JOURNAL of the American Animal Hospital Association 417
Prognosis for Dogs with Stage I or II
Splenic Hemangiosarcoma Treated by
Splenectomy Alone: 32 Cases (1991–1993)
Carrie A. Wood, DVM
Antony S. Moore, MVSc
John M. Gliatto, VMD
Lee A. Ablin, DVM
Robert J. Berg, DVM
William M. Rand, PhD
From the Departments of Medicine
(Wood, Moore, Berg) and
Tufts University School of
200 Westboro Road,
North Grafton, Massachusetts 01536; the
Angell Memorial Animal Hospital (Ablin),
350 South Huntington Avenue,
Boston, Massachusetts 02130; and
Tufts University Community Health
35 Kneeland Street,
Boston, Massachusetts 02111.
Doctor Wood’s current address is the
Department of Small Animal
College of Veterinary Medicine,
University of Minnesota,
C352 Veterinary Hospitals,
1365 Gortner Avenue,
St. Paul, Minnesota 55108.
Presented in part at the
14th Annual Conference,
Veterinary Cancer Society,
A retrospective analysis was performed on the case records of 32 dogs with Stage I
or II splenic hemangiosarcoma that were treated by splenectomy alone and that
survived the seven-day postoperative period. Median survival time for these 32
cases was 86 days (mean, 116 days; range, 14 to 470 days), and the one-year
survival rate was estimated to be 6.25%. Survival was not influenced by signalment,
presenting signs, stage of disease, or clinicopathological findings. The data provides
a basis from which to evaluate adjuvant chemotherapy for splenic hemangiosarcoma
that is confined to the spleen macroscopically.
J Am Anim Hosp Assoc 1998;34:417–21.
Hemangiosarcoma is a malignant neoplasm originating from the vas-
cular endothelium that is characterized by widespread metastases and
poor survival rates. Hemangiosarcoma may arise from any site in the
body, but the spleen is affected most commonly in dogs.
primary sites include the right atrium and auricular appendage, liver,
lungs, kidney, skin, and the subcutis.
Splenic hemangiosarcoma affects older dogs, and the German shep-
herd dog may be at greater risk than other breeds for developing the
Clinical signs vary; acute collapse may follow rupture of
the highly vascular primary tumor, while other cases may show non-
specific signs of weakness, pallor, or anorexia.
Causes of death in
dogs with primary splenic hemangiosarcoma include exsanguination
when the primary tumor ruptures, metastatic disease, secondary dis-
seminated intravascular coagulation (DIC), and cardiac arrythmias.
Clinical staging has been a variable indicator of prognosis.
Stage I splenic hemangiosarcomas are confined to the spleen with no
evidence of metastases, while Stage II tumors may have ruptured and
may or may not have regional lymph-node involvement [see Appen-
Splenic hemangiosarcomas have been treated traditionally through
splenectomy, with poor survival times reported.
little data is available characterizing the prognosis and clinical pro-
gression of those cases receiving surgical excision of the primary
tumors which have no evidence of gross metastatic disease.
Small sample numbers, inclusion of animals that died in the immedi-
ate postoperative period, and inclusion of animals that received adju-
vant chemotherapy or immunotherapy make conclusions unreliable.
The reported survival times vary from 19 to 143 days. Comparison
among studies results in uncertain conclusions, as many reports do
not distinguish between stage of disease, tumor location, or treatment
In a large study of 104 cases of hemangiosarcoma, eight cases were
reported as Stage I disease (median survival time, 151 days), and 15
cases were reported as Stage II disease (median survival time, 143
days). However, three of the Stage I disease cases and seven of the
Stage II disease cases received either a mixed bacterial vaccine alone
418 JOURNAL of the American Animal Hospital Association September/October 1998, Vol. 34
or a vaccine in combination with various chemothera-
pies. Survival data was not presented for Stage I or II
disease cases treated with surgery alone.
