Birt-Hogg-Dube Syndrome

Primary bars above (black) and secondary bars below (spheres).

Note:The Section Directory (below) provides access to other Sections on this site.


Case 5 of the Ochsner Clinicopathologic Skin Conference
presented and discussed by R.J. Reed, M.D.
First published on the late Bill Weemsís web site:
For additional information regarding the structure of most of the sections, see the Site Diagram in Section 4


Diagnosis: Branching or arborizing, trichogenic, compound stromal and epithelial (catagen-like, anagen-like, and mantle-like patterns) hamartoma (Trentís follicular hamartoma)



The stromal component of this unusual lesion is modified perifollicular connective tissue. The hair papilla is a special modification of the same mesenchyme. A natural representation of such mesenchyme might be found in a site in which a hair bulb has undergone cytolysis and the immediately adjacent tissue, including the papilla and the connective tissue sheath, has responded by becoming sclerotic (Fig 10).

At least a part of the lesion in question has features of a fibrofolliculoma as seen in the Birt-Hogg-Dube syndrome. Patterns of a characteristic example of the papular and polypoid lesions of the follicular hamartoma syndrome are represented in figs 11- 16. In Fig 11, three follicles are represented in cross section. The red arrow points to a follicle showing the pattern of fibrofolliculoma with mantle-like patterns and perifollicular fibrosis. The green arrows identify follicles that are not significantly distorted in the plane of the section but there is perifollicular fibrosis (perifolliculoma pattern). A sebaceous gland is associated with the follicle on the right. The lesion in fig 12 is slightly polypoid. The pattern is mantle-like with perifollicular fibrosis and sebaceous gland lobules (fibrofolliculoma pattern). The lesion in fig 13 is a plaque of fibrous tissue that presses upon the epidermis and is parallel to the surface of the skin. To the right there is nodular protrusion into the dermis that might prove to be a perifollicular component on additional section. A follicle is represented eccentrically to the left and shows perifollicular fibrosis. The combination of features in this lesion satisfy the criteria for the diagnosis of trichodiscoma. Sebaceous glands are associated with the follicle on the left. In fig 14, the plaque-like component is shown at higher magnification. The fibrous tissue is laminated and in areas the lamellae are perpendicular to the skin surface. Blue arrows outline a vessel showing incomplete concentric fibrous lamellae. Rete ridges are effaced. In fig 15, the interface between the epidermis and the dermis is straight. There are vacuolar changes at the dermal-epidermal interface. Basal clear cells protrude into the dermis in this area. The pattern of the fibrous tissue in this area is characteristic but it would not be possible to exclude a fibrous papule on the basis of the pattern in this field. In fig 16, vessels are dilated, clustered, and outlined by concentrically arranged fibrous lamellae. These vascular alterations may also be seen in fibrous papules. Clearly, fibrous papule and trichodiscoma are hamartomas of the same order: they are hamartomas of the adventitial dermis.

The lesion in this case is an organoid, polypoid tumor showing follicular and perifollicular components (Fig 1). The perifollicular component of specialized connective tissue sets the lesion apart from trichofolliculoma. In addition, the branching follicles are represented in a spectrum of patterns which more or less correspond to the phases of a follicular cycle (Figs 3-9); the patterns of follicular development are not synchronous. The distortion of follicular patterns sets the lesion apart from perifollicular fibroma. Fibrofolliculoma was originally a combined fibrous and follicular lesion showing arborizing thin cords of small squamous (squamoid) cells (note: sometimes characterized as basaloid cells) in the region of the infundibulum of a solitary follicle. The definition has been extended to encompass the variable patterns encountered in the Birt-Hogg-Dube syndrome in which the lesions are multiple. Clearly, the patterns are variable but the terminology of the solitary (presumably prototypic) lesions has been utilized to characterize the mixed patterns of the multiple variants.

Some of the solitary lesions in this category of follicular hamartomas are plaque-like and multi-follicular. For such lesions the stromal component often is fibromyxoid. In a single plaque-like lesion, patterns of fibrofolliculoma (in the broad definition), trichodiscoma, and acrochordon may be represented. It is difficult to pigeon-hole the variable patterns in the plaque-like and multiple follicular hamartomas to accommodate preconceived notions of what the respective lesions are supposed to show histologically.

