Fax: (608) 775-6627
Immunophenotypic and molecular features, clinical outcomes, treatments, and prognostic factors associated with subcutaneous panniculitis-like T-cell lymphoma†
A systematic analysis of 156 patients reported in the literature
Article first published online: 13 AUG 2004
Copyright © 2004 American Cancer Society
Volume 101, Issue 6, pages 1404–1413, 15 September 2004
How to Cite
Go, R. S. and Wester, S. M. (2004), Immunophenotypic and molecular features, clinical outcomes, treatments, and prognostic factors associated with subcutaneous panniculitis-like T-cell lymphoma. Cancer, 101: 1404–1413. doi: 10.1002/cncr.20502
Presented at the 40th Annual Meeting of the American Society of Clinical Oncology, New Orleans, LA, June 5–8, 2004.
- Issue published online: 1 SEP 2004
- Article first published online: 13 AUG 2004
- Manuscript Accepted: 15 JUN 2004
- Manuscript Revised: 9 JUN 2004
- Manuscript Received: 23 MAR 2004
- Gundersen Lutheran Medical Foundation
- subcutaneous panniculitis-like T-cell lymphoma;
- natural history;
Subcutaneous panniculitis-like T-cell lymphoma (SPTCL) is an uncommon type of skin lymphoma. The natural history, optimal treatment strategy, and prognostic factors associated with this malignancy are not well defined.
The authors performed a systematic analysis of all patients with SPTCL reported on in the English-language medical literature, with emphasis on specific clinical features, experiences involving the use of radiotherapy and systemic agents, and prognostic factors predictive of treatment response and clinical outcome.
One hundred fifty-six patients with SPTCL were identified in the literature. Hemophagocytic syndrome (HPS) was a presenting feature in 37% of patients, and > 90% of patients required treatment at diagnosis. Prednisone was used frequently as initial therapy in patients who had less aggressive disease at presentation; however, durable complete remissions (CR) were infrequent. Anthracycline-based chemotherapy regimens were the most commonly used and most effective systemic treatment options, producing long-term CR in ∼30% of patients. Among patients who received high-dose chemotherapy and stem cell transplantation (HDT-SCT) for refractory or recurrent disease, 92% achieved CR, with a median response duration of ≥ 14 months. The presence of HPS at diagnosis and expression of the γ/δ T-cell receptor (TCR) by tumor cells were associated with poor survival, whereas age was not. After a median follow-up of 24 months, 48% of patients died of disease. The median survival duration was 27 months.
SPTCL has an aggressive natural history. Nonetheless, a subgroup of patients with SPTCL can have long-term disease remission following anthracycline-based initial therapy or subsequent HDT-SCT. HPS and the TCR phenotype may be useful prognostic markers for patients with this malignancy. Cancer 2004. © 2004 American Cancer Society.
Subcutaneous panniculitis-like T-cell lymphoma (SPTCL) is a type of skin lymphoma that is characterized by infiltration of the subcutaneous tissue by pleomorphic T cells and benign macrophages that mimic lobular panniculitis (Fig. 1).1 This malignancy typically presents itself in the form of skin nodules that involve the extremities and can become ulcerated (Fig. 2). The clinical course of SPTCL may be either protracted, with occasional spontaneous regression of skin lesions, or rapidly progressive and fatal despite the use of aggressive chemotherapy. In the latter setting, SPTCL often is associated with hemophagocytic syndrome (HPS) (Fig. 3).
Since its recognition in 1991 as a new clinicopathologic entity,1 SPTCL has remained a rare malignancy, with fewer than 200 cases reported in the English-language literature.1, 3, 5–65 There have been several reviews addressing the issue of SPTCL in recent years.2–4 Most of these reviews, however, have focused on histopathologic diagnosis and clinical presentation, with limited discussion of specific therapeutic options, treatment outcomes, and prognostic factors. At present, there is no consensus regarding the optimal treatment of SPTCL. In addition, prognostic features have not been well characterized. Because of the relative rarity of SPTCL, a retrospective study of reported cases may be the only way to obtain a sample size adequate for gaining some amount of insight into these clinically relevant questions. The current article presents a systematic analysis of 156 patients with SPTCL who were reported on in the literature, with emphasis on specific clinical features (immunologic and T-cell receptor [TCR] phenotypes, incidence of HPS, and natural history), experiences involving the use of radiotherapy and systemic agents, and factors that predict treatment response and clinical outcome.
