Zebras and hen's teeth: recognition and management of rare T and NK lymphomas

The peripheral T-cell lymphomas (PTCLs) represent 10%-15% of non-Hodgkin lymphomas (NHLs) and are composed of >20 different entities. Many of our current treatment strategies are borrowed from approaches used in aggressive B-cell lymphomas, phase 2 studies, and retrospective case series. As the most common entities, PTCL, not otherwise specified (NOS), angioimmunoblastic T-cell lymphoma (AITL), and ALK-negative anaplastic large-cell lymphoma (ALCL), account for >60% of cases, the majority of our data is derived from these subsets. In making clinical decisions for the even more rare entities, we must rely upon our interpretation of the best data available and our own experience. In this section, we will focus of some of the most rare T-cell lymphomas that may be encountered by practicing physicians: enteropathy-associated T-cell lymphoma, (EATL), hepatosplenic T-cell lymphoma, (HSTCL), extranodal NK/T-cell lymphoma, (ENKTL), and subcutaneous panniculitis-like T-cell lymphoma (SPTL). We will focus on recognition of these entities and existing data regarding initial management with an emphasis on how the approaches are similar to, or divergent from, the ways we treat the more common subtypes of PTCL.

Enteropathy-associated T-cell lymphoma (types I and II) is a mature T-cell lymphoma that may present anywhere within the gastrointestinal tract but most commonly the jejunum or ileum. Presentation in the large intestine is rare, especially in EATL type I. EATL accounts for 6%-9% of all PTCL in North America and Europe with type I being highly associated with celiac disease and type II occurring sporadically.^1,2  Up to 2% of patients with celiac disease are reported to eventually develop EATL, although recent estimates of celiac disease prevalence make that number seem high.³  The highest risk is for those with refractory celiac disease not responsive to a gluten free diet. Occasionally a diagnosis of EATL may be the presenting feature of celiac disease.

The 2 types of EATL are distinguished primarily based on the specific immunophenotype: EATL type 1 characteristically is CD3+ and CD4−, CD8−, and CD56−, often with some expression of CD30. EATL type 2 typically is CD3+, CD4−, CD8+, and shows a cytotoxic phenotype including common expression of CD56 and TCRγδ.⁴ 

Retrospective series of patients with EATL report a very poor prognosis with median survival in the range of 7 months.^5,6  Many patients with EATL present with acute abdominal symptoms often requiring urgent or emergent surgical procedures to arrive at the diagnosis⁷  resulting in a high fraction of patients beginning therapy with poor performance status or too ill to receive chemotherapy. This was reflected in a series from the United Kingdom where 9/31 (29%) patients were either too sick to receive chemotherapy or died from complications of the first cycle of chemotherapy, mostly from bleeding or infection.⁵  The importance of patient fitness and tumor burden in outcome is further confirmed by the International T-cell lymphoma Project (ITCP) where poor performance status and high tumor burden correlated with a significantly worse outcome.¹  If we use ability to receive anthracycline-containing combination chemotherapy as a practical surrogate for treatment with curative intent, we see in the ITCP that ∼20% of those who received anthracycline containing chemotherapy were alive and failure-free at 2 years, whereas none among those who could not receive such therapy were alive without progression at 1 year. It is especially important to consider this subset of very unfit patients when looking at outcomes from prospective studies or series of more aggressive treatment strategies for which these ill patients would likely not have been eligible.

