OUP user menu

Simulators of Cutaneous Lymphoma
Where Should Our Efforts Go?

Philip E. LeBoit MD
DOI: http://dx.doi.org/10.1309/AJCP07TTPGFATGSO 414-417 First published online: 1 April 2013

This issue of the Journal contains the report of a workshop on cutaneous lymphoma sponsored by the Society for Hematopathology and the European Association for Haematopathology. Although consensus appears elusive regarding the spelling of the name of one of their subspecialties, the exercise is obviously worthwhile. Dermatopathologists and h(a)ematopathologists bring different perspectives to the study of cutaneous lymphoma, and the field needs input from and, indeed, continuing contact between both groups. Dermatopathologists have been the first to recognize some conditions, such as primary cutaneous follicular lymphoma and leg type B-cell lymphoma. These neoplasms have turned out to have counterparts in other organs, but they are recognized in the World Health Organization lymphoma classification.1 Hematopathologists can bring a broad perspective to the table and, in many cases, know better how to apply evolving technologies used to evaluate lymphomas in general.

The workshop, the highlights of which are delineated in this issue, concentrated on cutaneous lymphoid hyperplasia, drug reactions, lupus profundus, pityriasis lichenoides, Kikuchi disease, mucocutaneous ulcers, and pigmented purpuric dermatosis. The authors’ description of their experience notes some interesting and controversial cases among these entities. Certainly, anyone with an interest in this area would benefit from seeing these cases and hearing their discussion. But does it address the problems that pathologists who may encounter such cases have? Does it address the difficulties that consultant dermatopathologists have with this area? Are the issues important ones for the care of patients?

Troubling issues in the culture of medicine need to be solved to make the diagnosis of cutaneous lymphoma more accurate. Let’s start with dermatologists. Most dermatologists see more patients with cutaneous lymphoma during residency than in their careers. One senior physician who attended our cutaneous lymphoma clinic some years ago told me that he had never had a patient appear in his private office with mycosis fungoides (MF) (he followed dozens of such patients in our clinic, though). In contrast to epithelial tumors and melanocytic lesions, the clinical knowledge about cutaneous lymphoma and the comfort level in dealing with it may decrease, not increase, after graduation. An unfortunately prevalent custom among dermatologists is to fill out requisition sheets not with a description of clinical morphology but with a short list of “rule-outs.” A recent study, in which the final diagnosis made at a university department of dermatology was compared with proposed diagnoses written on the requisition, showed that in the majority of biopsies performed for an inflammatory skin disease, the correct diagnosis was not initially proposed by the dermatologist.2 When clinical digital images were also sent to the dermatopathologist, much higher accuracy could be achieved.

Some diagnoses in this arena require clinicopathologic correlation. Nearly everyone believes in clinicopathologic correlation, and legion are the pathology reports suggesting that it be performed. However, such correlation requires familiarity with the clinicopathologic entities that are or should be under consideration. In too many cases, an ambiguous pathologic diagnosis, or one in which the loop needs to be closed by such correlation, results in a referral to a hematologist-oncologist. Regrettably, the concept of primary cutaneous lymphoma (lymphoma that presents in the skin and is often limited to it for many years or indefinitely) is often unfamiliar outside of dermatology and dermatopathology, and mimics of cutaneous lymphoma are often completely outside the training of these physicians.

Frequently, more work needs to be done by the clinician, and the dermatologist may feel uncomfortable doing it. Most dermatologists, understandably, want to do a biopsy, get back a timely report, and tell the patient exactly what he or she has. A follow-up appointment may instead be along the following lines: “Mrs Jones, the pathologist isn’t quite sure of what you have, and I need to examine you more carefully [eg, perform a full-body skin examination, in the case of MF, including ‘double-clothed’ areas, and palpate for adenopathy], and then we will discuss the case and get back to you.” Couple this with a shaky knowledge base, and we can well understand why referring the patient to a hematologist-oncologist is a popular solution.

On the pathologic side, many general pathologists are not sufficiently familiar with inflammatory skin diseases, either with respect to their clinical presentations or their histopathologic appearances, to confidently wade into the field. Even among dermatopathologists , many are uncomfortable with cutaneous lymphoid infiltrates. This results in unnecessary worry about lichenoid keratosis and drug eruptions as well as extra immunostains. Many pathologists and dermatopathologists know that any uncertainty in a diagnosis is likely to result in a request for consultation or genotypic studies, rather than in a more detailed clinical exegesis.

So how does this play out in real life? Several problem areas come to mind. One is with regard to CD30+ lymphoproliferative disorders. A typical scenario is as follows. A biopsy is done, and the specimen shows an infiltrate with large lymphocytes. A CD30 stain follows. Many of the large cells are positive. Everyone panics, and the case goes for consultation, sometimes while the patient is sent off to the hematologist-oncologist. Here’s what needs to happen instead. The pathologist needs to assess if the CD30+ cells could be in an infection (such as those due to barnyard pox-virus) or inflammatory disease3 or whether the histopathologic features could be those of lymphomatoid papulosis (LyP), a tumor of MF, or CD30+ anaplastic large T-cell lymphoma (ALCL). The dermatologist needs to determine if the lesions are papules that come and go (eg, like those of LyP), perform a full-body skin examination, and make sure that no subtle patches of MF have been overlooked or if there is a fixed nodule, as in tumors of MF and ALCL. In some patients, only 1 lesion—a papule—is present, which is then removed by biopsy. In such cases, one has to wait and see if other papules arise; the best way to tell if they are those of LyP is to follow them for up to 12 weeks to see if spontaneous involution occurs. This requires teamwork and dedication (not just lip service) to clinicopathologic correlation. The reflexive responses—referral to a hematologist-oncologist, more immunostains, and gene rearrangement studies—will involve spending more resources and time and will not solve this problem.

