Outcome and prognostic factors for dogs with a histological diagnosis of splenic hematoma following splenectomy: 35 cases (2001–2013)

Abstract

Canine splenic hematoma can be indistinguishable from hemangiosarcoma on clinical presentation and grossly at the time of surgery. However, hemangiosarcoma represents an aggressive malignancy and a misdiagnosis of hematoma would forgo indications for chemotherapy. This study describes a long-term follow-up of cases with a histologic diagnosis of splenic hematoma following splenectomy to determine if the clinical course of the disease corroborated the diagnosis. Thirty-five dogs were evaluated to determine survival and prognostic associations with signalment and clinical data. Overall median survival time was 647 days (range: 0 to 3287 days). Statistically significant variables included a palpable abdominal mass during physical examination, sub-clinical coagulopathy, and metastasis. Four cases (11%) had reported evidence of metastasis at the time of euthanasia; 1 case was histologically confirmed. Overall prognosis for splenic hematoma appears excellent, as expected, but a small proportion of cases may have an undiagnosed malignant component.

Introduction

Differential diagnoses for a canine splenic mass include malignant tumors such as hemangiosarcoma, lymphoma, histiocytic sarcoma, and other sarcomas (1,2) as well as benign masses including splenic hematoma, and lymphoid hyperplasia. Hemangiosarcoma is a malignancy arising from endothelium and is the most common neoplasm of the canine spleen (2,3). One study led to the development of the so-called two-thirds rule: approximately 2/3 of splenic masses will be malignant, and 2/3 of those will be hemangiosarcoma; the remaining 1/3 of splenic masses would be expected to be benign (4). Other studies have reported a 50% chance of malignancy without the presence of hemoperitoneum, with hemangiosarcoma the most common malignancy (2,5). The most common benign splenic lesion reported in dogs is splenic hematoma (3,6). Canine splenic hematoma and hemangiosarcoma can have a similar clinical presentation and gross appearance at the time of surgery. This is because a hematoma is a component of most splenic hemangiosarcomas. However, the prognosis of these 2 conditions is very different, with a poor prognosis in canine splenic hemangiosarcoma due to early and aggressive metastasis (7). With splenectomy alone, the reported median survival time is 86 d with an estimated 1-year survival rate of 6.25% (8). A more recent study reported a median survival time of 1.6 mo with 11.1% alive at 1 y and 4% alive at 2 y following splenectomy alone (3). Histopathology is required to differentiate hemangiosarcoma from hematoma. In cases in which the signalment is commonly associated with hemangiosarcoma, older and large breed dogs (1), a histological diagnosis of splenic hematoma can still create doubt for the clinician as to whether or not hemangiosarcoma is present, but was not included on histological sectioning. This is especially true if a large component of the splenic mass was a hematoma grossly.

The objective of this study was to retrospectively determine outcome and prognostic factors associated with a histological diagnosis of splenic hematoma following splenectomy in dogs. Secondary aims were to evaluate the potential to predict when cases may have been misclassified as hematoma, by evaluating clinical data to look for prognostic factors, which may be used to increase the clinical suspicion of malignancy rather than a benign splenic hematoma.

Materials and methods

Medical records at our veterinary teaching hospital were reviewed, retrospectively, for 35 dogs with a histopathological diagnosis of splenic hematoma following splenectomy during the period February 2001 to February 2013. The entire spleen was formalin-fixed and available for histopathologic evaluation via standard hematoxylin and eosin staining methods. A board-certified pathologist performed all histopathology. Potential prognostic factors including signalment and clinical data including gender, breed, body weight, splenic mass volume, hemorrhagic peritoneal effusion, anemia, blood transfusion, coagulation parameters, and palpable mass were collected from the medical records and evaluated for associations with survival or whether or not evidence of confirmed or suspected metastasis was determined, employing both univariate and multivariate analyses. Anemia and thrombocytopenia were defined as values below the normal reference range. All blood values were determined at the time of admission. Splenic mass volume was calculated via linear measurements in 3 dimensions of the splenic mass on the gross specimen following splenectomy. Outcome was determined through follow-up with referring veterinarians and owners. Patients in which no follow-up data were available and/or were still alive at the time of this study were censored from the final statistical analyses. Survival time was determined as the time between surgery and death and was reported using Kaplan-Meier survival curves. A Cox Proportional Hazard Model was employed to compare variables with survival data. Clinical variables were analyzed as likelihood estimates of affecting median survival time using both univariate and multivariate analyses. The Wilcoxon test, which places more weight on early survival times, is reported for palpable abdominal mass. Otherwise the score test is reported. Statistical significance was defined as P ≤ 0.05.

