient 2) had a circumferential thickening of a section of jejunum, with wall thickness of between 9 and 10.5 mm, and jejunal diameter of 26.5 mm. Wall layering was indistinct, similar to adjacent small intestinal loops in this patient. The histopathologic diagnosis was jejunal adenocarcinoma, with associated lymphocytic-plasmacytic enteritis with mucosal hyperplasia of the affected jejunal segment and adjacent jejunal loops. Thus jejunal wall thickness was greater than range of 3.31–3.87 mm, and diameter greater than range of 9.00–12.53 mm (Fig. 7).

Figure 6. Postcontrast transverse CT image; patient is in dorsal recumbency. There is a gastric fundic stromal mass (arrow) wherein distinct wall layers are indistinguishable.

Figure 7. Postcontrast transverse CT image. There is heterogeneous contrast enhancement of a jejunal adenocarcinoma (arrow), without distinguishable wall layering.

Discussion

For gastrointestinal segment detection, abdominal CT allowed distinct identification of particular gastrointestinal segments in 62.8%, using serosa-to-serosa definition. However, 77.7% of the gastrointestinal segments were detected using serosa-to-mucosa identification of a single wall. The ileum, pyloric antrum, descending colon, pylorus, and transverse colon were difficult to identify in a few dogs because juxtaposition of adjacent mesentery and other gastrointestinal segments caused border effacement of the serosal margins. This was the most prominent at the lesser curvature of the stomach and the ileum. The descending and transverse colon were empty in one study each with surrounding contrast-enhancing mesentery, preventing identification of serosal margins. Similarly, ability to identify serosal margins of one wall is possible in an increased number of segments compared to ability to identify serosal margins of two walls.

Contrast enhancement was identified in 77.7% of gastrointestinal segments. Distinct wall layering was present in only 24% of these segments, although individual wall layers as described in sonographic evaluation could not be distinguished. We suspect that this is related to voxel size and spatial resolution. Computed tomography is not used in human patients to identify gastrointestinal wall layering due to reduced resolution compared to ultrasonography.[14-16]

Gastrointestinal diameter and wall thickness values were not significantly associated with patient age, consistent with previous studies.[4, 5] After an arbitrary division of patients into groups above and below 9 kg, weight was significantly associated with gastrointestinal wall thickness in the gastric fundus, gastric body, gastric pylorus, gastric pyloric antrum, ascending duodenum, jejunum, and ascending colon (Table 1), and with gastrointestinal diameter in the cranial duodenal flexure, descending duodenum, transverse duodenum, ascending duodenum, and jejunum (Table 2).

The gastrointestinal diameter values acquired on CT are within previously published radiographic reference ranges. Gastrointestinal wall thickness measurements were divided into previously published weight ranges used for ultrasonographic examination. Computed tomography measurements of wall thickness did not correlate well with the previously published ultrasonographic values, and as such reference ranges obtained from ultrasonography cannot be directly applied to CT. This could be secondary to volume averaging artifact, or inclusion of extraintestinal mesenteric contrast enhancing tissue in measurements of intestinal wall thickness.

The neoplastic lesions described in patient 1 and patient 2 led to a gastrointestinal wall thickness greater than the range of values obtained using CT in the unaffected population, with patient 1 having altered wall layering. Patient 2 had increased jejunal diameter.

Unfortunately, 22.3% of the total gastrointestinal segments were not identified using CT in this study. We suspect that certain gastrointestinal segments were collapsed or empty, and as such could not be identified. Intraluminal gas, intraluminal fluid, or intraluminal contrast medium may aid distinction of specific segments and their wall. In human patients, 1.3–2 l of water, methyl-cellulose solution, polyethylene glycol solution, or low concentration barium is administered orally over 1 h before the CT study.[17, 18] Administering similar contrast media to canine patients may be useful to improve identification of the gastrointestinal tract. Comparing images obtained pre- and postintravenous contrast medium administration would also be useful to assess if there is a significant difference in segments detected and measurement values obtained.

Measurements acquired during this study were obtained by one author (S.H.), at a single time point, who was aware of the final diagnosis/outcome. As such parameters, such as intraobserver and interobserver variance, was not investigated. The clinically unaffected patients in this study did not have histopathologic confirmation of normalcy. The patient base was also predominantly large breed and small breed dogs, reflecting our patient population.

Our results demonstrate the utility of CT for examination of the gastrointestinal tract, allowing identification of gastrointestinal segments, and allowing assessment of gastrointestinal diameter and wall thickness.

1. 1

   GE HiSpeed LXi, General Electric Healthcare, Milwaukee, WI.