Histologic and Immunohistochemical Review of Splenic Fibrohistiocytic Nodules in Dogs
This article is corrected by:
- Errata: Erratum Volume 27, Issue 1, 215, Article first published online: 11 January 2013
Abstracts of portions of this study were presented in part at the 28th Veterinary Cancer Society Annual Conference in 2008, and at the 2009 ACVIM Forum and Canadian Veterinary Medical Association Convention, Montreal, Quebec, Canada
Corresponding author: Antony S Moore, Veterinary Oncology Consultants, 379 Lake Innes Drive, Wauchope, NSW Australia; e-mail: firstname.lastname@example.org.
Splenic fibrohistiocytic nodules (SFHN) are commonly diagnosed. It is suspected that these represent a heterogeneous group of malignant and nonmalignant diseases, separation of which could improve the ability of clinicians to prognosticate for dogs.
Immunohistochemistry will differentiate histologic diagnoses within the group of SFHN; survival after splenectomy is associated with those histologic types.
Thirty-two dogs with SFHN treated by or under direction from veterinary oncologists.
Retrospective case record analysis from dogs followed from splenectomy until death. Clinical, histopathologic, and immunohistochemistry data analyzed for an association with survival time.
Thirty-two dogs had SFHN; grade 1 (2 dogs), grade 2 (9 dogs), and grade 3 (lymphoid percentage <40%; 21 dogs). Twenty-two dogs died, 10 were censored (9 alive median of 883 days after splenectomy). Median overall survival was 387 days, and grade 3 SFHN was positively associated with survival time as previously reported (P < .001). Of 31 available samples, dogs had diseases reclassified as nodular hyperplasia (13; 8 complex, 5 lymphoid including 2 marginal zone), lymphoma (4; 2 marginal zone lymphoma, 1 high grade B-cell lymphoma, and 1 marginal zone transitional to high grade B-cell lymphoma), 8 stromal sarcomas, and 6 histiocytic sarcomas. Dogs with histiocytic sarcoma had worse survival (median 74 days) than dogs with other diseases.
Conclusions and Clinical Importance
Splenic histiocytic sarcoma is an aggressive disease; however, some dogs with stromal sarcomas had long survival times. The term SFHN is no longer warranted for this group of disorders.
splenic fibrohistiocytic nodule
diffuse large B-cell lymphoma
marginal zone lymphoma
hemophagocytic histiocytic sarcoma
marginal zone hyperplasia
complex nodular hyperplasia
lymphoid nodular hyperplasia
The histiocytic disorders remain poorly defined in dogs. They represent a group of diseases that appear to fall into several categories, including reactive (cutaneous and systemic) histiocytosis, localized histiocytic sarcoma, and malignant histiocytosis (also termed systemic histiocytic sarcoma). Splenic fibrohistiocytic nodules (SFHN) have been characterized by a mixed population of histiocytoid, spindle cells, or both in varying proportions intermixed with hematopoietic elements, plasma cells, lymphocytes, or combinations; and appear to form a continuum between splenic lymphoid nodular hyperplasia, defined as lesions of >70% lymphocytes with interspersed fibrohistiocytic cells (grade 1 SFHN); to malignant splenic stromal neoplasms, which are primarily fibrohistiocytic cells with <40% lymphocytes (grade 3 SFHN).
Hematologic and biochemical abnormalities in dogs with SFHN were not consistent in 1 study, but included evidence of anemia, neutrophilia with left shift, and elevated serum alkaline phosphatase activity. That study found the lymphoid : fibrohistiocytic cell ratio in the lesion to be the most important prognostic indicator for dogs with this syndrome. In the original study, dogs with grade 3 SFHN had a 32% 1-year survival rate compared with an 87% 1-year survival rate for dogs with grade 1 or 2 SFHN.
