Lymphomatoid Papulosis: A T-Cell Dyscrasia with a Propensity To Transform into Malignant Lymphoma

Methods

We retrieved the records of all cases encoded as lymphomatoid papulosis from the pathology archives of the University of Texas M.D. Anderson Cancer Center and the University of Nebraska Medical Center. Our primary interest was to analyze and describe patients meeting one or both of the following clinical entry criteria: lymphomatoid papulosis misdiagnosed at the time of original presentation and lymphomatoid papulosis that later developed into lymphoma.

All slides had been initially reviewed by a hematopathologist at presentation. In all but three cases, histologic material was available for a second review and for classification into the two Willemze subtypes of lymphomatoid papulosis. Two hematopathologists reviewed the pathology slides from their respective institutions. The clinical criteria we used to diagnose lymphomatoid papulosis included the following: 1) multiple skin lesions consisting of papules or nodules (Figure 1); 2) spontaneous regression of the lesions followed by recurrence after several weeks of observation (waxing and waning) [In three cases, either because of surgical excision or topical therapy, the initial behavior of the skin lesions could not be observed. In each case, however, recurrent lesions showed the typical waxing and waning pattern.]; 3) no evidence that skin lesions progressively grew to a diameter greater than 3 cm during 3 months of observation without treatment; and 4) absence of any lymphadenopathy or systemic lesions associated with lymphoma before or concomitant with the development of the skin lesions. In patients who presented with lymphadenopathy, a biopsy was required to rule out lymphoma. We included two such patients in our study.

We histopathologically subclassified cases of lymphomatoid papulosis (as defined by the above clinical criteria) using the criteria of Willemze [2]. Willemze type A lesions (Figure 2) are characterized by large anaplastic tumor cells and Reed-Sternberg-like cells, whereas type B lesions (Figure 3) have cerebriform mononuclear cells and epidermotropism similar to or indistinguishable from those of mycosis fungoides. Intermediate or transitional forms of these types were encountered, and lesions of both types were seen in some patients. In two patients, the morphologic finding was that of large non-cleaved-cell lymphoma or immunoblastic lymphoma and was therefore not typical of either type A or B lymphomatoid papulosis.

Of 28 cases identified in our search, 21 fulfilled this histologic and clinical entry criteria. Two additional cases fulfilled four of the five criteria we required for the diagnosis of lymphomatoid papulosis; with short-term follow-up, however, it became clear that these patients did not have lymphomatoid papulosis but rather large-cell lymphoma. These two cases show the difficulties in prospectively establishing the diagnosis of lymphomatoid papulosis, a point not well emphasized in the literature. We describe one of these cases in case history 3. We excluded two other cases of classic lymphomatoid papulosis that did not fulfill either of the two entry criteria for inclusion in the study (no misdiagnosis and no evolution into lymphoma). The remaining cases constitute the group we analyzed to describe their features and clinical evolution.

Case Histories

Case History 1: Misdiagnosed Lymphomatoid Papulosis

A 77-year-old man noticed a skin rash in 1991. He had waxing and waning erythematous skin lesions for about 6 months until a biopsy was done in August 1991. His condition was diagnosed as anaplastic large-cell lymphoma of the Ki-1 type. A staging evaluation that included a bone marrow biopsy showed no evidence of disease outside of the skin. Electron-beam radiation was given, and the lesions promptly regressed. However, the cutaneous eruption recurred 2 months later, just a few weeks after radiation therapy was completed. In January 1992, the patient received four cycles of combined therapy with ifosfamide and etoposide. The lesions again promptly regressed with therapy but reappeared when the treatment was discontinued after the fourth cycle. The patient was scheduled to receive a regimen of methotrexate, doxorubicin, cyclophosphamide, vincristine, prednisone, and bleomycin (MACOP-B) when he was referred for a second opinion. When untreated, his skin lesions consistently progressed from small red spots to prominent erythematous lesions and then regressed spontaneously, leaving a small scar. The patient was afebrile, and other vital signs were normal. The only abnormal findings were the numerous erythematous, raised lesions on his lower extremities that he said were typical of what he had been experiencing. After reviewing the biopsy specimen obtained before referral and without the benefit of a clinical history, the hematopathologist diagnosed the condition as consistent with anaplastic large cell lymphoma. However, after discussion of the clinical symptoms with the pathologist, the diagnosis was revised to lymphomatoid papulosis. The patient has remained healthy with no problems except those related to his skin lesions.