Another report reviewed 92 cases of splenic hem-
angiosarcoma for which tissue samples were submit-
ted to a pathology laboratory service. Survival data
(median survival time, 19 days) was available for 59
cases. Data regarding clinical staging was not pre-
sented, and survival calculations included those ani-
mals that died in the immediate postoperative period.
A survey of 100 cases of splenectomies for various
indications reported survival data in four cases of
Stage I disease (median survival time, 91 days) and
one case of Stage II disease (survival time, 168 days).
The relatively small number of cases available makes
conclusions regarding overall survival times unreli-
able. Clinical staging was reported to be of no prog-
nostic significance; however, most cases that were
used for comparison received adjuvant chemo-
A more recent retrospective analysis of 1,480 cases
of canine splenic disease included data from 22 cases
of splenic hemangiosarcoma in which tissues were
submitted for histopathological evaluation. Mean sur-
vival time for these 22 cases was reported as 91 days.
Clinical stages of the disease were not reported, and
cases that died in the immediate postoperative period
were included in survival data calculations.
The primary purpose of this retrospective study
was to determine survival times for dogs with Stage I
or II splenic hemangiosarcomas that were treated by
splenectomy alone and survived the immediate post-
operative period. The secondary purpose was to de-
termine if the current staging system independently
predicted survival in this group of cases.
Materials and Methods
Medical records of 32 dogs seen at the Tufts Univer-
sity School of Veterinary Medicine (n=17) and the
Angell Memorial Animal Hospital (n=15) in the three-
year period (January 1, 1991 through December 31,
1993) were reviewed retrospectively. Cases were cho-
sen on the basis of a histopathological diagnosis of
hemangiosarcoma confined to the spleen (either in-
tact or ruptured) and no evidence of gross metastatic
disease. All cases were treated with surgery alone and
were excluded from the study if any adjunctive thera-
pies were given postoperatively. Only those cases
that survived the immediate postoperative period (i.e.,
seven days) were included. All histopathological
slides were reviewed, and the diagnoses were con-
firmed by one pathologist (Gliatto).
Details of age, breed, and sex were recorded along
with presenting history, duration of signs, and physi-
cal examination findings. Each case had a complete
blood count, serum biochemical profile, and thoracic
radiographs performed. Results of abdominal
radiographs (n=24), abdominal ultrasonographic ex-
amination (n=21), echocardiography (n=6), electro-
cardiography (n=19), or liver biopsy (n=20) were
recorded. Results of any additional adjunctive tests
were noted. Follow-up information was obtained
through telephone contact and a survey mailed to
owners and referring veterinarians. Owners confirmed
signalment and date of surgery. They also answered
questions regarding clinical signs following surgery,
current status (i.e., alive or dead) of the animal, and
date and cause of death.
Survival data was obtained from the medical
records and owner contact and was analyzed using
standard methods. Survival time was defined as the
interval from splenectomy until death. All computa-
tions were performed using a software package.
Kaplan-Meier product limit estimate of the survivor
function was generated, and standard parametric
survival distributions (i.e., exponential, Weibull,
Rayleigh, gamma, normal, and log normal) were fit-
ted to the data, with log normal calculated to meet the
goodness of fit criterion.
The Peto generalization of
Wilcoxon’s 2-sample rank sum test was used to ex-
amine the potential relationship between survival and
the following factors: Stage I or Stage II disease,
presenting signs of collapse or abdominal distention,
presence of anemia (i.e., hematocrit less than 32%),
DIC, postoperative ventricular arrhythmias, or ad-
ministration of a blood transfusion. The same test
also was used to assess the effects of gender and
Clinical Staging System for
T Primary tumor
T0—no evidence of tumor
T1—tumor confined to spleen
T2—tumor confined to spleen, but ruptured
T3—tumor invading adjacent structures
N Regional lymph nodes
N0—no regional lymph-node involvement
N1—regional lymph-node involvement
N2—distant lymph-node involvement
M Distant metastasis
M0—no evidence of distant metastasis
I—T0 or T1, NO, MO
II—T1 or T2, N0 or N1, M0
III—T2 or T3, N1 or N2, M1
September/October 1998, Vol. 34 Splenic Hemangiosarcoma 419
presence of anisocytosis, schistocytosis, or polychro-
masia on survival.