The recent descriptions of histiocytic diseases and histologic subtypes, and reclassification of splenic lymphoma in dogs suggests that it is incorrect to consider SFHN as a single entity, and that it reflects a more complex group of diseases that can be better distinguished with immunohistochemistry.[2-4] Our study retrospectively evaluated the clinical course in dogs with splenic fibrohistiocytic nodules following splenectomy, and retrospectively subjected available samples to immunohistochemistry and reclassification in an attempt to identify prognostic factors for survival.
Materials and Methods
Case records from adult dogs previously diagnosed on histopathology with SFHN were retrieved from a specialty veterinary practice (Animal Referral Hospital, Sydney) and from case consultations (Veterinary Oncology Consultants). Criteria for entry included histopathology sample/slide available for review, and survival beyond splenectomy. Before splenectomy, dogs were staged using a complete hematology, serum biochemical profile, urinalysis, and thoracic (3 view) radiographs as well as complete history and physical examination. The outcome evaluated was survival, with the endpoint being death due to any cause; the day of splenectomy was considered day 1 for statistical analysis. Dogs were censored from survival analysis at the last day of contact if still alive or if they were lost to follow-up.
All histology slides were reviewed by 1 pathologist (NS) and SFHN grading was performed according to published criteria. In brief, those lesions having 70–99% lymphocytes had grade 1 SFHN, those with 40–69% lymphoid cells had a grade 2 SFHN, and those with <40% lymphocytes had grade 3 SFHN. In addition to grading, mitotic index and percent necrosis were assessed. Other histopathologic criteria evaluated were presence of giant cells, erythrophagocytosis, hemosiderosis, and hematopoiesis; these criteria were evaluated on a subjective scale of 0–3. Further, immunohistochemistry was performed by another pathologist experienced in histiocytic and lymphoid lesions (PFM); for these analyses, CD3, CD20, CD79a, CD18, CD11d, and Ki-67, were used depending on the lesion and as determined by that pathologist; the same pathologist (PFM) reclassified the samples according to the results and previously published criteria.[2-4]
To establish that we were evaluating the same disease as originally published by Spangler and Kass, the Kaplan–Meier product limit method and Cox regression analysis were used to evaluate the effect on survival time for proportion of lymphoid:fibrohistiocytic cells (<40%, 40–69%, >69%); and for stromal sarcomas, the effects on survival time of histopathologic criteria listed above (mitotic index, percent necrosis, presence of giant cells, erythrophagocytosis, hemosiderosis, and hematopoiesis evaluated on a subjective scale of 0–3). Statistical analyses were performed using a software package.1 For the final analysis, values of P < .05 were considered significant.
Thirty-two dogs with SFHN diagnosed between 2003 and 2010 were entered into the study; 22 dogs reached an endpoint (died); 1 dog was lost to follow-up 38 days after splenectomy (grade 3 SFHN), and 9 dogs were still alive between 318 and 1767 days (median 883 days) after splenectomy.
There were 11 males (9 castrated) and 21 spayed females. The median age was 9 years (range 2–15 years) and median weight was 23 kg (range 3.7 kg–51.9 kg); 6 dogs weighed 10 kg or less. There were 17 pure-breeds represented in the population; the most common were Rottweiler (5 dogs) and Staffordshire Bull Terrier (3 dogs).
The most common clinical signs were lethargy (16 dogs), inappetance (15), vomiting (9), weight loss (7), and collapse (6). The diagnosis of an abdominal mass was made at ultrasonography in all dogs before surgery. Thoracic radiographs obtained at the time of surgery showed no evidence of metastatic disease in any dog.
Spangler & Kass Classification for SFHN
Two dogs had grade 1 SFHN, 9 dogs had grade 2 SFHN, and 21 dogs had grade 3 SFHN. Median overall survival was 387 days, and grade 3 SFHN (P < .001) was negatively associated with survival. For 21 dogs with grade 3 SFHN, the median survival was 113 days (95% CI 0-253 days); 1 year after splenectomy, 30.6% were alive, and 10.2% were alive 2 years after surgery. In comparison, there were 11 dogs with grade 1 or 2 SFHN of which 7 (64%) were alive; the median survival time could not be calculated, but the mean survival time was 1,118 days (95% CI 651–1,585 days). One year after splenectomy 90.1% of dogs were alive, and 52% were alive 2 years after surgery.
