Abstract
Evaluation of dogs with splenic masses to better educate owners as to the extent of the disease is a goal of many research studies. We compared the use of ultrasonography (US) and contrast-enhanced computed tomography (CT) to evaluate the accuracy of detecting hepatic neoplasia in dogs with splenic masses, independently, in series, or in parallel. No significant difference was found between US and CT. If the presence or absence of ascites, as detected with US, was used as a pretest probability of disease in our population, the positive predictive value increased to 94% if the tests were run in series, and the negative predictive value increased to 95% if the tests were run in parallel. The study showed that CT combined with US could be a valuable tool in evaluation of dogs with splenic masses.
Résumé
Évaluation comparative du foie chez des chiens présentant une masse splénique à l’aide d’échographie et de tomodensitométrie à contraste renforcé. L’évaluation des chiens présentant une masse splénique dans le but de renseigner les propriétaires sur l’évolution de la maladie constitue l’objet de plusieurs recherches. Nous avons comparé l’utilisation de l’échographie (E) et de la tomodensitométrie à contraste renforcé (TC) pour évaluer précisément la détection de néoplasies hépatiques chez des chiens présentant une masse splénique, indépendamment, en série ou en parallèle. Il n’y avait pas de différence significative entre l’E et la TC. Si la présence ou l’absence d’ascite, détectée par E, constituait un critère de probabilité de la maladie dans notre population, la valeur prédictive positive augmentait à 94 % si les examens étaient effectués en série et la valeur prédictive négative augmentait à 95 % si les examens étaient effectués en parallèle. L’étude a démontré que la TC combinée à l’E pouvait être une méthode intéressante d’évaluation des chiens présentant une masse splénique.
(Traduit par Docteur André Blouin)
Introduction
Splenectomy because of either generalized splenomegaly or focal splenic masses is commonly performed in dogs. Splenic masses may be either benign (hemangioma, leiomyoma, myelolipoma, hematoma, abscess, nodular hyperplasia, infarction, cyst) or malignant (hemangiosarcoma, fibrosarcoma, leiomysarcoma, lymphosarcoma) (1). The prognosis for splenic masses varies depending upon whether the mass is benign or malignant. Prior studies showed that focal splenic masses in dogs were nonneoplastic or benign in more than 50% of cases (2–7). Attempts have been made to distinguish between benign and malignant splenic masses preoperatively, in order to educate owners of the prognosis. Due to the grave prognosis, many owners opt to have their dog euthanized if evidence of metastatic lesions are found preoperatively [ultrasonographs (US), thoracic radiographs] or intraoperatively (gross metastasis) (2,7). Fine needle aspiration of splenic masses has been shown to have poor sensitivity (8). Seeding tumor cells along the aspiration tract is also a concern; therefore, aspiration of splenic masses is usually not recommended (1,8–9). The ultrasonographic appearance of splenic masses, including hemangiosarcoma and hematoma, is variable; therefore, the sensitivity and specificity of US to differentiate malignant from benign disease is poor (1,10–12).
The most prevalent malignant splenic neoplasm in dogs is hemangiosarcoma (6). Hemangiosarcoma is most common in dogs more than 8 y old (13,14), with a predilection for German shepherds (4,14,15) and male dogs (15–17). The prognosis for hemangiosarcoma is poor. Surgery as the sole treatment results in a survival time of 19–83 d (15). Survival time can be extended with the addition of follow-up chemotherapy (18–20). The most common cause of death in dogs with hemangiosarcoma is metastatic disease. Metastasis to the liver, spleen, peritoneum, omentem, mesentery, kidneys, lymph nodes, adrenal glands, diaphragm, and nervous tissue has been reported to occur via hematogenous spread (4,7,14,16,17,21,22). Hemangiosarcoma frequently spreads to the liver via splenic venous blood, lymphatic drainage, or transcoelomic metastasis (23).
The presence of hepatic masses concurrently with a splenic mass increases the likelihood of a malignant process (1). Ultrasonographic images of hepatic hemangiosarcoma lesions range from that of mixed echogenicity to small anechoic lesions interspersed in areas of ill-defined, coarse, hypoechoic parenchyma (10). Ultrasonographic examination of the liver may be incomplete in dogs due to impedance by the ribs or gas in the gastric lumen. Additionally, the size of the patient can occasionally make complete examination of the liver difficult (24). This may cause hepatic lesions to go undetected. Computed tomography has been investigated as a tool to differentiate malignant from benign disease in the spleen of dogs (24). Magnetic resonance imaging has recently been reported to have a 94% accuracy of identifying malignant versus benign disease in the liver and spleen in dogs (25).
