Epistaxis and Destructive Computed Tomographic Features in Canine Nasal Transitional Carcinoma: Frequent Cribriform Plate Destruction but Uncommon CT-Detected Brain Parenchymal Involvement

Simple Summary

Nasal tumors in dogs often present with non-specific clinical signs, making early recognition and characterization challenging. Computed tomography (CT) is valuable for defining lesion extent, regional bone involvement, and planning diagnostic procedures, but CT findings alone generally cannot establish tumor histotype. In this retrospective multicenter study, we evaluated the clinical signs and CT imaging features of histopathologically confirmed nasal transitional carcinoma (TC) in 16 dogs. Epistaxis was the most common clinical sign. On CT, tumors consistently involved the caudal nasal cavity and commonly showed destructive growth, regional extension to adjacent structures, cribriform plate lysis, intralesional non-enhancing fluid-attenuating regions, and peritumoral fluid accumulation. In contrast, CT-detected brain parenchymal involvement was uncommon despite frequent cranial vault invasion. These findings describe frequent CT patterns in canine nasal TC rather than features that can reliably distinguish it from other nasal neoplasms. Histopathological examination remains necessary for definitive tumor-type identification.

Abstract

Nasal transitional carcinoma (TC) is one of the most common epithelial nasal tumors in dogs, but systematic descriptions of its clinical and computed tomographic (CT) features are limited. This multicenter retrospective study characterized the clinical and CT features of histopathologically confirmed nasal TC in 16 dogs examined at four animal hospitals between July 2014 and January 2024. Signalment, clinical signs, CT findings, and follow-up information were summarized descriptively. Epistaxis was the most common clinical sign (12/16, 75%). All tumors involved the caudal nasal cavity, and extension to adjacent structures or most commonly affected the maxillary recess (14/16, 88%), orbit (11/16, 69%), and nasal choanae (10/16, 63%). Most tumors exhibited destructive growth (15/16, 94%) with frequent cribriform plate lysis (12/16, 75%). Cranial vault invasion was identified in 12 dogs (75%), whereas CT-detected brain parenchymal involvement was present in only 2 dogs (13%). Intralesional non-enhancing fluid-attenuating regions consistent with necrosis or cystic change and peritumoral fluid accumulation were observed in 12 (75%) and 10 dogs (63%), respectively, and intralesional calcification was absent in all dogs. These results provide a frequency-based CT reference pattern for canine nasal TC; however, the findings are not specific for this tumor type, and histopathological examination remains necessary for definitive diagnosis.

1. Introduction

Sinonasal neoplasms account for approximately 0.08–2% of all canine tumors, with those of epithelial origin being the most prevalent [1]. Adenocarcinoma is the most common, followed by transitional carcinoma (TC) and squamous cell carcinoma [2]. The clinical diagnosis of nasal tumors remains challenging owing to their frequently non-specific clinical signs [2]. These signs often overlap with those of other intranasal conditions, including fungal or bacterial infections, inflammatory rhinitis, intranasal foreign bodies, and developmental anomalies [3]. Therefore, a comprehensive diagnostic approach combining clinical examination, advanced imaging, and cytologic or histopathologic confirmation is essential.

Computed tomography (CT) is the preferred tool for evaluating nasal disease and differentiating neoplastic from non-neoplastic conditions because of its high spatial resolution and its ability to facilitate detailed assessment of lesion extent, bony involvement, and regional or distant metastasis [4]. CT serves as a valuable guide for biopsy procedures and radiation therapy planning. Although CT has substantially improved the diagnostic evaluation of nasal neoplasms, relatively few studies have systematically investigated its imaging features across distinct tumor types [4].

Among these tumors, nasal TC has been reported as the second most common canine nasal tumor; however, studies focusing on its clinical presentation and CT imaging features are scarce. The limited available literature suggests that nasal TC can affect middle-aged to older dogs and is often associated with non-specific clinical signs. Existing case-based reports also suggest that it is locally invasive and may present as an extensive soft-tissue mass with bilateral nasal cavity involvement, frontal sinus extension, retrobulbar invasion, marked regional bony destruction, and variable degrees of intracranial involvement [5,6,7,8]. However, no systematic study with a larger case series has comprehensively characterized the CT features of canine nasal TC.

Therefore, this retrospective study aimed to systematically characterize the clinical and CT imaging features of histopathologically confirmed nasal TC in dogs and to provide a frequency-based reference for lesion assessment and diagnostic planning. Because no comparison group of other nasal tumor types was included, the study was not designed to determine the specificity of the described CT findings for nasal TC.