Results

The mean age at time of surgery was 10.1 y (range: 6.0 to 18.8 y). There were 54% castrated males and 46% spayed females (no intact animals were evaluated). The mean body weight was 29.8 kg (range: 6.4 to 59.1 kg). Three small breed dogs (< 10 kg) were included. No specific breed predilections were determined to be statistically significant prognostic indicators, although German shepherds (4/35; 11.3%) and Labrador retrievers (2/35; 6%) were the most common breeds represented. A historical and associated traumatic event was reported in only 3/35 dogs (9%); all 3 events consisted of a fall down the stairs prior to identification of a splenic mass but it is unknown whether a splenic mass existed prior to the traumatic event in any of these 3 cases. Three-view thoracic radiographs were taken in 27/35 (77%) dogs before splenectomy with no evidence of pulmonary metastasis reported in any case. A board-certified radiologist reviewed all thoracic radiographs. Abdominal ultrasound examination was also conducted in 27/35 (77%) dogs before surgery and ultrasonographic findings in all cases were consistent with a large, cavitated splenic mass +/− abdominal effusion with no other significant abnormalities within the peritoneal cavity. Cardiac ultrasounds were performed in 7 cases prior to surgery, no evidence of neoplasia and/or metastasis was reported in any of these patients. Two patients were diagnosed with mild mitral valve regurgitation secondary to mitral valve disease.

Hemorrhagic peritoneal effusion was present in 11/27 dogs (40%; based on abdominal ultrasound findings) and anemia was present in 17/35 dogs (48%). A blood transfusion was administered in 5/35 dogs (17%). Of the 11 dogs presenting with a hemorrhagic peritoneal effusion, 4 (36%) were given a blood transfusion. All patients received packed red blood cells. One patient received a blood transfusion after surgery for a hemorrhagic effusion secondary to a reported surgical complication. Concurrent hemorrhagic peritoneal effusion and anemic state was found in 8/35 (22%) cases. A coagulopathy was identified in 9/20 dogs (45%). This was defined as either a low platelet count (3/20 dogs; 15%) [2 were considered mild thrombocytopenias, and 1 was marked (19 × 10⁹/L; reference interval: 117 to 418 × 10⁹/L)], or elevated partial thromboplastin time (PTT) (7/20; 35%). One patient was mildly thrombocytopenic and had an elevated PTT. No clinical signs associated with coagulopathy were reported for any cases.

The overall median survival time (MST) was 674 d, with a range of 0 to 3287 d. Four patients were lost to follow-up; subsequently their survival data were censored.