The use of US, CT, and magnetic resonance imaging (MRI) of the spleen and liver for detection of mass lesions in humans has been extensively described (26–32). When CT has been used, most hepatic metastases in humans have been described as hypodense on contrast-enhanced scans, depending upon the contrast phase (26). With injection of a venous contrast bolus, the arterial phase occurs within 20–30 s, with peak enhancement of normal hepatic parenchyma occurring 10–20 s later (portal hepatogram phase), because of the dual afferent hepatic circulation. The contrast then slowly equilibrates over the next 5 min (26). In humans, helical CT scanners allow the liver to be scanned during each of these phases (triple phase contrast-enhanced CT) (27). Recent studies have shown that contrast-enhanced US (CEUS) has similar efficacy to CT for the detection of hepatic metastasis in humans (25–29); however, CEUS is not widely used or available in veterinary medicine. The technique for detection of hepatic metastasis with the highest sensitivity in humans is intraoperative US (29).
The purpose of this study was to compare the accuracy of standard US and contrast-enhanced CT in detecting hepatic neoplasia in dogs with a splenic mass. Accurately identifying the presence or absence of hepatic neoplasia will allow clinicians to better educate owners as to the prognosis and, thus, whether or not to pursue surgery. Due to the occasional difficulty in imaging the entire liver with US and from reports of studies in humans, we hypothesized that contrast-enhanced CT would have a higher sensitivity and specificity than US in detecting hepatic metastasis in dogs with a splenic mass or masses.
Materials and methods
The inclusion criteria for this study included an abdominal US by 1 of 2 board certified ultrasonographers (JD,SF), an abdominal CT scan evaluated by a board certified radiologist (JS), and an exploratory laparotomy for a splenectomy and liver biopsy by either a board certified surgeon or a surgical resident under the supervision of a board certified surgeon. A histopathologic examination of the spleen and liver were required for all dogs. Only dogs with a focal splenic mass that had been diagnosed via US and confirmed at surgery were included in the study.
Dogs that presented to Garden State Veterinary Specialists between September 2003 and June 2004 with a complaint of collapse or a suspected abdominal mass were evaluated for inclusion in this study. All patients received a complete blood cell count and serum chemical profile. Patients were initially stabilized with IV fluids and supportive care. Packed red blood cell transfusions were administered according to each individual case need. Thoracic radiographs and an abdominal US, using an 8 MHz probe (GE Logiq 5 ultrasound; GE Medical Systems, Glendale Heights, Illinois, USA), were obtained.
All abdominal US findings were recorded. If hypoechoic or mixed echogenic hepatic nodules were identified, the dog was recorded as having hepatic neoplasia, based on US (Figure 1). If no hepatic nodules or if only 1 or 2 nodules were seen and the nodules were isoechoic, the dog was recorded as having either a normal liver or benign hepatic disease. The presence or absence of ascites, based on US, was also recorded for each patient.
Figure 1.
Ultrasonographic image of liver. Solid arrow — Hypoechoic hepatic parenchyma and suspected neoplasia.
Owners were informed of the US results. Some owners opted not to pursue surgery if multiple hepatic metastases were suspected on US and the prognosis was guarded. These dogs were euthanized and not included in the study. Dogs whose owners opted for surgical intervention were induced to anesthesia with propofol (Rapinovet; Schering-Plough Animal Health, Union, New Jersey, USA), 5 mg/kg bodyweight (BW), IV, to effect. They were intubated with a cuffed endotracheal tube and anesthesia was maintained with isoflurane (IsoFlo; Abbott, North Chicago, Illinois, USA) inhalant anesthetic. Dogs were positioned in a supine position within the CT machine (Siemens AR.T Helical scanner; Siemens AG D-80333, Munich, Germany). All dogs received 50% diatrizoate sodium (Hypaque; Amersham Health, Princeton, New Jersey, USA), 0.45 mL/kg BW, IV, to a maximum dose of 50 mL, as a slow bolus, immediately prior to initiation of the CT scan. Transverse slices (5 mm) of the entire abdominal cavity were obtained in axial mode, beginning at the cranial extent of the liver and ending at the level of the wings of the ilium, during the equilibration contrast phase. All CT scans were obtained and printed in a soft tissue window with an average width of 550 Hounsfield units (HU) and length of -10 HU for evaluation by the radiologist. The window was adjusted for each case in order to create a maximum contrast within the liver parenchyma.
When the CT scan had been completed, patients were brought to the surgical suite for a splenectomy and full abdominal exploration. Surgeons were aware of the US results at the time of surgery. The CT was not evaluated until after completion of the surgery; therefore, surgeons were unaware of the CT results. If the liver appeared grossly normal at surgery, a wedge biopsy was taken at the periphery of a convenient liver lobe. No attempt was made to biopsy deep liver tissue in patients whose hepatic parenchyma on US and at surgery appeared normal. Masses identified within the liver at surgery were excised if possible. If multiple hepatic masses were identified, 2 or 3 representative masses were excised for histopathologic examination. Any mass identified in the spleen at surgery was excised and submitted for histopathologic examination. All histopathologic examinations were performed by 1 of 2 commercial laboratories (22 cases to Antech Diagnostics, Lake Success, New York, USA and 3 cases to Idexx Veterinary Services, Westbrook, Maine, USA). The pathologists were unaware of the US and the CT results.
