The study was performed at the University of Tennessee John and Ann Tickle Small Animal Hospital, University of Tennessee College of Veterinary Medicine, Knoxville, TN
Comparison of Presumptive Echocardiographic and Definitive Diagnoses of Cardiac Tumors in Dogs
Article first published online: 19 JUL 2013
Copyright © 2013 by the American College of Veterinary Internal Medicine
Journal of Veterinary Internal Medicine
Volume 27, Issue 5, pages 1092–1096, September/October 2013
How to Cite
Rajagopalan, V., Jesty, S.A., Craig, L.E. and Gompf, R. (2013), Comparison of Presumptive Echocardiographic and Definitive Diagnoses of Cardiac Tumors in Dogs. Journal of Veterinary Internal Medicine, 27: 1092–1096. doi: 10.1111/jvim.12134
- Issue published online: 13 SEP 2013
- Article first published online: 19 JUL 2013
- Manuscript Accepted: 22 MAY 2013
- Manuscript Revised: 27 APR 2013
- Manuscript Received: 16 AUG 2012
Echocardiography is used for identification of cardiac tumors and presumptive diagnoses often are made based on the location of identified masses.
To determine the accuracy of echocardiographically based presumptive diagnoses of cardiac tumors when compared with clinicopathologic or histopathologic definitive diagnoses.
A total of 24 client-owned dogs having a cardiac mass on echocardiogram that was subsequently definitively diagnosed by cytology or histopathology.
Retrospective study. A Cardiac Veterinary Database search of animals seen at the University of Tennessee John and Ann Tickle Small Animal Hospital from 2006 to 2012 identified 24 dogs that fit the inclusion criteria.
The presumptive diagnosis of chemodectoma, ectopic thyroid carcinoma, or lymphoma in cases with heart base masses was correct in 7/9 cases. The presumptive diagnosis of hemangiosarcoma in cases with right atrial masses was correct in 4/8 cases. Seven cases had an open diagnosis because of the unusual presentation on echocardiogram (ECG); various neoplasms were diagnosed in these animals, but hemangiosarcoma, chemodectoma, ectopic thyroid carcinoma, and lymphoma accounted for 6 of them. Pericardial effusion was seen in 10/24 cases. ECG abnormalities were seen in 8/24 cases. Survival ranged from <1 to >150 days.
Conclusions and Clinical Importance
In this retrospective study, the presumptive diagnosis based on echocardiographic tumor location was only moderately accurate. Cardiac tumors that were considered unusual on echocardiogram were nonetheless frequently found to be the common cardiac tumor types seen in dogs.
ectopic thyroid carcinoma
Primary and metastatic tumors involving the heart are relatively uncommon in humans and dogs. The relative frequency of various tumors depends on the population studied. Based on 1 large medical record search, the reported incidence of cardiac tumors in dogs was 0.19%. In that study, hemangiosarcoma (HSA) was the most common cardiac tumor in dogs, representing 69% of all cardiac masses on histopathology. The next most common tumors were aortic body tumors (chemodectomas [CD], 8%; lymphoma [LSA], 4%; and ectopic thyroid carcinoma [ETC], 1%). In a pathology-based study of cardiac tumors in dogs, 63% were HSA and 18% were CD. In a report assessing only dogs with pericardial effusion in which cardiac tumors were diagnosed, 55% had HSA, 23% had mesothelioma, 14% had CD, 9% had ETC, and 5% had LSA. Other reported canine cardiac tumors are variable, and all are considered rare.[2, 3]
The prognosis for life expectancy varies widely depending on tumor type. Hemangiosarcoma is associated with a poor prognosis. Many studies cite survival times <2 months from diagnosis, even with pericardectomy.[5, 6] In fact, 1 study evaluating pericardectomy in dogs with HSA showed no improvement in survival time compared with no treatment. Although in selected cases, surgical debulking of a right auricular mass followed by chemotherapy may improve life expectancy slightly, reported median survival times with this approach still are <6 months.[5, 7, 8] Conversely, although dogs with heart base tumors do not have a much longer life expectancy than dogs with HSA if pericardectomy is not performed, pericardectomy confers a significant increase in survival time, extending median or mean survival times from <2 months to ≥22 months.[9, 10] Ideally, the clinician would be able to make an accurate presumptive diagnosis regarding cardiac tumor type before the decision is made to pursue pericardectomy or other treatment. Definitive diagnosis of cardiac tumors using minimally invasive techniques (eg, pericardial effusion analysis, fine-needle aspiration) can be challenging, however, because pericardial effusion cytology is usually nondiagnostic[7, 10-13] and because the mass may be inaccessible for fine-needle aspiration. For this reason, the clinician often has to make a presumptive diagnosis based on information provided solely by the echocardiogram. This case series reports on the echocardiographic presentation of 24 cardiac tumors in dogs with subsequent definitive diagnoses. The aim of this study was to determine the accuracy of echocardiographically based presumptive diagnoses of cardiac tumors when compared with clinicopathologic or histopathologic definitive diagnoses.