MATERIALS AND METHODS
We reviewed the English-language medical literature for reports of patients with SPTCL up to June 2003 using the MEDLINE and CANCERLIT databases. References cited by the retrieved articles also were scrutinized carefully for reports of patients with panniculitic T-cell lymphoproliferative disorders that would fit the current description of SPTCL. One author (S.M.W.) reviewed the pathologic description of each patient, when available, to confirm the diagnosis. The pertinent histologic features of SPTCL are described in the World Health Organization Classification of Tumors of the Hematopoietic and Lymphoid Tissues.66 These features include 1) neoplastic T cells with preferential involvement of the subcutaneous tissue and relative sparing of overlying epidermis and dermis, 2) rimming of neoplastic cells surrounding individual fat cells, and 3) clonal TCR gene rearrangement. We found 156 reports of patients with SPTCL.1, 3, 5–65 Individual clinical findings and treatment methods were described incompletely in 57 of these reports.22, 23, 35, 38, 47, 55, 59, 64 Three patients29, 53, 58 did not require radiation or systemic treatment, whereas three other patients19, 34, 37 died before receiving any therapy. Of the remaining 93 patients who received treatment, adequate treatment data (specific type of therapy and nature of response) were available for 80.1, 3, 5–9, 11–18, 20, 21, 24–27, 29, 30–32, 36, 39–44, 48–51, 53, 56–58, 60–63, 65 The latter group is the focus of the discussion below (see Treatment).
Clinical features, including patient age, the presence or absence of HPS, and the type of TCR, when reported, were analyzed for potential prognostic significance. Because not all reports fully described the essential clinical features of interest in the current review, the denominators involved in our analyses vary. The Fisher exact test was used to compare binary variables. The Kaplan–Meier method was used to analyze overall survival. Survival curves were compared using the log-rank test. All P values were two sided, and P < 0.05 was considered indicative of statistical significance. All calculations were performed using GraphPad Prism software (Version 3.0; GraphPad, San Diego, CA).
Microscopic descriptions and photographs were available for review in the vast majority of cases. A panniculitis-like pattern of neoplastic involvement of the subcutaneous tissue was described in almost all patients. Concomitant involvement of the overlying dermis was noted in 44 of 156 patients (28%). Although the extent of dermal invasion was variable, such invasion was minimal in most cases.1, 5, 7, 8, 9, 15, 17, 19, 21, 29, 33–35, 41, 55, 58, 60, 61, 63 Involvement of the epidermis was extremely rare. Neoplastic lymphocytes generally were reported to have an atypical morphologic appearance. The most common histology was a mixed lymphocytic proliferation composed of large cells intermixed with medium-size and/or small cells. The next most common pattern consisted of an infiltrate of small-to-medium lymphocytes. Large cell histology was the least common presentation. Admixed benign histiocytes were frequently present. Occasionally, vasculitis, sarcoid-like granulomas, foam cells, and fat necrosis were also documented. Patients who previously were reported to have SPTCL were disqualified from the current review if there was no indication that the neoplastic cells were of T cell lineage or if the atypical lymphocytes exhibited diffuse CD30 immunoreactivity.