The CT scans were evaluated without the radiologist knowing the US, surgical, or histopathological findings. The liver parenchyma was evaluated for any nodule that differed in density from that of the rest of the liver parenchyma. Hypodense lesions within the liver parenchyma were described by the radiologist as probably metastatic or neoplastic (Figure 2). Nodules that were relatively isodense, or slightly less dense than their surrounding smooth hepatic parenchyema, were recorded as benign.
Figure 2.
Computed tomography image of liver. Solid arrow — Hypodense hepatic parenchyma and suspected neoplasia.
The US and CT findings were compared with the gold standard histopathologic findings. Any hepatic neoplasm identified histologically as a malignant neoplasm was classified as hepatic neoplasia. The kappa statistic was used to estimate the percent agreement beyond chance among US and CT tests to detect hepatic neoplasia. McNemar’s test was used to evaluate the hypothesis that the observed agreement among tests was due to chance alone. If no hepatic neoplasia was identified histologically, the patient was classified as not having hepatic neoplasia. The sensitivities and specificities of US and CT, when used separately to detect hepatic neoplasia, were calculated. Test accuracy was also calculated for US and CT and a chi square analysis was used to determine whether the difference among tests was statistically significant.
The use of US and CT in combination, as opposed to either test alone, was also evaluated. Overall sensitivity and specificity was calculated for the use of US and CT in series or in parallel. When used in series, a dog was classified as having a hepatic neoplasm, if both the US and CT test results were positive. When used in parallel, the dog was classified as having hepatic neoplasia, if either test was positive. Overall sensitivity, specificity, and test accuracy were calculated for the combination of US and CT when used in series or in parallel.
Results
Twenty-five dogs (3 intact ♂, 10 castrated♂, and 12 spayed ♀) met the criteria for inclusion in the study. There were 10 mixed-breed dogs, 4 German shepherds, 3 golden retrievers, 2 Labrador retrievers, 1 pointer, 1 wheaten terrier, 1 miniature schnauzer, 1 Samoyed, 1 German shorthaired pointer, and 1 Doberman pinscher. Ages ranged from 7 to 15 y (mean 10.6 y). None of the dogs had evidence of thoracic metastasis, based on 3-view thoracic radiographs. All of the dogs had been diagnosed with a focal splenic mass based on US.
Sixteen of 25 (64%) dogs had a malignant neoplasm diagnosed within the spleen, based upon histopathologic examination. Fifteen of these (94%) were diagnosed with hemangiosarcoma and 1 was diagnosed with leiomyosarcoma. The age range of dogs diagnosed with malignant splenic neoplasia in this study was 8 to 13 y (mean 10.6 y). Patients diagnosed with a malignant splenic neoplasm included 7 castrated ♂, 7 spayed ♀, and 2 intact ♂ dogs. Of these 16 dogs with a malignant splenic neoplasm, 11 (68.8%) had hepatic neoplasia confirmed at surgery and on histopathologic examination.
Nine of 25 (36%) dogs were diagnosed with benign disease within the spleen, based upon histopathologic examination. Five dogs were diagnosed with splenic hematoma, 2 with nodular hyperplasia, 1 with nodular hyperplasia and hematoma, and 1 with myelolipoma. The age range of dogs diagnosed histologically with benign disease of the spleen was 7 to 15 y (mean 11 y). Five of these dogs were spayed ♀, 3 were castrated ♀, and 2 were intact ♂. None of the dogs diagnosed with a benign mass in the spleen were diagnosed with hepatic neoplasia on histopathologic examination.
Eleven of 25 (44%) dogs were diagnosed with hepatic neoplasia, based upon the histopathologic examination. Nine (60%) of the 15 dogs with splenic hemangiosarcoma had histopathologic conformation of metastasis within the liver at the time of surgery. One (7%) of the dogs with splenic hemangiosarcoma was diagnosed with hepatocellular carcinoma in the liver, based upon histopathologic examination. One dog had splenic leiomyosarcoma with metastasis to the liver at the time of surgery.
Fourteen of 25 (56%) dogs had no evidence of hepatic neoplasia based upon histopathologic examination. Hepatic histopathologic findings of these dogs included 6 with normal liver, 2 with nodular hyperplasia, 2 with lymphoplasmacytic hepatitis, 1 with a possible regenerative nodule, 1 with vaculolar hepatopathy, 1 with centrolobular congestion, and 1 with hepatic congestion (Table 1).
Table 1.