The effects of age, duration of
clinical signs, hematocrit, platelet count, presence
and number of nucleated red blood cells (nRBCs),
and white blood cell (WBC) count on survival also
were analyzed using the Cox proportional hazards
Nineteen female (18 spayed, one intact) and 13 male
(five neutered, eight intact) cases were diagnosed with
Stage I (n=13) or Stage II (n=19) splenic heman-
giosarcomas. The median age was 10 years (range,
six to 14 years). Seven different breeds were repre-
sented, with the most common being mixed-breed
dog (n=10) followed by the golden retriever (n=7)
and German shepherd dog (n=7). Of the 19 cases
diagnosed with Stage II disease, all had rupture of
their primary tumors and none had regional lymph-
Clinical signs at presentation were similar among
the cases. An acute onset of lethargy or weakness
sometimes was accompanied by collapse (n=10). The
median duration of clinical signs was two days (range,
zero to 60 days). Signs in some cases waxed and
waned over several days, while other cases were pre-
sented to veterinarians at the first occurrence of clini-
cal signs. Two cases were without clinical signs, and
the splenic masses were incidental findings.
On the initial presenting physical examinations,
many cases had pale mucous membranes (n=17) or a
palpable abdominal mass (n=20). Presence of a
splenic mass was diagnosed or confirmed either by
abdominal radiography (n=24) or abdominal ultra-
sonography (n=21). Presence of abdominal fluid pre-
cluded the use of abdominal radiographs or made the
results of abdominal radiographs nondiagnostic in
some cases. One case had neither abdominal radiog-
raphy nor abdominal ultrasonography performed;
however, emergency surgery confirmed rupture of a
The overall median hematocrit was 24% (range,
16% to 50%). Twenty-three of the cases were anemic
at presentation. Eleven cases received blood transfu-
sions, and one case received four additional
transfusions in the weeks following splenectomy. Ab-
normalities in red blood cell (RBC) morphology were
common. Anisocytosis (n=18) and nRBCs (n=15)
were observed most frequently; however, schisocytes
(n=6), acanthocytes (n=6), burr cells (n=6), poikilo-
cytes (n=5), spherocytes (n=3), Howell-Jolly bodies
(n=3), and target cells (n=1) also were present on
peripheral blood smears. The median platelet count
was 60,500/μl (range, 28,000 to 612,000/μl) at ad-
mission. Thirteen of the cases were thrombocytopenic
(platelet counts, less than 200,000/μl).
Coagulation profiles were performed in 19 cases.
Five cases were diagnosed with DIC based on meet-
ing three of the following five criteria: presence of
thrombocytopenia; fibrinogen degradation products
(FDPs) greater than 10 μ g/ml; prolongation of one or
more of the coagulation times by greater than 25% of
the control; presence of fragmented RBCs; and fi-
brinogen less than 80 mg/dl.
Cases diagnosed with
DIC were treated variably with heparin, whole blood,
fresh-frozen plasma, or a combination of these treat-
ments, and they were monitored continually by re-
peated coagulation profiles.
Preoperative electrocardiograms were performed
on 18 cases because of suspected cardiac arrhythmias.
