MALIGNANT FIBROUS HISTIOCYTOMA OF B O N E : A S T U D Y OF 35 CASES Edward F. McCarthy, M.D.,* Takeo Matsuno, M.D.,t and Howard D. Dorfman, M.D.~
Abstract Thirty-five cases of primary malignant fibrous histiocytoma of bone are reported. Twenty of these cases were collected from a retrospective analysis of other malignant bone tumors. The age range was fi'om 11 to 69 )'ears; the average age was 34 years. The tumor occurred most commonly in the distal femur and proximal tibia. The distinguishing histologic feature was a storiform arrangement of spindle cells. The differential diagnosis included fibrosarcoma, osteogenic sarcoma, malignant giant cell tumor, malignant l)'mphoma, and metastatic carcinoma. Follow-up of at least three ).ears was available in 21 cases. Of these,' nine patients were alive and free of metastases three and one-half to 12 years after treatment. Two were alive with solitary metastases at three years;, and 10 patients died between three months and three )'ears after treatment. :)'~ In four cases the lesions were multicentric at the time of'diagnosis mad in four cases were associated with bone infarction. This tumor must be recognized as an important complication of bone infarction and should b e suspected when a patient with a known history of bone infarction develops a change in symptoms. Because the prognosis of this tumor is significantly better than that in those tumors with which it had been previously grouped, and in view of its association with bone infarction, it deserves to be maintained as a distinct clinicopathologic entity. Amputation is the treatment of choice.
It Ires only been since 1972 that primary malignant fibrous histiocytoma of bone has been recognized as a distinct bone tumor. Since then there have been a total of 105 cases reported, v~5 Most of these were identified on the basis of their microscopic features from retrospective
studies of tumors previously classified as anaplastic fibrosarcomas or osteosarcomas. Although the association of tiffs neoplasm with bone infarction has been well documented,4, z-,a,H no clear pattern of its clinical behavior has emerged from the four large series and several case reports.
*Instructor in Pathology, The Johns Hopkius Medical lnstitt,tions. Assistant Pathologist, Sinai ltospital of Baltimore, Baltimore, Maryland. ?Staff, Department of Pathology, I lokkaido University School of Medicine, Sapporo, Japan. +Associate Professor of Pathology and Orthopedic Surgery, The Johns t lopkins Medical Institutions. Pathologist-in-Chief, Sinai Hospital of Bahimore, Baltimore, Maryland.
-57
HUMAN PATHOLOGY--VOLUME I0, NUMBER 1 Jamlmy 1979 O u r purpose in tiffs article is to present our experience with 35 cases of malignant fibrous histiocytoma of bone, four of which were associated with bone infarcts. O u r data suggest that the t u m o r is less aggressive than much of the previous literature suggests and the association with bone infarction is confirmed.
MATERIALS
AND
METHODS
fibrous histiocytoma complicating a bone infarct. 7 Fifteen cases, in patients seen in the last three )'ears, were diagnosed originally as primary malignant fibrous histiocytoma of bone. We have noted focal patterns of malignant fibrous histiocytoma in osteogenic sarcomas, sarcomas complicating Paget's disease, and in some dedifferentiated chondrosarcomas, but these hax'e been excluded f r o m the present study.
These 35 cases were collected f r o m the fles of the Hospital for Joint Diseases and Medical Center, Sinai Hospital of Baltimore, Johns Hopkins Hospital, and the personal consultation files of one of us (HDD). T w e n t y cases were collected in a retrospective analysis o f primary bone neoplasms, principally fibrosarcomas. T h i r t e e n originally had been diagnosed as anaplastic fibrosarcoma, one of which was associated with muhiple bone infarcts. Other diagnoses had been osteosarcoma, giant cell tumor, and malignant lymphoma. Reports of three cases had been previously published by one of us (HDD), two as anaplastic fibrosarcomas complicating bone infarcts ~G'tr and one as a malignant
109810
CLINICAL PRESENTATION In this series there were 23 males and 12 females. T h e age range was 11 to 69 )'ears and the average age was 34 years. Most patients were either in their teens or in their 40s (Fig. i). T h e location varied. However, 21 tumors occurred a r o u n d the knee with 12 in the proximal tibia, eight in the distal f e m u r , and one in the proximal fibula (Fig. 1). In two cases there was contiguous im'olvement of two bones at the time of initial diagnosis. Four cases were multicentric from the outset (Fig. 2). O f these, one case involved the f e m u r and spine; a n o t h e r invol~'ed the femur, spine,
"::::::::::
]!]i]i]i]i] MALES 23 "EMALES 12 i:!:i:!:!:! FEMALE~ -g iii!i!ilill TOTAL
......
5-
iF!i!i!i!i!
iiiiiiiiii iiiiiiiiiii -8
I "12" --1
":':':':':'tl'"f'"?'l
10
58
20
i
30 40 AGE
i
50
60
70
Figure 1. Age and sex distribution ill 35 cases, skeletal distribution of 31 solitary lesions. Asterisks denote sites associated with bone infarction.
,MALIGNANT FIBROUS HISTIOCYTOMA OF BONE--McCARTHY" ET AL. cases arose in association with bone infarction. O f the cases associated with b o n e infarction, two were in patients with occupational e x p o s u r e to compressed air environments and two were in patients with idiopathic infarcts. I n f o r m a t i o n relating to clinical symptoms were available in 23 cases. Bone pain was present in 22 cases for one m o n t h to five )'ears b e f o r e diagnosis. One patient was asymptomatic until the t u m o r was f o u n d in r a d i o g r a p h s taken following trauma.
RADIOGRAPHIC PATTERNS
Figure 2. Skeletal distribution in fimr muhicentric cases.
and a rib; a n o t h e r involved the ilium and ribs; and one patient had two separate lesions in the same f e m u r . F o u r patients presented with pathologic fractures and f o u r
Radiographs were available for stud)' in 33 cases. A radiologic diagnosis o f malignant disease was m a d e in 23 cases. T h e tumors p r e s e n t e d as ill defined, single or confluent areas o f radiolucency with endosteal erosion and minimal cortical expansion (Fig. 3). Occasionally the cortex was p e n e t r a t e d completely (Fig. 4). T h e r e was minimal reactive b o n e formation, and evidence o f t u m o r calcification and bone formation was not presen.t. Occ,asionally a highly aggressive m a l i g n a n t ' r a d i o g r a p h i c pattern resembling Ewiri~ s sarcoma o r tel-
Figure 3. Radiographs of maligtmnt fibrous histiocytoma showing a radiolucent defect with endosteal erosion :rod minimal cortical expansion.
59
HUMAN PATHOLOGY--VOLUME 10, NUMBER 1 January 1979
Figure 4. Radiographs of malignant fibrous histiocytoma showing a poorly circumscribed, radiolucellt defect with cortical penetration.
angiectatic osteosarcoma was seen (Figs. 5, 6). In a few cases the radiographic findings suggested a benign t u m o r such as nonossifying fibroma, giant cell t u m o r o f bone, or solitary b o n e cyst (Fig. 7). In the f o u r cases associated with b o n e infarction there were one or m o r e islands o f dense intraosseous calcification resembling typical b o n e infarcts. T h e involved area o f infarction was adjacent to or surr o u n d e d by m a r k e d radiolucency with cortical erosion and often with cortical destruction.
TREATMENT
60
AND
FOLLOW-UP
In this study o f 35 patients, follow-up data for at least three years were available in 21 patients. O f the remaining 14 cases, 13 were diagnosed witlfin the p r e c e d i n g three years and are not considered in the follow-up data. Follow-up- information was not available in one case. O f the 21 patients, nine were alive and well without metastases t h r e e and one-half to 12 years after treatment. Five patients were well longer than eight ),ears. T w o o t h e r patients were alive with solitary me-
tastases after t h r e e years, and 10 patients were dead. / Death o c c u r r e & " i o the 10 patients three m o n t h s to three years after treatinent. T h e age r a n g e was 15 to 60 years with no age g r o u p prevailing. T w o o f the patients who died had multicentric tumors initially. Seven o f the 10 patients were treated by a m p u t a t i o n and one had additional r a d i o t h e r a p y . One had a resection, and the two patients with muhicentric tumors were treated by radiation, one after a simple curettage. Six had lung metastases, which a p p e a r e d one m o n t h to two )'ears after treatment. In one patient t h e r e was regional s p r e a d to o t h e r bones at six months. T w o patients had lymph n o d e metastases discovered at the time o f autopsy. Autopsies were p e r f o r m e d in three patients, and in these cases the histologic features o f the metastases did not vary f r o m those o f the p r i m a r y tumor. T h e r e were no distinctive histologic features o f the primary t u m o r in the patients who died. O f the nine patients alive and free o f disease longer than three and o n e - h a l f years, all but o n e were treated by amputation and only one o f these had p r e o p e r a tive radiotherapy. T h e patient who did not
MALIGNANT
FIBROUS
HISTIOCYTOMA
O F BONE--McCaRTW~" ET ak.
