Of all examples of proliferating tricholemmal cystic carcinoma reported on under various names in the literature, eight of them have resulted in death.^{2,4,17,18,32–36} That being the case, the malignant nature of the neoplasm is incontrovertible. What follows is a précis of those eight neoplasms that eventuated in metastases with fatal outcome.

The patient of Caylor² who died of proliferating tricholemmal cystic carcinoma was described by him as follows:

Case XII–A woman, aged seventy-one, had had a cystic tumor of the right cheek for twenty years. Three years after it first appeared it was excised, but soon recurred . . . The tumor remained about 0.5 cm in diameter until three months before examination at the clinic, when it began to enlarge rapidly . . . A piece of tissue 3 by 3 by 2 cm was excised, and examination revealed squamous-cell epithelioma, grade 4, apparently arising in a sebaceous cyst. This woman died seven months after operation from carcinomatosis.

Peden's⁴ account of a patient who died consequent to the effects of proliferating tricholemmal cystic carcinoma read thus:

Case 2–A 79-year-old white male entered Barnes Hospital complaining of a recurrent lesion of the right ear removed by cautery six weeks previously . . . Histologic examination of the specimen showed a well-differentiated squamous-cell carcinoma originating in the wall of a sebaceous cyst . . . Seven months later the patient returned with an ulcerated, indurated mass below the right ear about 5 centimeters in diameter and fixed to the deeper structures. This lesion could be only incompletely removed due to fixation to the transverse processes of the cervical vertebrae. Microscopic examination of this specimen showed a squamous-cell carcinoma deep in the subcutaneous tissue. Lymphatic permeation was a prominent feature of the sections, particularly around nerves and one lymph node showed invasion by cancer.

Peden's patient died of "unresectable metastases."

Amaral and associates¹⁷ recorded the plight of a 52-year-old woman who had what they called a "proliferating pilar (tricholemmal) cyst." Her condition was described by them in these words:

" . . . an 8 cm mass involving the left inguinal region and proximal thigh . . . A resection of the lesion was performed but the surgeon reported that he left residual disease . . . The patient was well until December 1982, when she returned with severe dyspnea; a chest roentgenogram showed pleural effusion. A pleural biopsy specimen demonstrated anaplastic carcinoma. The patient died one month later . . . Autopsy disclosed metastatic tumor . . . involving the pleura, lungs, mediastinum, and liver."

The "proliferating pilar cyst" that killed this patient is pictured in their Fig. 8, and it fulfills criteria for proliferating tricholemmal cystic carcinoma. Amaral and coworkers stated in the legend to their figure 7, that there was "carcinomatous transformation" of a pre-existing cyst and that judgment may well have been correct.

In 1989, Arico and associates¹⁸ reported on a 42-year-old woman with a 2-cm tumor on the scalp that was thought to be a "proliferating tricholemmal tumor." The neoplasm was excised and no evidence of a pre-existing cyst could be detected by microscopy. This is what happened next:

"There was some recurrence in the same area of the scalp 6 months later when she was pregnant again, and this nodule was excised . . . In December 1983 there was a further recurrence, and by May 1985 the tumour was 3 cm in diameter. It was excised, and 4 months later an enlarged cervical lymph node on the left side of the neck was removed and found to contain tumour metastases. Further lymph nodes were removed from the left side of her neck, and three of these were found to contain epithelial metastases that showed evidence of tricholemmal keratinization . . . In December 1986 a further large tumour developed rapidly in the same area of her scalp, and on histology was that of PTT [proliferating tricholemmal tumor]."

Additional information about this patient was given by Noto and colleagues in 1997, to wit, that she had developed "regional metastases and many local relapses with a total of 8 surgical excisions and radiotherapy. In 1992, infiltration of the occipital periosteum was present due to a relapse involving the cervical area at the C1 level . . . [She] died of locoregional extension of the tumor in the cervical area, 10 years after nodal metastases, without any other distant spread."³²

Noto and coworkers also reviewed the literature pertinent to the subject and, based on that review and their own experience, came to the following conclusions:

"A benign and malignant form of proliferating tricholemmal cystic carcinoma are distinguished. The differential diagnosis between benign and malignant proliferating tricholemmal cystic carcinoma can be made on the basis of mitotic rate, cytological and architectural atypia, necrosis and stromal infiltration . . . However, cases with little or no cytological and architectural atypia may have an aggressive behaviour and vice versa . . . Concluding, we suggest that proliferating tricholemmal cystic carcinoma should always be considered as a low-grade squamous-cell carcinoma and an accurate follow-up is recommended in all cases, as its biological behavior seems to be actually, histologically unpredictable."