Clinical Features and Outcome
The median age at diagnosis was 39 years (range, 0.5–84.0 years). Seventy percent of patients were between age 18 years and age 60 years. There were roughly equal numbers of male and female patients reported on (n = 67 and n = 75, respectively). At the time of diagnosis, HPS was reported in 36 of 109 patients (33%). Immunophenotypic and molecular studies are summarized in Table 1. Malignant cells were predominantly cytotoxic T cells that expressed the α/β TCR. Almost all CD56-positive tumors (17 of 18) and all CD4-negative/CD8-negative tumors (8 of 8) expressed the γ/δ TCR. Clonal rearrangement of the TCR was detected in 85% of patients and was either not assayed or not found in the remaining patients. In only a minority of patients (6 of 61 [10%]), the presence of Epstein–Barr virus (EBV) was detected in tissue biopsies by polymerase chain reaction analysis or fluorescence in situ hybridization.16, 17, 23, 31 Five of six patients with EBV found in tumor tissue resided in Asia. Only 1 of 45 patients residing outside of Asia (2%) was positive for EBV infection.
|Immunophenotypic/molecular studies||No. of patients tested||Present (%)|
|TCR gene rearrangement (clonal)||100||85|
|EBV by PCR or FISH||61||10|
Among the 110 patients who reportedly received treatment, treatment was necessary at the time of diagnosis in most (> 90%). Five patients did not receive specific treatment for SPTCL. Three of those patients died before therapy could be initiated.19, 34, 37 Spontaneous partial or complete remission was infrequent and was described in only two patients.29, 58 Nevertheless, skin lesions eventually progressed or recurred in both patients, but neither required any treatment after 37 months and 60 months of follow-up, respectively. Follow-up data were available for 105 patients: Fifty patients (48%) died of disease, 32 (30%) were alive without disease, 20 (19%) were alive with disease, 2 (2%) died with disease, and 1 (1%) died without disease. The median survival of the entire group was approximately 27 months (Fig. 4). Among the patients who remained alive, the median follow-up duration was 24 months.
In general, radiotherapy; immunosuppressive agents, such as prednisone and cyclosporine; and low-dose chemotherapy involving a single agent, such as cyclophosphamide or methotrexate, represented the initial treatment options for patients who had a relatively indolent or localized presentation. For patients with more aggressive disease, upfront combination chemotherapy regimens (e.g., anthracycline-based regimens) were used most frequently. There also were several reports of patients who received high-dose chemotherapy followed by stem cell transplantation (HDT-SCT) as salvage therapy.
SPTCL generally is sensitive to radiation. The use of radiotherapy was reported in 14 patients.5, 12, 14, 25, 26, 29, 43, 56, 58, 60 In three patients, radiation was administered concomitantly with systemic treatments, which included combination chemotherapy,25 prednisone therapy,5 and interferon therapy.29 Most of the lesions were localized and involved the extremities. Among the 11 patients who were treated with radiotherapy alone, there were 4 complete responses (CRs) (36%) and 5 partial responses (PRs) (45%), for an overall response (OR) rate of 81%. Of the patients who achieved CR, 1 had durable remissions ranging in duration from ≥ 12 months to ≥ 23 months. One patient developed recurrent disease within a few months, whereas the duration of CR was not reported in another patient. All eight patients who did not achieve a CR or who developed recurrent disease subsequently received chemotherapy.
Because SPTCL does not usually involve extracutaneous sites at presentation and may begin its course in an indolent manner, one-third of all patients were treated initially with immunosuppressive agents with known activity against lymphoma. Prednisone was the most commonly used agent of this type. Prednisone alone or an equivalent corticosteroid was used as initial therapy in 20 patients, resulting in 6 CRs (30%) and 4 PRs (20%), for an OR rate of 50%.3, 8, 11, 15, 17, 27, 32, 41, 44, 49, 57, 58, 61 Most of those responses were relatively short, lasting < 6 months, and disease progression typically occurred once steroid tapering was initiated. No patient died of infection during treatment. Only 4 patients, all without HPS, had sustained remissions, which ranged in length from ≥ 18 months to ≥ 160 months (median, ≥ 36.5 months).3, 32, 44, 57 The other 16 patients (80%) subsequently required chemotherapy or died of disease. Cyclosporine was administered (with or without prednisone) to four patients24, 41, 45; two did not have responses, and the other two had brief PRs.