Figure 5. This malignant fibrous histioc)tonm has a permeative pattern and could be interpreted as representing Ewing's sarcotna.
Figure 6. Radiograph showing a destructive malignant fibrous histiocytoma of tile ischitml. T h e r e is no reactive bone formation. This" rar " ~," pattern resembles a tclangiectatic o s l e o s a r c o n l a . -
61
HUMAN PATHOLOGY--VOLUME 10, NUMBER 1 Janumy 1979
,'ql
t i/ i
t/ Figure 7. Occasionally a radiographically benign pattern may occur, This malignant fibr,,us histi,)cvtoma was t h o u g h t by radiographic examination to be a metaphyseal fii)rous defect. (Follow-ul): T h e l)atient; a 60 year old man, died of p u h n o n a r y metastases 1 ~/2 yCd[llSfollowing ampuhltion.)
have an amputation was alive three and one-half years after a simple curettage.
HISTOLOGIC FEATURES A N D DIFFERENTIAL DIAGNOSIS
62
The histologic features of malignant fibrous histiocytomas of soft tissue have been described in earlier reports.'8.19 The range of histologic expression of primary malignant fibrous histiocytoma of bone is as broad as that seen in soft tissues. We believe that these tumors are neoplastic proliferations of malignant histiocytes in which their capacity to act as facuitative fibroblasts is expressed. The histologic variations of the tumor reflect tile two fimctional states of tile neoplastic cell - - fibroblastic and histiocytic. Therefore, there is a morphologic spectrum ranging from sheets of histiocytes that may show phagocytosis to a storiform or matted appearance. These patterns may alternate within a given tumor in equal proportions, or one pattern may predominate. More often tile fibroblastic and histiocytic expressions of tiffs biphasic cell are randomly intermingled. Cytologic features of malignancy are
seen in both patterns, but nuclear atypicality and abnormal mitotic figures are more florid in the !)]~stiocytic areas. Also a number of special his.tologic patterns, to be discussed, may be seen in any tumor, and these may dominate the histologic picture, leading to difficulties in differential diagnosis. The storiform pattern is the most characteristic feature of malignant fibrous histiocytoma. Bundles of elongated or spindle cells are arranged in a pinwheel fashion ~Fig. 8.4). The nuclei are oval or elongated and vesicular and show mild to moderate pleomorphism. The cytoplasmic margins are indistinct. If tile spindle cell pattern predominates, the tumor may be misdiagnosed as a fibrosarcoma. However, the presence of the storiform pattern will help to identify the tumor as a fibrous histiocytoma. Tile other end of the spectrum of differentiation of the neoplastic cell may predominate and tile tumor will be composed of sheets of tmiform histiocytes. Occasionall),, when tiffs occurs, a diagnosis of histiocytic lymphoma is made. But these histiocytic cells will show a tendency to differentiate toward spindle cells, a feature that is not seen i n lymphoma (Fig. 8B).
MALIGNANT
FIBROUS
HISTIOCYTOMA
O F B O N E - - M c C A R T H Y ET AL.
9
.b.
Q
,...:- , , o9 ~o
9
9 .
.,..
I
o
D
10
9
"
I
Figure 8, :1, Storifi~rm arrangement of spindle cells, fl, Ncoplastic uniform histiocytes will show subtle differentiation to spindle cells. C, Sheets o[ bizarre plcom~rl)hic histiocytes. D. Bilobed mzclci with prominent mlcleoli may look like Reed-Sternbcrg cells. (:1, • 1S0. B, • 200. C, • I ! 0. D, •
63
HUMAN PATHOLOGY--VOLUME 10, NUMBER 1 JanuaD'1979 Sometimes, rather than being uniform, the histiocytes are quite pleomorphic and bizarre (Fig. 8C). In the past this pattern was interpreted as anaplastic fibrosarcoma. The histiocytes have large, oval or lobulated, vesicular nuclei, often with a prominent eosinophilic nucleolus. When the nuclei are bilobed, there may be a resemblance to Reed-Sternberg cells (Fig. 8/)). T h e cytoplasmic margins of these histiocytes are distinct, and the cytoplasm is sometimes so abundant as to look like the "strap" cells of rhabdomyosarcoma (Fig. 9A). Close observation reveals that the cytoplasm is foamy and cross striations are not seen. There are sometimes clusters of foam}' cells intermingled with the atypical histiocytes. This feature, along with lipid laden cytoplasm of the histiocytes, has led to the erroneous interpretation of liposarcoma (Fig. 9B). Occasionally even cholesterol clefts may be seen (Fig. 9C). Hemosiderin is also frequently seen within the cytoplasm of the neoplastic histiocytes. Abnormal mitotic figures are most prominent in fields of histiocytic differentiation (Fig. 9D). Occasionally the histiocytic and fibroblastic components are both present but distinct. This imparts a biphasic pattern to the tumor with clusters of epithelioid histiocytes separated by fascicles of elongated fibroblast-like cells (Fig. 10A). This epithelioid pattern can lead to the misdiagnosis of metastatic carcinoma. Muhinucleated giant cells are often a prominent feature of malignant fibrous histiocytoma. More often they are tumor giant cells in areas of histiocytic differentiation (Fig. 10B). However, osteoclast-like giant cells can sometimes be seen (Fig. 10C). For this reason the tumor can be mistaken for a giant cell tumor. However, if the background is fibrous and is made up of pleomorphic spindle cells that tend to form a storiform pattern, the lesion can be recognized as a malignant fibrous histiocytoma. Often this tumor has been interpreted as a malignant giant cell tumor. We prefer to reserve this 51iagnosis for malignant tumors that arise in a setting of a typical giant cell tumor of bone. One variant that is occasionally encountered is a diffnse infiltrate of inflammatory cells admixed with and sometimes
64
obscuring the neoplastic cells (Fig. 10D). This variant is analogous to that which occurs in soft tissues described by Kyriakos and Kempson. 2~ The inflammatory infiltrate consists of plasma cells, lymphocytes, and polymorphonuclear leukocytes. The polymorphous population of cells in this variant along with Reed-Sternberg-like cells can lead to the misdiagnosis of Hodgkin's disease. The spindle cells of malignant fibrous histiocytoma of bone may sometimes be separated by abundant extracellular ground substance, which imparts a focal myxomatous appearance to the tumor (Fig. 1 1A). This variant also has a soft tissue counterpart in the myxoid variant of malignant fibrous histiocytoma described by Weiss and Enzinger. 21 Another variant that occurs in bone is a pattern similar to that of hemangiopericytoma. T h e neoplastic cells are arranged in clusters or bundles in an organoid pattern around vascular spaces (Fig. 11B). Occasionally a single layer of neoplastic cells lines vascular spaces separated by a fibrous stroma. This pattern may mimic that of a hemangioendothelial sarcoma (Fig. 1 1C). Quite often dab collagen in the extracellular matrix lslvery coarse or lacelike and may be indisffnguishable from osteoid in demineralized sections (Fig. 11D). The additional presence of thin spicules of reactive bone creates the very difficult problem of distinguishing this tumor from an anaplastic osteosarcoma. Indeed, Dahlin and Unni 22 have expressed the view that this distinction is an arbitrary one. A filrther complication is that a certain small amount of osteoid can be present in malignant fibrous histiocytomas of soft tissue. However, since the sine qua non for osteoid is mineralization, we believe that if undecalcified sections of soft areas reveal large quantities of unmineralized matrix, this matrix is coarse collagen rather than osteoid and the tumor should be classified as a malignant fibrous histiocytoma. In all the various histologic patterns of malignant fibrous histiocytoma the important features to recognize are that the cells are histiocytic and can phagocytose iron or lipid and that these cells may differentiate into facultative fibroblasts that exhibit a storiform pattern.
MALIGNANT
'
~
"
,,
.
_
~
"
e~
FIBROUS HISTIOCYTOMA
,
o '~
,.
V~
8" '~
"h"
~ - -'
.
-.~
"
dPl"e'-
O F BONE--McC,,,RT~V," ET aL.
..
.
-.
"e"
9 ..'~'
_~k~.l~'~
L.
@
.t.
._
~t4tk
,
9
w O " ~ ; ".e'l-~,",'~'~:""., ~ " lk.
t *
~'.'~"g 5~" ; : . .
