Resnik, Kenneth S MD*; Kutzner, Heinz MD†

To the Editor:

As part of a recent publication in this journal,¹ we presented a case of CD30+ lymphoproliferative disorder associated with keratoacanthomatous hyperplasia. Subsequently, one of us (KSR) received correspondence regarding CD30+ cells associated with keratoacanthoma.^2,3 The subject of that correspondence raises an important issue, namely, how a histopathologist differentiates CD30+ lymphoproliferative disorder associated with keratoacanthomatous hyperplasia from authentic keratoacanthoma associated with incidental CD30+ lymphocytes.

It is clear that CD30+ lymphocytes may be associated with keratoacanthoma. Typically, a small minority of the associated infiltrate (usually low single-digit percentage) may be CD30+ cells.^2,3 This jibes with our experience in which incidental CD30+ lymphocytes may be seen in a host of entities, including, but not limited to, carcinomas, cutaneous infections (herpes virus, molluscum contagiosum, verruca, etc), allergic contact dermatitis, drug eruptions, etc. This is in marked contrast to the case we published,¹ in which most of the cells in the infiltrate (greater than 90%) were CD30 postive. Furthermore, the clinical history in the case we published¹ was consonant with lymphomatoid papulosis rather than multiple keratoacanthomas.

The thorny issue though for histopathologists is those instances in which a solitary lesion is characterized by a keratoacanthomatous proliferation associated with more than occasional, but not a preponderance of CD30-postive lymphocytes. The difficulty in such an instance turns on the fact that lymphomatoid papulosis may be characterized by an infiltrate of cells in which CD30-postive ones are in the minority. Is there a clear-cut answer to how a histopathologist should handle such a case? Simply put, the answer is no. Although CD30+ cells are a diagnostic feature for CD30+ lymphoproliferative disorders, they are not pathognomonic for those cutaneous lymphomas because a whole range of other entities may show CD30+ lymphocytes.⁴

In historical perspective, the story of CD30+ cells plays out in a similar fashion to all the other antibodies that when introduced are supposedly specific for an entity. In time, the “specific” antibody is revealed to be quite nonspecific. Interestingly, despite this pattern being repeated over time for multiple different antibodies, it does not seem to dampen the initial euphoria that tends to welcome the introduction of each next greatest “specific” antibody. In the 1980s, shortly after the antibody CD30 became available for routine immunostaining, CD30 positivity was heralded as an indicator of Hodgkin's disease in the skin.⁵ Subsequently, it has been shown that CD30 expression is not specific for lymphoid neoplasia and may be found in activated non-neoplastic lymphocytes associated with a plethora of inflammatory dermatoses. Atypical CD30+ lymphocytes commonly occur in conjunction with cutaneous viral infections,⁴ insect bite reactions,⁶ pseudolymphomas (personal experience HK), in scabies infestation,⁷ and may also be seen in superficial fungal infections of the skin,^6,8 cutaneous bacterial infections,⁶ as well as in mycobacterial infections.^9,10 This is not a complete list of inflammatory entities in which atypical CD30+ lymphocytes may be found; instead, it is simply illustrative of the extensive assortment of dermatoses that may be associated with this finding. Furthermore, the non-specificity of CD30 expression can be taken a step further because it has been demonstrated in neoplastic cells unassociated with lymphoid lineage, including various benign and malignant mesenchymal neoplasms (eg, leiomyoma, leiomyosarcoma, malignant fibrous histiocytoma, synovial sarcoma, rhabdomyosarcoma, etc),¹¹ embryonal carcinoma,¹² etc. Again, this is not a comprehensive list of this phenomenon.

It is well accepted that pseudocarcinomatous hyperplasia may be induced by conditions that show florid proliferation of neoplastic or inflammatory cells in the dermis, for example, granular cell tumor/schwannoma and deep fungal infection. The cellular infiltrates in these entities likely induce the pseudocarcinomatous hyperplasia via either cytokines or stimulatory growth factors, i.e., the epithelial hyperplasia is a secondary phenomenon. It is plausible to infer that the same applies to CD30+ lymphocytes in those instances of keratoacanthomatous hyperplasia associated with CD30+ cutaneous lymphoproliferative disorders.

A recent publication⁴ notes that various viruses are one of the most common causes of inducing clusters of CD30+ lymphocytes. Is this induction brought about by the virus itself, the associated pseudocarcinomatous hyperplasia, or do both come into play? In herpes virus infection, the answer is straightforward because that virus tends not to induce pseudocarcinomatous hyperplasia. In pox, parapox, and papilloma virus infections, does the virus, the associated pseudocarcinomatous hyperplasia, or both elicit clusters of CD30+ lymphocytes? The answer is not clear-cut because neoplastic epithelial proliferations (keratoacanthomatous squamous-cell carcinoma for instance) may induce clusters of CD30+ lymphocytes.^2,3 Furthermore, although speculation, we would not be surprised if one day lymphomatoid papulosis turns out to be an epiphenomenon of a viral infection.

In sum, the interplay between lymphocytes and cutaneous epithelium, and vice versa, is a documented phenomenon. CD30+ lymphocytes of cutaneous lymphoproliferative disorders may induce secondary epithelial proliferation (pseudocarcinomatous) and neoplastic epithelial proliferations such as keratoacanthoma may bring about secondary clusters of CD30+ lymphocytes. In the realm of CD30-positive staining of lymphocytes in cutaneous specimens, there seem to be two possibilities: (1) a lymphoma and (2) the epiphenomenon of CD30+ lymphocytes induced by numerous different primary conditions.

Kenneth S. Resnik, MD

Institute for Dermatopathology, Newtown Square, PA

Heinz Kutzner, MD

Dermatopathologie Friedrichshafen, Friedrichshafen, Germany

REFERENCES