Computed Tomographic Diagnosis of Nongastrointestinal Foreign Bodies in Dogs

Introduction

Foreign bodies are an important cause of recurrent abscesses, draining tracts, or pyogranulomas in dogs.^1–19 Types of nongastrointestinal foreign bodies include wood, plant material, microchips, retained surgical sponge, bone, needles, and hooks. The prognosis for affected dogs is good if the foreign bodies can be accurately localized and completely removed.^{1–3,8–10,13,18} In dogs, the initial diagnostic imaging tests most commonly recommended for localizing superficial foreign bodies are radiography, ultrasonography, and sinography.^6–8,13,15

In humans, computed tomography (CT) is recommended for treatment planning when foreign bodies are in complex anatomical regions and/or when they are surrounded by air or bone.^20–30 Computed tomography is a noninvasive imaging technique that uses X-rays and computers to create sectional images of structures.³¹ One advantage of CT over radiography is the ability to remove superimposition of overlying bony structures. Because images are acquired digitally, they can be displayed using various settings (windows) to best demonstrate the structure of interest. Reformatting software programs also allow creation of three-dimensional images that can be rotated in space. Tissue densities can be measured using numerical values called Hounsfield units, which are units of density relative to water.³¹ Only a few individual case reports could be found describing CT diagnosis of nongastrointestinal foreign bodies in dogs.^1,2,10 The purpose of this retrospective study is to describe the clinical and CT characteristics of confirmed foreign bodies in dogs.

Materials and Methods

Medical records were searched from January 1, 1995 to December 31, 2005 for dogs that had a final diagnosis of a nongastrointestinal foreign body and were examined with CT. Cases with a diagnosis of foreign body confirmed by direct visualization and/or histopathology were included in the study. Clinical characteristics retrieved from medical records included age, breed, sex, clinical signs, location of foreign body, method of final diagnosis, final diagnosis, and outcome. Complete digital CT image files for each case were reloaded onto a reformatting workstation^a and reviewed jointly by two people. The evaluators were aware of the location and type of foreign body present. Technical parameters, such as type of CT scanner, slice thickness and interval, and type of contrast procedure (if available) were recorded for each case. A consensus opinion was reached for describing CT characteristics, such as foreign body size, shape, opacity, and margination; mean CT density of the foreign body; and the CT display setting that best demonstrated the foreign body. Other findings recorded were the appearance of secondary reactions, appearance of the foreign body in post-contrast images (if available), presence or absence of adjacent bony lysis or proliferation, and usefulness of three-dimensional reformatting.

Results

Thirteen dogs met the criteria for inclusion in the study. Ages ranged from 1 to 13 years (median=4.9 years). Represented breeds were highly variable [Table 1]. Affected dogs included six males (two castrated) and seven females (four spayed). Dogs with nasal foreign bodies most commonly exhibited chronic nasal discharge (n=7). Other clinical signs included sneezing (n=3), pawing at the face (n=1), and stertor (n=1). Dogs with nonnasal foreign bodies had a localized soft-tissue swelling (n=2), chronic draining tract (n=1), and head tilt (n=1). Most foreign bodies were identified in the nasal cavity or nasopharynx (n=9). Other foreign bodies involved the thoracic wall (n=1), brain (n=1), and retropharyngeal region (n=2).

Most foreign bodies were of plant origin (n=10). Four dogs had small pieces of plant material confirmed primarily by histopathology. Large plant components were confirmed by surgery (n=1), blind nasal biopsy with forceps (n=1), or rhinoscopy (n=4) in the remaining six dogs. Three of the large plant foreign bodies were wooden sticks. Other large plant foreign bodies were grass blades (n=2) and evergreen sprigs (n=1). Nonplant foreign bodies were identified primarily by direct visualization (rhinoscopy or oropharyngeal examination), and these included toy stuffing material (n=1), tennis ball fibers (n=1), and a needle (n=1).

For those dogs with known outcomes (n=11), most had resolution of their clinical signs after treatment (n=8). One dog with a plant foreign body granuloma in the cerebellum was euthanized because of severe neurological deficits and a poor prognosis for recovery. One dog with a plant foreign body pyogranuloma of the thoracic wall suffered recurrent abscesses after surgical excision. After one abscess ruptured, a segment of juniper branch was found at the site. One dog with rhinitis secondary to a mixture of plant material, hair, and keratin was reexamined for nasal discharge 1 year after treatment. A second endoscopic examination and repeated nasal flushings did not yield any additional foreign material, and further outcome information was unavailable.