If virtual images of a continuum are to be manipulated, segments are required (and the phases of follicular development are arbitrary segments of a biologic and morphologic continuum) . We have taken the continuum of follicular growth and segmented it by assigning labels to commonly encountered, extreme patterns. With the images of taxonomically defined segments, the histologic patterns of the recurring cycles of normal follicles are recapitulated in the epithelial components of follicular hamartomas. The options include anagen, catagen, and telogen patterns. In addition some of the patterns can be characterized as mantle-like (Figs 1,3,7 & 9). The mantle-like patterns characterize the epithelial patterns of classic fibrofolliculoma.

If we examine the lesion in Ochsner case 5 (Trentís case), the basic quality is a distorted root from which follicles or epithelial strands project in branching or arborizing patterns; the hamartoma is a branching or arborizing variant. Clearly, some of the branches are recognizable as differentiating follicles in anagen-like patterns. For these branches, there is a hair bulb at their extremities, and inner and outer root sheaths are recognizable. In addition, there are small hair shafts that are delivered to the dilated lumen of the root (Fig 2). The root is attached to the epidermis.

The presence of small hair shafts in the lumens of some of the small branching follicles qualifies the lesion as a trichogenic variant. This quality of trichogenesis was even evident clinically. Trichogenesis in a lesion showing branching patterns was not emphasized in the original descriptions of fibrofolliculoma but is characteristic of trichofolliculoma (a branching, but purely epithelial hamartoma).

Mantle-like qualities initially were emphasized in the definition of fibrofolliculoma, although a definition of the mantle post-dates that of fibrofolliculoma. In turn, the concept of fibrofolliculoma antedates that of the Birt-Hogg-Dube syndrome. The conceptual rigidity of the original designation had to be softened to accommodate the varied patterns of the Birt-Hogg-Dube syndrome.

The follicular hamartomas can be divided into those which are pure epithelial distortions and those which are both epithelial and stromal (therefore compound) distortions. The stroma of a follicle is the perifollicular connective tissue sheath. The perifollicular sheath is an special adaptation of the adventitial dermis. In the compound follicular hamartomas, both epithelial and stromal distortions are represented. Are the two components equally expressive of a genetic derangement or is one primary and the other a response?----- Fibrofolliculoma is a organoid hamartoma.

Trichodiscoma is an eccentric, non-branching follicular hamartoma.. In it, the stromal response greatly overshadows the epithelial contribution. In some examples, the stroma of a trichodiscoma is myxoid ref. 2. Myxoid tissue might be characterized as the activated phase of fibrous tissue. From this approach, it is tempting to propose that the stromal component is primary and that the stromal component has an inductive influence on the differentiation of the epithelial component; one follows the other.

It might be tempting, in our tendency to segment continua and to crystallize patterns, to propose that the follicles of a follicular hamartoma are fixed in a particular stage of a cycle. They might all show anagen-like patterns and might not have the capacity to cycle. In Trentís case, some of the patterns at the extremities of some of the branches have catagen qualities. Some have anagen qualities and some have mantle-like qualities. One striking aspect of the case (Ochsner case 5; Trentís case) in question is the recapitulation of a variety of follicular patterns within a single lesion. Not all of the branching follicles are synchronized in patterns that recapitulate a single phase of a follicular cycle. In Trentís case, there is indisputable evidence that the branches are cycling and that the cycles among the follicles are not synchronous.

Some examples of the Birt-Hogg-Dube syndrome are characterized by multiple papular lesions, occasionally prominent on the butterfly area of the face (a site of predilection for follicular hamartomas). This anatomic favoritism is also evidenced in the distribution of adenoma sebaceum of the tuberous sclerosis complex and the respective fibrous papule (melanocytic angiofibroma).

Some of the lesions of the follicular hamartoma syndromes are also characterized by concentric fibrosis extending along vessels away from the main lesion. This pattern of perivascular fibrosis is similar to perivascular changes which may be encountered in fibrous papules. The perivascular connective tissue might be characterized as an extension of the adventitial dermis.

The follicles of fibrous papules and adenoma sebaceum are often small and not associated with significant sebaceous glands. Perifollicular and concentric perivascular fibrosis are common features. On the basis of the poor definition of normal boundaries in the zone of fibrosis at the papillary dermal-reticular dermal interface, it is difficult to characterize the fibrous papule as a hamartoma of the adventitial dermis; both the papillary dermis and the reticular dermis appear to be involved. On the other hand, the common perifollicular component of a fibrous papule is a clear marker for a contribution by a specialized component of the adventitial dermis.