" .
. . . . . . . ~. ~
-" -. ,
"
:,
9
".
B
w
r D
%. o
o
Figure 9. Variations in histologic pattern of malignant fibrous histiocytoma of bone. A, Some giant histiocytes have abundant eosinopllilic cytoplasm and can look like tile "'snap" cells of rhabdcmlyosarcoma. B, Clusters of foam cells occur in histiocytic areas. C, Cholesterol clefts may occasionally occur. D, Bizarre mitotic figures occur most frequently in areas of histiocyfic differentiation. (A and D, x400. B, x 130. C, x 120.)
65
HUMAN
PATHOLOGY--VOLUME
10, N U M B E R
1
Janumy 1979
(, .TJ p
-
._
~ .... 9 - : . . ~
,t ,
.':'l,b'""Y~; %
.--"
'" "'
Q
,m.
....
.-
).
qlqb
o -
"
"
9 . .',,- -/~,,~-"~--' m ~ .J~.
; ~#I.",~,-,,,"
~... - ~ . % , , ~ ~. ,~..~ ~',~,~-
,,o.-
~-"'., 1~
~
. ~ : - ~ . ~ . . ,/., '.I~',,~ ,"
~ , j .
"'"-
%~-~""T- t.s
~,..~
~
~ I ~ 1 ~.~
o
~
~,~
~
'~ -,,,.
q~,__~
e. 4
' :.-." ,f.
Figure 10. Variations in histol~3gic pattern of malignant fibrous histioc'stoma of bone. ,-I. A~!iacemgroups of hisfiocytes and spindle cells impart an organoid appearance. B, Bizarre tumor giant ceils occur frequendy. C, Osteoclast-like giant cells may occur, ttowever, unlike the giant cell tumor, the stroma is tibrous. D, T h e net>plastic histiocytes may be obsct,red I)y it heavy infiltration of lyml)hocytes and plasma cells. (A, x llO. B, x400. C, x 160. D, x400.)
66
MALIGNANT
FIBROUS HISTIOCYTOMA
9, b :~ ~ "
j,,,v
,,,
"-/:i'
" :;
:'!
/"
i.,,
- ..--, .." , ". :
O F B O N E - - M c C A R T H Y ET AL.
" -
/
.z i k
,
" "
1,
i.;~.~
. -,,,...
.i~,
-.-;.,,;\
..
Figure 11. Variations in histologic pattern o f malignant fibrous histiocytoma o f Ix)no. A, T h e r e may be abundant cxtracclhdar g r o u n d stl|)stl.lnce, which imparts a myxoid appeara,~ce. B, Vascular slits may occur, and the t u m o r may be mistaken fi~r a hemangiopericytoma. C, T h e vascular slits may be lined by anaph~stic t u m o r cells. D, Coarse collagen matrix is often mistaken fin" osteoid. This collagen will be seen to be unmineralized in tmdecalcified sections. (:t, x 100. B, x 130. (2 and D, x 170.)
67
HUMAN PATHOLOGY--VOLUME 10, NUMBER 1 Janumy 1979 ELECTRON MICROGRAPHIC FEATURES
Electron microscopic examination was performed in three cases in this series. Although tumor cells showed a wide spectrum of differentiation from fibroblastlike cells to histiocytic cells, the five distinct types of cells seen by Inada et al. 6 in their study of five cases of malignant fibrous histiocytoma of bone could not be distinguished. The fibroblastic cells were elongated and contained abundant rough endoplasmic reticulum. These cells were associated with abundant extracellular collagen. The histiocytic cells were round or polygonal and usually contained abundant phagocytosed lipid and iron. Most tumor cells showed intermediate differentiation between these types of cells and showed evidence of both protein synthesis and phagocytosis. Myofibrils, which have been reported in some malignant fibrous histiocytomas, were not seen. -~
DISCUSSION
68
In 1960 Stout za introduced the concept that neoplastic histiocytic proliferations can differentiate as fibroblasts both fimctionally and morphologically. Tissue culture studies by Ozzello et al. 25 subsequently demonstrated in vitro that certain neoplastic histiocytes can function as facultative fibroblasts. In retrospect such entities as dermatofibrosarcoma protuberans, juvenile xanthogranuloma, and sclerosing hemangioma were believed to be neoplastic proliferations in which the varying functional potential of the histiocyte found different morphologic expressions. By 1964 O'Brien and Stout 2G showed that some of these neoplastic histiocytic proliferations, called fibrous xanthomas, could behave as malignant.tumors, and in 1967 Stout and Lattes ~9suggested the name malignant fibrous histiocytoma for this group of neoplasms. The nltrastrt~ctural features of this group of tumors were reported in 1971 by Merkow et al., -~7in 1974 by Fu et al., -~ in 1977 by Taxy and Battifora, 29 and in 1978 by Alquacil-Garcia et al? ~ The latter three studies postulated that the cell of origin of this tumor might not be the histiocyte itself but the undifferentiated mes-
enchymal cell that develops along fibroblastic, histiocytic, and intermediate cell lines. Malignant fibrous histiocytomas were originally recognized as soft tissue sarcomas. It was not until 1972 that Feldman and Norman 3 reported 20 cases in which t h e r e was bone involvement by this neoplasm. Nine of these cases were said to be primary bone tumors. In 1977 Feldma'n and Lattes z added 14 cases of primary malignant fibrous histiocytoma of bone to the original group of nine. They described in detail the radiographic presentation of this tumor and reported follow-up information in 11 of their 23 cases. Seven patients had died between eight months and eight years after treatment and four were alive but had metastases. A poor prognosis with this bone tumor may be supported by the 11 cases reported by Spanier et al. 11 in 1975. In 1977 Spanier 12 added four cases. Of the total of 15 patients, 11 had died of the disease within four years (average, one and one-half years). Two of the surviving patients had been followed less than one ),ear. These authors stressed that metastatic foci could assdme a histiocytic or fibroblastic differentiation irrespective of the predominant, feffture in the primary tumor. Lymph node metastases were present in three patients. In 1976 Huvos 5 reported 18 cases of primary malignant fibrous histiocytoma of bone. Six of these had been followed for two years or less. Of the remaining 12 patients, four were alive without disease for five years or longer; one was alive with a recurrence at five years; and seven had died two months to eight years after diagnosis. Dahlin et al. ~ in 1977 reported 35 cases of this primary bone tumor. Most of these were collected from a retrospective study of fibrosarcomas and osteosarcomas. Follow-up information was available in 28 patients, 19 of whom had amputations. In this group of 19 patients there were six long term, disease free survivors. In the group of nine patients who did not tmderg.o amputation but were treated with radiauon or chemotherapy, six survived longer than five )'ears. Clinical cure was effected by radiation alone in two of their patients. The overall long term survival rate was 33 per cent.
MALIGNANT
FIBROUS
HISTIOCYTOMA
The follow-up information in our series of 35 cases indicates that the prognosis in primary malignant fibrous histiocytoma of bone is not as bleak as in the neoplasms with which this tumor had been formerly grouped, i.e., fibrosarcoma and osteosarcoma. Our stud), also supports the observations of Dahlin et al. that the prognosis is not as bad as earlier reports" and series of this tumor would indicate. Therefore, because it has distinctive histologic features with recognizable variants, we believe that its continued classification as a distinct primary bone tumor is justified. Nine of our 21 patients in whom we have follow-up data were alive and disease free for three and one-half ),ears or longer. Two are alive with solitary metastases at three years. All our 10 patients who died of this tumor died before three years. These data suggest that 50 per cent of the patients with primary malignant fibrous lfistiocytoma of bone will survive three ),ears. Five of our patients (25 per cent) have survived longer than eight years. Therefore, this tumor is less aggressive than osteosarcoma or fibrosarcoma. There was no correlation between the age of the patient or any of the histiologic expressions of the tumor, including degree of pleomorphism and the presence or absence of osteoid-like matrix, and survival. There is, therefore, no way to predict which patients will do well. Because all but one of the survivors had amputations, we must recommend amputation as the primary therapy for primary malignant fibrous histiocytoma of bone. Our data do not permit us to comment about the efficacy of radiotherapy. Four patients had multicentric lesions at the time of diagnosis. Their ages were 12, 18, 22, and 32. Except for one patient with two separate lesions in the same femur, the locations of the muiticentric lesions were the axial skeleton and femur. In this series four tumors arose in association with bone infarcts. T h e locations of these tumors were the distal femur, distal tibia (two cases), and proximal tibia. The tumors occurred directly adjacent to the bone infarcts. The ages of the patients ran.ged from 44 to 69 )'ears. Two of the patients had a history of occupational exposure to compressed air. The cause of the infarcts in the other two cases is not known. Two of these patients died and one
O F B O N E - - M c C A R T H y ET AL.
