Until the 1950s, the particular type of carcinoma that commands our attention here had not been identified with specificity, but was published as an example of squamous-cell carcinoma that developed in a pre-existing sebaceous cyst. Seff and Berkowitz have been given credit, consistently, for being the first to report, in their article in 1916 titled "Carcinomatous degeneration of sebaceous cysts," on a squamous-cell carcinoma that supposedly had arisen in a cyst and subsequently metastasized widely.¹ In our estimation, however, the description of histopathologic findings and illustrations of the findings by Seff and Berkowitz are not those of proliferating tricholemmal cystic carcinoma, but of sebaceous carcinoma. The first convincing examples of proliferating tricholemmal cystic carcinoma were presented by Caylor in 1925 under the heading "Epitheliomas arising in a sebaceous cyst."² Caylor described and pictured a cystic tumor of the right cheek that resulted in death from metastases of a 71-year-old woman (his Case XII). That neoplasm persisted at the local site following excision of it partially, as was the case for several other neoplasms of that particular type in his series of 12 patients. It is likely that all of those neoplasms were examples of proliferating tricholemmal cystic carcinoma. Caylor concluded spuriously that "all sebaceous cysts should be removed, for they may be the site of a malignant tumor."

Bishop, in his series of 11 patients with "epidermoid carcinoma in sebaceous cysts" published in 1931,³ and Peden, in his communication about "carcinoma developing in sebaceous cysts" reported on in 1948,⁴ also included examples of what seem to be proliferating tricholemmal cystic carcinoma that they, like Caylor,² believed represented "squamous-cell carcinoma arising in a sebaceous cyst." Collins, in his review in 1936 of "Carcinoma originating in sebaceous cysts," emphasized the "malignant potential" of those cysts and advised, too, that "all sebaceous cysts should be considered as pre-cancerous lesions, and they should be closely observed if the individual does not consent to their surgical removal."⁵

It was not until 1957, however, that Lund, in a fascicle that came from the Armed Forces Institute of Pathology, put forth a notion that persists to this day, to wit, that the condition under discussion here is a benign neoplasm for which he suggested the name "benign dermal acanthoma (subepidermal cystic acanthoma)."⁶ His concept of the condition was expressed in these sentences: "The solid nodules are sometimes classed as carcinoma arising in epidermoid cysts, but it is doubtful that they are truly carcinomas in the sense of metastasizing tumors. They are more accurately designated as benign dermal acanthomas, or, as Helwig suggests, acanthosis in epidermoid cysts. I feel that they arise in a manner similar to epidermoid cysts, but may be solid tumors from the start." In the legend to an illustration of benign dermal acanthomas (his Fig. 5), Lund stated that "They may become very large, but do not metastasize."

In 1958, Welch told of his experience with sections of tissue from specimens of 239 cysts that had been excised.⁷ He set forth criteria for separating sebaceous cysts from epidermoid cysts and he averred that each of those two types of cyst had a benign and a malignant counterpart. Of the 239 cysts identified by Welch, 134 were thought by him to be "sebaceous" and, of those, 128 were benign and 6 malignant. One hundred and six epidermoid cysts also were recognized by him, 103 of which were benign and 3 malignant. Figure 2 in his article illustrates an example of what he conceived to be a carcinoma within a sebaceous cyst, a neoplasm that we regard to be a proliferating tricholemmal cystic carcinoma.

In 1966, Edward Wilson Jones described, in detail, the clinical and histopathologic findings in a lesion he called "proliferating epidermoid cyst"⁸ and that he encountered in nine patients. These were his conclusions:

"This relatively uncommon benign tumor of squamous epithelium, which clinically resembles a keratinous or sebaceous cyst, usually occurs on or close to the scalp of elderly women . . . The name of proliferating epidermoid cyst is suggested because of the character of the epithelium and because at least some of these tumors originate in skin cysts. This term is preferred to proliferating sebaceous cyst in order to distinguish it from sebaceous gland tumors . . . Proliferating epidermoid cyst has not been well described before as a distinctive tumor. However, it is very likely that many examples of this tumor have been reported in the literature as either squamous-cell epitheliomas arising in cysts or as Malherbe's calcifying epithelioma . . . The presence of a benign dyskeratosis in this tumor can easily lead to confusion with carcinoma, but further progression into a true malignant epithelioma remains a possibility, as suggested by Nurnberger."

What Jones pictured as a proliferating epidermoid cyst in his figures 1, 4, 6, and 7 is a proliferating tricholemmal cystic carcinoma, fulfilling as it does all of the criteria for that malignant neoplasm. Numerous dyskeratotic cells are an expected finding in several types of cutaneous carcinomas, from keratoacanthomatous squamous-cell carcinoma to matrical carcinoma, but Jones interpreted those cells as representing "benign dyskeratosis." The concept of "benign" and "malignant" dyskeratosis was popular in circles in dermatopathology in the middle of the twentieth century but, having been discredited, now is passé.