Case History 2: Lymphomatoid Papulosis Developing into Lymphoma

Recurrent, spontaneously regressing cutaneous papular lesions developed on the extremities and trunk of a 25-year-old man. After 12 years of follow-up, a right submandibular lymph node mass developed; a biopsy specimen of the mass led to a diagnosis of a T-immunoblastic lymphoma. A careful staging evaluation, including bone marrow biopsy, showed no other sites of disease. After the patient received five cycles of chemotherapy with cyclophosphamide, doxorubicin, prednisone, bleomycin, vincristine, and prednisone (CAP-BOP), he achieved a complete remission. The cutaneous lesions of lymphomatoid papulosis dramatically improved with chemotherapy but subsequently recurred. Nine years after chemotherapy, the lymphoma has not recurred, but the patient continues to manifest the skin lesions of lymphomatoid papulosis.

Case History 3: A Lesion Histologically Consistent with Lymphomatoid Papulosis but Clinically Similar to Lymphoma

A 35-year-old woman noticed a single erythematous skin lesion on her right thigh that within 6 months gradually became a 12-cm soft-tissue mass. A few months before a biopsy specimen of the mass was obtained, the mass spontaneously regressed to 8 cm in diameter. This regression was only temporary, and the mass steadily grew again to the original size. The patient had not experienced any constitutional symptoms. A biopsy specimen obtained before referral had been interpreted as cutaneous T-cell lymphoma. Her initial serum lactate dehydrogenase level was 3.73 µkat/L (normal, 1.67 to 3.75 µkat/L). Physical examination did not show peripheral lymphadenopathy or any other significant abnormalities except for an inflamed 12-cm soft-tissue mass in the right thigh that appeared infected and ulcerated. Results of a complete staging work-up, including computed tomographic imaging of the abdomen and pelvis, chest radiograph, and bone marrow biopsy, were all within normal ranges.

A second skin biopsy specimen was interpreted as suggestive of anaplastic large-cell lymphoma, but lymphomatoid papulosis could not be histopathologically ruled out. Tumor cells were immunoreactive for CD30 (Ki-1), CD2, and CD4 but were nonimmunoreactive for CD3. The lesion was too large for lymphomatoid papulosis, but, because of the spontaneous regression, the patient was observed for another month. When the lesion did not regress any further, the patient was treated for anaplastic large-cell lymphoma. After chemotherapy was initiated with the cyclophosphamide, doxorubicin, vincristine, prednisone, and bleomycin (CHOP-BLEO) regimen, the patient attained a complete remission but relapsed 8 months later with an endobronchial lesion that showed the same histologic features as the primary lesion. Although salvage therapy was instituted with the etoposide, high-dose cytosine arabinoside, methylprednisolone, and platinum (ESHAP) regimen [5], the patient's clinical course continued to worsen. She died 12 months after chemotherapy was initiated.

Results

Twenty-one patients fulfilled all the clinical and histologic criteria for a diagnosis of lymphomatoid papulosis and also met one of our two study entry criteria. Nineteen of these 21 patients were initially misdiagnosed, including 16 patients whose condition was initially diagnosed as a malignant neoplasm but who have not developed lymphoma and 3 whose condition was initially misdiagnosed but later developed into lymphoma. The initial diagnoses, the type of lymphomatoid papulosis, and the treatment given to the 19 patients misinterpreted as having malignant lymphoma are summarized in Table 1.

Table 1 also shows that 8 of the 19 patients whose condition was misdiagnosed as malignant received either chemotherapy or radiation therapy. Of the five patients who received chemotherapy, four received combination regimens such as MACOP-B [6] and CHOP-BLEO [7].