Four dogs (11%) were euthanized for clinical signs and evidence of metastatic disease during the follow-up period. Clinical evidence of metastatic disease was reported in the referring veterinarian’s medical records at the time of euthanasia for 3 dogs. These 3 cases were not reassessed at our institution and histopathological confirmation of metastasis was not available. Two of these dogs had radiographic evidence of pulmonary metastasis and were euthanized by their primary care veterinarian. These cases included a 9-year-old castrated male golden retriever and a 9-year-old spayed female English setter. Their survival times were 138 and 329 d, respectively. The third case was a 13-year-old spayed female mixed breed dog (35 kg) which was presented to the primary care veterinarian with a hemorrhagic peritoneal effusion and had a hepatic mass based on abdominal ultrasound. This dog was also euthanized by the primary care veterinarian 522 d after splenectomy. Histopathologic confirmation of metastasis was not available for these 3 cases. The fourth case with evidence of metastatic disease re-presented to our hospital 168 d after splenectomy in hypovolemic shock with evidence of a marked tri-cavitary (peritoneal, pleural, and pericardial) hemorrhagic effusion. This was a 10-year-old castrated male mixed breed dog that underwent splenectomy, for an incidentally found splenic mass following staging for an unrelated soft tissue sarcoma recut of the left shoulder area. Staging prior to splenectomy included three-view thoracic radiographs, thoracic computed tomography (CT) scan and abdominal ultrasound, with no evidence of metastasis reported. Original histologic evaluation of the spleen provided a diagnosis of splenic hematoma and nodular hyperplasia. This patient was euthanized on the day it re-presented for tri-cavitary effusion and postmortem examination confirmed diffuse metastatic hemangiosarcoma both grossly and histologically. Of these 4 cases of metastatic disease (3 suspected), 3 presented initially with evidence of an anemic state; only 1 of the 3 was diagnosed with a hemorrhagic peritoneal effusion, and none of them received a blood transfusion. The other 2 cases were most consistent with anemia of chronic disease. Thoracic radiographs were obtained in 3 of the 4 cases prior to splenectomy with no evidence of pulmonary metastasis. The survival time of the patient not receiving thoracic radiographs prior to surgery was 522 d; therefore, if misdiagnosed it would not be anticipated that gross pulmonary metastasis would have been present pre-operatively. The remainder of the study population died of causes unrelated to splenic hematoma. One dog included in the study cohort was euthanized in the immediate post-operative period. Extensive and uncontrollable hemorrhage was noted during a second exploratory laparotomy following exploratory laparotomy and splenectomy the previous day. Postmortem examination revealed evidence of disseminated intravascular coagulopathy presumed secondary to pancreatitis.

Survival was not influenced by signalment. Presence of a palpable abdominal mass during physical examination was a negative prognostic factor, while evidence of a sub-clinical coagulopathy, and suspected or confirmed metastasis through univariate analyses were statistically significant positive prognostic factors for overall survival (Table 1). Subsequently, a palpable abdominal mass would be more indicative of a misdiagnosis of hematoma over a malignancy. Other evaluated parameters, such as splenic mass volume, hemorrhagic peritoneal effusion, anemia, body weight, and blood transfusion were not statistically significant prognostic factors. Representative Kaplan-Meier survival curves are depicted for the variables of whether a palpable abdominal mass was present during physical examination, a documented sub-clinical coagulopathy, presence of (presumed or confirmed) metastasis, as well as overall survival (Figure 1); gender, splenic mass volume, evidence of hemorrhagic peritoneal effusion, blood transfusion were also analyzed but not depicted as Kaplan-Meier survival curves. The development of presumptive/confirmed metastatic disease was not influenced by the same clinical variables. No statistically significant correlations were determined via multivariate analyses. All combinations of the collected data were evaluated.

Table 1.

Parameters evaluated for survival of dogs diagnosed with splenic hematoma. P-values, hazard ratios, and 95% confidence limits are reported. Statistical significance was determined for the examined variables palpable abdominal mass during physical examination, coagulopathy, and suspected/confirmed metastasis

				95% Confidence interval
	n	P-value	Hazard ratio	Lower	Upper
Gender (male/female)	35	0.7733	0.879	0.365	2.117
Splenic mass volumea (cm3)	24	0.842	0.950	0.735	1.229
Elevated ALP (yes/no)	32	0.4444	0.691	0.268	1.781
Hemorrhagic effusion (yes/no)	11	0.484	1.366	0.493	3.784
Anemia (yes/no)	35	0.972	0.953	0.403	2.254
Body weight (kg)	35	0.817	0.996	0.969	1.023
Transfusion (yes/no)	5	0.293	0.527	0.173	1.608
Coagulopathyb (yes/no)	20	0.0497	0.811	0.270	2.434
Palpable massb (yes/no)	35	0.0315	3.33	1.13	10.005
Metastasisb (yes/no)	4	0.005	0.204	0.060	0.692

ALP — alkaline phosphatase.