2. Materials and Methods

2.1. Case Selection

This multicenter, retrospective case series study included client-owned dogs from four animal hospitals in Taiwan, examined between July 2014 and January 2024. The included dogs underwent CT of the head and/or body, and nasal TC was diagnosed on histopathological examination. The exclusion criteria were incomplete medical records, inadequate imaging data, cytological examination only, or non-specific carcinoma diagnosis based on histopathology. No a priori sample size calculation was performed due to the retrospective nature of the study. All cases that met the inclusion criteria during the study period were included. No additional procedures were performed beyond those required for clinical diagnosis and treatment. Details of each veterinary hospital and its imaging equipment are provided in Table S1.

2.2. Medical Record Review

Signalment (age, sex, neutering status, weight, breed), medical history, clinical signs (types, duration), biopsy methods, bacterial cultures, antibiotic susceptibility tests, treatment, and follow-up information were collected from each dog’s medical records.

2.3. Imaging Acquisition

All dogs underwent CT in sternal recumbency under general anesthesia. Pre- and post-contrast head CT images were available for all dogs. In all dogs, post-contrast images were acquired 90–120 s after intravenous administration of 700 mg I/kg of iohexol (Omnipaque^® 350; GE Healthcare Ireland, Cork, Ireland) or ioversol (Optiray^® 350; Guerbet LLC, Plainville, CT, USA). Slice thickness ranged from 0.5 to 2.0 mm, depending on patient size and center-specific protocols. Images were reconstructed using soft-tissue and bone algorithms. Thoracic or abdominal CT scans obtained during the same session were evaluated, when available, using soft-tissue, lung, and bone algorithms. CT manufacturer/model, detector configuration, and case-specific acquisition parameters are provided in Tables S1 and S2.

2.4. Image Analysis

All CT images were reviewed and evaluated qualitatively and quantitatively by two authors (L.-K.N. and L.-S.L.) in consensus using commercial medical imaging software (OsiriX MD, version 12.5.3; Pixmeo SARL, Bernex, Switzerland). Because eligibility was established before image review, the reviewers were aware that all included dogs had histopathologically confirmed nasal TC. The reviewers did not use detailed histopathological findings, treatment information, or follow-up outcomes during CT feature assessment. The size of each nasal mass was measured using multiplanar reformatted images. The nasal cavity was divided into rostral, middle, and caudal portions according to a previously published classification [9]. The rostral portion extended from the nostrils to the first upper premolar, the middle portion from the first upper premolar to the rostral root of the fourth premolar, and the caudal portion from the rostral root of the fourth premolar to the cribriform plate. Tumor growth was categorized as turbinate destruction, permeative, expansile, or destructive, using previously described criteria [10].

Mass homogeneity was assessed subjectively on pre- and post-contrast images. To quantify contrast enhancement, three circular regions of interest, each with an area of 5 mm², were manually placed on six separate CT slices while avoiding vascular structures and non-enhancing regions. Mean Hounsfield unit (HU) values were calculated, and enhancement was categorized as mild (5–25%), moderate (25–75%), or strong (>75%) according to the percentage increase from pre-contrast images. Hyperattenuating foci on pre-contrast images were recorded as intralesional calcification. Non-enhancing fluid-attenuating regions within a mass were recorded as intralesional necrosis or cystic change, and peritumoral fluid accumulation and its extent were also recorded. CT-based modified Adams stage was assigned as follows: stage 1, tumor confined to a nasal passage, paranasal sinus, or frontal sinus without bone involvement beyond the turbinates; stage 2, bone involvement beyond the turbinates without orbital, subcutaneous, or submucosal extension; stage 3, orbital, nasopharyngeal, subcutaneous, or submucosal extension; and stage 4, cribriform plate lysis [11]. When more than one stage criterion was present, the highest applicable stage was assigned. When available, thoracic and abdominal CT images were evaluated for pulmonary nodules, lymphadenopathy, and other concurrent findings. For dogs with follow-up CT examinations, changes were determined by comparison with the initial examination.

2.5. Statistical Analyses

Statistical analyses were performed by LKN using a spreadsheet (Microsoft^® Excel for Mac, version 16.78; Microsoft Corp., Redmond, WA, USA) and commercial software (Prism, version 9.5.0; GraphPad Software Inc., San Diego, CA, USA). The normality of continuous variables was assessed using the Shapiro–Wilk test. Variables with normal distributions were summarized as mean ± standard deviation, whereas those with non-normal distributions were reported as median and interquartile range. Categorical variables were summarized as n/N (%).

