Elastofibroma Dorsi

The elastofi broma is a rare type of tumor, benign behavior, characterized by developing slowly from soft tissues. Virtually all develop in the scapular inferior pole, 75% unilateral and 25% bilateral. A predilection for female sex has been observed, in people over 50 years of age and it is extremely rare in the pediatric population. The etiology remains undetermined and continues to be a source of debate. Typically, symptoms may include a well-circumscribed, unmoving, painful, clicking mass with abduction and adduction of the affected scapula. We present a female patient of 63 years of age who was diagnosed as an elastofi broma dorsi by histopathological study. The patient presented a tumor in the left dorsal region of 12 months of evolution, not painful and of progressive growth. Physical examination revealed a soft, delimited mass of 5 cm in diameter, not mobile, in the dorsal region, apparently without infiltration to other tissues, it was suspected in a lipoma. A computed tomography of the chest was performed. There is evidence that surgical intervention is the therapeutic method of choice for elastofi broma dorsi, with immediate clinical improvement. A case report is made.

features are constant enough to warrant separa¬ tion of this lesion as a clinicopathological entity. Briefly, the mass is seen in the subscapular space of adult, often elderly, individuals. It enlarges slowly, displacing neighbouring structures, and is prone to be clinically confused with a sarcoma. The histopathology, however, establishes the benignity of the lesion, despite its lack of circumscription. The growth appears composed of mature fibrous tissue containing a number of fibres that behave tinctorially as elastic fibres. It is the purpose of this paper to describe the morphology and hypothetical origin of these lesions as revealed by light and electron microscopy; on the basis of these morphologic findings the possible nature of this process is considered.

Material and Methods
Light Microscopy Formalin-fixed, paraffin-embedded tissue from a case of elastofibroma dorsi observed by us and described below, and paraffin blocks from three similar cases previously reported by Barr1 were cut on a microtome and stained by the following techniques: hematoxylin and eosin, orcein, aldehyde fuchsin, resorcin fuchsin, Verhoeff, van Gieson, Congo red, von Kossa, acridine orange, MacManus' PAS and Masson's trichrome.
Unstained sections of paraffin-embedded tissue, frozen sections of formalin-fixed tissue, and sections stained by hematoxylin and eosin and by acridine orange were prepared in all cases and examined by fluorescent microscopy. Paraffin sections, 10 /x. thick, were subjected to microincineration. This was done by the simplified burner method, placing the slide on an asbestos board and applying heat from one side for 45 minutes or until the section almost dis¬ appeared. Breakage was prevented by covering the slide with an inverted metallic pan. Examin-From the Departments of Pathology, Queen's University, and the Kingston General Hospital, Kingston, Ontario. ?Assistant Professor of Pathology. tLecturer in Pathology. IProfessor of Pathology. Reprint requests to: Dr. F, Gonzalez-Crussi, Richardson Laboratory, Queen's University, Kingston, Ontario. ation was made with dark-field microscopy.
NaOH extraction: 0.5-cm. cubes of tissue from all cases fixed in 10% buffered formalin were prepared as follows: After being washed in running tap water for 24 hours, they were defatted in refluxing methanol for four hours and in cold acetone for another four hours. Sub¬ sequently the tissue fragments were placed in 0.1 N NaOH solution at 98°C. for intervals varying from 45 minutes to 2 hours. The tissue thus treated was macerated in a glass tissue grinder. Partly ground material was teased with microdissection needles and examined. The completely ground material was suspended in saline and also examined under direct and phase contrast illumination.
Electron Microscopy Fragments of formalin-fixed tissue, lxl c.mm., from the case described below were washed in phosphate buffer (pH 7.8) for 24 hours, fixed in glutaraldehyde and post-fixed in osmium tetroxide. The fragments were rinsed in veronal buffer and dehydrated in alcohol of ascending concentrations and propylene oxide. The tissues were embedded in Epon 812, cut on a Porter-Blum microtome and mounted on Formvarcoated grids. Staining was with saturated uranyl acetate solution in 70% ethanol or by phosphotungstic acid (1 g. in 100 ml. of 70% ethanol).
Examination was made with a Hitachi HU11C electron microscope. For the sake of comparison the aorta pf a newborn infant was processed and examined under similar conditions.