was alive after 11 )'ears. The other case is a recent one (1977). These four cases bring the total of reported cases of malignant fibrous histiocytomas associated with bone infarcts to 12. One case previously reported by Michael and Dorftnan 7 is included in this series, Mirra et al. s, 0 reported a total of four cases. Spanier reported that in her series of 15 cases two were associated with bone infarction. Six additional cases in that series were associated with other underlying bone lesions, such as Paget's disease, fibrous dysplasia, and enchondroma. In 1978 Galli et al. 4 reported two cases of malignant fibrous histiocytoma associated with bone infarction. One patient died of metastases at 19 months and the other was alive after four years. The)' snggested that the risk of developing sarcoma is greatest in infarcts with a large medullary component. In Dahlin's series, 1'2z three lesions developed in Paget's disease and four developed in bones that had been exposed to radiation. This suggests that the prolonged reparative process that accompanies these lesions, especially infarcts, may be an important factor in the pathogenesis of malignant fibrous histiocy.toma. Ahhough other malignant neoplasnhs may complicate bone infarcts, suclii(as fibrosarcoma and osteosarcoma, 4,s,17;3r malignant fibrous lfistiocytoma is the most common tumor to do so, and this tumor must be suspected when a radiographically malignant tumor is found associated with a bone infarct and when a patient with a known history of bone infarction develops a change in symptoms. Of the total of 12 reported malignant fibrous histiocytomas associated with bone infarction, follow-up information has been reported in nine cases. Of these, six died 13 to 20 months after treatment and three were alive after one and one-half to 11 )'ears. This suggests that the prognosis in malignant fibrous histiocytoma is unaffected by the association with bone infarction.
ACKNOWLEDGMENTS
Tile authors wish to thank the following physicians for contributing cases: John Pickren, Amnon Fried, Gist Farr, Frederic Dalldorf, William Ober, Aaron Perlman, Joseph Eggleston, Risa Mann, Anthony
69
HUMAN PATHOLOGY--VOLUME
10, N U M B E R 1 Jamtmy 1979
Prezyna, Sheldon Glusman, Andre Rigaud, Suzanne Spanier, Harlan Spjut, John Sutherland, S. H. Edberg, Nilo Herrera, James McAlister, Richard SIavin, Michael Bonfiglio, Sidney GeIiman, Douglas Riddell, and Helen Feiner. We also wish to thank Mrs. Marie Shipley for preparation of the mamtscript and Miss Harriet Sherr for preparation of the photographs. REFERENCES I. Dahlin, D. C., Unui, K. K., and Matsuno, T.: Malignant fibrous Ifistiocytoma of b o n e - f a c t or fancy? Cancer, 39:1508, 1977. 2. Feldman, F., and Lattes, R.: Primary malignant fibrous histiocytoma (fibrous xanthoma) of bone. Skel. Radiol., 1:145, 1977. 3. Feldman, F., and Norman, D.: Intra- and extraosseous malignant histiocytoma (malignant fibrous xanthoma). Radiology, 104:497, 1972. 4. Galli, S.J., Weintraub, H. P., and Proppe, K. H.: Malignant fibrous histiocytoma and pleomorphic sarcoma in association with medullary bone infarcts. Cancer, 41:607, 1978. 5. Hurts, A.: Primary malignant fibrous histiocytoma of bone. Clinicopathologic study of 18 patients. N.Y.J. Med., 76:552, 1976. 6. Inada, O., Yumoto, T., Furuse, K., and Tanaka, T.: Uhrastructural features of malignant fibrous histiocytoma of bone. Acta Path. Jap., 24:491, 1976. 7. Michael, R. H., and Dorfman, H. D.: Malignant fibrous histiocytoma associated with bone infarcts. Report of a case. Clin. Orthop., 118:180, 1976. 8. Mirra, J. M., Bullongh, P. G., Marcove, R. C., Jacobs, B., and Huvos, A. G.: Malignant fibrous histiocytoma and osteosarcoma in association with bone infarcts. Report of four cases, two in caisson workers. J. Bone Joint Surg., 56A:932, 1974. 9. Mirra, J. M., Gold, R. H., and Marafiote, R.: Malignant (fibrous) histiocytoma arising in association with a bone infarct in sickle-cell disease: coincidence or cat, se-and-effect? Cancer, 39:186, 1977. 10. Newland, R. C., Harrison, M. A., and Wright, R. G.: Fibroxanthosarcoma of bone. Pathology, 7:203, 1975. 11. Spanier, S. S.: Malignant fibrous histiocytoma of bone. Orth. Clin. N. Amer., 8:947, 1977. 12. Spanier, S. S., Enneking, W. F., and Enriquez, P.: Primary malignant fibrous histiocytoma of bone. Cancer, 36:2084, 1975. 13. Spector, G. F., and Ogura, 3. H.: Malignant fibrous histiocytoma of the maxilla. A report of an unusual lesion. Arch. Otolaryngol., 99:385, 1974. 14. Yumoto, T., Mori, Y., Inada, O., and Tanaka, T.:
7O
15. 16.
17.
18.
19.
20. 21. 22. 23.
24. 25.
26. 27.
28.
29. 30.
31.
Malignant fibrous histiocytoma of bone. Acta Path. Jap., 26:295, 1976. Webber, W., and Wienke, E. C.: Malignant fibrous histiocytoma of the mandible. Case report. Plast. Reconstr. Surg., 60:629, 1977. Dorfman, H. D., Norman, A., and Wolff, H.: Fibrosarcoma complicating bone infarction in a caisson worker. J. Bone Joint Surg., 48,4:528, 1966. Spjut, H.J., Dorfman, H. D., Feclmer, R. E., and Ackerman, L. V.: Tumors of bone and cartilage. Atlas of Tumor Pathology, Fasc. 5, ~nd Series. Washington, D.C., Armed Forces Institute of Pathology, 1971, pp. 194-196. Soule, E. H., and Enriquez, P.: Atypical fibrous histiocytoma, malignant fibrous histiocytoma, malignant histiocytoma, and epithelioid sarcoma. A comparative study of 65 tumors. Cancer, 30:128, 1972. Stout, A. P., and Lattes, R." Tumors of soft tissues. Atlas of Tumor Pathology, Fasc. 1, 2nd Series. Washington, D.C., Armed Forces Insti9 tute of Pathology, 1967, pp. 38-52. Kyriakos, M., and Kempson, R. L.: Inflammatory fibrous histiocytoma. An aggressive and lethal lesion. Cancer, 37:1584, 1976. Weiss, S. W., and Enzinger, F. M.: M)xoid variant of malignant fibrous hisfiocytoma. Cancer, 39:1672, 1977. Dahlin, D. C., and Unni, K. K.: Osteosarcoma and its important recognizable varieties. Am. J. Surg. Path., 1:61, 1977. Churg, A. M., and Kahn, L. B.: Myofibroblasts and related ceils in malignant fibrous and fibrohistiocytic tur~ors. Human Path., 8:205, 1977. ~l~." Stout, A. P.: Fibre'us. tumors of the soft tissues. Minerva Med.; 43:455, 1960. Ozzello, L., Stovt, A. P., and Murray, M. R.: Cultural characteristics of malignant histiocytomas and fibrous xanthomas. Cancer, 16:331, 1963. O'Brien, J. E., and Stout, A. P.: Malignant fibrous xanthomas. Cancer, 17:1445, 1964. Merkow, L. P., Frich,J. C., Slifkin, M., Kyreages, C. G., and Pardo, M.: Ultrastructure of the fibroxanthosarcoma (malignant fibroxanthorna). Cancer, 28:372, 1971. Fu, Y., Babbiaui, G., Kaye, G. I., and Lattes, R.: Malignant soft tissue tumors of probable histiocytic origin (malignant fibrous histiocytoma). General considerations and electron microscopic and tissue culture studies. Cancer, 35:176, 1975. Taxy, J. B., and Battifora, H.: Malignant fibrous histiocytoma. An electron microscopic stud)'. Cancer, 40:254, 1977. Alguacil-Garcia, A., Unni, K. K., and Goellner, J. R.: Malignant fibrous histiocytoma. An ultrastructural study of six cases. Am. J. Clin Path., 69:121, 1978. Furey, J. G., Ferrer-Torells, M., and Reagan, J. W.: Fibrosarcoma arising at the site of bone infarcts. A report of two cases. J. Bone Joint Surg., 42A:802, 1960.