3. Results

3.1. Study Population

A total of 16 dogs met the inclusion criteria: 5 Shiba Inus, 3 mixed-breed dogs, 2 Chihuahuas, and 1 each of Pembroke Welsh Corgi, Cardigan Welsh Corgi, Siberian Husky, Miniature Poodle, Doberman Pinscher, and Shih Tzu. The dogs included 6 spayed females, 2 sexually intact females, 3 neutered males, and 5 sexually intact males. Their mean age was 11.5 ± 2.8 years (range, 7–16 years), and their mean body weight was 13.0 ± 7.8 kg (range, 3.1–30.0 kg).

3.2. Clinical Presentation and Clinical Examination

Signalment and clinical findings are summarized in Table 1, with detailed individual case information provided in Table S3.Epistaxis was the most prevalent clinical sign (12/16, 75%), followed by abnormal breathing sounds, nasal discharge, and facial deformity (each 6/16, 38%). The duration of clinical signs before presentation was available for 9 dogs and ranged from 7 to 270 days (mean, 112.7 ± 96.3 days). Bacterial cultures were performed in 5 dogs; 4 yielded positive results and 1 was negative.

Table 1.

Signalment and clinical findings in dogs with histopathologically confirmed nasal transitional carcinoma (n = 16).

Domain	Variable	Value
Signalment	Age, years	11.5 ± 2.8 (range, 7–16)
	Body weight, kg	13.0 ± 7.8 (range, 3.1–30.0)
	Spayed female	6/16 (38%)
	Sexually intact female	2/16 (13%)
	Neutered male	3/16 (19%)
	Sexually intact male	5/16 (31%)
Clinical signs	Epistaxis	12/16 (75%)
	Abnormal breathing sounds	6/16 (38%)
	Nasal discharge	6/16 (38%)
	Facial deformity	6/16 (38%)
	Nasal obstruction	5/16 (31%)
	Sneezing	4/16 (25%)
	Exophthalmos	2/16 (13%)
	Anorexia	2/16 (13%)
	Blindness, lethargy, respiratory distress, or reverse sneezing	1/16 (6%)
Other clinical information	Duration of clinical signs, days (n = 9)	112.7 ± 96.3 (range, 7–270)
	Bacterial culture performed	5/16 (31%)
	Positive bacterial culture among tested dogs	4/5 (80%)

Values are n/N (%) unless otherwise stated.

3.3. CT Imaging Findings

Computed tomographic findings are summarized in Table 2. Nasal TCs measured 6.36 ± 1.69 cm in length, 2.99 ± 1.53 cm in width, and 3.72 ± 1.15 cm in height. All tumors were diffusely distributed within the nasal cavity and involved the caudal portion (16/16, 100%). Middle and rostral nasal cavity involvement was observed in 15 (94%) and 9 (57%) dogs, respectively. Bilateral involvement was present in 10 dogs (63%), and all bilateral masses extended across the nasal septum.

Table 2.

Summary of computed tomographic findings in dogs with histopathologically confirmed nasal transitional carcinoma.

Domain	CT Finding	Dogs Affected
Tumor distribution	Caudal nasal cavity involvement	16/16 (100%)
	Middle nasal cavity involvement	15/16 (94%)
	Rostral nasal cavity involvement	9/16 (57%)
	Bilateral nasal cavity involvement	10/16 (63%)
Extension to adjacent structures	Maxillary recess involvement	14/16 (88%)
	Sphenoidal sinus involvement	10/16 (63%)
	Frontal sinus effusion	9/16 (56%)
	Frontal sinus mass invasion	5/16 (31%)
	Orbital extension	11/16 (69%)
	Nasal choanal extension	10/16 (63%)
	Paranasal soft-tissue extension	8/16 (50%)
	Oral cavity extension	7/16 (44%)
	Parafrontal extension	5/16 (31%)
	Nasopharyngeal extension	2/16 (13%)
Growth pattern and osseous changes	Destructive growth pattern	15/16 (94%)
	Turbinate-only destruction	1/16 (6%)
	Endoturbinate osteolysis	16/16 (100%)
	Ethmoid bone, vomer, or palatine bone osteolysis	13/16 (81%)
	Nasal septum osteolysis	12/16 (75%)
	Cribriform plate lysis	12/16 (75%)
Mass attenuation and contrast enhancement	Heterogeneous pre-contrast attenuation	11/16 (69%)
	Heterogeneous post-contrast enhancement	16/16 (100%)
	Moderate enhancement	10/16 (63%)
	Strong enhancement	6/16 (38%)
	Intralesional calcification	0/16 (0%)
	Intralesional non-enhancing fluid-attenuating regions	12/16 (75%)
	Peritumoral fluid accumulation	10/16 (63%)
Intracranial, nodal, and thoracic findings	Cranial vault invasion	12/16 (75%)
	Adjacent meningeal enhancement without CT evidence of brain parenchymal involvement	7/16 (44%)
	CT-detected brain parenchymal involvement	2/16 (13%)
	Midline shift	2/16 (13%)
	Regional lymphadenopathy	12/16 (75%)
	Pulmonary nodules (n = 14)	8/14 (57%)
Modified Adams stage	Stage 1	1/16 (6%)
	Stage 2	3/16 (19%)
	Stage 3	0/16 (0%)
	Stage 4	12/16 (75%)

Values are n/N (%) unless otherwise stated. CT, computed tomography.