Case History
A 74-year-old white woman sought medical attention because of the subjective feeling of a mass beneath her left scapula. Examination at that time was entirely negative, but when re-examined three months later a "fist-sized" mass was discovered, located immediately beneath the inferior angle of the left scapula. At operation the tumour appeared unencapsulated and firmly attached to the adjacent seventh and eighth ribs and intercostal muscle. It was not adherent to the overlying latissimus dorsi or rhomboideus muscles. Total removal of this mass was possible and the patient had an uneventful recovery. There was residual functional impair¬ ment, and on follow-up one year after operation there was no evidence of recurrence. Retinal angioid streaks, abnormal skin folds and other stigmata of pseudoxanthoma elasticum were absent. Grossly the specimen consisted of a roughly spherical mass, measuring 8 cm. in diameter, and appeared to be composed of intermingled white, firm, fibrous and adipose tissue. It had a rough external surface and was poorly delimited. Scattered areas of cystic degeneration measuring up to 2 cm. in diameter were present within the mass. These areas were at times multilocular, contained serous fluid and were lined by finely granular white tissue. The microscopic findings are described along with those of the rest of the cases, as the histopathology is strikingly uniform.
Histologically these lesions were composed of mature, poorly cellular fibrous collagenous tissue forming broad bands and solid areas. The tissue was poorly vascularized, and the few capillaries present were inconspicuous, owing to compression by surrounding collagen fibres. No mitoses were seen, and the general impression was that of a scarred area rather than an active neoplasm. In the more loosely textured areas, however, the vessels were clustered in small aggregates and the degree of cellularity was inereased. This was noted particu¬ larly towards the margins of the lesions, but the cytologic characteristics were never sufficiently atypical to raise concern about the possibility of malignant neoplasm.
The salient histological feature, as illustrated in were always oriented in the same plane as collagen bundles. When viewed in profile, some of the thick fibres revealed a central core and a pattern of evenly spaced eosinophilic bands. When viewed "en face", the appearance of some fragmented fibres suggested a crystalline foreign body.3 Examination with darkfield microscopy after microincineration revealed no significant calcium deposits in the fibres or in the fragmented globular material. The tinctorial properpjg. 3._Altered elastic fibre showing banded appear¬ ance (A), distinct ramifications (B) and basophilic central core (C). Note comparatively greater cellularity, and failure to follow the plane of collagen bundles (D).
(Masson's trichrome, original magnification X 81b.) ties of the fibres resembled in general those associ¬ ated by Lansing4 with "senile" elastin, except for the absence of extensive calcification. Thus they were yellow-orange with Congo red, and stained variably with hematoxylin and densely with resorcin fuchsin, orcein and aldehyde fuchsin. When tissue was teased with microdissection needles after partial NaOH treatment, a three-dimensional arrangement of different fibres and bundles was noted. Typical elastic fibres were seen, appearing yellow in the un¬ stained state. Microscopically these fibres were long, Feb. 22, 1969, vol branching slender cylinders of high refractile index and smooth surfaces (Fig. 4). Altered elastic fibres were swollen and had irregular "moth-eaten" contours. A more marked change of swollen fibres was the presence of alternating areas of constriction, giving a banded appearance. By progressive maceration and grinding, individual fibres could be dissected out. It was then apparent that the changes were sometimes segmental, involving part of a fibre, (Unstained slide mounted in 10% aqueous ferric chloride solution. Phase contrast. Original magnification X 630.) or localized to one branch of a typical fibre (Fig.  5). Wide unbranched fibres, seemingly a part of collagen bundles, had similar refractility, so that it was difficult to distinguish them from altered elastic fibres. Their collagenous nature was sug¬ gested by their location within the rest of the bundle, their parallel orientation and their wide non-branched appearance (Fig. 6). These fibres, however, were intensely orceinophilic and more Fig. 6..Wide unbranched flbre dissected from a col¬ lagen bundle by progressive maceration and NaOH ex¬ traction. Note numerous globular fragments with marked tendency to clump. (Unstained slide mounted in 10% ferric chloride aqueous solution. Phase contrast. Original magnification X 630.) numerous in foci rich in fragmented elastic-like masses. Finally, there were globular segments suggesting an origin from broken segmented fibres. Intense orceinophilia was again noted. Although the factors responsible for the known fluorescence of elastic tissue are still poorly understood, it is of interest that all the above-described types of fibres exhibited yellow autofluorescence in paraffin sections. This was also seen after hematoxylin and eosin staining and was enhanced by acridine orange staining. The least fluorescence was present in the wide non-branching fibres. The rest of the fibres fluoresced as intensely as did sections of newborn pulmonary artery used as controls (Fig. 7). The appearance of the orceinophilic fibres and globular masses was found by electron microscopy to be quite variable. Occasional elastic fibres were electron-lucent, but by and large the majority appeared composed of an amorphous electron-dense substance. An interesting pattern of development is suggested by the changes present in Fig. 8. In the centre of a longitudinally sectioned fibre is an electron-lucent core, with only a faint suggestion of fibrillar structural organization; this is the appear¬ ance of "young elastin". Surrounding this central core is an electron-dense substance that adds great bulk to the fibre. The irregular edges suggest growth by accretion to the surface of larger fibres. The structural organization of the dark substance was difficult to discern in view of the variability of the "staining" of the fibres. However, collagen fibres were included in this substance. No "protofibrils" or microfibrils were clearly identified, the predominant pattern being amorphous. Two types of alteration were observed in the collagen bundles: (1) decrease in their ability to show in the electron-micrographs, so that they appeared pale and devoid of striations; (2) the appearance of a dense amorphous or finely granular substance which was deposited between collagen fibres and bundles, merged with them and ulrimately became homogeneous. At this stage the fibres were no longer recognizable as collagen, except for a few isolated fibres which retained their cross-banding. As these masses enlarged, relatively well preserved collagen fibres and even entire bundles were "engulfed" or incorporated into the dense masses. Areas of ground substance were also incorporated, adding to the bulk of the mass.