The clinical features of these 21 patients are summarized in Table 2. The most commonly involved sites were the extremities and the trunk (chest or abdomen). The median size of the largest lesion observed was 1.6 cm, and no patients had lesions larger than 3 cm. According to the criteria for lymphomatoid papulosis, lesions could be no larger than 3 cm. The median age at onset of symptoms was 47 years, but very young (age, 14 years) and old (age, 77 years) patients were also affected. Table 2 also lists immunophenotypic features and the results of molecular studies in the few patients examined with these techniques at the time of diagnosis of lymphomatoid papulosis.

Comparison of Patients with and without Progression to Lymphoma

The frequency of evolution to lymphoma in our series (5 of 21 patients [24%]) is similar to that described in the literature. In our 21 patients, the median age at onset of lymphomatoid papulosis was 47 years (range, 14 to 77 years). The median age at the time of diagnosis of lymphomatoid papulosis for the 5 patients who later developed lymphoma was 36 years (range, 14 to 51 years) compared with 51 years (range, 28 to 77 years) for those who did not (P < 0.05). Of the 11 patients who first developed lymphomatoid papulosis when they were 45 years of age or younger, 4 have developed lymphoma. In contrast, only 1 of the 10 patients whose disease developed when they were older than age 45 years has developed lymphoma (P = 0.2). Because younger patients have a longer life expectancy and thus are expected to be at a higher risk for developing lymphoma with time, we analyzed the time elapsed from the first sign of lymphomatoid papulosis to the development of frank lymphoma and compared it with the time elapsed between the first sign of lymphomatoid papulosis and the last follow-up date in patients who have not developed lymphoma. The median time from the onset of lymphomatoid papulosis to diagnosis of lymphoma was 12 years (range, 5 to 18 years), whereas the median duration from onset of lymphomatoid papulosis to the last date of follow-up for patients who have not developed lymphoma was only 2.1 years. This finding suggests that the risk for developing lymphoma increases with time and that the difference in age between patients who developed lymphoma and those who did not could be explained by the greater risk younger patients have because of their longer life expectancy.

In the literature on the risk for transformation of lymphomatoid papulosis to lymphoma, the cumulative risk has not been considered in terms of time but rather in terms of the absolute frequency [8, 9]. For example, if a patient has had lymphomatoid papulosis for 1 year, the risk for developing lymphoma would be expected to be lower than that for a patient who has had the disease for 10 years. Nevertheless, investigators have only calculated the absolute frequency of developing lymphoma, giving no attention to the actuarial cumulative risk with time. To determine the actuarial cumulative risk for developing lymphoma with time, we plotted a reverse Kaplan-Meier curve that shows the percentage of patients who developed lymphoma as a function of time beginning with the onset of lymphomatoid papulosis. This technique considers how long the patient has been at risk for developing a given phenomenon, in this case, a malignant lymphoma. Included in this curve are the 21 patients whose condition met all the criteria for lymphomatoid papulosis and who also met the entry criteria for the study. The cumulative risk after 15 years approaches 80% ± 18% (Figure 4). The first patient to develop lymphoma did so 60 months after lymphomatoid papulosis was diagnosed, and the last patient to develop lymphoma did so 216 months after diagnosis.

To determine if the development of lymphoma correlated with the Willemze histologic type of lymphomatoid papulosis, we compared patients who developed lymphoma with those who did not. Of the 16 patients who did not develop lymphoma, 14 had archival histologic material available for review and classification. Four of these patients presented with type A, 2 with types A and B in different skin lesions, 7 with type B, and 1 with neither type A nor type B. The latter patient had lesions histologically similar to nonanaplastic large-cell lymphoma that did not have the typical features of type A. Biopsy specimens from 4 of the 5 patients who developed lymphoma were available for a second review. Examination of the specimens showed the following: One patient presented with type A, 2 with type B, and 1 with neither A nor B (the lesion histologically resembled immunoblastic lymphoma). Thus, it does not appear that in this small cohort the Willemze type of lymphomatoid papulosis is strikingly correlated with the subsequent development of lymphoma.

