Computed tomographic characteristics of cavitary pulmonary adenocarcinoma in 3 dogs and 2 cats

Abstract

Cavitary pulmonary lesions can result from the localized breakdown of pulmonary parenchyma or be caused by the expulsion of a necrotic part of a mass. The objectives of this study were to describe the clinical and computed tomographic characteristics for cases of cavitary pulmonary adenocarcinoma and find associations between the features and those identified in human pulmonary cavitary soft tissue lesions. Five cases were identified that had a cavitary pulmonary mass on thoracic computed tomography (CT) and histopathology of the lesions. Three dogs and 2 cats had cavitary pulmonary adenocarcinoma. Common features of CT in these cases included lesions in the caudal lung lobes, lobular and spiculated lesion margins, air bronchograms within the mass, pleural tags, heterogeneous contrast enhancement, and ground glass opacity in the surrounding parenchyma. The findings of this case series suggest there are similarities in the CT characteristics of malignancy in human and animal cavitary pulmonary masses.

Introduction

A cavitary pulmonary lesion is defined as a gas-filled space within a zone of pulmonary consolidation or within a mass or nodule (1). This lesion is caused by the expulsion of a necrotic part of the lesion via the bronchial tree (2). There are both infectious and neoplastic etiologies for the lesions, including neoplasms, granulomas, paragonimus cysts, and abscesses (3,4). The proposed etiology behind these lesions in neoplastic disease is vascular inadequacy to the center of the mass, resulting in necrosis followed by central cavitation (5). Differential diagnoses for other similarly appearing lesions include emphysema, pneumatoceles, bullae, and sub-pleural blebs which can mimic true cavitary pulmonary lesions due to the presence of air-filled areas, but lack the surrounding zone of pulmonary consolidation (3,4).

Although the discussion of cavitary pulmonary lesions usually centers around infectious causes in human medicine, cavitation can also be seen in malignant neoplastic lesions (1–3,5). Primary pulmonary neoplasia in dogs and cats is relatively uncommon, with an incidence rate of 4.2 cases per 100 000 dogs and 2.2 cases per 100 000 cats (6,7). In a recent study in 2014, the prevalence of cavitation in primary pulmonary neoplasia in dogs was 11% (1). Across numerous studies of these malignant neoplasms in humans, rates of cavitation were similar, ranging between 10 and 15% (1,8,9). In a 2015 study of cats with primary pulmonary neoplasia, air-containing cavities were identified in 63% of the patients, although there was no differentiation in the type of cavities, such as an air bronchogram or cavitation (10).

Although cavitary lesions may be seen on thoracic radiographs, with the increased accessibility of computed tomography (CT) in veterinary practice, these cases often undergo CT scans to further characterize lesions. The CT appearance of cavitary pulmonary lesions is well-documented in human medicine, and lesions can be classified based on CT appearance, shortening the list of possible differential diagnoses. Computed tomographic characteristics of these lesions have yet to be described in dogs and cats. The objectives of this study were to describe the clinical and CT characteristics for cases of cavitary pulmonary adenocarcinoma and to compare CT features identified to those described for human pulmonary cavitary soft tissue lesions.

Materials and methods

The diagnostic imaging picture archiving and communication system (PACS) for the Ohio State Veterinary Medical Center (OSU-VMC) was searched to identify dogs and cats with cavitary pulmonary masses presenting between August 24, 2018, and January 6, 2020. Search terms included “cavitary,” “cavitated,” and “cavity,” and the database was filtered for thoracic CT studies. Criteria for inclusion in the study were cases identified by a radiologist as having cavitary pulmonary soft tissue lesions on thoracic CT. All included cases had been administered intravenous contrast. Cases were excluded if histopathology of the lesion(s) was not performed. Medical records for included cases were reviewed, and information recorded included; signalment, clinical history, blood abnormalities, date of diagnosis, imaging studies and radiographic findings, treatments, and lesion histopathology.