Tumor extension to adjacent structures most commonly involved the maxillary recess (14/16, 88%), orbit (11/16, 69%), nasal choanae (10/16, 63%), and sphenoidal sinus (10/16, 63%). Frontal sinus effusion was detected in 9 dogs (56%), whereas mass invasion of the frontal sinus was identified in 5 dogs (31%). Other sites of extension included the paranasal soft tissues (8/16, 50%), oral cavity (7/16, 44%), parafrontal region (5/16, 31%), and nasopharynx (2/16, 13%).

Facial deformity was identified on CT in 11 dogs (69%), including exophthalmos in 4 dogs (25%) and mixed deformities in 7 dogs (44%). Most tumors (15/16, 94%) exhibited a destructive growth pattern; the remaining tumor was characterized by turbinate-only destruction (dog 14). Osteolysis of the endoturbinates was present in all dogs (16/16, 100%), and the ethmoid bone, vomer, and palatine bone were each affected in 13 dogs (81%). The nasal septum and cribriform plate were each affected in 12 dogs (75%). Hyperostosis or sclerosis of the incisive bone was observed in 2 dogs (13%), and mixed bony changes involving the frontal bone were observed in 2 dogs (13%). Representative CT images illustrating tumor growth patterns and bony destruction are shown in Figure 1. Detailed location-by-location bony changes are provided in Table S4.

Figure 1.

Representative bone-window CT images showing growth patterns and regional osteolysis. Images are shown in transverse (A–C) and sagittal (D) planes. (A) Nasal TC in the left nasal cavity with turbinate-only destruction in dog 14 (white arrowhead). (B) Destructive growth pattern with marked osteolysis of adjacent structures in dog 2 (white arrows). (C) Extensive osteolysis involving the orbital plate of the ethmoid bone, endoturbinates, vomer, nasal septum, and perpendicular plate of the palatine bone in dog 1 (dotted white arrows). (D) Lysis of the ventral cribriform plate (left dotted white arrow) and presphenoid bone (right dotted white arrow) in dog 4. Window width, 1500 HU; window level, 500 HU; slice thickness, 0.5–1.5 mm. CT, computed tomography; HU, Hounsfield unit; TC, transitional carcinoma.

Heterogeneous attenuation on pre-contrast CT was observed in 11 dogs (69%). All tumors enhanced after contrast administration and had heterogeneous post-contrast enhancement; moderate and strong enhancements were observed in 10 (63%) and 6 (38%) dogs, respectively (Figure 2A–D). Intralesional calcification was absent in all dogs. Intralesional non-enhancing fluid-attenuating regions consistent with necrosis or cystic change were recorded in 12 dogs (75%), and peritumoral fluid accumulation was recorded in 10 dogs (63%) (Figure 2E,F). In 6 of the 10 dogs with peritumoral fluid accumulation, the fluid extended into the paranasal or frontal regions.

Figure 2.

CT images showing variable degrees of contrast enhancement of nasal transitional carcinoma (A–D), intralesional non-enhancing fluid-attenuating regions consistent with necrosis or cystic change (E), and peritumoral fluid accumulation (F). Images are shown before contrast administration (A,C) and after contrast administration (B,D–F). (A,B) Moderate enhancement in dog 9. (C,D) Strong enhancement in dog 11. (E) A non-enhancing fluid-attenuating intralesional region in dog 2 (black dashed arrow). (F) Abundant peripheral non-enhancing fluid-attenuating material extending dorsally into the frontal subcutis in dog 16 (white dashed arrow). Window width, 400 HU; window level, 40 HU; slice thickness, 0.5–2.0 mm.

Cranial vault invasion was identified in 12 dogs (75%). Adjacent meningeal enhancement without CT evidence of brain parenchymal involvement was observed in 7 dogs (44%). CT-detected brain parenchymal involvement and associated midline shift were each identified in 2 dogs (13%) (Figure 3). According to the CT-based modified Adams staging system, 12 dogs (75%) were classified as stage 4, 3 (19%) as stage 2, and 1 (6%) as stage 1; no dogs were classified as stage 3. Regional lymphadenopathy was observed in 12 dogs (75%), most commonly affecting the mandibular and medial retropharyngeal lymph nodes. Only 1 dog (6%) had heterogeneous enhancement of an enlarged regional lymph node. Pulmonary nodules were detected in 8 of 14 dogs (57%) with thoracic CT images available. Most were mineral-attenuating foci (144–2780 HU on pre-contrast CT) considered compatible with pulmonary osseous metaplasia; at the initial CT examination, two dogs had enhancing soft-tissue-attenuating nodules (39–69 HU on post-contrast CT). Additional thoracic and abdominal findings are summarized in Table S5.