Establishing Evolution to Lymphoma

In three of the five patients whose lymphomatoid papulosis developed into lymphoma, the diagnosis was established on the basis of a biopsy specimen from a new lymph node. The diagnoses in these three patients were Hodgkin disease of the nodular sclerosing type in one patient (the lesions were CD15-positive and CD30-positive, whereas the cutaneous lymphomatoid papulosis lesions were CD15-negative and CD30-positive) and T-immunoblastic lymphoma in two patients. In two patients, lymphoma evolved in the skin. In these two patients the diagnosis was established on clinical grounds alone because the histologic features were not distinguishable from those of lymphomatoid papulosis. In one of the two patients with cutaneous involvement, a 4 × 6 cm ulcerated skin lesion developed that did not regress despite treatment with psoralen and ultraviolet A irradiation (PUVA) (Figure 5). On a clinical basis, the patient's condition was considered to have evolved into lymphoma. Examination of the skin biopsy specimen showed findings compatible with either Willemze type B or mycosis fungoides, whereas the patient's original biopsy specimen had shown evidence of Willemze type A lymphomatoid papulosis. Marker analysis of the lymphoma sample showed a CD2+, CD3+, CD5+, CD30+, CD45RO^-, and CD20^- immunophenotype. Studies of DNA showed a rearranged T-β-receptor gene and germline immunoglobulin genes. Immunophenotypic studies done before transformation were limited to L-26, UCHL-1, and the leukocyte common antigen, all of which were negative. Markers were not available from the original biopsy specimen obtained in 1978.

The other patient with cutaneous transformation developed skin plaques that were clinically characteristic of mycosis fungoides and differed from the lymphomatoid papulosis lesions previously observed. The lesions did not wax and wane. Histologic features were similar to those at the time of diagnosis and were consistent with either Willemze type B lymphomatoid papulosis or mycosis fungoides. Immunophenotypic studies at the time of transformation to lymphoma indicated T lineage (CD2+, CD3^-, CD30+, and CD22^-).

Response to Treatment and Survival

For this analysis, we separately evaluated the five patients who developed lymphoma. These five patients have survived for at least 6, 15, 36, 48, and 108 months each since lymphoma developed. All five attained a complete remission, although the patient who developed Hodgkin disease required high-dose chemotherapy and autologous bone marrow transplantation. This patient had achieved only a partial response to front-line chemotherapy but achieved a complete response after high-dose chemotherapy and autologous bone marrow transplantation. Three years after Hodgkin disease developed, the patient is alive, and her illness has not recurred. The other four patients remain in complete remission.

Response of Lymphomatoid Papulosis to Chemotherapy

In 8 of the 12 patients who received chemotherapy or radiotherapy, including those who were treated after the lymphomatoid papulosis had evolved into lymphoma, the lymphomatoid papulosis lesions appeared to respond transiently but then recurred later, usually after therapy had been discontinued. In three patients, the lymphomatoid papulosis lesions disappeared after chemotherapy and have not recurred; in one patient they appeared to worsen while the patient received chemotherapy. In two patients, the lymphoma remains in remission, but for several years the lymphomatoid papulosis lesions have continued to wax and wane.