Materials and Methods
A search of the University of Tennessee Cardiology service's echocardiogram database was made for all dogs with cardiac masses from 2006 to 2012. Three board-certified cardiologists, 2 cardiology residents in training, and 1 cardiology specialty intern performed the initial echocardiograms (the residents and intern were under the supervision of the cardiologists). The inclusion criterion was a definitive histopathologic or cytologic diagnosis of the mass made by a board-certified anatomic or clinical pathologist. Cytologic diagnosis was made on fine-needle aspiration of the mass; pericardial fluid analysis never yielded a definitive cytologic diagnosis of tumor type. The exclusion criterion was lack of a histopathologic or cytologic diagnosis. All histopathologic sections were reviewed by a single anatomic pathologist (LEC) and all echocardiograms were reviewed by a single cardiologist (SAJ) for manuscript preparation; each was blinded to the results of the other at the time of the review. Some of the cases had only an echocardiographic tumor location reported in the medical record, whereas other cases had echocardiographic tumor location and presumptive diagnosis reported. For the cases that did not include a presumptive diagnosis, a presumptive diagnosis was made post hoc when the case was reviewed. The cardiologist that made the post hoc presumptive diagnoses (SAJ) was blinded to the definitive diagnosis made by the pathologist. The presumptive diagnosis in all cases was based mainly on location of the mass on echocardiogram (although additional characteristics such as echo texture and invasiveness were taken into account as well), such that right atrial masses were presumed to be HSA, heart base masses were presumed to be chemodectoma, ectopic thyroid carcinoma, or lymphosarcoma (CD/ETC/LSA), and masses in locations other than the right atrium or heart base were given an open diagnosis. The electronic medical record system at the hospital was used to collect all relevant information for each patient, including signalment, clinical signs at presentation, echocardiographic findings, the presence or absence of pericardial effusion, presumptive diagnosis based on echocardiogram, survival time from the diagnosis of a cardiac mass, ECG findings, cytologic or histopathologic diagnosis, and pertinent cardiac tumor-related necropsy findings if applicable.
A calculation of the percentage of correct presumptive diagnoses when compared with the histopathologic or cytopathologic definitive diagnoses was performed.
A search of the echocardiogram database at the University of Tennessee for dogs with cardiac tumors between 2006 and 2012 yielded 100 cases. Applying the inclusion criterion yielded 24 cases with a definitive diagnosis of the cardiac mass based on histopathology or cytology. The 24 dogs had varying signalments, clinical signs, duration of clinical signs, and survival (Table 1). Breeds included 3 Labrador Retrievers, 3 Golden Retrievers, 3 Shetland Sheepdogs, 2 English Bulldogs, 1 Samoyed, 1 Flat-Coated Retriever, 1 Toy Poodle, 1 Boxer, 1 Weimaraner, 1 German Shepherd, and 7 mixed breed dogs. Ten were male (1 intact) and 14 were female (1 intact). Age ranged from 3 to 13 years old. The duration of clinical signs ranged from <1 day to 3 months before presentation. Presenting concerns included respiratory signs (11), lethargy (5), neurologic signs (3), partial or complete inappetence (3), abdominal distention (2), collapse (1), epistaxis (1), and no clinical signs (4). Some dogs were presented with multiple clinical signs. The patients with no clinical signs had a number of presenting concerns from referring veterinarians, including an enlarged cardiac silhouette on survey radiographs, an abdominal mass on palpation, and pancytopenia.