^aRepresents log transformed data.

^bRepresents statistically significant variables (P < 0.05).

Figure 1.

Kaplan-Meier survival curves for the clinical variables A — suspected/confirmed metastasis; B — palpable mass during physical examination; and C — coagulopathy. Each variable was determined to be a statistically significant influence on survival. D — Overall survival is also depicted.

Discussion

The reported MST of 674 d in our study, for dogs with a histopathological diagnosis of splenic hematoma, is longer than those previously reported. Two previous studies reported MST for splenic hematomas following splenectomy to be 338 d (9) and 252 d (5). Therefore, overall long-term survival with a diagnosis of splenic hematoma is excellent. It appears that a proportion of these cases may in fact be malignant cancer, with 4/35 (11%) in this series presenting with reported or confirmed clinical evidence of pulmonary (n = 2), hepatic (n = 1), or widespread (n = 1) metastasis. Unfortunately, histopathological confirmation of metastasis was not available in 3 of these cases and these reported lesions may have been the result of an alternate pathology. However, we believe that it is important to warn owners of the risk of misdiagnosis, and at the same time confer with the pathologists regarding the index of suspicion for a diagnosis of malignant splenic disease as this may warrant further histopathological investigation and scrutiny involving additional sectioning of tumor to definitively rule out a malignant component, especially given the association with concurrent (and potentially large) hematomas. In this study, the risk of this occurrence appears to be around 11%.

A previous study of splenic hematoma and hemangiosarcoma reported that 3/125 dogs originally diagnosed with splenic hematoma were reclassified to hemangiosarcoma following a review of the slides by 2 independent pathologists (9). The 4 dogs in our study that died of presumptive or confirmed metastatic disease had survival times of 138, 168, 329, and 522 d.

In the current study, a palpable abdominal mass was a negative prognostic factor, while evidence of a sub-clinical coagulopathy, and development of confirmed or presumed metastasis were both positive prognostic factors for overall survival in univariate analysis. A previous study characterized the prevalence of splenic hemangiosarcoma (76.1%) and hematoma (23.9%) in a population of 71 dogs presenting with anemia and hemorrhagic peritoneal effusion requiring a transfusion and found that the need for a transfusion was strongly associated with a histological diagnosis of hemangiosarcoma (10). In our study, 3 of 4 dogs with death due to suspected or confirmed metastatic disease initially presented in an anemic state, 1 dog had evidence of hemorrhagic peritoneal effusion and none of them were treated with a blood transfusion. Subsequently, in contrast to this previous study, our results did not indicate that a blood transfusion acted as a significant factor in the suspected determination of malignancy versus a benign hematoma. This may have been the result of a smaller patient population in the present study, not allowing statistical significance to be reached. Another recent study showed that splenic hematomas had significantly higher mass-to-splenic volume ratio and a significantly increased splenic weight as a percentage of total body weight compared to malignant disease (11). Our study found that splenic mass volume was not significantly associated with overall survival. Absolute splenic mass volume was compared in the current study rather than mass-to-spleen volume. Splenic volume was not routinely recorded as part of the medical record.

Not every case included in the final analysis of this study had a coagulation profile performed, but of the 20 that did, 9 had evidence of a coagulopathy (thrombocytopenia, or prolonged PTT). A coagulation profile was not performed in the 4 cases with evidence of metastatic disease. Evidence of a sub-clinical coagulopathy cannot be correlated with the risk of a misdiagnosed malignant splenic disease but rather acts as a positive prognostic indicator for “true” splenic hematoma. Coagulopathy was a positive predictor of survival in univariate analysis only in this study. It is not clear whether or not splenic hematoma was a cause or an effect of the coagulopathies noted.