The use of biologic agents was reported only infrequently. Interferon-α as a single agent was used in two patients as salvage treatment after failure of immunosuppressive agents and combination chemotherapy, respectively.8, 57 One patient achieved a PR of 4 months' duration, whereas the other patient had no response. Another group reported using the combination of interferon-α with radiotherapy to treat a patient who did not have a response to methotrexate.29 The patient achieved a CR but developed recurrent disease when interferon-α was discontinued. Reinstitution of interferon-α resulted in a PR. The use of denileukin diftitox was described recently.50 Unfortunately, the patient who was treated using this agent developed a severe vascular leak syndrome and died before any response was noted.
Anthracycline-based and anthracenedione-based chemotherapy
With the exception of a few patients in whom SPTCL behaved more indolently and who had a sustained response to either radiotherapy or corticosteroids, most patients reported on in the literature eventually received combination chemotherapy regimens that are appropriate for aggressive nodular lymphomas. The use of aggressive chemotherapy is summarized in Table 2.
|Chemotherapy||No. of patients||CR||PR||OR rate (%)||Response duration (mos)||References|
|CHOP||28||11||6||≥2 to ≥72||1, 7, 9, 12, 13, 15, 16, 18, 21, 25, 36, 39, 40, 41, 43, 49, 51, 57, 58, 61, 62, 65|
|CNOP||4||0||1||≥28||31, 56, 58|
|ProMACE-CytaBOM||3||1||0||3||1, 20, 30|
|ESHAP||3||1||0||NR||43, 61, 62|
|High-dose therapy and stem cell transplantation||13||12||0||92||3 to ≥70||18, 21, 25, 27, 29, 30, 39, 43, 46, 47, 62, 65|
Most of the combination chemotherapy regimens used as initial therapy were anthracycline-based or anthracenedione-based, with combined cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP) or CHOP-like combinations used the most frequently. Altogether, CHOP or CHOP-like therapy was used in 38 patients, resulting in an OR rate of 53% (20 of 38), with a 32% CR rate and a 21% PR rate. Response durations ranged from ≥ 2 months to ≥ 72 months. Almost all responses (17 of 20) occurred in patients who received CHOP. None of the five patients who received anthracenedione-based regimens achieved a CR. When CHOP alone was considered, the CR rate increased to 39% (11 of 28). The reported median duration of CR was approximately ≥ 11.5 months. More intensive anthracycline-based third-generation non-Hodgkin lymphoma chemotherapy and other similar regimens (Table 2) were used in seven patients. There were four CRs and one PR. The duration of remission ranged from 3 months to ≥ 24 months. At the time of the last reported follow-up, only 3 of 16 patients experiencing CR after receiving anthracycline-based chemotherapy had developed recurrent disease. All patients with recurrent disease underwent HDT-SCT and continued to experience CR as of the most recent follow-up report.21, 30, 39
Purine nucleoside analogues
There were two reported patients who received fludarabine-based chemotherapy as their initial treatment (Table 2). Both patients achieved CRs, which lasted for ≥ 6 months and ≥ 15 months, respectively.
There were 13 reports on the use of salvage chemotherapy, including cisplatin-based, melphalan-based, ifosfamide-based, and purine nucleoside analogue–based regimens (Table 2). Five patients had responses, for an overall response rate of 38%. A CR was attained in three patients. However, all but 2 responders, who immediately underwent HDT-SCT, developed progressive disease within 6–9 months.39, 62
Thirteen patients underwent HDT-SCT (Table 2). All but one patient underwent autologous SCT. The conditioning regimens used were variable, and 12 patients (92%) achieved CR, with a median response duration of ≥ 14 months at last follow-up. The lone patient who underwent allogeneic SCT had the longest follow-up period.49 This patient's disease remained in CR after 70 months. Only 1 patient developed recurrent disease, which was diagnosed 3 months after SCT.22
We analyzed the prognostic significance of age (≥ 40 years vs. < 40 years), the presence or absence of HPS, and the type of TCR expressed by the tumor in all patients for whom adequate, specific data were available. Further analyses were performed for patients who received anthracycline-based and anthracenedione-based chemotherapy regimens as initial therapy with the objective of investigating the impact of HPS and TCR phenotype on the CR rate.