|Echo Location of Mass||Age (y)||Clinical Signs||Duration of Clinical Signs||Survival|
|RA (8)||8–13|| |
Abdominal distention (1)
|1 d–3 m||1–83 d|
|HB (9)||3–13|| |
|1 d–6 m||1 d–>150 d|
|Other (7)||6–12|| |
Abdominal distention (1)
Head tilt (1)
|1 d–1 w||1–44 d|
On echocardiography, 9/24 dogs were described as having a predominant heart base mass and were presumed to have CD/ETC/LSA (1 also had right atrial involvement, and 1 also had biatrial and left ventricular involvement). Of these cases, all 6 that had a necropsy performed were confirmed to have a predominant heart base mass on gross necropsy. Based on histopathology or cytology, 7/9 were confirmed to have a diagnosis of CD/ETC/LSA (78% accuracy) whereas the other 2 cases were diagnosed as undifferentiated carcinoma. On echocardiography, 8/24 dogs were described as having a predominant right atrial or right auricular mass and were presumed to have HSA (2 also had a heart base location; 2 also had right ventricular involvement; and 1 also had left atrial involvement). Of these cases, 6/8 were confirmed to have a predominant right atrial mass on gross necropsy (the other 2 cases had a predominant heart base mass with extension to the right atrium). Based on histopathology, 4/8 were confirmed to have HSA (50% accuracy), whereas the other 4 cases were diagnosed with CD, ETC, thrombus, and inflammation. On echocardiography, 7/24 dogs were described as having a mass predominantly involving structures other than the heart base or right atrium and therefore had an open diagnosis (2 right ventricular outflow tract masses; 1 right ventricular, tricuspid valve, and right atrial mass; 1 right and left ventricular mass; 1 right ventricular mass; 1 left atrial mass; and 1 left ventricular and mitral valve mass). Of these cases, 6/7 were confirmed to have predominant masses in locations other than the heart base or right atrium on gross necropsy. The remaining case had a predominant right atrial mass with extension to the right ventricle. Based on histopathology, 1/7 had a diagnosis other than CD/ETC/LSA or HSA (chondrosarcoma). The other 6/7 cases were diagnosed with ETC (3), HSA (2), and LSA (1). These data are presented in Table 2.
|Definitive Diagnosis||ECG Findings||Pericardial Effusion||Echocardiographic Anatomy||Presumptive Diagnosis||Necropsy Anatomy|
|HSA (6)|| |
|CD/ETC/LSA (13)|| |
High 2° AVB (1)
|Other (5)|| |
Ten of the 24 cases had varying degrees of pericardial effusion ranging from mild to severe, and 5 cases had cardiac tamponade. One case had pericardiocentesis performed elsewhere 1 week before presentation. Of the cases with a final diagnosis of HSA, 4/6 had pericardial effusion. Of the cases with a final diagnosis of CD/ETC/LSA, 3/13 had pericardial effusion. Of the cases with a diagnosis other than HSA, CD, ETC, or LSA, 3/5 had pericardial effusion. These data are presented in Table 2.
ECG abnormalities included ventricular arrhythmias in 5/24, supraventricular tachycardia in 2/24, high 2° atrioventricular (AV) block in 1 case, and a right bundle branch block in 1 case. Of the cases with a final diagnosis of HSA, 2/6 had ECG abnormalities. Of the cases with a final diagnosis of CD/ETC/LSA, 5/13 had ECG abnormalities. Of the cases with a final diagnosis other than HSA, CD, ETC, or LSA, 1/5 had ECG abnormalities. These data are presented in Table 2.
In 23/24 dogs, survival ranged from <1 day to >150 days after the initial echocardiogram (1 dog was lost to follow-up). Analysis of survival time between groups of dogs was not performed because of the small number of dogs with each tumor type, because treatment was not pursued in any dog, and because decisions to euthanize were based on factors other than tumor type.