Routine hematoxylin and eosin stained slides were available for review for all 32 dogs, but only 31 had blocks available for immunohistochemical staining.
The most common diagnoses were stromal sarcoma (8 dogs), complex nodular hyperplasia (CNH; 8 dogs); histiocytic sarcoma (6 dogs; 3 of which were hemophagocytic); lymphoid nodular hyperplasia (LNH; 3 dogs); diffuse marginal zone hyperplasia (2 dogs); and 4 dogs had lymphoma; (1 diffuse large B-cell lymphoma [DLBCL]; mitotic rate 50 per 10 400x fields; and 3 marginal zone lymphoma [MZL]). Transition of MZL to DLBCL was evident in 1 dog; the proliferative fraction (Ki-67 + ) varied regionally between 5 and 40%.
The two grade 1 SFHN were reclassified as MZL and LNH. The 9 grade 2 SFHN were reclassified as marginal zone hyperplasia (1), marginal zone lymphoma (1), CNH (2), LNH (1), and stromal sarcoma (3) and 1 sample was unavailable for immunostaining. The 21 grade 3 SFHN were reclassified as marginal zone hyperplasia (diffuse histiocytic and stromal hyperplasia of red pulp cords; sample also negative for CD45 and CD31; 1), DLBCL (1) and MZL transitional to DLBCL (1), histiocytic sarcoma (6), CNH (6), LNH (1), and stromal sarcoma (5).
Histiocytic sarcoma: The cellular infiltrate obliterated splenic architecture and formed masses. Infiltrating cells were discrete, round, or occasionally spindle shaped. They were large and contained single or multiple nuclei with abundant amphophilic or foamy cytoplasm. These large infiltrating cells were often accompanied by a sparse to abundant lymphocytic infiltrate. Large cytologically atypical cells expressed abundant CD18, but lacked expression of CD11d.
Hemophagocytic histiocytic sarcoma: The cellular infiltrate diffusely expanded the splenic red pulp and encroached on the white pulp; mass formation was not observed. Infiltrating cells were discrete and round, and infiltrated the red pulp cords and sinuses. They were large and possessed mostly single nuclei; anisokaryosis was marked. These cytologically atypical cells manifested prominent erythrophagocytosis and were accompanied by discrete foci of extramedullary hematopoiesis. The large, cytologically atypical cells expressed abundant CD11d and CD18.
Splenic stromal sarcoma: A focal splenic mass composed of sheets of cohesive polygonal to spindle shaped cells. These cells exhibited anisokaryosis and often had large vesicular oval-to-round nuclei; occasional multinucleated cells were observed. Large areas of necrosis were frequently encountered. Diffuse or nodular lymphoid aggregates and foci of extramedullary hematopoiesis were often observed within these tumors. Stromal sarcomas did not express CD18. One stromal sarcoma was negative for CD31, but further classification of these stromal sarcomas to subtype was not attempted.
Diffuse marginal zone lymphoid hyperplasia: The splenic marginal zone was expanded by intermediate sized lymphocytes, which possessed single, large nucleoli. These cells did not obliterate the periarteriolar lymphoid sheath (PALS – T cell domain) or the follicle (B-cell domain). Also, merging of white pulp domains was not observed. Marginal zone lymphocytes expressed CD20 and CD79a. The domains were clearly evident by HE stains in all but 1 sample in which preservation of the PALS was confirmed by the observation of abundant CD3 expressing T cells.
Marginal zone lymphoma (MZL): The splenic marginal zone was expanded by intermediate-sized lymphocytes, which possessed single, large nucleoli. These cells infiltrated the PALS and the germinal center of the follicle, effectively obliterating the white pulp. Invasion of the red pulp enabled coalescence of adjacent white pulp domains. The marginal zone expansion was diffuse, mass forming, or both. Marginal zone lymphocytes expressed CD20 and CD79a. Mitotic figures were rare, and the proliferative fraction (Ki-67 labeling index) was low (<5%).