Figure 3.

Post-contrast CT images showing variable cranial vault invasion and intracranial changes. Images are sagittal in (A–C) and transverse in (D). (A) Cranial vault invasion without CT evidence of brain parenchymal involvement in dog 6 (dashed white arrow). (B) Cranial vault invasion with adjacent meningeal enhancement in dog 4 (black arrows). (C,D) Cranial vault invasion with CT-detected brain parenchymal involvement (black arrowheads, (C)) and rightward midline shift (black dashed arrow, (D)) in dog 1. Window width, 400 HU; window level, 40 HU; slice thickness, 0.5–1.5 mm. CT, computed tomography; HU, Hounsfield unit; TC, transitional carcinoma.

3.4. Follow-Up CT Examination

Overall, 7 dogs underwent 1–4 follow-up CT examinations (mean interval, 9.9 months; range, 82–673 days). Initial CT examinations and biopsies were inconclusive in 2 dogs, suggesting cellulitis in dog 11 and chronic lymphoplasmacytic rhinitis in dog 14; follow-up CT examinations and repeat biopsies subsequently confirmed nasal TC (Figure 4). Among the remaining dogs, dog 5 underwent follow-up CT because of recurrent clinical signs after radiation therapy, whereas dogs 3, 7, 10, and 15 underwent CT for treatment-response assessment. Tumor size increased in 5/7 dogs (71%), regional osteolysis became more severe in 6/7 dogs (86%), and cranial vault invasion increased in 4/7 dogs (57%). At follow-up, none of the five dogs with cranial vault invasion had CT evidence of brain parenchymal involvement; adjacent meningeal enhancement was observed in four of these dogs (Figure 5). Regional lymph node size increased in dogs 3 and 5. Among the four dogs with follow-up thoracic CT examinations, the number of pulmonary nodules increased in dog 14, decreased in dog 10, and was unchanged in dog 3. A pulmonary nodule was identified at follow-up in dog 5; however, because thoracic CT was not performed at the initial examination, interval development could not be determined. Detailed comparisons are provided in Table S6.

Figure 4.

Post-contrast transverse CT images of dogs 11 and 14: (A,C) initial; (B,D) follow-up (A,B). Follow-up scan of dog 11 shows a marked increase in tumor size with a broader margin (black arrowheads, (B)) and more extensive osteolysis of the surrounding bony structures (black arrow, (B)) relative to the initial CT scan. (C,D) Follow-up CT image of dog 14 reveals an enlarged nasal TC occupying the entire left nasal cavity (black dashed arrows (D)), compared with the initial CT image. Window width, 400 HU; window level, 40 HU; slice thickness, 1.0–2.0 mm. CT, computed tomography; TC, transitional carcinoma; HU, Hounsfield unit.

Figure 5.

Sagittal post-contrast CT images of dog 3 show increased severity of cranial vault invasion without apparent brain parenchymal involvement: (A) initial; (B) follow-up. (A,B) Follow-up CT revealed more extensive cranial vault invasion with adjacent meningeal enhancement but no CT evidence of brain parenchymal involvement (black arrows, (B)). Window width, 400 HU; window level, 40 HU; slice thickness, 0.5 mm. CT, computed tomography; TC, transitional carcinoma; HU, Hounsfield unit.

3.5. Treatment and Follow-Up

Treatment categories were not mutually exclusive. Five dogs (31%) received systemic chemotherapy using various protocols; dog 7 also underwent surgical debridement. Three dogs (19%) received piroxicam alone, and 1 dog underwent rhinoscopy-guided mass reduction combined with medical therapy. One dog received intratumoral para-toluenesulfonamide under rhinoscopic guidance, and 1 dog received radiation therapy. Three dogs (19%) received palliative treatment, whereas 2 dogs (13%) received no tumor-directed treatment.

Follow-up information was available for 9 dogs, including 8 dogs that died and 1 dog (dog 16) that remained alive at the final follow-up and was censored at 34 days after diagnosis. Survival times for the 8 dogs that died ranged from 7 to 728 days. Seven dogs (44%) were lost to follow-up, regardless of treatment status. Among the 8 dogs that died, 5 (63%) were euthanized, 2 (25%) died of unspecified causes reported by their owners, and 1 (13%) died of septicemia following cholecystectomy. The treatment protocols, survival times, and clinical outcomes for dogs with available data are provided in Table S7.