Discussion

Lymphomatoid papulosis was first described by Macauley in 1968 [10]. In his classic article, he discussed a case in which “repeated biopsies of. skin lesions consistently reveal an alarming infiltrate of large pleomorphic hyperchromatic cells which expert histopathologists and hematologists. variously classified as highest grade malignant lymphoma, malignant reticulosis, metastatic carcinoma, malignant melanoma, undifferentiated malignant tumor.” As shown in our study and by the experience of others [1], this disorder continues to be frequently misdiagnosed, which leads to an underestimation of its incidence. In our study, we could not determine the exact frequency of the misdiagnosis of lymphomatoid papulosis. The tertiary nature of the two institutions that participated in the study probably led to a selection bias favoring patients who were referred because of a cancer diagnosis. However, it is evident from the literature that this problem is not unusual. In at least two reports of cutaneous anaplastic large-cell lymphoma, some of the patients fulfilled the diagnostic criteria for lymphomatoid papulosis [11, 12]. Moreover, the most significant feature differentiating lymphomatoid papulosis from lymphoma is the tendency of the former to regress spontaneously. Yet, in 1986, Kadin and colleagues [13] reported a case of cutaneous Ki-1-positive anaplastic large-cell lymphoma with lymphadenopathy in which skin lesions spontaneously regressed but shortly thereafter grew again [13]. According to Willemze and colleagues [14], 25% of cases of primary cutaneous anaplastic large-cell lymphoma regress spontaneously. This adds further confusion and makes spontaneous involution an imperfect discriminator between lymphomatoid papulosis and anaplastic large-cell lymphoma. Case history 3 is an example of a patient with anaplastic large-cell lymphoma who had a transient, spontaneous regression. Despite this regression, the large size and solitary nature of the soft-tissue mass indicated malignant lymphoma from the onset.

Macauley [15] suggested that this illness forms part of a clinical and histologic spectrum of skin disorders known as “rhythmic paradoxical eruptions” that includes lymphomatoid papulosis, pityriasis lichenoides et varioliformis acuta, regressing atypical histiocytosis, and, perhaps, primary cutaneous Hodgkin disease [16]. All of these disorders probably represent variants of the same illness. Lymphomatoid papulosis can become malignant, and identical gene rearrangements have been shown in both the benign and malignant phases, suggesting that the latter derives from the former [4, 17]. Two types of lymphomatoid papulosis have been histologically described. Type A is the large-cell type and cannot be distinguished morphologically from anaplastic large-cell lymphoma. As shown in our study, the considerations for adifferential diagnosis include not only large-cell lymphoma but also melanoma, squamous carcinoma, metastatic undifferentiated carcinoma, and Hodgkin disease metastatic to the skin. Type B is the small-cell type that closely resembles mycosis fungoides. Because of these histologic similarities, lymphomatoid papulosis cannot be diagnosed without knowledge of the clinical findings. Features associated with high-grade malignant lymphomas such as anaplasia, monoclonality as shown by restriction analysis, aberrant T-cell phenotype, aneuploidy, and other chromosomal abnormalities all coexist in this “clinically benign, histologically malignant” disorder [10]. Consequently, only the clinician, in collaboration with the pathologist, can ultimately decide whether the disorder represents a benign or a malignant lesion.

The incidence of transformation to lymphoma is widely quoted as ranging from 15% to 20%. These figures are basically derived from two series in which the incidence of transformation ranged from 12% to 20% [8, 9]. These series included 56 cases in which 9 (16%) became malignant. Unfortunately, the investigators in the two series did not consider the actuarial cumulative risk for developing lymphoma as a function of time. Many of the patients described in these two series were followed for only a few months, so the reported risk for transformation would be expected to be low. When we calculated the cumulative risk in our series, it was clear that the risk for transformation begins 5 years after the first sign of lymphomatoid papulosis and continues to increase with time (Figure 4). Thus, the frequently quoted 15% to 20% rate of transformation is an underestimation of the real risk. On the other hand, our estimate might be biased by the exclusion of patients who did not develop lymphoma and whose condition was not misdiagnosed as lymphoma. The tertiary nature of our institutions does not allow us to determine the real denominator of all cases of lymphomatoid papulosis. Most of our patients are referred because of either the diagnosis or a strong suspicion of cancer. The real estimate of the risk for developing lymphoma lies somewhere between the 20% absolute frequency quoted in the literature and the 80% cumulative frequency in our series.

As a corollary, we can also conclude that because of their longer life expectancy, young patients with lymphomatoid papulosis are at a higher risk for developing lymphoma. Thus, it is not surprising that the median age of our patients who developed lymphoma was 39 years compared with 51 years in those whose condition has not transformed.