Hepatic and splenic histopathologic results
| # | Breed | Sex | Age (y) | Hepatic biopsy findings | Splenic biopsy findings | Ascites |
|---|---|---|---|---|---|---|
| 1 | German shepherd | CM | 8 | Hemangiosarcoma | Hemangiosarcoma | + |
| 2 | Golden retriever | SF | 10 | Hemangiosarcoma | Hemangiosarcoma | + |
| 3 | German shepherd | SF | 13 | Hemangiosarcoma | Hemangiosarcoma | + |
| 4 | German shepherd | CM | 13 | Hemangiosarcoma | Hemangiosarcoma | + |
| 5 | German shepherd | CM | 9 | Hemangiosarcoma | Hemangiosarcoma | + |
| 6 | Mixed breed | M | 10 | Hemangiosarcoma | Hemangiosarcoma | + |
| 7 | Mixed breed | SF | 12 | Hemangiosarcoma | Hemangiosarcoma | + |
| 8 | Mixed breed | CM | 12 | Hemangiosarcoma | Hemangiosarcoma | + |
| 9 | Mixed breed | SF | 12 | Hemangiosarcoma | Hemangiosarcoma | − |
| 10 | Mixed breed | SF | 13 | Hepatocellular carcinoma | Hemangiosarcoma | + |
| 11 | Golden retriever | CM | 10 | Leiomyosarcoma | Leiomyosarcoma | − |
| 12 | Pointer | SF | 9 | Normal | Hemangiosarcoma | + |
| 13 | Min schnauzer | SF | 9 | Nodular hyperplasia | Hemangiosarcoma | + |
| 14 | Mixed breed | CM | 11 | Normal | Hemangiosarcoma | + |
| 15 | German short. pointer | M | 9 | Nodular hyperplasia | Hemangiosarcoma | − |
| 16 | Golden retriever | CM | 9 | Normal | Hemangiosarcoma | + |
| 17 | Mixed breed | SF | 8 | Hepatic congestion | Nodular hyperplasia | − |
| 18 | Labrador retriever | SF | 14 | Possible regenerative nodule | Hematoma | − |
| 19 | Mixed breed | CM | 10 | Normal | Hematoma | − |
| 20 | Wheaten terrier | SF | 15 | Normal | Hematoma | + |
| 21 | Samoyed | SF | 11 | Normal | Nodular hyperplasia/hematoma | + |
| 22 | Mixed breed | CM | 7 | Vacuolar hepatopathy | Nodular hyperplasia | − |
| 23 | Doberman pinscher | M | 7 | Centrolobular congestion | Hematoma | − |
| 24 | Mixed breed | CM | 12 | Mild lyphoplasmacytic hepaitis | Hematoma | − |
| 25 | Labrador retriever | SF | 15 | Lymphoplasmacytic hepatitis | Myelolipoma | − |
CM — castrated male
SP — spayed
M — male
Ultrasonography correctly identified 5/11 (45%) dogs with hepatic neoplasia. There were 2 dogs that were suspected to have hepatic neoplasia in which hepatic neoplasia was not histopathologically identified. Ultrasonography also correctly identified 12/14 (86%) dogs for which no histopathologic evidence of hepatic neoplasia was found, but failed to identify hepatic neoplasia in 6 dogs (Table 2).
Table 2.
Histopathologic, ultrasonographic, and computed tomographic results
| Ultrasonography
|
Computed tomography
|
||||
|---|---|---|---|---|---|
| Histopathologic results | Positive | Negative | Positive | Negative | Total |
| Hepatic neoplasia | 5 | 6 | 8 | 3 | 11 |
| No hepatic neoplasia | 2 | 12 | 3 | 11 | 14 |
Computed tomography correctly identified 8/11 (73%) dogs with hepatic neoplasia and misclassified 3 dogs as having hepatic neoplasia for which no histopathologic diagnosis of neoplasia was made. Eleven dogs that were classified by CT as free of hepatic neoplasia were later confirmed as true negatives histologically. Computed tomography misclassified 3 dogs as having normal livers when histopathologic examination identified hepatic neoplasia (Table 2).
The sensitivity of US in detecting hepatic neoplasia in our population was 45% (95% CI 16%, 75%) and the specificity was 86% (95% CI 67%, ~1). The sensitivity of CT in detecting hepatic neoplasia in our sample of dogs was 73% (95% CI 46%, 99%) and the specificity was 79% (95% CI 57%, ~1). The test accuracies for US (68%) and CT (76%) were not significantly different, P = 0.53. The agreement between US and CT in detecting hepatic neoplasia in our study is shown in the 2 × 2 table (Table 3). The estimated percent of agreement using McNemar’s test (49%) beyond that expected to occur by chance alone was considered moderate and was not statistically significant, P = 0.06.
Table 3.
Ultrasonography vs computed tomography for the detection of hepatic neoplasia
| Ultrasonography
|
|||
|---|---|---|---|
| Positive | Negative | ||
| Computed tomography | Positive | 6 | 5 |
| Negative | 1 | 13 | |
Thirteen of the 16 (81.25%) patients that were diagnosed with malignant tumor within the spleen had ascites at the time or surgery, including all but 1 of the 5 dogs with hemangiosarcoma in the spleen but not the liver. Of the 9 dogs diagnosed with benign splenic disease, only 2 (22%) had ascites at the time of surgery. A higher percentage of patients with ascites detected by US had a histopathologic diagnosis of hepatic neoplasia. For patients with ascites, 9/15 (60%) had evidence of hepatic neoplasia; for patients without ascites, only 2/10 (20%) had evidence of hepatic neoplasia.