Department of Pathology Sinai Hospital of Baltimore Baltimore, Maryland 21215 (Dr. McCarthy)
Abstract Thirty-five cases of primary malignant fibrous histiocytoma of bone are reported. Twenty of these cases were collected from a retrospective analysis of other malignant bone tumors. The age range was fi'om 11 to 69 )'ears; the average age was 34 years. The tumor occurred most commonly in the distal femur and proximal tibia. The distinguishing histologic feature was a storiform arrangement of spindle cells. The differential diagnosis included fibrosarcoma, osteogenic sarcoma, malignant giant cell tumor, malignant l)'mphoma, and metastatic carcinoma. Follow-up of at least three ).ears was available in 21 cases. Of these,' nine patients were alive and free of metastases three and one-half to 12 years after treatment. Two were alive with solitary metastases at three years;, and 10 patients died between three months and three )'ears after treatment. :)'~ In four cases the lesions were multicentric at the time of'diagnosis mad in four cases were associated with bone infarction. This tumor must be recognized as an important complication of bone infarction and should b e suspected when a patient with a known history of bone infarction develops a change in symptoms. Because the prognosis of this tumor is significantly better than that in those tumors with which it had been previously grouped, and in view of its association with bone infarction, it deserves to be maintained as a distinct clinicopathologic entity. Amputation is the treatment of choice.
It Ires only been since 1972 that primary malignant fibrous histiocytoma of bone has been recognized as a distinct bone tumor. Since then there have been a total of 105 cases reported, v~5 Most of these were identified on the basis of their microscopic features from retrospective
studies of tumors previously classified as anaplastic fibrosarcomas or osteosarcomas. Although the association of tiffs neoplasm with bone infarction has been well documented,4, z-,a,H no clear pattern of its clinical behavior has emerged from the four large series and several case reports.
*Instructor in Pathology, The Johns Hopkius Medical lnstitt,tions. Assistant Pathologist, Sinai ltospital of Baltimore, Baltimore, Maryland. ?Staff, Department of Pathology, I lokkaido University School of Medicine, Sapporo, Japan. +Associate Professor of Pathology and Orthopedic Surgery, The Johns t lopkins Medical Institutions. Pathologist-in-Chief, Sinai Hospital of Bahimore, Baltimore, Maryland.
-57
HUMAN PATHOLOGY--VOLUME I0, NUMBER 1 Jamlmy 1979 O u r purpose in tiffs article is to present our experience with 35 cases of malignant fibrous histiocytoma of bone, four of which were associated with bone infarcts. O u r data suggest that the t u m o r is less aggressive than much of the previous literature suggests and the association with bone infarction is confirmed.
MATERIALS
AND
METHODS
fibrous histiocytoma complicating a bone infarct. 7 Fifteen cases, in patients seen in the last three )'ears, were diagnosed originally as primary malignant fibrous histiocytoma of bone. We have noted focal patterns of malignant fibrous histiocytoma in osteogenic sarcomas, sarcomas complicating Paget's disease, and in some dedifferentiated chondrosarcomas, but these hax'e been excluded f r o m the present study.
These 35 cases were collected f r o m the fles of the Hospital for Joint Diseases and Medical Center, Sinai Hospital of Baltimore, Johns Hopkins Hospital, and the personal consultation files of one of us (HDD). T w e n t y cases were collected in a retrospective analysis o f primary bone neoplasms, principally fibrosarcomas. T h i r t e e n originally had been diagnosed as anaplastic fibrosarcoma, one of which was associated with muhiple bone infarcts. Other diagnoses had been osteosarcoma, giant cell tumor, and malignant lymphoma. Reports of three cases had been previously published by one of us (HDD), two as anaplastic fibrosarcomas complicating bone infarcts ~G'tr and one as a malignant
109810
CLINICAL PRESENTATION In this series there were 23 males and 12 females. T h e age range was 11 to 69 )'ears and the average age was 34 years. Most patients were either in their teens or in their 40s (Fig. i). T h e location varied. However, 21 tumors occurred a r o u n d the knee with 12 in the proximal tibia, eight in the distal f e m u r , and one in the proximal fibula (Fig. 1). In two cases there was contiguous im'olvement of two bones at the time of initial diagnosis. Four cases were multicentric from the outset (Fig. 2). O f these, one case involved the f e m u r and spine; a n o t h e r invol~'ed the femur, spine,
"::::::::::
]!]i]i]i]i] MALES 23 "EMALES 12 i:!:i:!:!:! FEMALE~ -g iii!i!ilill TOTAL
......
5-
iF!i!i!i!i!
iiiiiiiiii iiiiiiiiiii -8
I "12" --1
":':':':':'tl'"f'"?'l
10
58
20
i
30 40 AGE
i
50
60
70
Figure 1. Age and sex distribution ill 35 cases, skeletal distribution of 31 solitary lesions. Asterisks denote sites associated with bone infarction.
,MALIGNANT FIBROUS HISTIOCYTOMA OF BONE--McCARTHY" ET AL. cases arose in association with bone infarction. O f the cases associated with b o n e infarction, two were in patients with occupational e x p o s u r e to compressed air environments and two were in patients with idiopathic infarcts. I n f o r m a t i o n relating to clinical symptoms were available in 23 cases. Bone pain was present in 22 cases for one m o n t h to five )'ears b e f o r e diagnosis. One patient was asymptomatic until the t u m o r was f o u n d in r a d i o g r a p h s taken following trauma.
RADIOGRAPHIC PATTERNS
Figure 2. Skeletal distribution in fimr muhicentric cases.
and a rib; a n o t h e r involved the ilium and ribs; and one patient had two separate lesions in the same f e m u r . F o u r patients presented with pathologic fractures and f o u r
Radiographs were available for stud)' in 33 cases. A radiologic diagnosis o f malignant disease was m a d e in 23 cases. T h e tumors p r e s e n t e d as ill defined, single or confluent areas o f radiolucency with endosteal erosion and minimal cortical expansion (Fig. 3). Occasionally the cortex was p e n e t r a t e d completely (Fig. 4). T h e r e was minimal reactive b o n e formation, and evidence o f t u m o r calcification and bone formation was not presen.t. Occ,asionally a highly aggressive m a l i g n a n t ' r a d i o g r a p h i c pattern resembling Ewiri~ s sarcoma o r tel-
Figure 3. Radiographs of maligtmnt fibrous histiocytoma showing a radiolucent defect with endosteal erosion :rod minimal cortical expansion.
59
HUMAN PATHOLOGY--VOLUME 10, NUMBER 1 January 1979
Figure 4. Radiographs of malignant fibrous histiocytoma showing a poorly circumscribed, radiolucellt defect with cortical penetration.
angiectatic osteosarcoma was seen (Figs. 5, 6). In a few cases the radiographic findings suggested a benign t u m o r such as nonossifying fibroma, giant cell t u m o r o f bone, or solitary b o n e cyst (Fig. 7). In the f o u r cases associated with b o n e infarction there were one or m o r e islands o f dense intraosseous calcification resembling typical b o n e infarcts. T h e involved area o f infarction was adjacent to or surr o u n d e d by m a r k e d radiolucency with cortical erosion and often with cortical destruction.
TREATMENT
60
AND
FOLLOW-UP
In this study o f 35 patients, follow-up data for at least three years were available in 21 patients. O f the remaining 14 cases, 13 were diagnosed witlfin the p r e c e d i n g three years and are not considered in the follow-up data. Follow-up- information was not available in one case. O f the 21 patients, nine were alive and well without metastases t h r e e and one-half to 12 years after treatment. Five patients were well longer than eight ),ears. T w o o t h e r patients were alive with solitary me-
tastases after t h r e e years, and 10 patients were dead. / Death o c c u r r e & " i o the 10 patients three m o n t h s to three years after treatinent. T h e age r a n g e was 15 to 60 years with no age g r o u p prevailing. T w o o f the patients who died had multicentric tumors initially. Seven o f the 10 patients were treated by a m p u t a t i o n and one had additional r a d i o t h e r a p y . One had a resection, and the two patients with muhicentric tumors were treated by radiation, one after a simple curettage. Six had lung metastases, which a p p e a r e d one m o n t h to two )'ears after treatment. In one patient t h e r e was regional s p r e a d to o t h e r bones at six months. T w o patients had lymph n o d e metastases discovered at the time o f autopsy. Autopsies were p e r f o r m e d in three patients, and in these cases the histologic features o f the metastases did not vary f r o m those o f the p r i m a r y tumor. T h e r e were no distinctive histologic features o f the primary t u m o r in the patients who died. O f the nine patients alive and free o f disease longer than three and o n e - h a l f years, all but o n e were treated by amputation and only one o f these had p r e o p e r a tive radiotherapy. T h e patient who did not
MALIGNANT
FIBROUS
HISTIOCYTOMA
O F BONE--McCaRTW~" ET ak.
