Metastatic lesions in hemangiosarcoma (HSA) can result from hematogenous spread or intracavitary implantation, with the liver, peritoneum, and lungs being the most common sites. Diagnosis of suspected HSA is essential, with abdominal ultrasound (AUS) and 3-view thoracic radiographs being the primary imaging methods. The prevalence of visible lung metastases in dogs with HSA ranges from 19% to 52%, typically presenting as a coalescing unstructured to structured miliary interstitial pattern.
CT is more sensitive than radiology for detecting pulmonary nodules, especially in larger dogs or when abdominal sonography is hindered by intraluminal gas or food. This retrospective study aimed to evaluate the CT features of pulmonary nodules suspected to be metastases in dogs with confirmed HSA at various sites.
This is the first veterinary study to describe CT features of lung metastases from HSA. The majority of pulmonary metastatic lesions were small, numerous, and generally distributed, often exhibiting a halo sign and feeding vessels. These findings align with human literature, where metastatic pulmonary angiosarcoma typically presents as multiple nodules with sharp margins.
The study also found a statistically significant correlation between fewer (2–10) lung lesions and a more frequent peripheral distribution. This pattern, commonly seen in human cases, is thought to result from hematogenous spread, with emboli forming at the most peripheral parts of the pulmonary vessels. When numerous nodules are present, they are more randomly distributed, with varying sizes due to repeated embolization or different growth rates.
The presence of a feeding vessel—often a pulmonary artery branch leading to a nodule—was also observed, consistent with other neoplastic and non-neoplastic processes involving hematogenous spread. A 72.7% prevalence of the halo sign was noted, comparable to the 58–72.7% prevalence in human patients with angiosarcoma pulmonary metastases.
Most lung lesions in this study were small (<1 cm) and demonstrated homogeneous enhancement (78.8%), in line with findings in human cases where nodule size correlates with enhancement patterns. Gaseous collections within nodules were observed in some human studies, with a prevalence of 21–58%.
Of the dogs in this study, 30.3% had lung metastases as the sole site of spread. The remaining 69.7% had metastases in multiple extrathoracic organs, including the liver (42.4%), spleen (33.3%), muscles (30.3%), and peritoneum/retroperitoneum (18.2%). The SPLASH sign, or comparable contrastographic characteristics, is unique to HSA and should be prioritized in differential diagnoses.
Mineralization in extrapulmonary metastases was rare, occurring in one case of muscular and peritoneal/retroperitoneal nodules. Interestingly, all three dogs had at least one osseous lesion in the proximal humerus, consistent with findings from a recent paper. However, limitations in this study include the absence of histological confirmation for all metastatic lesions and the lack of primary cardiac involvement, particularly in the right atrium or auricle.
In conclusion, lung HSA metastases are characterized by numerous, small, generalized nodules, frequently exhibiting the halo sign and feeding vessels. These findings should aid radiologists and clinicians in diagnosing metastatic HSA, particularly when lesions cannot be sampled. Whole-body CT remains critical in staging dogs with primary masses compatible with HSA.