When the tests were used in parallel, the overall negative predictive value changed from 68% (US) or 78% (CT) to 85% (Table 4). The sensitivity increased from 45% (US) or 73% (CT) to 85% when the tests were run in parallel. The specificity increased from 86% (US) or 79% (CT) to 97% when the tests were run in series. The test accuracy was highest for the tests run in parallel (77%).
Table 4.
Comparison of ultrasonography (US) and computed tomography (CT) when used alone, in series, and in parallel to diagnose the presence of hepatic neoplasia among dogs with splenic masses (n = 25).
| Test | Sensitivity | Specificity | Positive predictive value | Negative predictive value | Test accuracy |
|---|---|---|---|---|---|
| US | 45% | 86% | 71% | 67% | 68% |
| CT | 73% | 79% | 73% | 79% | 76% |
| Series | 33% | 97% | 90% | 65% | 65% |
| Parallel | 85% | 68% | 68% | 85% | 77% |
The study group was split into patients with and patients without ascites based on US. Ascites changed the pretest probability of a patient having hepatic neoplasia. When these estimates were substituted for the pretest probability of hepatic neoplasia, the positive and negative predictive values for US and CT in the presence of ascites were 83% and 51% and 84% and 66%, respectively (Table 5a). In the absence of ascites, the positive and negative predictive values for US and CT were 45% and 86% and 46% and 92%, respectively. When the tests are run in series and the patient has ascites, the positive and negative predictive values are 94% and 49%, respectively. When ascites is absent, the positive and negative predictive values are 73% and 85%, respectively. For parallel testing, the positive and negative predictive values when the patient has ascites are 80% and 75%, respectively. When ascites is absent, positive and negative predictive values are 40% and 95%, respectively.
Table 5a.
Comparison of ultrasonography (US) and computed tomography (CT) when used alone, in series, and in parallel to diagnose the presence of hepatic neoplasia among dogs with splenic masses and ascites (n = 15)*
| Test | Sensitivity | Specificity | Positive predictive value | Negative predictive value |
|---|---|---|---|---|
| US | 45% | 86% | 83% | 51% |
| CT | 73% | 79% | 84% | 66% |
| Series | 33% | 97% | 94% | 49% |
| Parallel | 85% | 68% | 80% | 75% |
Pretest probability of hepatic neoplasia in the presence of ascites 9/15 (60%)
Discussion
Preoperative diagnostics to identify dogs with malignant disease are helpful when treatment decisions are being made in dogs with splenic masses. In 1988, US was shown to be the best tool, when compared with radiography or laboratory testing, for detecting canine splenic hemangiosarcoma and the resulting liver metastasis (10). Studies have shown CT to be superior to standard US for detection of hepatic metastasis in humans (30,36–38). Contrast enhanced CT is considered essential to the examination of human patients with suspected hepatic metastasis (27–29,33,34), although, recently, MRI has been shown to be preferable (35). Magnetic resonance imaging has limited availability and can be cost prohibitive in veterinary medicine. Future studies to evaluate the detection of hepatic metastasis should consider including the use of CEUS or intraoperative US, as these techniques become more readily available.
Small sample size and exclusion of the dogs that did not have surgery due to US findings were limitations of the study. The wide confidence intervals around the estimates of sensitivity and specificity and the moderate agreement between the 2 tests are a reflection of the relatively small sample size. If a larger number of dogs or if all dogs diagnosed with a splenic mass, including those that were euthanized, had been evaluated with CT and necropsy, estimates of sensitivity and specificity may have been affected; however, the dogs that were subjected to CT were likely representative of those that would be evaluated in a practice setting.
In this study, only post bolus contrast CT scans of the abdomen were obtained and evaluated by the radiologist. In order to limit the duration of anesthesia and closely represent the clinical situation, where the equipment needed for triple phase contrast CT is often not available, this study evaluated the use of bolus contrast CT scans of the abdomen. With the technique used, the CT scan of the liver occurred during the equilibration phase. In human studies, hepatic lesions were most clearly identified in the portal-venous or equilibration phase post contrast injection as regions lacking contrast enhancement (27). Malignant masses (hemangiosarcoma) in spleens of dogs have been shown to have significantly lower HU values than nonmalignant masses (hematomas and nodular hyperplasia) in both pre- and post-contrast studies (24). A further limitation of this study was the radiologist’s use of a perceived difference in solid nodule density from a contrast-enhanced background to determine malignancy rather than HU.
The test accuracy for US and CT were not statistically different in this study; therefore, one test cannot be recommended over the other for the detection of hepatic neoplasia in dogs with a splenic mass. By using the tests in parallel the sensitivity and negative predictive value were increased and the specificity and positive predictive value were decreased when compared with either test separately. If the presence of ascites was accounted for, both the positive and negative predictive value of the tests run parallel increased. The increase in negative predictive value to 95% when no ascites was present increases the confidence that hepatic neoplasia is not present before treatment decisions and surgery are recommended. Likewise, a positive predictive value of 94% when ascities is present can also provide practitioners and owners with information to help decide whether or not to pursue surgical intervention.