Figure 5. This malignant fibrous histioc)tonm has a permeative pattern and could be interpreted as representing Ewing's sarcotna.
Figure 6. Radiograph showing a destructive malignant fibrous histiocytoma of tile ischitml. T h e r e is no reactive bone formation. This" rar " ~," pattern resembles a tclangiectatic o s l e o s a r c o n l a . -
61
HUMAN PATHOLOGY--VOLUME 10, NUMBER 1 Janumy 1979
,'ql
t i/ i
t/ Figure 7. Occasionally a radiographically benign pattern may occur, This malignant fibr,,us histi,)cvtoma was t h o u g h t by radiographic examination to be a metaphyseal fii)rous defect. (Follow-ul): T h e l)atient; a 60 year old man, died of p u h n o n a r y metastases 1 ~/2 yCd[llSfollowing ampuhltion.)
have an amputation was alive three and one-half years after a simple curettage.
HISTOLOGIC FEATURES A N D DIFFERENTIAL DIAGNOSIS
62
The histologic features of malignant fibrous histiocytomas of soft tissue have been described in earlier reports.'8.19 The range of histologic expression of primary malignant fibrous histiocytoma of bone is as broad as that seen in soft tissues. We believe that these tumors are neoplastic proliferations of malignant histiocytes in which their capacity to act as facuitative fibroblasts is expressed. The histologic variations of the tumor reflect tile two fimctional states of tile neoplastic cell - - fibroblastic and histiocytic. Therefore, there is a morphologic spectrum ranging from sheets of histiocytes that may show phagocytosis to a storiform or matted appearance. These patterns may alternate within a given tumor in equal proportions, or one pattern may predominate. More often tile fibroblastic and histiocytic expressions of tiffs biphasic cell are randomly intermingled. Cytologic features of malignancy are
seen in both patterns, but nuclear atypicality and abnormal mitotic figures are more florid in the !)]~stiocytic areas. Also a number of special his.tologic patterns, to be discussed, may be seen in any tumor, and these may dominate the histologic picture, leading to difficulties in differential diagnosis. The storiform pattern is the most characteristic feature of malignant fibrous histiocytoma. Bundles of elongated or spindle cells are arranged in a pinwheel fashion ~Fig. 8.4). The nuclei are oval or elongated and vesicular and show mild to moderate pleomorphism. The cytoplasmic margins are indistinct. If tile spindle cell pattern predominates, the tumor may be misdiagnosed as a fibrosarcoma. However, the presence of the storiform pattern will help to identify the tumor as a fibrous histiocytoma. Tile other end of the spectrum of differentiation of the neoplastic cell may predominate and tile tumor will be composed of sheets of tmiform histiocytes. Occasionall),, when tiffs occurs, a diagnosis of histiocytic lymphoma is made. But these histiocytic cells will show a tendency to differentiate toward spindle cells, a feature that is not seen i n lymphoma (Fig. 8B).
MALIGNANT
FIBROUS
HISTIOCYTOMA
O F B O N E - - M c C A R T H Y ET AL.
9
.b.
Q
,...:- , , o9 ~o
9
9 .
.,..
I
o
D
10
9
"
I
Figure 8, :1, Storifi~rm arrangement of spindle cells, fl, Ncoplastic uniform histiocytes will show subtle differentiation to spindle cells. C, Sheets o[ bizarre plcom~rl)hic histiocytes. D. Bilobed mzclci with prominent mlcleoli may look like Reed-Sternbcrg cells. (:1, • 1S0. B, • 200. C, • I ! 0. D, •
63
HUMAN PATHOLOGY--VOLUME 10, NUMBER 1 JanuaD'1979 Sometimes, rather than being uniform, the histiocytes are quite pleomorphic and bizarre (Fig. 8C). In the past this pattern was interpreted as anaplastic fibrosarcoma. The histiocytes have large, oval or lobulated, vesicular nuclei, often with a prominent eosinophilic nucleolus. When the nuclei are bilobed, there may be a resemblance to Reed-Sternberg cells (Fig. 8/)). T h e cytoplasmic margins of these histiocytes are distinct, and the cytoplasm is sometimes so abundant as to look like the "strap" cells of rhabdomyosarcoma (Fig. 9A). Close observation reveals that the cytoplasm is foamy and cross striations are not seen. There are sometimes clusters of foam}' cells intermingled with the atypical histiocytes. This feature, along with lipid laden cytoplasm of the histiocytes, has led to the erroneous interpretation of liposarcoma (Fig. 9B). Occasionally even cholesterol clefts may be seen (Fig. 9C). Hemosiderin is also frequently seen within the cytoplasm of the neoplastic histiocytes. Abnormal mitotic figures are most prominent in fields of histiocytic differentiation (Fig. 9D). Occasionally the histiocytic and fibroblastic components are both present but distinct. This imparts a biphasic pattern to the tumor with clusters of epithelioid histiocytes separated by fascicles of elongated fibroblast-like cells (Fig. 10A). This epithelioid pattern can lead to the misdiagnosis of metastatic carcinoma. Muhinucleated giant cells are often a prominent feature of malignant fibrous histiocytoma. More often they are tumor giant cells in areas of histiocytic differentiation (Fig. 10B). However, osteoclast-like giant cells can sometimes be seen (Fig. 10C). For this reason the tumor can be mistaken for a giant cell tumor. However, if the background is fibrous and is made up of pleomorphic spindle cells that tend to form a storiform pattern, the lesion can be recognized as a malignant fibrous histiocytoma. Often this tumor has been interpreted as a malignant giant cell tumor. We prefer to reserve this 51iagnosis for malignant tumors that arise in a setting of a typical giant cell tumor of bone. One variant that is occasionally encountered is a diffnse infiltrate of inflammatory cells admixed with and sometimes
64
obscuring the neoplastic cells (Fig. 10D). This variant is analogous to that which occurs in soft tissues described by Kyriakos and Kempson. 2~ The inflammatory infiltrate consists of plasma cells, lymphocytes, and polymorphonuclear leukocytes. The polymorphous population of cells in this variant along with Reed-Sternberg-like cells can lead to the misdiagnosis of Hodgkin's disease. The spindle cells of malignant fibrous histiocytoma of bone may sometimes be separated by abundant extracellular ground substance, which imparts a focal myxomatous appearance to the tumor (Fig. 1 1A). This variant also has a soft tissue counterpart in the myxoid variant of malignant fibrous histiocytoma described by Weiss and Enzinger. 21 Another variant that occurs in bone is a pattern similar to that of hemangiopericytoma. T h e neoplastic cells are arranged in clusters or bundles in an organoid pattern around vascular spaces (Fig. 11B). Occasionally a single layer of neoplastic cells lines vascular spaces separated by a fibrous stroma. This pattern may mimic that of a hemangioendothelial sarcoma (Fig. 1 1C). Quite often dab collagen in the extracellular matrix lslvery coarse or lacelike and may be indisffnguishable from osteoid in demineralized sections (Fig. 11D). The additional presence of thin spicules of reactive bone creates the very difficult problem of distinguishing this tumor from an anaplastic osteosarcoma. Indeed, Dahlin and Unni 22 have expressed the view that this distinction is an arbitrary one. A filrther complication is that a certain small amount of osteoid can be present in malignant fibrous histiocytomas of soft tissue. However, since the sine qua non for osteoid is mineralization, we believe that if undecalcified sections of soft areas reveal large quantities of unmineralized matrix, this matrix is coarse collagen rather than osteoid and the tumor should be classified as a malignant fibrous histiocytoma. In all the various histologic patterns of malignant fibrous histiocytoma the important features to recognize are that the cells are histiocytic and can phagocytose iron or lipid and that these cells may differentiate into facultative fibroblasts that exhibit a storiform pattern.
MALIGNANT
'
~
"
,,
.
_
~
"
e~
FIBROUS HISTIOCYTOMA
,
o '~
,.
V~
8" '~
"h"
~ - -'
.
-.~
"
dPl"e'-
O F BONE--McC,,,RT~V," ET aL.
..
.
-.
"e"
9 ..'~'
_~k~.l~'~
L.
@
.t.
._
~t4tk
,
9
w O " ~ ; ".e'l-~,",'~'~:""., ~ " lk.
t *
~'.'~"g 5~" ; : . .
" .
. . . . . . . ~. ~
-" -. ,
"
:,
9
".