Of the 18 cases, 12 had ventricular arrhythmias at-
tributable to their splenic disease and were treated
with intravenous lidocaine intraoperatively and post-
Two cases required additional oral
antiarrhythmic medication following discharge from
the hospital. Cardiac ultrasonographic examinations
were performed in six cases either because of
arrhythmias or suggestion of a cardiac abnormality
on thoracic radiographs; however, no ultrasono-
graphic evidence of a mass was noted in any case. In
one case, cardiac ultrasonographic examination con-
firmed a previously diagnosed dilatative cardiomy-
opathy that was responsive to treatment with digitalis
glycosides, and two other cases were diagnosed with
compensated mitral valvular regurgitation secondary
to valvular endocardiosis.
Nineteen cases had hemoperitoneum at surgery.
Liver biopsy was performed in 20 cases due to a
grossly abnormal appearance of the livers at surgery.
No evidence of metastatic disease was found in any
of the liver biopsies, and the most common histo-
pathological diagnosis was nodular hyperplasia. One
case had a mesenteric lymph-node biopsy with no
evidence of metastatic disease.
Follow-up information was obtained in 31 cases,
and all were dead at the time of reporting. One case
moved out of state 241 days following the splenec-
tomy and was lost to follow-up. Of the 31 cases with
follow-up, one case died 162 days postsplenectomy
from complications associated with surgical correc-
tion of gastric dilatation-volvulus. No gross meta-
static disease was observed at the time of surgery. Of
the remaining 30 cases, four cases died acutely and
the others were euthanized due to a return of clinical
signs (e.g., weakness, lethargy, collapse, abdominal
distention) presumed to be caused by tumor me-
tastases. No cases were necropsied, but one case had
evidence of hepatic metastases on abdominal ultra-
sonography at the time of euthanasia. The median
survival time was 86 days (mean, 116 days; range, 14
to 470 days) [Figure 1]. The one-year survival rate
was 6.25%. No variables were found to be significant
420 JOURNAL of the American Animal Hospital Association September/October 1998, Vol. 34
prognostic indicators of survival. Stage of disease did
not correlate with survival time in this group of cases.
Population characteristics and clinical histories were
very similar to those previously reported. The patient
group was composed primarily of older, large-breed
dogs with golden retrievers and German shepherd
dogs being the predominant purebreds represented.
The median survival time of 86 days (mean, 116
days) is longer than previously reported survival
In the absence of necropsies for these cases, the
authors have presumed that the cause of the deaths
was related to hemangiosarcoma in all but the case
which died during an unrelated surgery. The signs of
abdominal swelling and collapse were compatible
with hemangiosarcoma metastases and resultant
Hemangiosarcoma is an aggressive neoplasm
which appears to have microscopic metastases present
at the time of surgery and which causes death rapidly
in most affected dogs. Stage of disease did not seem
to be a predictor of survival. Cases in this study with
Stage II hemangiosarcoma did not have macroscopic
evidence of lymph-node metastases. It has been
thought previously that rupture of the tumor may re-
sult in “seeding” of the peritoneal cavity with meta-
static tumors. It appears that early in the course of the
disease, Stage I versus Stage II status does not affect
survival. However, under conditions of appropriate
treatment, long-term survival of splenic hemangiosar-
coma does appear to be affected by stage. In a recent
report comparing the use of liposome-encapsulated
muramyl tripeptide phosphatidylethanolamine (L-
MTP-PE) as an adjuvant therapy to splenectomy to
the use of combination chemotherapy (i.e., doxorubi-
cin and cyclophosphamide) as adjuvant therapy to
splenectomy, a significant difference in survival times
was found between cases with Stage I disease and
cases with Stage II disease. Cases with clinical Stage
I disease had significantly prolonged disease-free in-
tervals, less metastases, and longer overall survival
times as compared to cases with Stage II disease.
Lesions in the liver that were thought to be com-
patible with metastatic disease were seen at surgery
in 20 cases. Similar lesions were identified on ultra-
sonography. Histologically, most of these lesions were
hyperplastic nodules and none were metastatic le-
sions. Therefore, care should be taken in presumptive
diagnosis and staging without confirmatory histopa-
Despite the poor survival for most of these cases,
owners were pleased with their decisions to treat their
pets surgically. Quality of life was deemed good for
all cases following discharge from the hospital, and
all cases returned to normal health until the develop-
ment of clinical metastatic disease.