The margin of a small globular orceinophilic mass is shown in Fig. 9. There is a "stippled" appearance in the centre, presumably due to the microfibrillary component of the elastic tissue. This appearance is commonly present in elastic structures of newborn infants. A dense amorphous coating containing altered collagen fibres surrounds this clear zone.

Comment
Proliferative lesions of the connective tissue present considerable difficulty in interpretation, Furthermore, reviews of tumorous conditions of the soft somatic tissues generally do not include elastofibroma dorsi;5-6 this may be due to its rarity and relatively recent identification, since the lesion was first described in 1961 by Jarvi and Sax£n,7 and few cases have been reported since.
The available descriptions in the literature and the morphologic characteristics of the cases that we have examined suggest to us that the lesion is a "reactive fibromatosis'' rather than a true tumour. Fibromatosis has been defined as "a benign fibroblastic overgrowth of an infiltrative rather than expansile nature pursuing a benign course except when vital structures are affected".8 Much confusion has arisen from the often mentioned tendency to recur; such tend¬ ency varies with the different types of fibromatoses, and results, in the vast majority of cases, from incomplete surgical excision of a poorly delimited lesion. Elastofibroma dorsi shares common morphologic and clinical charac¬ teristics with other fibromatoses but has a uniformly benign prognosis.
A recent attempt to classify benign connec¬ tive tissue proliferations is that of Bonenfant.9 In this schema desmoid tumours and congenital fibrosis of the sternocleidomastoid muscle are grouped under the designation of benign connec¬ tive tissue proliferations of deep sclerosing type. Under the same general designation ("conjonctivoses b&rignes proliferes") is distinguished a deep fibromatous type, which comprises such heterogeneous lesions as palmar and plantar fibromatosis, Peyronie's disease, juvenile aponeu-rotic fibromatosis, mesenteric, periosteal and tendinous fibromatosis. Lesions exhibiting an im¬ portant inflammatory component such as retro¬ peritoneal fibrosis and pseudosarcomatous fascitis are classified under a separate "granulomatous" category. The predominant histologic pattern of elastofibroma dorsi is that of ramifying bundles of collagenous tissue, with little or no tendency to the formation of whorls. This appearance is in keeping with the general character of fibromatoses. In many cases the abundance of the fibrillary component and the scarcity of cells would seem to indicate the "sclerosing" nature of the process. However, the degree of cellu¬ larity may vary from case to case, or in different areas of the same lesion, as mentioned above. Furthermore, its anatomie origin in close rela¬ tionship with musculoaponeurotic structures, the well-documented tendency to bilaterality10 and the lack of circurnscription are features in which it resembles the "fibromatoses". Until now only one case has been reported as occurring outside the shoulder area. This is Barr's1 case originating in the trochanteric region, and is included in the present study; no distinctive morphologic charac¬ teristics were seen in this case. The clinical manifestations are purely local, related to the physical presence of the mass. Although defini¬ tive statements will have to await descriptions of a larger number of cases than is now avail¬ able, there is nothing to indicate a malignant potential. What appears to set this lesion apart, clearly separable from other fibromatoses, is a great abundance of elastic tissue. No morphologic counterpart is present in the other closely allied fibrous tissue growths, even those subject to frequent trauma or mechanical deformity. Specifically, a search for similar fibres in cases of palmar and plantar fibromatosis was un¬ successful.1 The usefulness of this morphologic feature as a major diagnostic point for the pathologist can¬ not be overemphasized. In a critical review of the problem of fibromatosis affecting the shoulder area, Enzinger and Shiraki11 found evidence of local recurrence in 57% of 30 cases followed up for 10 years or longer. This aggres¬ sive clinical course is in sharp contrast to the benign nature of elastofibroma. Enzinger and Shiraki11 also emphasized the lack of correla¬ tion between histology and clinical behaviour in musculo-aponeurotic fibromatoses of the shoulder. Elastofibroma exhibits a distinctive histologic architecture, the hallmark of which is the unusually prominent elastotic component.