B
w
r D
%. o
o
Figure 9. Variations in histologic pattern of malignant fibrous histiocytoma of bone. A, Some giant histiocytes have abundant eosinopllilic cytoplasm and can look like tile "'snap" cells of rhabdcmlyosarcoma. B, Clusters of foam cells occur in histiocytic areas. C, Cholesterol clefts may occasionally occur. D, Bizarre mitotic figures occur most frequently in areas of histiocyfic differentiation. (A and D, x400. B, x 130. C, x 120.)
65
HUMAN
PATHOLOGY--VOLUME
10, N U M B E R
1
Janumy 1979
(, .TJ p
-
._
~ .... 9 - : . . ~
,t ,
.':'l,b'""Y~; %
.--"
'" "'
Q
,m.
....
.-
).
qlqb
o -
"
"
9 . .',,- -/~,,~-"~--' m ~ .J~.
; ~#I.",~,-,,,"
~... - ~ . % , , ~ ~. ,~..~ ~',~,~-
,,o.-
~-"'., 1~
~
. ~ : - ~ . ~ . . ,/., '.I~',,~ ,"
~ , j .
"'"-
%~-~""T- t.s
~,..~
~
~ I ~ 1 ~.~
o
~
~,~
~
'~ -,,,.
q~,__~
e. 4
' :.-." ,f.
Figure 10. Variations in histol~3gic pattern of malignant fibrous histioc'stoma of bone. ,-I. A~!iacemgroups of hisfiocytes and spindle cells impart an organoid appearance. B, Bizarre tumor giant ceils occur frequendy. C, Osteoclast-like giant cells may occur, ttowever, unlike the giant cell tumor, the stroma is tibrous. D, T h e net>plastic histiocytes may be obsct,red I)y it heavy infiltration of lyml)hocytes and plasma cells. (A, x llO. B, x400. C, x 160. D, x400.)
66
MALIGNANT
FIBROUS HISTIOCYTOMA
9, b :~ ~ "
j,,,v
,,,
"-/:i'
" :;
:'!
/"
i.,,
- ..--, .." , ". :
O F B O N E - - M c C A R T H Y ET AL.
" -
/
.z i k
,
" "
1,
i.;~.~
. -,,,...
.i~,
-.-;.,,;\
..
Figure 11. Variations in histologic pattern o f malignant fibrous histiocytoma o f Ix)no. A, T h e r e may be abundant cxtracclhdar g r o u n d stl|)stl.lnce, which imparts a myxoid appeara,~ce. B, Vascular slits may occur, and the t u m o r may be mistaken fi~r a hemangiopericytoma. C, T h e vascular slits may be lined by anaph~stic t u m o r cells. D, Coarse collagen matrix is often mistaken fin" osteoid. This collagen will be seen to be unmineralized in tmdecalcified sections. (:t, x 100. B, x 130. (2 and D, x 170.)
67
HUMAN PATHOLOGY--VOLUME 10, NUMBER 1 Janumy 1979 ELECTRON MICROGRAPHIC FEATURES
Electron microscopic examination was performed in three cases in this series. Although tumor cells showed a wide spectrum of differentiation from fibroblastlike cells to histiocytic cells, the five distinct types of cells seen by Inada et al. 6 in their study of five cases of malignant fibrous histiocytoma of bone could not be distinguished. The fibroblastic cells were elongated and contained abundant rough endoplasmic reticulum. These cells were associated with abundant extracellular collagen. The histiocytic cells were round or polygonal and usually contained abundant phagocytosed lipid and iron. Most tumor cells showed intermediate differentiation between these types of cells and showed evidence of both protein synthesis and phagocytosis. Myofibrils, which have been reported in some malignant fibrous histiocytomas, were not seen. -~
DISCUSSION
68
In 1960 Stout za introduced the concept that neoplastic histiocytic proliferations can differentiate as fibroblasts both fimctionally and morphologically. Tissue culture studies by Ozzello et al. 25 subsequently demonstrated in vitro that certain neoplastic histiocytes can function as facultative fibroblasts. In retrospect such entities as dermatofibrosarcoma protuberans, juvenile xanthogranuloma, and sclerosing hemangioma were believed to be neoplastic proliferations in which the varying functional potential of the histiocyte found different morphologic expressions. By 1964 O'Brien and Stout 2G showed that some of these neoplastic histiocytic proliferations, called fibrous xanthomas, could behave as malignant.tumors, and in 1967 Stout and Lattes ~9suggested the name malignant fibrous histiocytoma for this group of neoplasms. The nltrastrt~ctural features of this group of tumors were reported in 1971 by Merkow et al., -~7in 1974 by Fu et al., -~ in 1977 by Taxy and Battifora, 29 and in 1978 by Alquacil-Garcia et al? ~ The latter three studies postulated that the cell of origin of this tumor might not be the histiocyte itself but the undifferentiated mes-
enchymal cell that develops along fibroblastic, histiocytic, and intermediate cell lines. Malignant fibrous histiocytomas were originally recognized as soft tissue sarcomas. It was not until 1972 that Feldman and Norman 3 reported 20 cases in which t h e r e was bone involvement by this neoplasm. Nine of these cases were said to be primary bone tumors. In 1977 Feldma'n and Lattes z added 14 cases of primary malignant fibrous histiocytoma of bone to the original group of nine. They described in detail the radiographic presentation of this tumor and reported follow-up information in 11 of their 23 cases. Seven patients had died between eight months and eight years after treatment and four were alive but had metastases. A poor prognosis with this bone tumor may be supported by the 11 cases reported by Spanier et al. 11 in 1975. In 1977 Spanier 12 added four cases. Of the total of 15 patients, 11 had died of the disease within four years (average, one and one-half years). Two of the surviving patients had been followed less than one ),ear. These authors stressed that metastatic foci could assdme a histiocytic or fibroblastic differentiation irrespective of the predominant, feffture in the primary tumor. Lymph node metastases were present in three patients. In 1976 Huvos 5 reported 18 cases of primary malignant fibrous histiocytoma of bone. Six of these had been followed for two years or less. Of the remaining 12 patients, four were alive without disease for five years or longer; one was alive with a recurrence at five years; and seven had died two months to eight years after diagnosis. Dahlin et al. ~ in 1977 reported 35 cases of this primary bone tumor. Most of these were collected from a retrospective study of fibrosarcomas and osteosarcomas. Follow-up information was available in 28 patients, 19 of whom had amputations. In this group of 19 patients there were six long term, disease free survivors. In the group of nine patients who did not tmderg.o amputation but were treated with radiauon or chemotherapy, six survived longer than five )'ears. Clinical cure was effected by radiation alone in two of their patients. The overall long term survival rate was 33 per cent.
MALIGNANT
FIBROUS
HISTIOCYTOMA
The follow-up information in our series of 35 cases indicates that the prognosis in primary malignant fibrous histiocytoma of bone is not as bleak as in the neoplasms with which this tumor had been formerly grouped, i.e., fibrosarcoma and osteosarcoma. Our stud), also supports the observations of Dahlin et al. that the prognosis is not as bad as earlier reports" and series of this tumor would indicate. Therefore, because it has distinctive histologic features with recognizable variants, we believe that its continued classification as a distinct primary bone tumor is justified. Nine of our 21 patients in whom we have follow-up data were alive and disease free for three and one-half ),ears or longer. Two are alive with solitary metastases at three years. All our 10 patients who died of this tumor died before three years. These data suggest that 50 per cent of the patients with primary malignant fibrous lfistiocytoma of bone will survive three ),ears. Five of our patients (25 per cent) have survived longer than eight years. Therefore, this tumor is less aggressive than osteosarcoma or fibrosarcoma. There was no correlation between the age of the patient or any of the histiologic expressions of the tumor, including degree of pleomorphism and the presence or absence of osteoid-like matrix, and survival. There is, therefore, no way to predict which patients will do well. Because all but one of the survivors had amputations, we must recommend amputation as the primary therapy for primary malignant fibrous histiocytoma of bone. Our data do not permit us to comment about the efficacy of radiotherapy. Four patients had multicentric lesions at the time of diagnosis. Their ages were 12, 18, 22, and 32. Except for one patient with two separate lesions in the same femur, the locations of the muiticentric lesions were the axial skeleton and femur. In this series four tumors arose in association with bone infarcts. T h e locations of these tumors were the distal femur, distal tibia (two cases), and proximal tibia. The tumors occurred directly adjacent to the bone infarcts. The ages of the patients ran.ged from 44 to 69 )'ears. Two of the patients had a history of occupational exposure to compressed air. The cause of the infarcts in the other two cases is not known. Two of these patients died and one
O F B O N E - - M c C A R T H y ET AL.