The CT studies were reviewed by a radiology resident (SLJ) at the OSU-VMC, and findings were confirmed by a Board-certified radiologist (ETH). The radiology resident assessed pulmonary lesions in the CT images for location, wall regularity, presence of air bronchograms, characteristics of the surrounding pulmonary parenchyma, margin characteristics including the notch sign and spiculations, pleural tags, a ground glass opacity (focal unstructured interstitial pattern), contrast enhancement characteristics, mineral attenuation within the mass, lesion diameter and location, and presence of tracheobronchial lymphadenopathy. Definitions from the human medical literature were used to define features not commonly described in veterinary studies. The notch sign was defined as an abrupt bulging of the contour of the lesion or an indentation in the border of a solid lung mass at the location of a blood vessel that supplies the tumor (11). A spiculation (also called corona radiata) was considered a linear opacity that extends from the mass, and can also be defined as a pleural tag if the linear opacity extends from the mass to the visceral pleura (12,13). Pleural tags are linear opacities which extend from the mass to the visceral pleura and can represent fibrotic tissue that extends from the nodule to the visceral pleura that causes inward retraction of the visceral pleura (11,12,14). Air bronchograms referred to one or more linear branching tubular hypoattenuations representing bronchi or bronchioles passing through densely opacified lung parenchyma or masses (15).

Results

In total, 6 animals were identified as having cavitary pulmonary lesions on thoracic CT. Of those 6, 3 dogs and 2 cats met the criteria for inclusion in the study. Of the 3 dogs, all were medium to large breed dogs with ages ranging from 9 to 12 years of age, 2 were males and 1 was female. The 2 cats were adult spayed female domestic shorthair cats aged 12 and 15 y old.

Although 4 of the 5 cases were presented with several clinical complaints, all 5 cases had experienced coughing. In 4 cases, decreased appetite and weight loss were present. History and physical examination findings at presentation are summarized in Table 1. A complete blood (cell) count (CBC) and serum chemistry were performed in all 5 animals at presentation, and in 4 of the 5 animals no clinically significant abnormalities were detected (Table 1). Additional pre-operative tests are shown in Table 1.

Table 1.

Demographic and clinical testing information for 3 dogs and 2 cats with cavitary pulmonary soft tissue lesions.

Case	Age (y)	Gender	Breed	Clinical signs	Physical examination findings	CBC, chemistry	Abdominal ultrasound	Cytology	Other diagnostics performed
1	15	FS	Domestic short-haired cat	Coughing	NSF	Creatinine: 132 μmol/L	NSF	Carcinoma	FeLV/FIV: Negative
2	12	FS	Domestic short-haired cat	Weight loss, coughing	NSF	Mild neutrophilia, eosinophilia	Diffuse mild thickening of small intestines, no loss of layering	N/A	Free T4: Within normal limits
3	12	MN	Basset Hound	Weight loss, coughing, decreased appetite and activity	Grade IV/VI heart murmur	NSF	N/A	N/A	N/A
4	9	FS	Australian shepherd	Coughing, lethargy, weight loss, left forelimb lameness	Mild ocular discharge	NSF	N/A	Atypical cells	Fungal urine antigen test: Negative Baermann fecal examination: Negative
5	11	MN	Labrador retriever	Coughing, weight loss	Mild inspiratory stridor (diagnosed with laryngeal paralysis)	HCT: 34.4% ALP: 564 U/L (consistent with prior blood test)	N/A	N/A	N/A

CBC — Complete blood (cell) count; NSF — No significant findings; N/A — Not applicable; FS — Female spayed; MN — Male neutered; HCT — Hematocrit; ALP — Alkaline phosphatase; FeLV/FIV — Feline leukemia virus/Feline immunodeficiency virus.

Three-view thoracic radiographs were taken in all 5 cases, with 3 of the 5 cases having imaging completed before presentation to OSU-VMC by the referring veterinarian. Three of the 5 cases were identified as having cavitation, or lucent regions, within the masses on these radiographs.

On thoracic CT, multiple pulmonary cavitary soft tissue lesions were noted in 4 of the 5 cases. The primary (largest) cavitary lesion was located in the left caudal lobe in 4 cases and in the right caudal lobe in 1 case. Two cases (Case 1 and Case 4) had additional lesions in other lung lobes. Lesion diameters for the largest cavitary lesion were 2.0 cm and 1.5 cm in the cats and 5.6 cm, 6.0 cm, and 4.7 cm in the dogs. The walls were lobular in 4 cases and smooth in 1 case. Spiculations were noted in all 5 cases (Figure 1). Pleural tags were seen in all 5 cases (Figure 2). In 1 case, multiple masses had a large amount of contact with the pleural margin. Air bronchograms were present within the mass in all 5 cases. Ground glass opacity (unstructured interstitial pattern) was noted in the parenchyma adjacent to the masses in all 5 cases (Figure 3). The notch sign was present in all dogs but could not be accurately evaluated in the 2 cats, because of their small size. All 5 cases had mixed heterogeneous contrast enhancement of the cavitary pulmonary mass, with 1 case containing a non-cavitary metastatic nodule with homogenous contrast enhancement. In the case with a non-cavitary metastatic nodule, this nodule was not within the cavitary mass and was in a separate location within the lung parenchyma (right cranial lung lobe). Mineral attenuation was present in 4 out of 5 cases. Tracheobronchial lymphadenopathy was noted in 1 of 5 cases. The specific CT findings for each case are outlined in Table 2.