All 4 German shepherds included in the study had hemangiosarcoma in both the spleen and the liver at the time of surgery. This is consistent with previous findings showing that German shepherds have a predilection for hemangiosarcoma (4,14,15). This study did not find a sex predilection for hemangiosarcoma, as had been previously reported (15–17).
Most splenic hematomas are a result of necrosis of hyperplastic nodules; however, some may be a result of underlying neoplasia (6). The current study differs from previous reports that state at least 50% of splenic masses are benign (2–7). The actual percentage of dogs that presented for evaluation with malignant splenic neoplasms was most likely higher than the 64% in our study because dogs diagnosed with US to have both a splenic mass and multiple hepatic nodules were often euthanized by the owners and were not included in the study. The exclusion of these patients led to a bias in the study sample where dogs that either had no hepatic neoplasia or had hepatic neoplasia that was difficult to identify on US were overrepresented. This may have resulted in the actual sensitivity for both US or CT being underestimated, if all the results of these dogs had been incorporated in the study. Had all dogs diagnosed with a splenic mass received a CT scan and histopathologic examination of the liver, so that these dogs’ results could have been included in the study, it would have been ideal.
Both internists have years of experience with ultrasonography in a referral veterinary hospital, but another limitation of this study was the lack of an on-site radiologist to perform and interpret the ultrasonographs. This may affect the reliability of the results when used in a different practice setting.
Computed tomography findings were not available at the time of surgery; therefore, potential lesions deep within the liver parenchyma that could not be detected by palpation or by observation at surgery may not have been directly biopsied. This limitation may have led to an increase in false positive CT results, leading to an underestimation of sensitivity and specificity for CT. In this study, 3 dogs were classified as having a false positive CT result. All 3 dogs were diagnosed with a benign splenic mass (2 hematoma and 1 myelolipoma). On histopathologic examination of the liver, all 3 dogs had benign changes diagnosed (possible regenereative nodule, centrolobular congestion, and lyphoplasmacytic hepataitis). Both the splenic and hepatic biopsy results in these patients supported the absence of hepatic neoplasia; therefore, based on the described criteria, the calculated sensitivity and specificity for CT is accurate.
Even though this study did not demonstrate a statistically significant difference between the accuracy of US and contrast enhanced CT for the detection of hepatic neoplasia, it did demonstrate that CT can have a role in the preoperative diagnostic evaluation of the abdomen of dogs with a splenic mass. This study did demonstrate that US and CT, when used in series with the presence of ascites had a 94% positive predictive value and, therefore, would be useful to rule-in hepatic neoplasia or metastasis. The negative predictive value of 95%, when the tests were used in parallel without the presence of ascites strongly suggests the absence of hepatic neoplasia in dogs with a splenic mass. Computed tomography did demonstrate a good ability to evaluate the hepatic parenchyma of dogs with hemoabdomen and splenic disease. Computed tomography has additionally been shown to have benefit in evaluation of primary splenic neoplasms (17). Computed tomography still needs to be evaluated for its benefits in evaluating the rest of the abdominal cavity and potentially the thoracic cavity.
Further investigation of preoperative abdominal CT evaluation of patients is warranted. Studies evaluating hepatic parenchyma, using helical CT while infusing contrast material (wash in/wash out) may improve the sensitivity for detecting hepatic neoplasia. It is possible to use CT scan, in combination with US, as a screening tool for patients with splenic masses to try and differentiate those with hepatic neoplasia from those without. In addition, if the thoracic cavity were scanned, it might be possible to detect pulmonary and cardiac metastasis. As availability of CT increases and CT becomes more cost efficient, it should become a more important preoperative screening tool for intraabdominal metastasis and neoplasia.
This study suggests that contrast-enhanced CT has similar sensitivity and specificity to US for detection of hepatic metastasis in dogs with a splenic mass. Additional studies are needed to confirm this finding. This study did show that contrast-enhanced CT and US used in serial or parallel can increase the sensitivity and specificity of either test alone. Results of the present study provide evidence that contrast-enhanced CT when combined with US can be as valuable a preoperative tool to evaluate the hepatic parenchyma for evidence of metastasis from splenic pathology. Further studies involving CT evaluation of the abdomen, as well as the possible use of CT to look for thoracic or cardiac metastasis, are indicated. CVJ
Table 5b.