When all patients were analyzed as a group, the presence of HPS and the expression of γ/δ TCR in tumor cells predicted inferior survival (Fig. 5). In contrast, age was not prognostic of poorer outcome (data not shown). The potential association between HPS and the type of TCR expressed by malignant cells also was investigated. Adequate information regarding HPS status and the specific type of TCR was present for 44 patients. HPS was equally common in patients with the α/β type (9 of 28 [32%]) and patients with the γ/δ type (4 of 16 [25%]) of SPTCL (P = 0.74). In the group of patients who received anthracycline-based or anthracenedione-based chemotherapy regimens, chemotherapy outcomes were similar regardless of HPS status or TCR phenotype. However, there was a trend toward poorer response in patients with γ/δ-type SPTCL. Eight of 18 patients (44%) with HPS achieved a CR, whereas 7 of 23 patients (30%) without HPS experienced remission (P = 0.36). A CR was achieved by 1 of 7 patients (14%) and by 6 of 11 patients (55%) who had tumors that expressed the γ/δ receptor and the α/β receptor, respectively (P = 0.15). Data were inadequate for analyzing the impact of TCR phenotype on patients who underwent HDT-SCT.
SPTCL is predominantly a disease that affects young adults, with a median age of onset that is approximately 2 decades earlier compared with the median ages of patients with other types of primary cutaneous T-cell lymphoma.67, 68 Although only ∼7% of normal T cells are of γ/δ lineage, up to 30% of all reported SPTCL lesions expressed the γ/δ receptor, a finding that is consistent with the overrepresentation of this phenotype in SPTCL.69 Two histopathologic features of SPTCL deserve further mention. We found that dermal involvement in SPTCL (28%) may be more frequent than was previously believed. However, in most patients, dermal invasion is minimal. Our findings are consistent with 2 large SPTCL series, in which dermal involvement incidence rates of 20–83% were reported.35, 58 On a clinical note, dermal involvement in primary cutaneous lymphoma may not be of prognostic significance.64 although there are case reports of the presence of EBV in patients with SPTCL; almost all of those patients resided in Asia, where EBV infection is common. In the rest of the world, EBV typically is not present in patients with SPTCL. This suggests that EBV does not play a pathogenic role in SPTCL. The natural history of SPTCL frequently takes an aggressive course, resulting in almost 50% disease-related mortality and a median survival of ∼2 years. Because of short follow-up periods for many of the documented patients, it is expected that the actual mortality rate is much higher. A smoldering disease pattern usually is transient and frequently terminates in more aggressive disease that requires systemic therapy. However, patients who experienced an indolent disease course over several years also have been reported.3, 53, 56 Spontaneous disease remission can occur but is extremely rare.
Radiation can produce long-term remissions in patients with localized disease, especially when HPS is absent. In addition, it also can be considered for palliative purposes in patients with disseminated SPTCL when chemotherapy options are exhausted or not tolerated. Of all the immunosuppressive agents, it has been shown that corticosteroids are the most effective. Notable responses are found in patients with limited skin lesions and absent systemic symptoms. Nevertheless, responses usually are short lived, with frequent disease recurrences after corticosteroid tapering. There are insufficient data for making any statements regarding the efficacy of biologic agents. However, because of the expected slow onset of response, it is possible that such agents should be considered only in patients who have SPTCL with a less aggressive biology or as adjunctive treatment.