Figure 1.

Transverse plane CT images of lung parenchyma containing spiculations in A — Case 1 (cat), B — Case 4 (dog), and C — Case 5 (dog). Multiple hyperattenuating linear striations extend from the margins of the mass. In C, there is also a ground glass opacity that partially obscures the appearance of the spiculations. Window width and window level were adjusted to 1500/−600 (lung window), the slice thickness was 1.25 mm, and the images were acquired with 120 kVp (A–C).

Figure 2.

Transverse plane CT image of lung parenchyma containing pleural tags in Case 3 (dog). Notice the indentations along the pleural surface (black arrows). Two of the 3 pleural tags identified were also classified as spiculations in this image. Window width and window level were adjusted to 1500/−600 (lung window), the slice thickness was 1.25 mm, and the image was acquired with 120 kVp.

Figure 3.

Transverse plane CT images of lung parenchyma exhibiting ground glass opacity sign in Case 4 (dog). The pulmonary parenchyma dorsal to the cavitated pulmonary mass is hyperattenuating when compared to the parenchyma of the contralateral lung. This finding is described as a ground glass opacity, and when surrounding a pulmonary mass is referred to as the halo sign in human literature. Window width and window level were adjusted to 1500/−600 (lung window), the slice thickness was 1.25 mm, and the image was acquired with 120 kVp.

Figure 4.

Transverse plane CT image of a pulmonary mass exhibiting a notch sign in Case 3 (dog). Branches of the caudal pulmonary artery and anomalous pulmonary vein (arrows) course along the mass with an indentation of the margin of the mass along these vessels, indicating a notch sign. A soft tissue window width and window level were adjusted to 660/230, the slice thickness was 1.25 mm, and the image was acquired with 120 kVp.

Table 2.

Computed tomography findings for 3 dogs and 2 cats with cavitary pulmonary soft tissue lesions.

Case number (species)	Lesion location (lung lobe)	Wall regularity	Air bronchograms	Surrounding parenchyma characteristics	Contrast enhancement
1 (cat)	Right caudal	Lobular margin Spiculations	Present	Ground glass opacity adjacent to the mass Pleural tags	Heterogeneous
2 (cat)	Left caudal	Lobular margin Spiculations	Present	Patchy ground glass opacity adjacent to the mass Pleural tags	Heterogeneous
3 (dog)	Multiple Largest in left caudal (Smaller in accessory)	Lobular margin Spiculations Notch sign	Present in both	Patchy ground glass opacity adjacent to both masses. Pleural tags	Heterogeneous
4 (dog)	Multiple Largest and multiple in left caudal Additional in right cranial and caudal	Smooth margin (all) Spiculations (one smaller mass) smooth, internal is irregular. Notch sign	Present in the largest and multiple, but not all, masses	Ground glass opacity adjacent to all masses Pleural tags present in all, with a lot of pleural contact	Heterogeneous
5 (dog)	Multiple Left caudal mass and 2 nodules	Lobular margin Spiculations Notch sign Nodules: smooth margin	Present	Patchy ground glass opacity adjacent to the margin of the mass Pleural tags	Heterogeneous (mass), homogeneous (small nodule)

Four out of 5 animals underwent surgery to remove the masses, whereas palliative care was pursued in 1 animal. In 4 out of 5 cases, histopathologic sampling was via surgical biopsy and in 1 case via post-mortem tissue sampling. All 5 cases had pulmonary adenocarcinoma of varying degrees of differentiation and metastasis. The 2 cats had well-differentiated papillary adenocarcinoma in the right and left caudal lung lobes, respectively, and both had complete surgical excision of these masses. One cat had additional sampling from the accessory lung lobe, which had multifocal bronchoalveolar hyperplasia. All 3 dogs showed metastatic disease, and in the 2 surgical cases, incomplete excision of the primary tumors was reported. The cavitary mass of Case 3 was determined to be pulmonary carcinoma, whereas additional masses were revealed to be lymphoma. In Case 4, lesions in the left caudal lung lobe as well as the right cranial and caudal lung lobes were excised. All were characterized as aggressive adenocarcinoma, with incomplete excision of the right cranial lung mass. In Case 4, a biopsy of a cutaneous carpal lesion was performed at the same time as the thoracic surgery, and this was consistent with carcinoma presumed to be a metastasis of the pulmonary lesion. Case 5 exhibited invasive pulmonary adenocarcinoma, which was incompletely excised at the hilus. Common histopathologic descriptions of the parenchyma adjacent to the mass included alveolar histiocytosis (alveolar macrophages within alveolar spaces) in 4 cases, lymphocytic or neutrophilic inflammation in 3 cases, fibrosis/fibroplasia in 3 cases, type II pneumocyte hyperplasia in 2 cases, and neoplastic cells in 2 cases.