Comparison of ultrasonography (US) and computed tomography (CT) when used alone, in series, and in parallel to diagnose the presence of hepatic neoplasia among dogs with splenic masses without the presence of ascites (n = 10)*
| Test | Sensitivity | Specificity | Positive predictive value | Negative predictive value |
|---|---|---|---|---|
| US | 45% | 86% | 45% | 86% |
| CT | 73% | 79% | 46% | 92% |
| Series | 33% | 97% | 73% | 85% |
| Parallel | 85% | 68% | 40% | 95% |
Pretest probability of hepatic neoplasia in the absence of ascites 2/10 (20%)
References
- 1.Neer TM. Clinical approach to splenomegaly in dogs and cats. Compend Contin Educ Pract Vet. 1996;18:35–46. [Google Scholar]
- 2.Johnson KA, Powers BE, Withrow SJ, et al. Splenomegaly in dogs — predictors of neoplasia and survival after splenectomy. J Vet Intern Med. 1989;3:160–166. doi: 10.1111/j.1939-1676.1989.tb03092.x. [DOI] [PubMed] [Google Scholar]
- 3.Hosgood G. Splenectomy in the dog: A retrospective study in 31 cases. J Am Anim Hosp Assoc. 1987;23:275–283. [Google Scholar]
- 4.Frey AJ, Betts CW. A retrospective survey of splenectomy in the dog. J Am Anim Hosp Assoc. 1977;13:730–734. [Google Scholar]
- 5.Prymak C, McKee LJ, Goldschmidt MH, et al. Epidemiologic, clinical, pathologic, and prognostic characteristics of splenic hemangiosarcoma and splenic hematoma in dogs: 217 cases (1985) J Am Vet Med Assoc. 1988;193:706–712. [PubMed] [Google Scholar]
- 6.Spangler WL, Culbertson MR. Prevalence, type, and importance of splenic diseases in dogs: 1,480 cases (1985–1989) J Am Vet Med Assoc. 1992;200:829–834. [PubMed] [Google Scholar]
- 7.Brown NO, Patnaik AK, MacEwen EG. Canine hemangiosarcoma: Retrospective analysis of 104 cases. J Am Vet Med Assoc. 1985;86:56–58. [PubMed] [Google Scholar]
- 8.O’Keefe DA, Couto CG. Fine-needle aspiration of the spleen as an aid in the diagnosis of splenomegaly. J Vet Intern Med. 1987;1:02–109. doi: 10.1111/j.1939-1676.1987.tb01997.x. [DOI] [PubMed] [Google Scholar]
- 9.Hosgood G. Canine hemangiosarcoma. Compend Contin Educ Pract Vet. 1991;13:1065–1075. [Google Scholar]
- 10.Wrigley RH, Park RD, Konde LJ, et al. Ultrasonographic features of splenic hemangiosarcoma in dogs: 18 cases (1980–1986) J Am Vet Med Assoc. 1988;192:1113–1117. [PubMed] [Google Scholar]
- 11.Hanson JA, Pennick DG. Ultrasonographic evaluation of traumatic splenic hematoma and literature review. Vet Radiol Ultrasound. 1994;35:463–466. [Google Scholar]
- 12.Feeney DA, Johnston GR, Hardy RM. Two-dimensional, gray-scale ultrasonography for assessment of hepatic and splenic neoplasia in the dog and cat. J Am Vet Med Assoc. 1984;184:68–81. [PubMed] [Google Scholar]
- 13.Chun R. Feline and canine hemangiosarcoma. Compend Contin Educ Pract Vet. 1999;21:622–629. [Google Scholar]
- 14.Pearson CR, Head KW. Malignant hemangioendothelioma in the dog. J Small Anim Pract. 1976;17:737–745. doi: 10.1111/j.1748-5827.1976.tb06937.x. [DOI] [PubMed] [Google Scholar]
- 15.Hammer AS, Couto CG, Filippi J, et al. Efficacy and toxicity of VAC chemotherapy (vincristine, doxorubicin and cyclophosphamide) in dogs with hemangiosarcoma. J Vet Intern Med. 1991;5:160–166. doi: 10.1111/j.1939-1676.1991.tb00943.x. [DOI] [PubMed] [Google Scholar]
- 16.Kleine LJ, Zook BC, Munson TO. Primary cardiac hemangiosarcoma in dogs. J Am Vet Med Assoc. 1970;157:326–337. [PubMed] [Google Scholar]
- 17.Oksenen A. Hemangiosarcoma in dogs. J Comp Pathol. 1978;88:585–595. doi: 10.1016/0021-9975(78)90012-9. [DOI] [PubMed] [Google Scholar]
- 18.Ogilvie GK, Powers BE, Mallinckrodt CH, et al. Surgery and doxorubicin in dogs with hemangiosarcoma. J Vet Intern Med. 1996;10:376–384. doi: 10.1111/j.1939-1676.1996.tb02085.x. [DOI] [PubMed] [Google Scholar]
- 19.Sorenmo KU, Jeglum KA, Helfand SC. Chemotherapy of canine emangiosarcoma with doxorubicin and cyclophosphamide. J Vet Intern Med. 1993;7:370–376. doi: 10.1111/j.1939-1676.1993.tb01033.x. [DOI] [PubMed] [Google Scholar]