In this study, there were almost equal numbers of right atrial masses, heart base masses, and masses in other locations. This differs from other reports, in which right atrial masses or HSA was much more common than any other cardiac tumor in dogs.[2-4] The relatively high proportion of masses for which a presumptive diagnosis could not be made (7/24) likely reflects selection bias in which dogs were more likely to be submitted for necropsy if the echocardiographic appearance of the mass was unusual. We also found that although the echocardiographic description of the cardiac mass location in most cases agreed with the gross pathologic description (86% agreement based on 21 cases on which necropsy was performed), the presumptive diagnosis only agreed with the definitive diagnosis in 65% of the 17 cases with a presumptive diagnosis. In particular, the presumption that a right atrial mass represented HSA was only correct in 50% of cases, whereas the presumption of a heart base mass being CD/ETC/LSA was correct in 78% of cases. Finally, we found that tumors in other locations were nonetheless likely to be one of the 4 most common cardiac tumors in dogs (HSA, CD, ETC, LSA). In a previous report, the sensitivity and specificity of echocardiography performed by a board-certified cardiologist or supervised cardiology resident in training to detect cardiac masses in dogs with pericardial effusion were quite high (82% and 100%, respectively). That study also found that the differentiation between right atrial and heart base masses was very accurate, and that although most right atrial masses were HSA, other tumors also were found in this location, and most of the heart base masses were CD or ETC. The lack of pericardial effusion in most cases in this series may have decreased our ability to accurately delineate tumor origin and extension, although comparison between cases with and without pericardial effusion showed very similar rates of accuracy.
Despite low incidence, cardiac tumors can be clinically important in dogs, causing cardiac tamponade,[5, 7, 9, 11, 12] metastasis to or from other organs,[4, 5, 7-9] arrhythmias,[8, 9] and occasionally obstruction to blood flow[14-16] or heart failure. In addition, as with any neoplastic process, cardiac tumors can be associated with inappetence and lethargy. In this series, 10/24 cases had pericardial effusion and 50% of those were in cardiac tamponade at the time of the 1st echocardiogram. Pericardial fluid analysis for the most part has not been found to be useful in differentiating tumor types,[7, 10-13] although 2 more recent studies have shown that plasma cardiac troponin I concentration is increased in dogs with presumed cardiac HSA in comparison with dogs with pericardial effusion presumed not to be caused by HSA.[18, 19] We did not measure plasma cardiac troponin I concentration in these patients, however. In this series, 8/24 of the cases had ECG abnormalities. This serves as a reminder of the possible association between cardiac structure and electrophysiology. In fact, in some cases, it was the detection and characterization of an arrhythmia that led to the recommendation that an echocardiogram be performed.
This study has limitations. First, the case series is small and these 24 dogs are unlikely to be representative of the entire population of dogs with cardiac masses, especially given the bias of this study toward more unusual cases being submitted for necropsy. Also, dogs with cardiac tumors that are metastatic (and for which the primary tumor site dictates clinical signs, treatment, and prognosis) and cardiac tumors that are incidental would have been less likely to have had an echocardiogram and been included in this report. Second, the accuracy applies to only a few cardiologists at 1 institution. Other cardiologists at other institutions may have achieved higher or lower accuracy with regard to their presumptive diagnoses. Accuracy potentially also was affected by the low number of patients with pericardial effusion. Third, we used histopathology or cytology as the gold standard for the definitive diagnosis, but it is possible that the histopathologic or cytologic diagnosis was not 100% accurate. Finally, the clients in this study did not elect for treatment of any kind (pericardectomy, chemotherapy) and thus survival was skewed. Financial constraint and perception of the patient's quality of life at the time of diagnosis were the most common reasons that owners elected not to pursue pericardectomy in dogs with heart base masses.
Many of the 100 cases with cardiac masses noted on echocardiogram were lost to follow-up and died without a definitive diagnosis. Necropsy should be offered to clients of dogs with cardiac tumors when euthanasia is elected in the hopes of improving our understanding of the varied presentations of cardiac tumors.
Echocardiography is the most likely means to identify cardiac tumors in dogs antemortem. Although presumptive diagnoses are commonly made based on echocardiographic location of the mass (in addition to other characteristics such as echo texture and invasiveness), this case series illustrates that it is important to consider that the presumptive diagnosis may only be moderately accurate. These cases therefore might warrant a more extensive diagnostic evaluation, including pericardial fluid analysis with cardiac troponin I concentration if pericardial effusion is present, fine-needle aspiration, serum T4 concentration, cross-sectional imaging, and thorascopy or thoracotomy. Cardiac masses in dogs that were echocardiographically unusual in the present case series were nonetheless diagnosed frequently as HSA, CD, ETC, and LSA.
The authors gratefully acknowledge the primary clinicians who cared for these patients. This study was not supported by a grant. This study was not presented at a meeting.
Conflict of Interest Declaration: Authors disclose no conflict of interest.