A dendritic fibrocyte is characteristic of the fibrous component of a fibrous papule. In occasional lesions of the Birt-Hogg-Dube syndrome, some of the dermal changes of a fibrous papule or sebaceous adenoma are encountered among the involved follicles. All the follicular and perifollicular distortions of lesions in the Birt-Hogg-Dube syndrome appear to be distortions of one or more components of the adventitial dermis. They might be characterized as adventitial dermal, and epithelial hamartomas.

The lesion in Trentís case shows variations in expressions of phases of follicular differentiation. On at least one section of the tumor, the preponderant pattern was that of thin, interconnected cords of small squamous cells and a widened, fibrotic perifollicular sheath (the mantle patterns of a fibrofolliculoma). If this had been the only section examined the lesion would be easily characterized as a fibrofollicular hamartoma of a type in keeping with those emphasized in the Birt-Hogg-Dube syndrome. On the other hand, the clinical presentation of the lesion with visible hairs at the extremity of the lesion would be a feature against the interpretation that only fibrofolliculoma patterns were represented. Clearly, a portion of the lesion was trichogenic and the fibrofolliculoma component is not, by current custom, trichogenic.

The Hornstein-Knickenberg syndrome is clinically similar to the Birt-Hogg-Dube syndrome but is characterized by perifollicular fibromas. Perifollicular fibromas are characterized as pure hamartomas of the perifollicular connective sheaths to the exclusion of any significant distortions of the follicular component. The validity of distinctions between the Birt-Hogg-Dube syndrome and the Hornstein-Knickenberg syndrome recently have been questioned ref. 1. From the perspective of a cyclic continuum in the differentiation of stromal-follicular hamartomas, it may be appropriate to move away from the classical characterizations and adopt a more inclusive view.

Certainly, the follicular-stromal hamartomas can be characterized by giving attention to the predominance of patterns corresponding to those of normal follicular cycles. Those showing resting (mantle-like) patterns would then more or less correspond to fibrofolliculomas. Those without significant mantle-like patterns are likely to be characterized as perifolliculomas. Those in which the inductive influences are weak are likely to be disordered overgrowths of stroma with only rare, abortive follicles; they are the counterpart of trichodiscoma. The fibrous papule is mature (differentiated) and is compound in the sense of involvement of two components of the adventitial dermis; both the papillary dermis and the perifollicular connective tissue sheaths contribute.

In our approach to compound follicular hamartomas, a predominance of anagen-like patterns at the extremities of its epithelial branches might be characterized as fibrotrichofolliculloma. Surely, it is simply a matter of time before such a characterization will be added to the list of follicular hamartomas. Currently, there are too many designations for the stromal-follicular hamartomas.

Our designations should emphasize the anatomic compartments which are represented in hamartomas. I am not sure that the presence or absence of sebaceous gland lobules is a distinguishing characteristics requiring qualifications in the application of designations in the stromal-epithelial category (with the exception of lesions that are strikingly sebaceous in character). On the other hand, in the pure epithelial dysplasias, degree of differentiation in sebaceous components is basic to prognostic evaluations. If some follicular hamartomas are associated with apocrine glands as they are in the pure epithelial dysplasias of nevus sebaceus, then certainly the contribution should be recognized in assigning a designation.



I. Epithelial variants (exclusive of the basal cell dysplasias, follicular abnormalities of cellular blue nevi, and perifollicular myxomas with follicular abnormalities {so-called trichogenic myxomas; i.e., Carneyís syndrome})

A*. nevus verrucosus (follicular potentials expressed in striking basal cell hyperplasia; i.e., accentuation of the basal layer of the epidermis)
B*. the organoid nevi, including nevus sebaceus of Jadassohn (a complex dysplasia expressed in degrees of maturity and a variety of phenotypes (including apocrine types), all of which are predisposed to neoplastic transformation, particularly in the immature examples. Many of the associated neoplasms are of apocrine type)ASIDE.
C*. trichofolliculoma and giant pore (The former is defined in a manner that gives recognition to trichogenic and branching patterns. The latter is defined in a manner that gives recognition to poorly developed, branching, mantle-like patterns. Could one merely represent the cyclic variant of the other? Probably not. The clinical course of trichofolliculoma might suggest that the branching follicles are neoplastically committed to permanent trichogenesis. Is this possible without an opportunity for periodic renewal?--- Such a process would require a most efficient, continuous apoptotosis). Pilar sheath acanthoma might be characterized as the follicular counterpart of dermal duct tumor (nests of cells sit among collagen bundles of the reticular dermis without a significant stromal component).