Another example is that of patch-stage MF and its histopathologic simulators, of which, in common practice, pigmented purpuric dermatitis, drug eruptions, lymphomatoid contact dermatitis, the inflammatory stage of morphea, and extragenital lichen sclerosus pose common problems. Although a variety of scoring systems have been devised to pry apart cases of patch-stage MF from those of inflammatory skin disease, these can be misleading if the clinical context is not known to the pathologist or if the pathologist is not knowledgeable enough to appreciate it. Image 1 presents 4 images (admittedly taken out of context), only 1 of which represents a patch of MF. I doubt that any pathologist or dermatopathologist can confidently identify the case of MF except by exploiting his or her test-taking skills.

Image 1

Stripped away from the clinical context are flattish, slightly erythematous, and poorly circumscribed patches that are usually in sun-protected sites. The histopathologic simulants of mycosis fungoides (MF) can be virtually indistinguishable from the authentic condition, especially if viewed at high magnification. Only 1 of these 4 images is that of mycosis fungoides. A, Biopsy specimen from a perianal patch of lymphomatoid contact dermatitis secondary to scented wipes. B, A specimen of lichen sclerosus et atrophicus. C, A specimen of lichenoid purpura. D, A specimen of MF lacking significant epidermotropism from a patient in whom a lichenoid interface reaction is prevalent. Nonspecific interface and spongiotic features can occur in areas in patients with MF, and multiple biopsy specimens are often necessary to avoid this pitfall.

The most important question with regard to the simulators of cutaneous lymphoma pertains to which misdiagnoses have consequences beyond poor form and embarrassment. I find it questionable whether the distinction between cutaneous lymphoid hyperplasia (CLH) and primary cutaneous follicular lymphoma (PCFL) is much more than academic. The survival in PCFL is excellent. Although diagnosis leads to more avid lesional ablatement than in CLH (by excision, injection, or radiotherapy), it is unproven whether this has an effect on the course of the disease. The main downstream consequence pertains to the workup that the patient gets. However, a good case can be made that a patient with 1 or 2 lesions in 1 area of the skin containing a lymphoid infiltrate that appears to be PCFL (especially if there is a BCL2–immunophenotype) should not routinely get imaging studies or a bone marrow biopsy.

The same case can be made with regard to the diagnosis of patch-stage MF. There is little evidence that patients with stable, limited disease (less than 10% body surface area, or stage Ia) have a much altered life expectancy. Although toxic therapies such as topical nitrogen mustard have been used in the past, high-potency topical corticosteroids suffice to efface most lesions. Any treatment that results in treatment adequate to make the lesion resolve is just fine. One caveat is that if the diagnosis of MF is under consideration, agents that modulate immunity need to be used with caution.4

In contrast, some diagnoses, such as natural killer (NK)/T-cell lymphoma, nasal-type and γ/δ T-cell lymphoma (which can both simulate a panniculitis), subcutaneous panniculiticlike T-cell lymphoma, angiotropic lymphoma, CD8+ aggressive angiotropic T-cell lymphoma, and secondary cutaneous involvement by a variety of systemic lymphomas, are dire, and the consequences of missing them can be catastrophic for the pathologist, whether or not recognition would result in a different outcome. The consideration of lupus profundus as a mimic of the subcutaneous lymphomas is an important aspect of the workshop. I would add that staining for CD56 (to exclude NK/T-cell lymphoma, nasal type) and either βF1, GM3, or both (to exclude γ/δ T-cell lymphoma) are important adjuncts if there is any doubt. The recently described type D form of lymphomatoid papulosis, with pagetoid epidermotropism and ulceration, can mimic CD8+ aggressive angiotropic T-cell lymphoma, but the treatments are quite different.5

With respect to the spectrum of CD30+ lymphoproliferative disorders, the conditions not to miss are tumors of MF with large cell transformation and secondary cutaneous ALCL. The first is considerably more common than the second. Many years ago, I recall seeing a patient in our cutaneous lymphoma clinic diagnosed with “biopsy-confirmed” primary cutaneous ALCL. No one had looked at the man’s buttocks. He was about to leave clinic when I suggested that we do so. We found a thin, subtly scaly patch of MF. The prognosis went from that of primary cutaneous ALCL (with more than a 90% ten-year survival) to that of tumor stage MF with large cell transformation, with a much less favorable prognosis.

My prescription for improving our diagnostic accuracy in this area is both simple and daunting. Dermatologists need to keep up with this rapidly changing area and be more willing to expend time and energy when confronted with a patient in whom “clinicopathologic correlation is suggested.” Pathologists need to learn more about inflammatory skin diseases and cutaneous lymphomas and even a bit about their clinical morphology. Last, our efforts should go to the differential diagnoses that matter. If the energy and money that went into performing immunoperoxidase stains on suspected cases of patch-stage MF (where they are unlikely to be decisive) went into better communication between clinicians and pathologists or dermatopathologists, the world would be a better place. And at the apex of our field, bringing expert dermatopathologists, hematopathologists, and cutaneous lymphoma clinicians together is a laudable endeavor—but not as important as addressing the above-mentioned problems.


View Abstract