4. Discussion

To our knowledge, this is the first multicenter retrospective series focusing specifically on the CT features of histopathologically confirmed nasal TC in dogs. Previous information has largely been limited to individual case reports or studies that pooled several sinonasal tumor histotypes, precluding assessment of whether particular findings recur in nasal TC. By systematically evaluating tumor distribution, growth pattern, regional extension, bony changes, and CT-detected intracranial changes in 16 dogs, this study provides a frequency-based imaging reference for this uncommon tumor type. The principal contribution is not the identification of CT findings specific for nasal TC, because no comparison group was included, but the recognition of a recurring pattern: epistaxis as the predominant clinical sign, universal caudal nasal cavity involvement, predominantly destructive growth, frequent cribriform plate lysis, and comparatively uncommon CT-detected brain parenchymal involvement. This pattern may assist with lesion mapping, selection of biopsy targets, and planning of additional imaging, while histopathological examination remains essential for definitive tumor type identification.

The mean age of 11.5 years in this cohort is consistent with the predominance of older dogs among those affected by canine nasal epithelial neoplasms. Because the study lacked an unaffected comparison group, age could not be evaluated as an independent risk factor for nasal TC; however, the findings support consideration of this tumor in older dogs with persistent epistaxis, nasal discharge, abnormal respiratory sounds, or facial deformity. All tumors involved the caudal nasal cavity, and most extended into the middle and, less commonly, rostral nasal cavity. Although this distribution may be consistent with a transitional epithelial zone between rostral squamous and caudal respiratory epithelium, the site of tumor origin could not be determined with certainty from a single CT examination [2]. The frequent caudal involvement may also reflect progressive tumor extension into a region with greater space and less structural obstruction than the rostral nasal cavity [12]. Bilateral nasal cavity involvement was considered more likely to reflect local extension across the nasal septum than a truly multifocal origin, particularly given the high frequency of septal lysis and aggressive local growth. The maxillary recess and sphenoidal sinus may provide additional space for expansion, which may account for their frequent involvement [13].

Extension into the orbit was common in this cohort, emphasizing the locally invasive behavior of nasal TC. This observation is consistent with previous case reports of extensive orbital involvement [6]. In humans, both the anatomical origin and histologic type of sinonasal tumors influence the likelihood of orbital involvement [14]. In dogs, epithelial-origin neoplasms are more frequently associated with orbital involvement than non-epithelial neoplasms [15]. This histologic distinction may contribute to the high frequency of orbital extension observed in the present series.

Most nasal TCs in this study had a destructive growth pattern with extensive regional osteolysis, consistent with the locally aggressive behavior of epithelial-origin nasal neoplasms. CT growth patterns may help distinguish neoplastic from non-neoplastic nasal disease and may vary among tumor histotypes in both veterinary and human medicine [16,17]. Epithelial-origin tumors, including adenocarcinoma and TC, and some mesenchymal tumors, such as chondrosarcoma, can exhibit destructive growth [6,18,19,20]. The turbinate-only destruction pattern in dog 14 likely reflected an earlier or less locally extensive lesion. However, destructive growth is not exclusive to nasal TC. In the absence of a comparison group, the present findings should be interpreted as frequent CT features in nasal TC rather than as features that discriminate this tumor type from other nasal neoplasms.

Cribriform plate lysis was identified in 75% of dogs. This proportion appears higher than that reported in some previous studies of canine nasal carcinoma, but comparison should be made cautiously because those reports included mixed carcinoma populations and used different imaging technologies and protocols [11,20]. The observed difference may therefore reflect case composition and CT detection capability, although a tumor-specific biological difference cannot be excluded.

Despite frequent cribriform plate lysis, CT-detected brain parenchymal involvement and midline shift were each identified in only 13% of dogs. This frequency is lower than the 22–54% reported for other canine nasal neoplasms [21,22,23,24]. Because these previous studies included different tumor types and imaging modalities, and because CT has limited sensitivity for intracranial soft-tissue assessment, the low frequency in the present study may reflect underdetection, a truly lower frequency of brain parenchymal extension in nasal TC, or both. Magnetic resonance imaging provides superior soft-tissue contrast and may be more sensitive for intracranial extension, tumor-associated edema, and secondary changes such as ventricular compression [25]. Comparative CT and MRI studies in dogs with nasal TC and suspected intracranial extension, ideally with histopathological correlation, are needed to establish the true extent of intracranial involvement.