Diffuse large B-cell lymphoma (DLBCL): The splenic architecture was obliterated by sheets of large lymphoid cells, which had centroblastic and immunoblastic morphology. These lymphocytes had large nuclei (3–4 red cell diameters) with single (immunoblastic) or multiple nucleoli (centroblastic). The large lymphocytes expressed CD79a; small lymphocytes scattered within the mass expressed CD3.
Nodular hyperplasia – lymphoid: A focal splenic mass was observed. The mass was composed of discrete to merging nodular aggregates of lymphocytes. The lymphocytes had mixed cytological features; foci of mantle cell and marginal zone lymphoid morphology were observed. Splenic stromal proliferation was not observed.
Nodular hyperplasia – complex: A focal splenic mass with a pleo-cellular basis was observed. The lymphoid component resembled simple lymphoid nodular hyperplasia described above. In addition, proliferation of splenic stromal elements (fibroplasia and/or smooth muscle hyperplasia) and histiocytic hyperplasia were observed. Foci of extramedullary hematopoiesis and expanded populations of plasma cells were optional elements.
Outcome for Dogs Based on Revised Classification
Of 8 dogs with stromal sarcoma, 5 had died and median survival was 488 days (95% CI 0-1104; range 16–650 days) with 56.3% alive 1 year after splenectomy and 18.8% alive 2 years after splenectomy. Only 2 dogs received chemotherapy (carboplatin, actinomycin-D, vincristine, procarbazine, and prednisolone), with survival times of 488 and 197 days. Mitotic rate, percent necrosis, presence or absence of giant cells or erythrophagocytosis, hemosiderosis, and hematopoiesis were not statistically associated with survival.
Of 6 dogs with histiocytic sarcoma, all had died and median survival was 74 days (95% CI 22–109; range 2–146 days). Four dogs received chemotherapy (the same multiagent protocol was started in all dogs, but short survival time meant that 1 dog received only a single dose of CCNU and prednisolone, whereas the other dogs received CCNU, prednisolone, procarbazine, actinomycin-D, vincristine, cyclophosphamide, and doxorubicin). One dog had lymph node and liver metastases at diagnosis.
Of 8 dogs with complex nodular hyperplasia, 4 had died and median survival was 387 days (95% CI 330–444; range 3–1057 days) with 70% alive 1 year after splenectomy and 23.3% alive 2 years after splenectomy. One dog received chemotherapy and lived 397 days. The shortest survivor had brain signs and was euthanized without the cause being diagnosed.
Of 5 dogs with lymphoid nodular hyperplasia, only 2 (40%) had died, so median survival was not available, but mean survival was 570 days (95% CI 230–910; range 2–883 days) with 60% alive 1 year and 2 years after splenectomy. Two of the dogs had marginal zone hyperplasia and lived 200 days and 586 days; the latter dog received prednisolone. One dog was presented with severe peritonitis and acute respiratory distress syndrome and died 2 days after surgery.
Of 4 dogs with lymphoma, two had MZL and lived 99 and 586 days. One dog that had MZL that was transitional to DLBCL was euthanized after 5 days because of poor prognosis offered by grade 3 SFHN, and 1 dog with DLBCL received combination chemotherapy and was alive 466 days after splenectomy.
We confirmed in this group of dogs with SFHN that prognosis was determined by the proportion of lymphocytes in the tumor; dogs with grade 3 SFHN have a poor prognosis. This confirms that we were looking at a similar group of dogs to that originally described. However, when tissues from this group of dogs were subjected to immunohistochemistry, SFHN proved to include a heterogeneous group of diseases that did not appear to correspond to the percentage of lymphocytes or SFHN grade, and had quite different probabilities for survival. As there were so many different final diagnoses, it is difficult to make meaningful statements about prognosis; however, dogs with histiocytic sarcoma had a poor median survival time of 74 days.