Fibromatoses of the shoulder tend to be located anteriorly, frequently involving the supraclavicu-lar fossa or the axillary space. There is also a high incidence in young subjects.11 Characteristically, elastofibroma arises posteriorly in the subscapular space adjacent to the scapular apex.
All of the patients so far described have been more than 45 years of age.
Whether or not elastofibroma is to be inter¬ preted as a modality of fibromatosis or desmoid tumour is a problem that further uncovers the limitations of current nosologic classifications of fibrous tissue proliferations. In view of the pre¬ ceding considerations it appears necessary to retain the present nomenclature. The term elastofibroma is non-committal as to pathogene¬ sis, and identifies the lesion under discussion.
Major differences in prognostic implications emphasize the importance of distinguishing elastofibroma from other fibromatoses, but the differential diagnosis should not be difficult in the majority of cases.
There is general agreement that elastic fibres never develop in the absence of collagen fibres, the latter appearing to be a prerequisite for the formation of the former.12 It is currently be¬ lieved that the same cell responsible for the formation of collagen fibres also causes the pro¬ duction of elastic fibres. The second possibility, that under special circumstances elastic fibres may derive from preformed collagen, did not receive much support until after the histochemical studies of Gillman et al.ls It had been known for over 60 years that in senile elastosis of the skin, the apparent increase in elastic fibres occurs at the expense of collagen fibres. Atypical elastic fibres in senile elastosis were called "ellacin" by Unna.14 Using a laborious methodology, Keech15 demonstrated the in vitro transformation of collagen to elastic-like struc¬ tures. This could be observed by electron micro¬ scopy in samples of skin of patients with various diseases involving the connective tissue. There is disagreement regarding the occurrence of this phenomenon, "elastotic degeneration of the collagen", in various pathologic situations. The problem is compounded by the well-known lack of specificity of standard tinctorial methods for elastin, incomplete knowledge of the biochemical structure of the fibres and the realization that there are species differences,16 age differ¬ ences17, 18 and even regional differences in elastic tissue composition from different ana¬ tomie sites. 16 Concerning the connective tissue change in elastofibroma, our morphologic find¬ ings favour a close interdependence between collagen and elastic fibres. That there is a breakdown of pre-existing elastic fibres is immediately apparent from the ordinary histologic prepara¬ tions. Excessive wrinkling and distortion are frequently seen, suggesting breakage followed by a recoil phenomenon.
Light-microscopy examination of isolated fibres after NaOH extraction revealed transitional stages between "typical" elastic fibres and swollen fibres with serrated edges. It was also possible to examine altered individual collagen bundles. Progressive degradation appeared to be accompanied by increasing admixture with elastotic material. Incorporation of collagen fibres into irregular elastotic masses was also present at the ultrastructural level. Similarly, the globular masses of elastotic material closely resemble the more ad¬ vanced stages of elastase-digestion. Previous re¬ ports on elastofibroma have generally failed to emphasize these similarities. However, chemical studies are needed before the specific role of any enzyme can be ascertained in the patho¬ genesis of elastofibroma, or other pathologic processes affecting the elastic tissue.^E lastofibroma dorsi is a distinct clibummary nico_path0logic entity probably be¬ longing to the category of the fibromatoses. From a practical standpoint it would be helpful to retain the present designation. Individualization of this lesion is imperative in view of the serious prognostic and therapeutic implications carried by other closely allied fibrotic growths of the soft somatic tissues. Although the lesion is uncommon and most reports describe isolated cases,2* 3» 10» 24 its clinical and pathologic features are distinctive enough to preclude diagnostic difficulties once the surgeon and the pathologist are aware of the existence of this process.
The unique patterns of reaction present in cases of elastofibroma provide an opportunity to study the relationship between collagen and elastic fibres.
Our observations indicate a close admixture of components of the two types of fibre. There are definite differences between the morphology of these lesions and other diseases involving degrada¬ tion of elastin, such as pseudoxanthoma elasticum. Whether or not these morphologic differences signify different pathogenesis remains unanswered. A hypothetical enzymatic flaw interfering with the aggregation of building materials for elastin and collagen or impairing the integrity of formed fibres merits consideration. This hypothetical deficit would reside in the ground substance at a local level. The resultant disaggregation and lack of coherence of the fibres would occur at such sites in the molecule that no loss of fluorescence and no inereased cal¬ cium salts-binding would ensue. Finally, the known close kinship of the "building blocks" or formative material of elastin and collagen25 would permit the close physical mixing observed here. An additional question that remains unsolved is the nature of the initial insult giving rise to the fibroelastic prolifera¬ tion.