We agree with Noto and associates that a benign counterpart of proliferating tricholemmal cystic carcinoma exists, namely, proliferating tricholemmal cystic acanthoma, but it is exceedingly rare. Unlike Noto, et al., however, we do not consider proliferating tricholemmal cystic carcinoma to be a "low-grade" neoplasm, because it may result in metastases and death.

Another patient who died of a "proliferating tricholemmal cystic carcinoma" was recalled by Mori and colleagues³³ in these words:

"A 58 year-old Japanese woman had a bean-sized papule in the right occipital region for 10 years [and which was excised] . . . Within 3 months there was local recurrence of the lesion . . . Physical examination showed a 5 x 5 x 5 x 7 cm exophytic red-blue tumor. The surface of the tumor was ulcerated and hemorrhagic. A sessile tumor measuring 3 cm in diameter was present at some distance from the larger mass . . . Wide excision, with surgical margins of 5 cm from the clinical limits of the tumor, was performed. Two weeks after operation, the patient developed a dry, hacking cough and left pleuritic chest pain. Radiograph of the chest showed blunting of the left costophrenic angle and a metastatic nodule 2 to 3 cm in diameter in the left middle pleura . . . A pleural biopsy specimen was composed of pleomorphic sarcomatous tumor cells similar to those in the primary lesion on the scalp. Left cervical lymph nodes were also palpable . . . The patient's condition worsened rapidly, and she died within 5 months of the initial operation."

The proliferating tricholemmal cystic carcinoma that killed the patient of Mori and coworkers³³ was described by them as follows: "The case showed a partial transformation to spindle cell carcinoma, and transition zones between squamous epithelium and spindle cells were present . . . Histologically, the recurrent tumor was composed of only spindle-shaped tumor cells . . . Thus, proliferating tricholemmal cystic carcinoma seems to have malignant potential. To our knowledge, this is the first case of spindle cell carcinoma arising from proliferating tricholemmal cystic carcinoma." We, unlike Mori and associates, view the entire neoplasm as a squamous-cell carcinoma, which was the original diagnosis rendered in their patient, as indicated by Mori and colleagues in these words: "Although the histological appearance in this case was characteristic of proliferating tricholemmal cystic carcinoma, the presence of atypia in the epithelium led to a diagnosis of squamous-cell carcinoma."

Park and colleagues recounted the history of a 32-year-old Korean man who had a tumor on the scalp that recurred repeatedly at the local site.³⁴ The clinical course was detailed by them as follows:

"A hard, non-tender fixed mass, 4 cm in diameter, in the right side of the anterior chest wall and a palpable supraclavicular lymph node were detected. Computed tomographic scan of the chest showed a soft tissue mass in the right paratracheal area; other masses were invading the superior vena cava and encasing the right descending pulmonary artery. There were also multiple lymph-node enlargements and metastatic nodules in the lung. A metastatic soft-tissue mass was destroying the rib and right lower anterior chest wall. A biopsy specimen of the chest wall mass showed tricholemmal carcinoma . . . He later complained of pain in the thigh, which proved to be a pathological fracture of the femur seen as an osteolytic lesion radiographically, suggesting metastasis. Then the patient was lost to follow-up."

Park and associates noted that "the involved lymph nodes had been replaced almost completely by tumor cells with morphological characteristics similar to the primary tumor." The authors were impressed, too, by the fact that "Each time the lesions were excised, histologic specimens demonstrated a proliferating tricholemmal tumor with increasing nuclear atypia . . . Ki-67 positivity showed a significantly increasing tendency as the extent of the tumor advanced; and, for each specimen, the positivity was generally constant. Tricholemmal keratinization with cyst formation was preserved in the tumor metastatic to lymph nodes but was lost where there were overt histologically malignant features, as in the latter two specimens."

Throughout their article, Park and coworkers referred to the neoplasm by a variety of designations that they used deliberately in an effort to reflect its progressive worsening, namely, "proliferating tricholemmal tumor," "metastatic proliferating tricholemmal tumor," "tricholemmal carcinoma" and "malignant proliferating tricholemmal tumor."