Discussion

There has been minimal description and characterization of the imaging appearance of cavitary pulmonary soft tissue masses in small animal species. Although there are few reports detailing the radiographic findings of cavitary pulmonary masses, there has not been an assessment of cavitary pulmonary soft tissue lesions with CT findings associated with histopathological diagnosis (1,3,16). In all 5 cases included in this case series, CT of the thorax was performed following radiographic findings of pulmonary masses, and all had cavitary pulmonary masses with similar CT findings. The CT findings in these cases included lesions present in the caudal lung lobes, lobular and spiculated lesions margins, presence of air bronchograms within the mass, pleural tags, heterogeneous contrast enhancement, and a ground glass opacity in the surrounding parenchyma.

On radiographs, pulmonary cavitary neoplastic lesions have characteristic findings including irregular margins, a predisposition for the caudal lung lobes, thick walls, and an increase in density of the surrounding parenchyma compared with normal pulmonary tissue (3,16). There are similarities between these radiograph findings and those seen on CT; including, the irregular margins, lesion location, and increased density of the surrounding parenchyma. However, CT is required to identify or better visualize some characteristic findings such as pleural tags, spiculations, intralesional air bronchograms, and contrast enhancement. Benign lesions in contrast to malignant lesions have, on both plain radiographs and CT, thin walls with smooth, regular margins, sharply defined borders, and minimal changes to the surrounding pulmonary parenchyma (3,17). The characteristic findings in this case series included caudal lung lesion location, wall irregularity, air bronchograms, pleural indentations/tags, contrast enhancement, and the characteristics of the surrounding parenchyma. Upon detailed review of the literature, several findings were not included in this study based on their subjectivity and lack of specificity. Although these markers are commented on in the human medical literature, they can be seen in various types of pulmonary lesions, including those from benign or infectious causes (5,11,18,19). The markers include mineral attenuation, presence of metastasis, large lesional diameter, and tracheobronchial lymphadenopathy. In addition, the notch sign was subjective and could not be adequately evaluated in cats.

In this case series, 4 out of 5 cases had an irregular, lobular appearance to the walls of the cavitated mass. This finding is similar to malignancy and pulmonary carcinoma in humans, in whom lobular margins are more commonly identified than smooth margins, due to an uneven rate of growth (11,12,14). In addition, spiculations, sometimes referred to as corona radiata, were documented in all 5 cases. Fine spiculations have been reported to have a positive predictive value of 90% for primary pulmonary adenocarcinoma in humans (10–12,18). Spiculations are associated with interlobular septal thickening and fibrosis secondary to the airway, septa, pulmonary vessel, or lymphatic obstruction by neoplastic cells, although they can also be associated with benign causes of growth of fibrotic tissue (11,12,14,20). In one study evaluating cavitary pulmonary lesions on CT in humans, 81% of cavitary lesions with an irregular wall were malignant, whereas just 19% of cases with irregular margins were of a benign etiology (21,22). Benign lesions typically have inner walls with a smooth, regular margin (3). Another widely accepted explanation for this difference in appearance between malignant and benign lesions is that the malignant nodules are variable in size, and when they undergo necrosis, they do so in a patchy distribution (19).

Lesion location is not specific for malignant versus benign neoplasms, although it is documented that pulmonary adenocarcinoma has an affinity for the caudal lung lobes, specifically the left caudal lung lobe (1). Of the 5 cases included in the study, all 5 had cavitary lesions in the caudal lung lobes. These results are consistent with the affinity previously described for the caudal lung lobes, and more specifically, the left caudal lung lobe. In other recent veterinary case reports that evaluated cavitary pulmonary lesions due to different underlying etiologies, specifically those associated with Aspergillus fumigatus infection and bronchogenic cysts, the lesions were reported in the cranial lung lobes of these patients (16,23). Migrating intrathoracic grass awns have been documented to have an affinity for the caudal and accessory lobes. These are not specifically associated with cavitary lesions but instead they are associated with unstructured interstitial and alveolar patterns, as well as pleural effusion and pneumothorax (24).