The biological behavior of hemangiosarcoma is a
pattern of micrometastases present at the time of sur-
gery. Since micrometastatic tumors generally are
thought to have a higher growth fraction resulting in
an increased sensitivity to most adjuvant therapies,
cases with splenic hemangiosarcoma theoretically
should benefit from postoperative chemotherapy. Che-
motherapy most likely is to be successful when the
total body burden of tumor is low or microscopic.
Adjuvant chemotherapy after surgical excision of vis-
ible tumor would appear to give the best prognosis
since it generally is accepted that grossly visible pri-
mary tumors are not curable with drug therapy alone.
Cases with Stage I or II splenic hemangiosarcoma
would appear to be suited ideally to receive adjuvant
chemotherapy, yet the role of commercially available
chemotherapy protocols have not been validated fully
in the treatment of this neoplasm. Reports of adjuvant
chemotherapy have included tumors in different loca-
tions and with different clinical stages and therefore
may represent very different manifestations of the
same histological disease.
Two similar chemotherapy protocols based on
doxorubicin and cyclophosphamide, with and without
the inclusion of vincristine, have been reported for
the treatment of hemangiosarcoma in dogs. In a group
treated with vincristine, doxorubicin, and cyclophos-
phamide (VAC protocol), the authors reported a me-
dian survival time of 124 days for three cases with
Stage II splenic disease. No cases with Stage I splenic
disease were treated.
Similar results were achieved when vincristine was
excluded from the treatment regime (i.e., doxorubicin
and cyclophosphamide, AC protocol). Four cases with
Figure 1—Kaplan-Meier survival curve duration estimates for
dogs with Stage I and Stage II splenic hemangiosarcoma treated
with surgery alone.
September/October 1998, Vol. 34 Splenic Hemangiosarcoma 421
Stage II hemangiosarcomas (both splenic and cutane-
ous) had a median survival time of 172 days. No
cases with Stage I splenic disease were treated.
taneous and subcutaneous hemangiosarcomas have
been shown to behave differently than visceral hem-
angiosarcomas, particularly when gross metastatic
disease is absent. Some cases have been reported to
have cutaneous hemangiosarcomas present 1.2 to 2.7
years before surgical removal. It also has been sug-
gested that hemangiosarcoma in the skin may repre-
sent a tumor associated with solar-induced changes,
accounting for its prolonged course.
other less-aggressive forms of the disease within a
treatment group may imply that adjuvant therapies
are more efficacious than may be warranted.
Interestingly, when doxorubicin was used as a
single agent, 27 cases that were rendered free of vis-
ible hemangiosarcoma had a median survival time of
172 days. Twelve of these cases had primary splenic
disease, but their survival times were not reported
independent of other locations. Cases with cutaneous
tumors were not included in this treatment group.
Inclusion of locations other than the spleen in evalu-
ating survival times makes inferences regarding the
role of doxorubicin difficult, because this may repre-
sent different manifestations of the same disease.
The largest reported group of dogs with primary
splenic disease treated with commercially available
chemotherapeutics consisted of seven cases with
Stage I disease and nine cases with Stage II disease.
These cases were treated with the AC protocol as a
control to evaluate the role of L-MTP-PE in the treat-
ment of canine hemangiosarcoma. Median survival
time reported in the seven Stage I cases was 166 days,
and median survival time reported in the nine cases
with Stage II disease was 96 days.
Splenic hemangiosarcoma is an aggressive disease
that warrants a grave prognosis. No prognostic
factors for survival following splenectomy were iden-
tified. Complications (i.e., DIC or ventricular ar-
rhythmias) do not appear to affect the course of the
disease after the immediate postoperative period. Che-
motherapeutic drugs and immune-response modifiers
may improve survival times significantly in dogs with
hemangiosarcoma macroscopically confined to the
spleen; however, clinical trials with appropriate se-
lection of patients are needed to define the role of
such modalities clearly.