Also in 1966, Reed and Lamar reported on 14 patients who had a "peculiar, keratinizing invasive epithelial growth occurring on the scalp."⁹ They called the lesion "invasive pilomatricoma" or "invasive hair matrix tumor of the scalp" and remarked that "the pattern of growth in our tumors is reminiscent of that seen in pilomatricoma or calcifying epithelioma of Malherbe, and the term invasive pilomatricoma is proposed for the tumors of the scalp that we are reporting." In our judgment, some of their photomicrographs (their Figs. 3, 4, 5, and 12) show proliferating tricholemmal cystic carcinoma. This, however, was the point of view of those coworkers: "Although several of our tumors were large and deeply invasive, metastases have not occurred and all have been controlled by local excision or irradiation . . . In the absence of documented evidence of metastasis from tumors of the type reported in this series, there is no justification for classifying them as epidermoid carcinoma."

In 1968, in an article headed "Tumors of lower hair sheath: Common histogenesis of certain so-called 'sebaceous cysts,' acanthomas and 'sebaceous carcinomas,'" Holmes presented seven examples of a malignant neoplasm she termed "trichochlamydocarcinoma."¹⁰ Because tricho means "hair" and chlamydo means "cloak" (both derived from Greek), trichochlamydocyst was deemed rightly by Holmes to be a "hair sheath cyst." By extension, the term trichochlamydocarcinoma must mean "hair sheath carcinoma," and Holmes, in fact, thought that those neoplasms arose from the lower part of the outer sheath. This is what she said about them:

"Trichochlamydocarcinoma is the term proposed for those solid tumors of the scalp of hair sheath origin which clinically have the appearance of wens until late in the course of the disease . . . Jones,⁸ in 1966, apparently was the first to recognize that all of his proliferating epidermoid cysts (trichochlamydocarcinomas) were derived from a sebaceous type of epithelium and that they usually occurred on the scalp of elderly women. He reported nine cases of the less aggressive type of trichochlamydocarcinoma. It is the purpose of the present report to present arguments for the lower outer hair sheath histogenesis of these solid tumors of the scalp and to call attention, although most are indolent carcinomas, to the existence of the more aggressive types with the potential of metastasis."

Some of Holmes' photomicrographs of trichochlamydocarcinomas, in particular her figure 8, exhibit what we interpret to be proliferating tricholemmal cystic carcinoma; she and we agree, in large measure, about the essential character of the neoplasm. Case 7 in her series was a tumor on the scalp of a 65-year-old woman that, following surgical procedures, persisted twice at the local site, eroded the outer table of the skull, and metastasized to at least a pre-auricular lymph node. Holmes understood full well that trichochlamydocarcinomas not only were follicular in nature, but that they were malignant neoplasms, about which she commented thus:

"Certainly metastasis is rare in trichochlamydocarcinoma . . . It may be tempting to draw a dividing line between those solid tumors with, and those without, cellular anaplasia; however, of the present series case 4 without anaplasia recurred and case 5 with anaplasia did not. Furthermore, the huge tumor of case 6 with the greatest anaplasia of all did not metastasize as did the same sized lesion of case 7 having less anaplasia . . . The chance of metastasis may not be very great but the possibility does exist."

In 1971, Dabska shared her experience with 12 specimens of what she called "giant hair matrix tumor."¹¹ Many of the tumors were unusually large, ranging in size from 3.5 to 25 cm in diameter. This is what she said about them: "They have clear differentiation toward pilosebaceous structures especially to the outer sheaths of hair follicle. Predilection to the skin of scalp and nucha in our cases, great sizes of tumors (up to 25 cm), and long terms of duration—all distinguish the 'giant hair matrix tumors.'" Dabska believed the neoplasms to be "benign but locally invasive . . . [and] not [to] metastasize but may recur if not completely excised." That she failed to recognize the malignant nature of the tumors was conveyed by this terse statement of hers: "The diagnosis of epidermoid carcinoma is not justified on the basis of available evidence." For us, the giant hair matrix tumor of Dabska is proliferating tricholemmal cystic carcinoma.

Although difficulty in distinguishing proliferating tricholemmal cyst from a type of carcinoma has been mentioned episodically,¹² except for Holmes, who got it right, the truly malignant nature of that so-called cyst was not appreciated until relatively recently. For example, Connors and Ackerman, in 1976, like nearly all other authors until that time, considered proliferating tricholemmal cyst to be a benign neoplasm.¹³ In an article published that year titled "Histologic pseudomalignancies of the skin," they wrote about proliferating tricholemmal cyst as follows: "The benign nature that these microscopic features suggest is corroborated by the gross and clinical features . . . This lesion may be erroneously interpreted as a squamous-cell carcinoma arising in a cyst."