Air bronchograms can be seen in various interstitial pulmonary diseases. When they are seen, and particularly if they are tortuous or ectatic in appearance and are located within a pulmonary mass, both solid and cavitated, these findings point towards a neoplastic rather than benign etiology (25,26). All 5 cases in this series had air bronchograms within their cavitary lesions. In cases of primary solitary nodules in humans, an air bronchogram is more frequently associated with malignant nodules and reportedly present in 73% of neoplastic lesions (12,14,17,24). It is thought that in malignant lesions, the tumor surrounds the bronchi and invades the bronchial walls causing strangulation and obliteration of the bronchial tracts leading to the presence of air bronchograms which are often torturous or ectatic in appearance (26).

There are additional supportive imaging findings of malignancy, including changes to the surrounding pulmonary parenchyma, pleural tags, and contrast enhancement characteristics. These findings are not specific for a diagnosis of neoplasia, but as a group, help to support a diagnosis of malignancy. In a veterinary study, it is well-documented that the surrounding pulmonary parenchyma in malignant cavitary pulmonary lesions has an increased density compared to the normal lung parenchyma (16). On thoracic CT in the human medical literature, the presence of ground glass opacity of the surrounding parenchyma of a lesion is a commonly documented finding in patients with primary pulmonary adenocarcinoma and is characterized by an increased attenuation of the lung parenchyma surrounding a pulmonary nodule or mass (11). This ground glass opacity is associated with various infectious causes, organizing pneumonia, vasculitis, and neoplasia and thought to result from hemorrhage, inflammation, or the growth of adenocarcinoma (12). All 5 cases in this study had ground glass opacity adjacent to the primary mass and, if present, to the secondary masses. Histopathology in these cases showed similar results in the surrounding parenchyma with alveolar histiocytosis, inflammation, fibrosis, neoplastic invasion, and type II pneumocyte hyperplasia. In human cases of peripheral pulmonary carcinoma, pleural tags have been reported in 60 to 80% of cases, although they are also associated with granulomas and metastases (12). In this series, all 5 cases had an extension of the lesion to the pleural margin, and in Case 4, multiple masses had a large amount of contact with the pleural margin. Heterogeneous contrast enhancement is a well-documented finding in malignant masses, likely due to the variable blood supply within tumors (15). Mixed heterogeneous enhancement was reported in all 5 cases in this series, although a non-cavitary nodule in Case 5 was homogenously contrast enhancing.

Although the maximum wall thickness of the lesion is considered to be among the best criteria for determining malignancy in cavitary pulmonary lesions in humans, this finding has not been used in veterinary medicine (27). In the current series, lesions had subjectively thickened walls compared to those reported in cysts, bullae, or abscesses, although we did not measure wall thickness specifically due to the lack of established methodology (17,23). These thicker walls surrounding the cavity give the appearance of a more solid mass, more consistent with a soft tissue tumor. In the human medical literature, wall thickness is measured at the thickest part of the wall of the cavity (21,22). In a study of 96 human patients presenting with cavitary pulmonary lesions, a wall thickness of > 24 mm was considered 100% specific for malignancy, whereas a wall thickness of ≤ 7 mm was 96.7% specific for non-malignant lesions (27,28). In a follow-up study, 84.2% of patients with lesions with a wall thickness of ≥16 mm had a malignant neoplasm on histopathologic follow-up examination following surgical excision of the masses (21,22). A larger study would be necessary to create similar guidelines for wall thickness in small animals. Additionally, the protocol for where and how to measure maximal wall thickness would need to be established. In 2 veterinary case reports, 1 detailing the CT findings of a bronchogenic cyst and the other from a solitary pulmonary lesion secondary to Aspergillus fumigatus infection, the walls of these lesions were reported to be thin and sharply delineated (17,23). In the present case series, the lesions had subjectively thickened walls compared to those routinely seen in cysts or bullae, but further study would be required to evaluate this characteristic as a marker of malignancy.

In conclusion, cases presented herein shared some of the previously identified findings characteristic of malignancy in human cavitary pulmonary lesions. The malignant cavitary pulmonary masses were predominantly present in the caudal lung lobes and had the following CT findings: lobular and spiculated margins, presence of air bronchograms within the mass, pleural tags, heterogeneous contrast enhancement, and ground glass opacity in the surrounding parenchyma. These findings can be used by clinicians and inform future studies in larger patient populations to investigate the differentiation of benign and malignant pulmonary lesions. Although thoracic CT findings can help narrow the differential diagnosis, histopathology is still necessary for definitive diagnosis.