The discordance between frequent cribriform plate lysis and less frequent CT-detected brain parenchymal involvement may also be influenced by the dura and other anatomic barriers at the skull base [26]. In human sinonasal undifferentiated carcinoma, dural invasion is reported more often than direct brain involvement [27]. Extension into brain tissue is biologically complex and may depend on tumor–cell interactions with brain microvascular endothelial cells, as well as cellular motility, adhesion, proteolytic activity, and angiogenesis [28]. These mechanisms were not evaluated in the present study; therefore, no biologic explanation for the observed imaging pattern can be established from these data.

The CT findings that raised suspicion for meningeal involvement were cribriform plate lysis and adjacent meningeal enhancement [29,30]. However, meningeal enhancement on CT is not specific for neoplastic invasion and may also represent reactive pachymeningitis or inflammation. Similar imaging findings were described by Strohmayer et al. [31], in which dural involvement was considered equivocal on CT but was absent on histopathology. Accordingly, the adjacent meningeal enhancement in the present cohort could have represented reactive change, true meningeal involvement, or both. Because none of the dogs underwent post-mortem examination or intracranial histopathological sampling, this distinction could not be established definitively.

Discrepancies were also observed between CT extent and neurological signs. For example, one dog with cranial vault invasion without obvious adjacent meningeal enhancement later developed seizures, whereas several dogs with extensive tumor involvement near skull-base foramina had minimal neurological deficits. Previous studies have reported both brain involvement without overt neurological signs and neurological signs that do not precisely reflect the extent of intracranial invasion [20,25,32]. Potential explanations include differences in the location and severity of neural tissue compression or invasion, as well as underreporting of subtle neurological abnormalities in retrospective records.

None of the nasal TCs in this study demonstrated intralesional calcification on CT images, likely reflecting the absence of chondroid or osteoid matrix production typical of sarcomas [18]. Although intralesional calcification is occasionally seen in canine squamous cell carcinomas, it was not observed in nasal TCs, suggesting a fundamental histologic distinction for diagnostic purposes [33]. Hyperostosis and/or sclerosis of the surrounding bones were noted in 4 dogs in this study. In human medicine, focal hyperostosis is typically associated with benign sinonasal tumors and sclerosis in the context of chronic inflammatory conditions or specific malignancies [34]. In veterinary literature, Burk et al. reported hyperostosis in the nasal cavity of dogs, but it was not specific to neoplasia [35]. Instead, chronic irritation or concurrent infection was proposed as a cause. In our study, bacterial culture was performed for only 1 of the 4 dogs with hyperostosis, which yielded a positive result for Micrococcus spp. This limited our ability to determine whether infection played a primary role. These findings suggest that the hyperostosis observed in some cases may reflect secondary changes resulting from chronic inflammation or infection rather than being directly attributable to nasal TC. Further studies incorporating microbiological and histopathological evaluations are needed to clarify the pathogenesis of this finding.

Intralesional non-enhancing fluid-attenuating regions, interpreted as compatible with necrosis or cystic change, were observed in most cases [2]. Nasal TC is typically characterized by a stronger inflammatory response than adenocarcinoma, which may contribute to necrosis and cyst formation [36]. However, a direct causal relationship has not been established. The diagnosis of intralesional necrosis or cyst formation relies on identifying non-enhancing regions within the mass. However, previous studies have demonstrated that fluid retained in the nasal cavity due to rhinitis may present similarly on CT images [37,38]. CT may also not reliably differentiate between fluid accumulation and mucosal thickening [39]. These limitations suggest that the frequency of reported cyst formation or necrosis may be overestimated, underscoring the need for future studies to incorporate MRI or histopathological correlation. Peritumoral fluid accumulation may result from intratumoral necrosis, inflammation, and concurrent rhinitis. Fluid may extend into the frontal or paranasal regions, potentially due to choanal obstruction and disrupted sinus drainage [38,40].

Positive bacterial culture results were identified in 4 of the 5 dogs tested. In dogs with nasal neoplasia, positive bacterial cultures are generally considered more likely to reflect secondary colonization or concurrent inflammation than a primary infectious process [41]. Although such findings may have clinical relevance, their significance in the present study remains uncertain because cultures were performed in only a small subset of cases, and the sampling source and collection method were not consistently documented.

Two dogs initially diagnosed with inflammatory conditions (cellulitis and chronic lymphoplasmacytic rhinitis) were later confirmed to have nasal TC. The diagnostic discrepancies likely stemmed from biopsy limitations, including small sample size, tumor secretions, peritumoral inflammation, and instrumentation constraints, which can lead to misdiagnosis [3,42,43]. In 14–29% of cases, a second biopsy is often required to reach a definitive diagnosis [21,42]. Additionally, their medical histories suggest a possible association between chronic inflammation and carcinogenesis, as reported in both human and veterinary literature [44,45].