Number Cruncher statistical system, Version 5.5 - Survival Analysis; JL
Heintze, Kaysville, UT
1. Brown NO, Patnaik AK, MacEwen EG. Canine hemangiosarcoma:
retrospective analysis of 104 cases. J Am Vet Med Assoc 1985;86:56–8.
2. Okansen A. Haemangiosarcoma in dogs. J Comp Path 1978;88:922–6.
3. Pulley LT, Stannard AA. Tumors of the skin and soft tissue. In: Moulton
JE, ed. Tumors in domestic animals. 3rd ed. Berkeley, CA: Univ Calif
4. Johnson KA, Powers BE, Withrow SJ, et al. Splenomegaly in dogs.
Predictors of neoplasia and survival after splenectomy. J Vet Int Med
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
6. Hirsch VM, Jacobsen J, Mills JH. A retrospective study of canine
hemangiosarcoma and its association with acanthocytosis. Can Vet J
7. Spangler WL, Culbertson MR. Prevalence, type, and importance of
splenic disease in dogs: 1,480 cases (1985–1989). J Am Vet Med Assoc
8. Hosgood G. Splenectomy in the dog: a retrospective study of 31 cases.
J Am Anim Hosp Assoc 1987;23:275–83.
9. Hargis AM, Feldman BF. Evaluation of hemostatic defects secondary to
vascular tumors in dogs: 11 cases (1983–1988). J Am Vet Med Assoc
10. Hammer AS, Couto CG, Fillpi J, et al. Efficacy and toxicity of VAC
chemotherapy (vincristine, doxorubicin, and cyclophosphamide) in dogs
with hemangiosarcoma. J Vet Int Med 1991;5:160–6.
11. Sorenmo KU, Jeglum A, Helfand SC. Chemotherapy of canine heman-
giosarcoma with doxorubicin and cyclophosphamide. J Vet Int Med
12. Vail DM, MacEwen EG, Kurzman ID, et al. Liposome-encapsulated
muramyl tripeptide phosphatidylethanolamine adjuvant immunotherapy
for splenic hemangiosarcoma in the dog: a randomized multi-institu-
tional clinical trial. Clin Cancer Res 1995;1:1165–70.
13. Owen LN. TNM classification of tumors in domestic animals. Geneva:
World Health Organization, 1980.
14. DeVita VT, Jr. Principles of chemotherapy. In: DeVita VT, Hellman S,
Rosenberg SA, eds. Cancer: principles and practice of oncology. 4th ed.
Philadelphia: JB Lippincott, 1993:276–92.
15. Lee ET. Statistical methods for survival data analysis. Belmont, CA:
Lifetime Learning Publications, 1980:21–74, 131–5, 306–18.
16. Hammer AS, Couto CG, Swardson C, Getzy D. Hemostatic abnormali-
ties in dogs with hemangiosarcoma. J Vet Int Med 1991;5:11–4.
17. Keyes ML, Rush JE, Couto CG, et al. Ventricular arrhythmias in dogs
with splenic masses. Vet Emerg Crit Care 1994;1:33–8.
18. Knapp DW, Aronsohn MG, Harpster NK. Cardiac arrhythmias associ-
ated with mass lesions of the spleen. J Am Anim Hosp Assoc 1993;29:
19. Hargis AM, Ihrke PJ, Spangler WL, et al. A retrospective clinicopatho-
logic study of 212 dogs with cutaneous hemangiomas and hemangio-
sarcomas. Vet Path 1992;29:316–28.
20. Ogilvie GK, Powers BE, Mallinckrodt CH, et al. Surgery and doxorubi-
cin in dog with hemangiosarcoma. J Vet Int Med 1996;6:379–84.