Whether there is a histological diagnosis of hemangiosarcoma or of hematoma, several breed predilections are reported, including Labrador and golden retrievers, German shepherds, and standard poodles (8,12,13). Though breed was not statistically significant as a prognostic factor, both Labrador retrievers and German shepherds were most commonly represented in our study. Therefore, breed does not appear to be useful in predicting a diagnosis of hematoma versus malignant splenic disease, such as hemangiosarcoma.

In our study, 11% of the cases had reported evidence of presumptive or confirmed metastatic disease at the time of death. A recent study reported a median survival time of 5.5 mo for stage 1 (confined to the spleen with no evidence of hemoabdomen) hemangiosarcoma (3). The survival times for 2 of the dogs with presumptive metastasis, treated with splenectomy alone, were 329 and 522 d. One of these dogs had pulmonary metastasis and 1 had evidence of a liver mass and hemorrhagic peritoneal effusion. This may suggest that either the 2 dogs without histological confirmation had a primary tumor other than splenic hemangiosarcoma, or that even if these were misdiagnosed cases of hemangiosarcoma they represent a more indolent course than what has classically been understood for hemangiosarcoma. The other 2 dogs with evidence of presumptive or confirmed metastatic disease had survival times of 138 and 168 d. This is more consistent with previously reported median survival times for hemangiosarcoma. Contrast harmonic ultrasonography has been evaluated as a means to differentiate benign from malignant splenic disease and has failed to correlate with the eventual histologic diagnosis (14). Contrast harmonic ultrasonography has, however, proven useful in the differentiation of malignant versus benign hepatic masses (14,15).

The limitations of this study are that it was retrospective and that a single pathologist did not review the slides. Review of the slides may not have been helpful in determining if hemangiosarcoma cases had been misclassified as splenic hematoma cases, because in general, misclassification is more likely to be a factor of the portion of the mass selected in sampling and slide production, rather than incorrect interpretation of the diagnostic material available on the slide. Multiple recuts of the fixed specimen may be required in cases in which the clinical index of suspicion is that the case is of malignant splenic disease, but the material that is examined is consistent with hematoma. Additional recuts were not possible in this retrospective study because only the slides were available for review. It is unfortunate that postmortem evaluation was not available for the 3 dogs that were euthanized as a result of suspected pulmonary or liver metastasis. Although it is possible that these events were unrelated, it is interesting that these 3 dogs died of suspected metastasis to sites that are commonly affected by splenic hemangiosarcoma.

One of the aims of this study was to determine factors that might raise the clinician’s index of suspicion that a case diagnosed with splenic hematoma is actually a malignancy such as hemangiosarcoma. A palpable mass during physical examination was a statistically significant negative prognostic factor with respect to survival and as such increases suspicion of a malignant process rather than a benign splenic hematoma. Sub-clinical coagulopathy and evidence of metastasis conferred a statistically significant protective effect with respect to survival. However, all 3 variables were only significant in univariate analysis. We did not determine any factors that were predictive for the development of metastatic disease. It is possible that we were not able to achieve this because, in our study, this only occurred in 4/35 (11%) of cases. With only 4 cases of presumptive or confirmed metastatic disease, our power was likely too low. No factors were significant prognostic factors in multivariate analysis. Cox proportional hazard modeling produced hazard ratios indicative of a protective effect of both the presence of metastatic disease and a coagulopathy. These results, obviously counterintuitive, are likely the result of low statistical power and it is likely that a greater sample number would dilute these results. Low statistical power was also the likely cause of the lack of statistical significance for the multivariate analyses performed in this current study.

In conclusion, the median survival time for canine splenic hematoma in this study was 674 d. This is longer than previously reported and indicates that in most cases, the long-term prognosis is excellent, even in breeds that are predisposed to splenic hemangiosarcoma. A small percentage of cases (11%) may in fact represent malignant disease. We were not able to identify any prognostic factors or factors that may predispose cases to metastatic disease in multivariate analysis in this study. Every histological diagnosis should be interpreted in light of all features of the case and more exhaustive sectioning and histological scrutiny may be warranted in certain cases but the criteria for this require additional studies. CVJ