 II. Stromal-branching (arborizing) epithelial variants (a category which finds diversity in presumed relationships between minor histologic variations and biologic associations. The Birt-Hogg-Dube syndrome and the Hornstein-Knickenberg syndrome are examples.

A. Organoid variants
1. arborizing epithelium in anagen patterns (trichogenic)
a. solitary and polypoid or papular
b. solitary and plaque-like (a mixture of epithelial and stromal patterns including those of so-called trichodiscoma, fibrofolliculoma, and perifollicular fibroma)
c. multiple and papular (?)
2. arborizing epithelium in catagen patterns (trichogenic; small hair shafts often represented in some of the epithelial islands.)
a. solitary and polypoid or papular
b. solitary and plaque-like
c. multiple and papular (?)
3. arborizing epithelium in mantle-like patterns (branching cords of cells in an expanded, specialized matrix in the pattern of fibrofolliculoma*)
a. solitary and polypoid or papular
b. solitary and plaque-like
c. multiple and papular (so-called Birt-Hogg-Dube syndrome. Sebaceous gland representation of questionable significance in establishing distinctions between fibrofolliculoma* and perifollicular fibroma*)

 III. Stromal-non-branching (non-arborizing) epithelial hamartomas

A. perifollicular fibroma* (a non-branching follicular component in an expanded specialized perifollicular stroma. In the common definition, the follicular component lacks a sebaceous component, a feature of questionable significance [1]).
1. solitary and polypoid or papular
2. solitary and plaque-like (?)
3. multiple papules (the Hornstein-Knickenberg syndrome; see ref. 1)
B. Compound variant (adventitial fibroma: a compound lesion involving both the papillary dermis and the perifollicular connective tissue: follicular representation variable)
1. solitary and polypoid or papular (fibrous papule)
2. solitary and plaque-like (shagreen patch; this lesion affects the reticular dermis and provides evidence of dysfunction affecting more than the adventitial dermis).
3. multiple papules (adenoma sebaceum: The epithelial component of the follicles within the perifollicular connective tissue component generally does not arborize. The follicles generally are small and may not be associated with sebaceous components. Hyperplastic sebaceous glands may be encountered in fibrous papules.
C. Simplex variant (stromal preponderance of perifollicular origin: follicular representation sparse)
1. solitary and polypoid or papular (?)
2. solitary and plaque-like (trichodiscoma*: For some myxoid lesions, follicular changes in the pattern of small hair bulbs attached to the overlying epidermis are a feature. The myxoid lesions are likely to be grouped with the myxomas rather than a myxoid variant of trichodiscoma.* In the category of multiple lesions, most examples are, and are likely to be dismissed as, examples of trichodiscoma;* they are fibrous lesions. In practice, many such lesions seem to be nothing more that stroma-rich variants of the so-called fibrofolliculoma* or perifollicular fibroma*).
3. multiple papules (multiple trichodiscomas*)

 An asterisk marks the accepted variants This classification is anticipatory and there are too many question marks. Part of the problem in attempting to relate the current classification with the classification proposed herein is that the cyclic variations of follicles, even dysplastic follicles, have been ignored in the common characterizations.

The patterns observed in random sections of branching follicular hamartomas have more to do with the preponderant phase of the component of branching follicles than with neoplastic determinants.


1. Schulz T, Hartschuh W: Birt-Hogg-Dube syndrome and Hornstein-Knickenberg syndrome are the same. Different sectioning technique as the cause of different histology. J Cutan Pathol 1999; 26:55-61.

2. Pinkus H : Trichodiscomas . J Invest Dermatol 1974;63:21-218.

The 3 clusters to the left are composed of secondary bars: Access to other sections or web site.     Primary bars to the right: Chapters of  section 4.

 Next page             Back a page               Up a page              Down a page