- 20.Moulton JE. Tumors in Domestic Animals. 2. Berkeley, California: Univ California Pr; 1978. pp. 35–36. [Google Scholar]
- 21.Fees DL, Withrow SJ. Canine hemangiosarcoma. Compend Contin Educ Pract Vet. 1981;3:1047–1052. [Google Scholar]
- 22.Jubb KVF, Kennedy PC, Palmer N. Pathology of Domestic Animals. 3. Orlando, Florida: Academic Pr; 1985. pp. 195–200. [Google Scholar]
- 23.Wrigley RH, Konde LJ, Park RD, et al. Clinical features and diagnosis of splenic hematomas in dogs: 10 cases (1980–1987) J Am Anim Hosp Assoc. 1989;25:371–375. [Google Scholar]
- 24.Fife WD, Samii VF, Drost T, et al. Comparison between malignant and nonmalignant splenic masses in dogs using contrast-enhanced computed tomography. Vet Radiol Ultrasound. 2004;45:289–297. doi: 10.1111/j.1740-8261.2004.04054.x. [DOI] [PubMed] [Google Scholar]
- 25.Clifford CA, Pretorius ES, Weisse C, et al. Magnetic resonance imaging of focal splenic and hepatic lesions in the dog. J Vet Intern Med. 2004;18:330–338. doi: 10.1892/0891-6640(2004)18<330:mriofs>2.0.co;2. [DOI] [PubMed] [Google Scholar]
- 26.Friedman AC, Dachman AH, editors. Radiology of the Liver, Biliary Tract, and Pancreas. St. Louis: Mosby-Year Book; 1994. pp. 244–249. [Google Scholar]
- 27.Dietrich CF, Kratzer W, Strobel D, et al. Assessment of metastatic liver disease in patients with primary extrahepatic tumors by contrast-enhanced sonography versus CT and MRI. World J Gastroenterol. 2006;12:1699–1705. doi: 10.3748/wjg.v12.i11.1699. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Leen E, Ceccotti P, Kalogeropoulou C, et al. Prospective multicenter trial evaluating a novel method of characterizing focal liver lesions using contrast-enhanced sonography. Am J Roentgenol. 2006;186:1551–1559. doi: 10.2214/AJR.05.0138. [DOI] [PubMed] [Google Scholar]
- 29.Quaia E, D’Onofrio M, Palumbo A, et al. Comparison of contrast- enhanced ultrasonography verses baseline ultrasound and contrast- enhanced computed tomography in metastatic disease of the liver: Diagnostic performance and confidence. Eur Radiol. 2006;16:1599–1609. doi: 10.1007/s00330-006-0192-7. [DOI] [PubMed] [Google Scholar]
- 30.Ferrozzi F, Bova D, Draghi F, et al. CT findings in primary vascular tumors of the spleen. Am J Roentgenol. 1996;166:1097–1101. doi: 10.2214/ajr.166.5.8615251. [DOI] [PubMed] [Google Scholar]
- 31.Stark DD, Wittenberg J, Butch RJ, et al. Hepatic metastases: Randomized controlled comparison of detection with MR imaging and CT. Radiology. 1987;165:399–406. doi: 10.1148/radiology.165.2.3310094. [DOI] [PubMed] [Google Scholar]
- 32.Bernadino ME, Thomas JL, Barnes PA, Lewis E. Diagnostic approaches to liver and spleen metastases. Radiol Clin North Am. 1982;20:469–485. [PubMed] [Google Scholar]
- 33.Malt RA. Medical intelligence: Current concepts — surgery for hepatic neoplasms. N Engl J Med. 1985;313:1591–1596. doi: 10.1056/NEJM198512193132507. [DOI] [PubMed] [Google Scholar]
- 34.Finlay IG, Meek DR, Gray HW, et al. Incidence and detection of occult hepatic metastases in colorectal carcinoma. BMJ. 1982;284:803–805. doi: 10.1136/bmj.284.6318.803. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Wegener OH, Mutzel W, Souchon R. Contrast media for computed tomography of the liver. Acta Radiol Diagn. 1980;21:239–247. doi: 10.1177/028418518002102a16. [DOI] [PubMed] [Google Scholar]
- 36.Pamilo M, Suramo I, Lahde S. Comparison of ultrasound and CT images of liver metastasis. Acta Radiol Diagn. 1983;24:139–143. doi: 10.1177/028418518302400206. [DOI] [PubMed] [Google Scholar]
- 37.Alderson PO, Adams DF, McNeil BJ. Computed tomography, ultrasound, and scintigraphy of the liver in patients with colon or breast carcinoma: A prospective comparison. Radiology. 1983;149:225–230. doi: 10.1148/radiology.149.1.6611929. [DOI] [PubMed] [Google Scholar]
- 38.Knoph DR, Torres WE, Fajman WJ, et al. Liver lesions: Comparative accuracy of scintigraphy and computed tomography. Am J Roentgenol. 1982;138:623–627. doi: 10.2214/ajr.138.4.623. [DOI] [PubMed] [Google Scholar]