was alive after 11 )'ears. The other case is a recent one (1977). These four cases bring the total of reported cases of malignant fibrous histiocytomas associated with bone infarcts to 12. One case previously reported by Michael and Dorftnan 7 is included in this series, Mirra et al. s, 0 reported a total of four cases. Spanier reported that in her series of 15 cases two were associated with bone infarction. Six additional cases in that series were associated with other underlying bone lesions, such as Paget's disease, fibrous dysplasia, and enchondroma. In 1978 Galli et al. 4 reported two cases of malignant fibrous histiocytoma associated with bone infarction. One patient died of metastases at 19 months and the other was alive after four years. The)' snggested that the risk of developing sarcoma is greatest in infarcts with a large medullary component. In Dahlin's series, 1'2z three lesions developed in Paget's disease and four developed in bones that had been exposed to radiation. This suggests that the prolonged reparative process that accompanies these lesions, especially infarcts, may be an important factor in the pathogenesis of malignant fibrous histiocy.toma. Ahhough other malignant neoplasnhs may complicate bone infarcts, suclii(as fibrosarcoma and osteosarcoma, 4,s,17;3r malignant fibrous lfistiocytoma is the most common tumor to do so, and this tumor must be suspected when a radiographically malignant tumor is found associated with a bone infarct and when a patient with a known history of bone infarction develops a change in symptoms. Of the total of 12 reported malignant fibrous histiocytomas associated with bone infarction, follow-up information has been reported in nine cases. Of these, six died 13 to 20 months after treatment and three were alive after one and one-half to 11 )'ears. This suggests that the prognosis in malignant fibrous histiocytoma is unaffected by the association with bone infarction.
ACKNOWLEDGMENTS
Tile authors wish to thank the following physicians for contributing cases: John Pickren, Amnon Fried, Gist Farr, Frederic Dalldorf, William Ober, Aaron Perlman, Joseph Eggleston, Risa Mann, Anthony
69
HUMAN PATHOLOGY--VOLUME
10, N U M B E R 1 Jamtmy 1979
Prezyna, Sheldon Glusman, Andre Rigaud, Suzanne Spanier, Harlan Spjut, John Sutherland, S. H. Edberg, Nilo Herrera, James McAlister, Richard SIavin, Michael Bonfiglio, Sidney GeIiman, Douglas Riddell, and Helen Feiner. We also wish to thank Mrs. Marie Shipley for preparation of the mamtscript and Miss Harriet Sherr for preparation of the photographs. REFERENCES I. Dahlin, D. C., Unui, K. K., and Matsuno, T.: Malignant fibrous Ifistiocytoma of b o n e - f a c t or fancy? Cancer, 39:1508, 1977. 2. Feldman, F., and Lattes, R.: Primary malignant fibrous histiocytoma (fibrous xanthoma) of bone. Skel. Radiol., 1:145, 1977. 3. Feldman, F., and Norman, D.: Intra- and extraosseous malignant histiocytoma (malignant fibrous xanthoma). Radiology, 104:497, 1972. 4. Galli, S.J., Weintraub, H. P., and Proppe, K. H.: Malignant fibrous histiocytoma and pleomorphic sarcoma in association with medullary bone infarcts. Cancer, 41:607, 1978. 5. Hurts, A.: Primary malignant fibrous histiocytoma of bone. Clinicopathologic study of 18 patients. N.Y.J. Med., 76:552, 1976. 6. Inada, O., Yumoto, T., Furuse, K., and Tanaka, T.: Uhrastructural features of malignant fibrous histiocytoma of bone. Acta Path. Jap., 24:491, 1976. 7. Michael, R. H., and Dorfman, H. D.: Malignant fibrous histiocytoma associated with bone infarcts. Report of a case. Clin. Orthop., 118:180, 1976. 8. Mirra, J. M., Bullongh, P. G., Marcove, R. C., Jacobs, B., and Huvos, A. G.: Malignant fibrous histiocytoma and osteosarcoma in association with bone infarcts. Report of four cases, two in caisson workers. J. Bone Joint Surg., 56A:932, 1974. 9. Mirra, J. M., Gold, R. H., and Marafiote, R.: Malignant (fibrous) histiocytoma arising in association with a bone infarct in sickle-cell disease: coincidence or cat, se-and-effect? Cancer, 39:186, 1977. 10. Newland, R. C., Harrison, M. A., and Wright, R. G.: Fibroxanthosarcoma of bone. Pathology, 7:203, 1975. 11. Spanier, S. S.: Malignant fibrous histiocytoma of bone. Orth. Clin. N. Amer., 8:947, 1977. 12. Spanier, S. S., Enneking, W. F., and Enriquez, P.: Primary malignant fibrous histiocytoma of bone. Cancer, 36:2084, 1975. 13. Spector, G. F., and Ogura, 3. H.: Malignant fibrous histiocytoma of the maxilla. A report of an unusual lesion. Arch. Otolaryngol., 99:385, 1974. 14. Yumoto, T., Mori, Y., Inada, O., and Tanaka, T.:
7O
15. 16.
17.
18.
19.
20. 21. 22. 23.
24. 25.
26. 27.
28.
29. 30.
31.
Malignant fibrous histiocytoma of bone. Acta Path. Jap., 26:295, 1976. Webber, W., and Wienke, E. C.: Malignant fibrous histiocytoma of the mandible. Case report. Plast. Reconstr. Surg., 60:629, 1977. Dorfman, H. D., Norman, A., and Wolff, H.: Fibrosarcoma complicating bone infarction in a caisson worker. J. Bone Joint Surg., 48,4:528, 1966. Spjut, H.J., Dorfman, H. D., Feclmer, R. E., and Ackerman, L. V.: Tumors of bone and cartilage. Atlas of Tumor Pathology, Fasc. 5, ~nd Series. Washington, D.C., Armed Forces Institute of Pathology, 1971, pp. 194-196. Soule, E. H., and Enriquez, P.: Atypical fibrous histiocytoma, malignant fibrous histiocytoma, malignant histiocytoma, and epithelioid sarcoma. A comparative study of 65 tumors. Cancer, 30:128, 1972. Stout, A. P., and Lattes, R." Tumors of soft tissues. Atlas of Tumor Pathology, Fasc. 1, 2nd Series. Washington, D.C., Armed Forces Insti9 tute of Pathology, 1967, pp. 38-52. Kyriakos, M., and Kempson, R. L.: Inflammatory fibrous histiocytoma. An aggressive and lethal lesion. Cancer, 37:1584, 1976. Weiss, S. W., and Enzinger, F. M.: M)xoid variant of malignant fibrous hisfiocytoma. Cancer, 39:1672, 1977. Dahlin, D. C., and Unni, K. K.: Osteosarcoma and its important recognizable varieties. Am. J. Surg. Path., 1:61, 1977. Churg, A. M., and Kahn, L. B.: Myofibroblasts and related ceils in malignant fibrous and fibrohistiocytic tur~ors. Human Path., 8:205, 1977. ~l~." Stout, A. P.: Fibre'us. tumors of the soft tissues. Minerva Med.; 43:455, 1960. Ozzello, L., Stovt, A. P., and Murray, M. R.: Cultural characteristics of malignant histiocytomas and fibrous xanthomas. Cancer, 16:331, 1963. O'Brien, J. E., and Stout, A. P.: Malignant fibrous xanthomas. Cancer, 17:1445, 1964. Merkow, L. P., Frich,J. C., Slifkin, M., Kyreages, C. G., and Pardo, M.: Ultrastructure of the fibroxanthosarcoma (malignant fibroxanthorna). Cancer, 28:372, 1971. Fu, Y., Babbiaui, G., Kaye, G. I., and Lattes, R.: Malignant soft tissue tumors of probable histiocytic origin (malignant fibrous histiocytoma). General considerations and electron microscopic and tissue culture studies. Cancer, 35:176, 1975. Taxy, J. B., and Battifora, H.: Malignant fibrous histiocytoma. An electron microscopic stud)'. Cancer, 40:254, 1977. Alguacil-Garcia, A., Unni, K. K., and Goellner, J. R.: Malignant fibrous histiocytoma. An ultrastructural study of six cases. Am. J. Clin Path., 69:121, 1978. Furey, J. G., Ferrer-Torells, M., and Reagan, J. W.: Fibrosarcoma arising at the site of bone infarcts. A report of two cases. J. Bone Joint Surg., 42A:802, 1960.
Department of Pathology Sinai Hospital of Baltimore Baltimore, Maryland 21215 (Dr. McCarthy)
