Carbamazepine-Induced, CD30+, Primary, Cutaneous, Anaplastic Large-Cell Lymphoma

Anticonvulsant agents may cause nodal and extranodal lymphoproliferative disorders, including dermatopathic lymphadenitis, atypical lymphoid proliferation, and cutaneous pseudolymphoma.¹ In most reported cases, regression follows the discontinuation of treatment with the causative drug, but in rare cases a true lymphoma may develop.^{2 - 3} Herein, we describe the clinicopathologic features of the first case of a CD30⁺, primary, cutaneous, anaplastic large-cell lymphoma (ALCL) possibly correlated to carbamazepine administration.

A 13-year-old girl experienced lipothymic episodes during her first menses. Electroencephalography findings revealed anomalous waves, and she was administered carbamazepine (Tegretol; Novartis Farma, Origgio ([Va], Italy) in increasing doses until the dose reached 600 mg daily. After 8 months of treatment, the patient was admitted to our dermatology department for an erythematous macular eruption clinically diagnosed as pityriasis rosea. The macular eruption regressed, but 1 month later she suddenly developed multiple painless reddish skin nodules that grew and ulcerated quickly (Figure 1); the nodules, located on the neck, trunk, and arms, varied in size from 0.5 to 6 cm, and only a few partially self-regressed. The patient was healthy; neither lymphadenopathy nor splenomegaly was detected; findings of the routine blood workup were unremarkable; cultures of swabs from cutaneous lesions yielded no bacterial pathogens. Paul-Bunnell test results were negative, and viral titers showed no clinically relevant rise.

Results of a histologic examination of a skin nodule revealed an area of epidermic pseudoepitheliomatous hyperplasia overlying a dermohypodermic diffuse lymphoid infiltrate of large anaplastic cells, scattered and in large cohesive clusters, the clusters more prominent in the deeper part of the lesion. The cellular infiltrate effaced the dermal architecture; the anaplastic cells showed a moderate degree of epidermotropism and, in some areas, a perivascular array. Almost all the anaplastic cells expressed the CD30/Ki-1 antigen (Figure 2), the TIA-1 antigen (a molecule associated with cytotoxic granules), and various T-cell antigens (CD3, CD43, and CD45RO), but they were negative for the monoclonal antibody ALK-1 (provided by Prof Brunangelo Falini, MD) specific for the p80 NPM-ALK. The CD30⁺ cells were intermingled with a reactive population of lymphocytes.

Results of molecular analysis corroborated the T-cell lineage of the cellular infiltrate, showing a monoclonal rearrangement of the γ subunit of the T-cell receptor. Findings from chest radiography, abdomen ultrasonography, and total-body computed tomography scan excluded lymphoma extracutaneous localizations. Combining these clinicopathologic findings we diagnosed, CD30⁺, primary cutaneous ALCL in accordance with the World Health Organization classification of hematological malignancies and the European Organization for Research and Treatment of Cancer classification for primary cutaneous lymphomas. Subsequently, new nodules similar to the preexisting lesions appeared on the abdomen. Carbamazepine treatment was tapered and then stopped; the patient received 8 sessions of radiotherapy with regression of 5 cutaneous lesions. Other untreated lesions self-regressed after 4 months, with considerable scarring. Presently the patient, 3 years after the lymphoma diagnosis, is healthy and her lymphoma is in complete remission.

About one third of carbamazepine adverse reactions affect the skin, mainly causing dermatitis, but also pseudolymphomas; recently a carbamazepine-induced pseudolymphoma with CD30⁺ cells has been described.⁴ However, carbamazepine-related lymphomas are unusual, and detailed information on the clinical features and outcome of such malignancies has been provided in 2 cases only. The first² was a nodal, large-cell, non-Hodgkin lymphoma developed after 4 months of carbamazepine administration; the other³ was a nodal, immunoblastic, T-cell, non-Hodgkin lymphoma diagnosed after 1 month of carbamazepine therapy. Notably, in both cases the lymphadenopathies spontaneously resolved after treatment with carbamazepine was discontinued, suggesting an atypical lymphoid proliferation instead of a true malignancy.³

However, a clinical comparison between ours and the above-mentioned cases might be partially inappropriate because of the exclusively extranodal locations of the disease in our patient and the different therapeutic approaches. In our case the possibility of a nonneoplastic, reactive CD30⁺ infiltrate has therefore been considered, including mainly other subtypes of CD30⁺, primary, cutaneous lymphoproliferative disorders⁵ with specific reference to lymphomatoid papulosis and lesions with borderline histologic characteristics between lymphomatoid papulosis and ALCL. In our patient, the clinical features of the skin lymphoma lesions (multiple, large nodules only partially regressing) and histomorphological characteristics (large, often cohesive sheets of CD30⁺ anaplastic cells) excluded a diagnosis of lymphomatoid papulosis, whereas findings of histomorphological analysis and chiefly molecular biology findings (monoclonal rearrangement for T-cell receptor γ) pointed more to ALCL than to borderline lesions.

Even if no conclusive relationship between carbamazepine and CD30⁺, primary, cutaneous ALCL may be drawn from a single case study, we deem it important for clinicians to be aware of the possible serious adverse effects of carbamazepine treatment, as the discontinuation of the drug treatment can play a crucial role in the regression of the clinical picture, sparing aggressive and harmful treatments.