Most of the pulmonary nodules identified at the initial CT examination were presumed to represent pulmonary osseous metaplasia, which is associated with type I pneumocyte osteoid formation and age-related calcification. Enhancing soft-tissue-attenuating nodules were identified at the initial CT examination in two dogs (dogs 1 and 10). In dog 14, the initial CT examination showed mineral-attenuating foci considered compatible with pulmonary osseous metaplasia, whereas multiple additional pulmonary nodules were detected on follow-up CT. These newly detected nodules could not be characterized without histopathological confirmation. The resolution of the pulmonary nodule in dog 10 during follow-up CT may align with previous observations that solitary nodules may resolve, indicating a treatment response or non-malignant etiology [46,47]. Regional lymphadenopathy was common in our study, and regional lymphatic metastasis has been reported in dogs with nasal neoplasms [2,21]. However, because no lymph nodes were confirmed cytologically or histopathologically, the enlarged nodes in our study could not be definitively classified as metastatic or reactive/inflammatory.

This study has several limitations. The small sample size limited generalizability. Its retrospective multicenter design resulted in heterogeneous CT equipment and acquisition protocols, incomplete clinical and follow-up information, and heterogeneous treatment protocols. Although the image reviewers used a consensus approach, they were not blinded to the diagnosis of nasal TC, and formal interobserver agreement for subjective imaging features was not assessed. CT-based assessment was also limited by difficulty distinguishing tumor tissue from retained secretions, mucosal thickening, reactive inflammation, intralesional necrosis or cystic change, and peritumoral fluid. No comparison group of other nasal tumor types was included; therefore, specificity of the described CT findings for nasal TC could not be assessed. The site of tumor origin could not be determined reliably from a single time point, and CT may have underestimated meningeal or brain parenchymal involvement relative to MRI. The absence of post-mortem examinations and intracranial tissue sampling prevented histopathological confirmation of intracranial extension. Finally, heterogeneous treatment protocols and incomplete follow-up precluded meaningful survival analysis. Larger prospective studies with standardized CT and MRI protocols, appropriate comparison groups, and consistent follow-up are warranted.

5. Conclusions

All nasal transitional carcinomas in this cohort involved the caudal nasal cavity, and most extended into the middle nasal cavity. They were characterized by a predominantly destructive CT growth pattern with extensive regional extension. Frequent CT findings included cribriform plate lysis, extension into the maxillary recess and paranasal sinuses, intralesional non-enhancing fluid-attenuating regions consistent with necrosis or cystic change, peritumoral fluid accumulation, and absence of intralesional calcification. CT-detected brain parenchymal involvement was uncommon despite frequent cranial vault invasion. These observations provide a frequency-based CT reference pattern for lesion assessment and diagnostic planning; however, these findings are not specific for nasal transitional carcinoma, and histopathological examination remains necessary for definitive tumor type identification.

Abbreviations

The following abbreviations are used in this manuscript:

CT	Computed tomography
HU	Hounsfield unit
MRI	Magnetic resonance imaging
TC	Transitional carcinoma

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/ani16142202/s1, Table S1: Veterinary hospitals and CT equipment details; Table S2: CT acquisition parameters for each dog; Table S3: Comparative clinical and diagnostic characteristics of dogs with histopathologically confirmed nasal transitional carcinoma. Cases are displayed as columns to facilitate comparison across the cohort; Table S4: Distribution and type of bony changes on CT (n = 16); Table S5: Comparative additional thoracic and abdominal CT findings at the initial CT examination. Cases are displayed as columns; “not available” indicates that thoracic and/or abdominal CT was not available for review; Table S6: Variable-centered summary of changes between initial and follow-up CT examinations (n = 7); Table S7: Treatment and outcome summary.

Author Contributions

Conceptualization, L.-S.L.; methodology, L.-K.N. and L.-S.L.; investigation, L.-K.N.; data curation, L.-K.N., C.-S.C., K.-S.C., C.C., H.-C.C. and L.-S.L.; formal analysis, L.-K.N. and L.-S.L.; writing—original draft preparation, L.-K.N.; writing—review and editing, L.-K.N. and L.-S.L.; supervision, L.-S.L.; project administration, L.-S.L.; validation, L.-K.N. and L.-S.L. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement

Ethical review and approval were waived for this retrospective study using pre-existing clinical and imaging data, in accordance with the policies of the participating hospitals.

Informed Consent Statement

Owner consent for the use of anonymized clinical and imaging data for research was obtained for all dogs.

Data Availability Statement

The original contributions presented in this study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding authors.

Conflicts of Interest

The authors declare no conflicts of interest.

Funding Statement

This research received no external funding.