In 1979, Connors and Ackerman, in a chapter about "Pseudomalignancies" in Cancer Dermatology, addressed the matter of proliferating tricholemmal cyst in these words: "The microscopic features that most clearly distinguish a proliferating tricholemmal cyst from a squamous-cell carcinoma are evident origin from an antecedent quiescent cyst and sharp circumscription of both the total lesion and its component cellularity."¹⁴ Figure 22–5 from the second publication by Connors and Ackerman shows a proliferating tricholemmal cystic carcinoma, despite a legend that reads: "Proliferating tricholemmal cyst. At first glance, this lesion resembles squamous-cell carcinoma. However, upper portion is not the epidermis, but rather the margin of this unusual type of follicular cyst. Note also that aggregates of atypical keratinocytes appear to advance by pushing, rather than by infiltrating stroma." In 1980, Janitz and Wiedersberg recorded the results of their study of 16 examples of "tricholemmal pilar tumor," a neoplasm that they thought "arose from" the outer root sheath of the hair follicle.¹⁵ We interpret tricholemmal pilar tumor to be proliferating tricholemmal cystic carcinoma.

In 1993, Ackerman and colleagues, in the first edition of Neoplasms with Follicular Differentiation, stated that "fundamentally, it ["proliferating tricholemmal cyst"] is a neoplasm, not a cyst, and that distinction has important biologic implications."¹⁶ That assessment prompted the authors to propose that it be designated "proliferating follicular cystic neoplasm." They acknowledged how difficult it could be to distinguish the neoplasm in question from squamous-cell carcinoma and even raised the possibility that some examples of proliferating follicular cystic neoplasm might, indeed, be squamous-cell carcinoma. This is how they said it:

"This then is the conundrum. Are seemingly authentic proliferating follicular cystic neoplasms with striking cytologic atypia and many mitotic figures truly benign proliferating cystic neoplasms; may they transform into squamous-cell carcinoma, or are some of them, from the outset, true squamous-cell carcinomas that only simulate the benign cystic neoplasm? The answer is not known with certainty, but our current position is that true proliferating follicular cystic neoplasms have the silhouette of a benign neoplasm and cytologic features that are consistent with benignancy, to wit, little or no nuclear atypia and few, if any, mitotic figures. Deposits of calcium in compactly arranged corneocytes often are copious. In contrast, there is a type of squamous-cell carcinoma whose architectural pattern resembles that of a proliferating follicular cystic neoplasm and that, on occasion, even originates in a pre-existing proliferating follicular cystic neoplasm. The carcinoma, however, is not enveloped completely by epithelial lining, and is poorly circumscribed in at least several foci; it is characterized by striking nuclear atypia and many mitotic figures, and calcium, if present at all, is scant. In sum, histopathologists should be aware of the pitfall of misdiagnosing a squamous-cell carcinoma with 'proliferating follicular cystic' features as an authentic proliferating follicular cystic neoplasm."

In an effort to avoid the pitfall of misdiagnosing a carcinoma as a proliferating follicular cystic neoplasm, Ackerman and associates diagnosed the proliferating tricholemmal cystic carcinoma illustrated in their figure 24–11 as squamous-cell carcinoma. They cautioned, in the legend to figure 24–11, that the carcinoma pictured "could be misinterpreted as a proliferating follicular cystic neoplasm;" in actuality, that neoplasm is a proliferating tricholemmal cystic carcinoma. Ackerman and coworkers believed that some of those neoplasms, if left in place, eventuated in squamous-cell carcinoma, but they were wrong about the carcinoma being squamous-cell in nature. This is what they wrote then:

"Some proliferating follicular cystic neoplasms, however, are exceedingly difficult to diagnose because, despite a silhouette that may imply benignancy, they are associated with prominent nuclear atypia, numerous mitotic figures, some of them abnormal, and many dyskeratotic cells. We have seen neoplasms, left in place because of a diagnosis of 'proliferating tricholemmal cyst' or one of its synonyms, recur locally as unquestionable squamous-cell carcinoma and, in some instances, metastasize."

That Ackerman and coworkers were not alone in being confused about the true character of proliferating tricholemmal cystic carcinoma is evident from the host of designations given to it, namely, "epidermoid carcinoma in sebaceous cysts,"² "subepidermal acanthoma,"⁶ "proliferating epidermoid cyst,"⁸ "trichochlamydocarcinoma,"¹⁰ "invasive hair matrix tumor of the scalp,"⁹ "giant hair matrix tumor,"¹¹ "proliferating pilar (tricholemmal) cyst,"¹⁷ "proliferating tricholemmal tumor,"¹⁸ and "proliferating follicular cystic neoplasm."¹⁶ Brownstein and Arluk,¹² in their article in 1981 titled "Proliferating tricholemmal cyst: A simulant of squamous cell carcinoma," agreed fundamentally with the findings of Connors and Ackerman^13,14 , but they stated that although proliferating tricholemmal cyst resembles squamous-cell carcinoma, it was benign.

In 1998, Mones and Ackerman, in an article published in Dermatopathology: Practical and Conceptual,¹⁹ and in 2001, Ackerman, Reddy, and Soyer, in a text given to the subject of Neoplasms with Follicular Differentiation,²⁰ asserted that proliferating tricholemmal cystic carcinoma really was a squamous-cell carcinoma of one type and not an authentic "follicular" neoplasm that shows tricholemmal (isthmic-catagen) differentiation. Moreover, they were equally assertive in regard to the origin of the carcinoma, to wit, it did not arise in a pre-existing isthmic-catagen cyst, but rather de novo. Those two sets of authors, Ackerman being the denominator in common of both, were wrong on both counts.