The survival rate of the 21 patients described in this series, as well as that reported in the literature, is excellent. All 21 patients in this series are alive. Therefore, the inappropriate administration of toxic chemotherapy regimens or radiation to these patients would be a mistake. On the other hand, it is crucial that malignant transformation be identified in a timely way because the response to treatment and clinical outcome in these cases appears favorable, in contrast to low-grade lymphomas, which at transformation are usually refractory to chemotherapy. All five patients in this series whose lymphomatoid papulosis transformed into lymphoma responded favorably to treatment and have remained free of recurrent disease for 6 to 108 months (median, 48 months).

Unfortunately, no single clinical characteristic at presentation can be used to distinguish lymphomatoid papulosis from lymphoma. Neither is any single criterion available to establish evolution to malignant lymphoma. At clinical onset, the presence of a solitary skin lesion greater than 3 cm in diameter, persistence without spontaneous regression, and the presence of significant lymphadenopathy all indicate that the condition is probably malignant. A biopsy specimen from any enlarged lymph node should be obtained to rule out lymphoma. Later in the disease course, a rapidly growing skin lesion that fails to regress spontaneously or becomes resistant to topical therapy such as PUVA and that exceeds 3 cm in diameter usually signals transformation to lymphoma. Again, it is not possible to histologically diagnose malignant evolution with absolute certainty when the disease is confined to the skin. Kadin [18] suggested that features indicating lymphoma include a high ratio of atypical cells to inflammatory cells, infiltration of atypical cells into the subcutis, and a change in tumor cell immunophenotype with further loss of T-cell antigens [18]. On the other hand, when extracutaneous dissemination occurs, as shown by case history 2, a diagnosis of lymphoma is straightforward. When any doubt exists, only careful and close clinical follow-up determines the diagnosis.

The differential sensitivity to chemotherapy shown by lymphomatoid papulosis and by secondary malignant lymphoma is another feature that has not received sufficient attention [19]. The complete response to chemotherapy shown by the secondary malignant lymphoma in contrast to the initial regression but later recurrence manifested by the lesions of lymphomatoid papulosis is intriguing. One of the explanations proposed for the spontaneous regressions observed in lymphomatoid papulosis is that the host's immune system can control the growth of these cells [20]. However, this does not completely explain the natural history of this disorder and specifically does not explain the differential sensitivity to chemotherapy shown by the malignant lymphoma. Cancer is known to be a multistep process that is usually preceded by a series of genetic lesions. The fact that lymphomatoid papulosis shows many biological abnormalities characteristic of malignancy suggests that this disorder is an incomplete lymphoma or a “form fruste” that has accumulated nearly all the genetic lesions required for malignant behavior but is missing one or a few critical lesions that would allow it to behave as a fully developed lymphoma. It appears likely that with time, a subclone of cells acquires one or more tumorigenic abnormalities that provide it with a proliferative advantage over the rest of the cells. This proliferative advantage may also impart a selective sensitivity to chemotherapy. An alternative hypothesis is that lymphomatoid papulosis is a low-grade T-cell lymphoma that can progress to a higher-grade malignancy. This would resemble low-grade B-cell lymphomas, which are known to transform to a more aggressive lymphoma with time. In many cases, however, the morphologic features of lymphomatoid papulosis are already those of a high-grade malignancy. Nevertheless, the progression to a higher-grade malignancy in these cases could occur at the molecular rather than at the histologic level. Kadin and colleagues [21] described a patient with lymphomatoid papulosis that transformed into lymphoma. The malignant cells in that case lost their tumor growth factor-β receptors, and this phenomenon rendered them incapable of responding to this growth factor, which normally inhibits growth [21]. If loss of tumor growth factor-β receptors mediates the transformation to a fully developed lymphoma, measurement of these receptors in lymphomatoid papulosis might replace the crude and ill-defined measures we currently use to determine malignant transformation in this unique clinical setting.