In 1998, Takenaka and coworkers recounted their experience with a proliferating tricholemmal cystic carcinoma that resulted ultimately in the death of the patient.³⁵ The neoplasm was situated on the scalp of a woman who was 41 years of age at the time that a diagnosis was made at first of squamous-cell carcinoma. Over the course of the ensuing 15 years, the carcinoma recurred again and again at the local site despite efforts surgically to remove it completely, and despite other measures designed to eradicate it, such as radiation and chemotherapy. The carcinoma eventually extended into the cranial bones of the occiput, the brain, and the subdural space and, at last, a diagnosis was made of "malignant proliferating tricholemmal tumor." The findings histopathologically were described by Takenaka, et al. in these sentences:

"Biopsy showed a dermal lobular and trabecular proliferation of squamous epithelium, in some areas constituted by clear cells, well demarcated from surrounding tissue by a hyaline membrane. These lobes exhibited tricholemmal keratinization in the central area. Tumour cells showed severe cytological and nuclear atypia. Atypical mitoses and dyskeratotic cells were frequently present."

Takenaka and collaborators then tried a novel approach to therapy; they injected ethanol directly into the carcinoma in an effort to reduce its size and to achieve hemostasis, but, despite these measures, the woman died 8 months later as a consequence of extension of the carcinoma into the brain.

In a letter to the editor of the Journal of Dermatology in 2000, You Chan Kim and colleagues recorded their experience with a neoplasm they characterized as a "giant proliferating tricholemmal tumor with malignant transformation."³⁶ The patient was a 61-year-old woman who had a very large neoplasm, that is, 16 x 16 x5 cm in diameter, positioned on the right parietal aspect of the scalp for the past 20 years but recently had enlarged even further and became ulcerated. The attributes of the neoplasm histopathologically were described by Kim et al. in the following manner:

"Microscopically, the ulcerated tumor was composed of variably sized lobules of squamoid epithelium with tricholemmal differentiation. Some of the tumor lobules showed peripheral palisading, clear cell formation, and tricholemmal keratinization. Masses of eosinophilic amorphous keratin were noted in the center of the lobules. The tumor cells in several areas showed dyskeratosis, numerous mitoses, and nuclear atypia."

Based on a diagnosis of "malignant proliferating tricholemmal tumor," the patient received radiation therapy which caused the neoplasm to become smaller in size, but the patient forbid any further therapy. The following year she developed paralysis of both lower extremities and severe headaches, and died. Kim and colleagues inferred that her death "may have been due to direct invasion or metastasis of the tumor to the brain or spine." Based on our review of both the specimen as it appeared grossly and of the photomicrograph presented by the authors, the neoplasm fulfills criteria for proliferating tricholemmal cystic carcinoma and must have been that from its inception, rather than a "proliferating tricholemmal tumor" that underwent "malignant transformation," as was the thesis of the authors. Of course we concur that that malignant neoplasm was responsible for the demise of the patient.

In addition to the eight patients who died from metastases of proliferating tricholemmal cystic carcinoma, 15 other patients developed metastases that, as of the time of publication of the communication, had not killed them^{10,21,37–50} (Table 25–2). Saida and colleagues attempted to explain that phenomenon thus:

"There seem to be three stages in the oncological development of this peculiar tricholemmal type of tumor. TC [tricholemmal cyst] is a tumor of the benign adenomatous stage and proliferating tricholemmal cystic carcinoma is considered to be an epitheliomatous tumor. Malignant proliferating tricholemmal cystic carcinoma or malignant proliferating tricholemmal tumor is the carcinomatous stage of this kind of tumor. In the present case, proliferation of the cyst wall was observed in some of the TCs, suggesting the development of proliferating tricholemmal cystic carcinoma from TC and malignant transformation occurred in one of the cysts after trauma."³⁷ Because a follicular is not a gland, a carcinoma that differenciates toward any part of it cannot rightly be termed "adenomatous.

In short, Saida and coworkers invoked a multistep theory of tumor progression (i.e., from cyst to carcinoma through a stage of proliferating cyst) in an effort to explain the different morphologic expressions of what we think is a single pathologic process, namely, proliferating tricholemmal cystic carcinoma that begins usually in an isthmic-catagen (tricholemmal) cyst.

Other examples of proliferating tricholemmal cystic carcinoma but without long-term follow-up or evidence of metastases also have been reported on.^51–65