Histiocytic sarcoma is an uncommon disease. One report characterized tumors as localized histiocytic sarcoma if there was a history of initially single site involvement and metastasis to the regional lymph node. In that study, 19 of 39 dogs met the criteria, and their primary tumors arose mostly in the spleen, skin, and subcutis. Although malignant fibrous histiocytoma has been used to describe many types of splenic sarcoma, immunohistochemistry can clearly differentiate histiocytic sarcoma.[2, 5, 6] Our results herein suggest that at least some dogs diagnosed with malignant fibrous histiocytoma may have stromal sarcoma and some have histiocytic sarcoma.
One study reporting lomustine treatment of histiocytic sarcoma (of which 39% had any splenic involvement either alone or in addition to other sites) found dogs with splenic involvement were less likely to achieve a remission, and had a median survival time of 58 days. Our study supports the poor prognosis for dogs with splenic histiocytic sarcoma. Other studies have not separately reported dogs with splenic histiocytic sarcoma, so median survivals are not available.[8, 9] Recent literature suggests that there may be a role for chemotherapy in treating histiocytic sarcoma,[8, 9] but those studies did not specifically separate dogs with splenic histiocytic sarcoma from dogs that had histiocytic sarcoma of other visceral and soft-tissue sites. In this small group of dogs with confirmed histiocytic sarcoma, 3 dogs received multiagent chemotherapy including CCNU), one received CCNU alone, and the longest survivor did not receive chemotherapy. In the authors' opinions, additional studies of active agents in larger numbers of dogs with splenic histiocytic sarcoma as a separate entity are warranted.
There were 8 dogs with stromal sarcomas diagnosed in this study, and they had a long median survival time of 488 days; only 2 dogs received chemotherapy. This is in contrast to the reported survival times for larger studies that included dogs with nonvascular, nonlymphomatous splenic sarcomas. In 3 series totaling 137 dogs with sarcomas of the spleen, the most common diagnosis was undifferentiated sarcoma (28%) followed by leiomyosarcoma (19%).[10-12] Other tumors, in order of frequency were fibrosarcoma, osteosarcoma, mixed mesenchymal sarcoma, myxosarcoma, and liposarcoma. One of these studies described 27 dogs with splenic sarcomas that underwent splenectomy; 11 dogs had obvious metastatic disease at the time of surgery. Overall median survival was 2.5 months. The median survival improved to 9 months for dogs without evidence of metastasis at the time of surgery. Another found that dogs with tumors that had a mitotic index (number of mitoses per 10 high power fields) of <9 had a median survival of 7 months, compared to all other groups (1–2 months). Our data did not show an effect of mitotic index on survival, but the numbers of dogs are very small.
One dog that had MZL transitional to DLBCL was euthanized 5 days after diagnosis because of the poor prognosis offered by having a grade 3 SFHN. The other dog with DLBCL responded to combination chemotherapy and survived more than 15 months; a revised diagnosis may have altered the outcome for the former dog, emphasizing the importance of correct classification.
As an entity, splenic fibrohistiocytic nodules were described in 1998, before the finer distinction of histiocytic sarcoma types by immunohistochemistry, and before the description of MZL in dogs. The authors feel that the definition of disease has evolved beyond the usefulness of the original description, and therefore propose that the term, splenic fibrohistiocytic nodules, to describe a specific disease entity is not warranted. Careful morphological assessment coupled using immunohistochemistry reveals a plethora of other diseases with widely different outcomes and requiring different treatment strategies. Additional studies should be performed to examine outcomes in each of these specific diseases.
The contribution of referring veterinarians to animal care and provision of follow-up data is greatly appreciated.
Disclosure of Interest Declaration: This study was completed by Veterinary Oncology Consultants in Australia; histopathology slides were reviewed at IDEXX Laboratories, Australia, immunohistochemistry was performed at University of California, Davis.
SPSS 10, Statistical Analytical Software, Chicago, IL