Computed tomography scans were performed with a fourth-generation scanner^b in eight dogs and a spiral scanner^c in five dogs. Slice thicknesses ranged from 2 to 5 mm, with intervals ranging from 1.5 to 5 mm. Intravenous contrast-enhanced CT was performed in three dogs, and CT sinography was performed in one dog. Small pieces of plant were not identified on CT in four dogs with confirmed plant foreign bodies [Table 2]. The secondary reaction in these dogs appeared similar to fungal disease or neoplasia [Figures 1A–1C].^32,33 Larger plant foreign bodies were distinguishable in six dogs, with recognition based primarily on shape and/or internal architecture. Most of the large plant foreign bodies were best seen with lung or soft-tissue window settings.³¹ In one dog (case no. 9), an intranasal wooden stick exhibited a “target” shape in the transverse plane and a tubular shape in the sagittal and dorsal planes [Figures 2A–2C]. In another dog (case no. 10), wooden stick foreign bodies in the retropharyngeal region exhibited a striated internal architecture [Figure 3A]. The locations of the foreign bodies and associated sinus tract, relative to palpable bony landmarks, were best delineated in three-dimensional, reformatted images [Figures 3B, 3C]. In case no. 12, cloth stuffing material appeared as a homogenous, nonenhancing, focal area of increased soft-tissue opacity and was similar to an area of fluid accumulation [Figure 4A].³¹ Tennis ball fibers in one dog (case no. 11) were visible as fusiform to oblong areas of air opacity within a region of soft-tissue opacity, and these were best seen with the lung window setting [Figure 4B].³¹ Awareness of the final diagnoses and the unusual shapes helped to differentiate these foreign bodies from normal gas pockets. A retropharyngeal needle foreign body was clearly distinguishable in one dog (case no. 13). The linear shape of the needle was best appreciated with a wide, bone window setting [Figure 5A].³¹ Radiating streak artifacts and the secondary soft-tissue swelling were best seen in a soft-tissue window setting [Figure 5B].³¹ The orientation of the needle, relative to the hyoid bones and other palpable bony landmarks, was best appreciated in three-dimensional reformatted images [Figure 5C].

Discussion

The goal of this study was to gain a better understanding of the CT appearances of nongastrointestinal foreign bodies in dogs; therefore, the people evaluating the CT scans were aware of the final diagnoses. Even with this prior knowledge, however, foreign bodies were not distinguishable in five dogs, and the secondary reactions that occurred in four dogs were considered similar to neoplastic or fungal disease. These similarities to neoplastic or fungal disease were in agreement with previous reports in humans and dogs.^10,32,33 Petersen et al. reported that foreign body reactions in humans can exhibit a pseudotumor appearance.³² Leskovar et al. described a suspected neoplasm that was ultimately diagnosed as a plant foreign body granuloma on the lateral margin of the medulla and cervical spinal cord of a dog.¹⁰ Saunders et al. described the presence of nasal foreign bodies in two of 35 dogs with mycotic rhinitis.³³

For those foreign bodies that were identified with CT, the use of multiple window settings was helpful. The observation that some foreign bodies were best seen with the lung window setting was consistent with previous case reports in humans, where wood was found to mimic focal gas pockets in soft tissues.^28,34–39 Identification of the wood foreign bodies in these cases was based primarily on a linear shape or orientation along a plane that did not match a known anatomical structure. The characteristic architecture of wooden sticks in some dogs was best seen using soft-tissue window settings, which was also similar to previous reports in humans.^1,25,32 In these human cases, wooden sticks appeared to have increased opacity because of their high inherent density or absorption of body fluids. Diagnoses were based primarily on the presence of internal striations or a “target” shape in transverse images.^1,25,32,33 The target shape was believed to be caused by inflammatory tissue formation on the surface of the wood.³³ Streak artifacts from a needle foreign body in one dog (case no. 13) obscured visualization of its shape in all but the widest available bone window setting. This finding was similar to a previous case series of known metal foreign bodies in the human chest.⁴⁰

For one dog (case no. 10) with chronic draining tracts from retropharyngeal wooden foreign bodies, the relationship of sinus tracts to palpable surgical landmarks was best demonstrated using CT sinography and three-dimensional reformatted images. These findings were consistent with a previous case report of a dog with a bronchoesophageal fistula.⁴¹ Authors of this case report concluded that CT sinography with three-dimensional reformatting assisted surgical planning by demonstrating the relationship between the subcutaneous neck abscess, left cranial lobar bronchus, adjacent cervical vertebrae, and thoracic wall.

Conclusion

Based on the 13 dogs in this study, some foreign bodies may not be distinguishable with CT. Secondary reactions from foreign bodies may also mimic neoplastic or fungal disease. Foreign bodies with a characteristic shape or architecture are more likely to be distinguished with CT, especially if images are viewed with soft-tissue, bone, or lung window settings. Three-dimensional CT reformatting was a useful tool for demonstrating the relationship between foreign bodies, related sinus tracts, and palpable bony landmarks.