Anthracycline-based combination regimens (in particular, CHOP) represent the most effective type of chemotherapy. Such regimens should be the option of choice when aggressive systemic treatment is considered. It is notable that ∼33% of patients will achieve long-term remission. Experience with purine nucleoside analogs (in particular, fludarabine), although limited, suggests that this class of agent may be useful in the treatment of SPTCL. Resistance to initial anthracycline-based therapy generally predicts a poor outcome, despite the use of second-line or salvage chemotherapy regimens. In contrast, a durable response is expected in most patients who attain a CR. The effectiveness of HDT-SCT in patients with SPTCL, as reported in the literature, is very impressive. Only ∼7% of patients had developed recurrent disease as of the most recent follow-up report. Consequently, HDT-SCT should be considered as soon as possible in patients who have refractory disease.
In terms of prognosis, the presence of HPS at diagnosis and the expression of γ/δ receptors by tumor cells are two factors that may be associated with inferior overall survival. Earlier case series suggested these findings as well.1, 12, 64, 70 However, we found that HPS was not seen more frequently in γ/δ-type SPTCL. There is a trend toward poorer response to anthracycline-based chemotherapy regimens in patients who have the γ/δ tumor phenotype. It is possible that γ/δ-phenotype cutaneous lymphoma, whether SPTCL or another entity, may be a distinct disease and therefore deserve its own classification.
Like any retrospective study with review of the literature, there are several factors that potentially may limit the usefulness of our data, such as relatively small sample sizes, inaccurate diagnoses, reporting biases, incomplete follow-up, and nonuniform interpretation of treatment responses. In addition, we may have overlooked some patients reported in the literature, especially those reported before the introduction of the term SPTCL. However, we made considerable efforts to reduce the impact of these limitations. To our knowledge, the current report includes an analysis of the largest number of patients with SPTCL reported to date in the literature. To minimize misdiagnoses, one of the authors (S.M.W.) reviewed all pathologic data, including photomicrographs, when provided, and confirmed diagnoses using World Health Organization criteria.66 Finally, to include all potential cases of SPTCL before its description as a distinct entity, we also reviewed reports of subcutaneous T-cell lymphoma associated with panniculitis published before 1990. However, SPTCL with a more indolent course may be underreported and, thus, underrepresented in this review.
In conclusion, the current study demonstrates that SPTCL has clinical features that are distinct from those of other primary cutaneous T-cell lymphomas. Evidence from the literature suggests that anthracycline-based chemotherapy regimens are the most effective form of initial treatment. HDT-SCT is an important option in the primary refractory or recurrent setting. However, even when this treatment strategy is used, long-term success is limited. The γ/δ phenotype is particularly aggressive, and early HDT-SCT can be explored for patients in the subgroup with this phenotype. Future investigations should explore more novel therapeutic agents that are active against T-cell malignancies, such as purine nucleoside analogs (fludarabine and pentostatin),71 monoclonal antibodies (alemtuzumab),72 and radioimmunoconjugates or exotoxin immunoconjugates (90Y-T101, denileukin diftitox, and anti-Tac[Fv]-PE38).70, 73, 74
The authors thank John P. Farnen, M.D., and Robert S. Witte, M.D. (Department of Internal Medicine, Gundersen Lutheran Medical Center), as well as Cathy L. Indenberg, M.S. (Gundersen Lutheran Medical Foundation), for reviewing the current article.
- 22Cutaneous angiocentric T cell lymphoma and subcutaneous panniculitic T cell lymphoma are distinct entities [abstract]. Mod Pathol. 1996; 9: 109A., , .
- 28Subcutaneous panniculitis-like T-cell lymphoma associated with a hemophagocytic syndrome in a patient with polycythemia vera and interferon alpha therapy [abstract]. J Cutan Pathol. 1998; 25: 508., .
- 47Successful treatment of aggressive, recurrent, subcutaneous panniculitis-like T cell lymphoma (SPTCL) with high-dose chemotherapy followed by allogeneic peripheral stem cell transplantation [abstract]. Blood. 2001; 98: 377b., , , , , .
- 66Subcutaneous panniculitis-like T-cell lymphoma. In: JaffeES, HarrisNL, SteinH, VardimanJW, editors. World Health Organization classification of tumors: pathology and genetics of tumors of hematopoietic and lymphoid tissues. Lyon: IARC Press, 2001: 212–215., .