Among such alternate strategies are clues that more intensive therapies may lead to better outcomes. The European Society for Blood and Marrow Transplantation (EBMT) reported results on 44 subjects who received autologous stem cell transplant (ASCT) as a consolidation (n = 31) or salvage (n = 13) therapy. Although all subjects had to have adequate response and performance status to undergo ASCT, the overall survival curves for those who underwent ASCT as consolidation in first complete remission (CR) or partial remission (PR) was 66% at 4 years.⁸  The largest prospective multicenter trial of an intensified approach is the Nordic study where patients with major subtypes of PTCL were treated with CHOP with etoposide (CHOEP) followed by ASCT. Among the 160 subjects treated, 21 had EATL. On an intent to treat basis, those with EATL fared similarly to those with PTCL-NOS and AITL with a 5 year progression-free survival (PFS) and overall survival (OS) of 38% and 48%, respectively. A very intensive upfront regimen IVE (ifosfamide, epirubicin, etoposide)-methotrexate followed by ASCT (IVE/MTX-ASCT) was piloted by the Scotland and Newcastle Lymphoma Group.⁷  Importantly, eligibility criteria included de novo EATL and ability to tolerate high-dose treatment as determined by a multidisciplinary meeting. When the IVE/MTX-ASCT patients were compared with historical controls, they had similar median age of 56 years versus 57 years, and similar impaired performance status (ECOG 2-4) of 77% versus 88%. With all the caveats of historical controls and patient selection, the more aggressive regimen appeared to give better results. Compared with the control patients who received an anthracycline containing regimen, patients treated with IVE/MTX-ASCT had a higher overall response rate ORR; 69% vs 42%), higher 5 year PFS (52% vs 22%), and higher 5 year OS (60% vs 22%).

Although the data is limited, it appears that EATL type II carries a poorer prognosis.⁹  Given that EATL type II shares features with NK/T-cell lymphomas and other cytotoxic T-cell lymphomas, alternate strategies, such as L-asparaginase containing regimens have been tried.⁹  However, although there is some logic here, the results are not clearly superior and the experience is too small to recommend this routinely.

There are 2 indolent lymphoproliferative processes of the GI tract that can at times be confused with EATL: indolent T-cell lymphoproliferative disease of the gastrointestinal tract and NK enteropathy.¹⁰  Neither of these disorders typically requires or responds to systemic therapy and both carry a very indolent or benign course.

Among the uncommon T-cell lymphomas, discussed here, EATL is the one we treat most like the more common forms of PTCL. Although in EATL, as in other subtypes of PTCL, CHOP often leads to suboptimal results, there is evidence, albeit inconclusive evidence, that more aggressive approaches do result in better long-term outcomes. As such, we tend to follow the largest phase 2, Nordic Trial, as a treatment paradigm and use CHOEP followed by ASCT as our most commonly used approach. We are intrigued by the results of IVE/MTX-ASCT and at least consider it in very young and fit patients.

Hepatosplenic T-cell lymphoma is a very rare lymphoma, representing 2%-3% of all T-cell lymphomas, typically infiltrating the spleen, liver and bone marrow. Patients frequently present with cytopenias, B symptoms, and elevated LDH.¹¹  In the majority of cases the malignant cells are gamma/delta expressing T cells however an alpha/beta phenotype has been reported.¹²  They express CD3 and are usually negative for CD4 and CD8.¹³  Clonal rearrangement of the TCR-γ gene is usually present, and many cases show an isochromosome 7q [I(7)(q10)].¹⁴  HSTCL typically occurs in young men and is strongly associated with immunosuppression such as occurs with anti-tumor necrosis factor-α or thiopurine agents used for inflammatory bowel disease or other immunosuppression post-organ transplant.^15,16 

HSTCL has an aggressive course, with a median survival of 16 months.¹⁷  The optimal therapy for HSTCL is not known. In initial reports, the majority of subjects received CHOP with poor results. In one series including 21 subjects; 19 were treated with a CHOP-like regimen, none of whom were alive at 4 years from diagnosis.¹⁷  The only 2 surviving patients were induced with a cisplatin-ara-c based regimen followed by ASCT. Based on the overwhelmingly poor results with CHOP, we have adopted a strategy of a non-CHOP induction, such as ifosfamide, carboplatin, etoposide (ICE), or ifosfamide, etoposide, cytarabine with intrathecal methotrexate (IVAC), followed promptly by hematopoietic stem cell transplant (HSCT), either allogeneic or autologous, in CR1.¹⁸  When we analyzed our results, we found a 50% OS at 5 years among 14 patients treated. The results were even better with a 63% OS among those treated with ICE or IVAC as initial therapy with the goal of HSCT in CR1. Although these are very small numbers from which to draw any definitive conclusions, the relatively high rate of surviving patients does stand in distinction to several series in the literature with alternative approaches.¹⁹  The potential importance of allogeneic stem cell transplantation in HSTCL was further supported by a recent EBMT analysis of 25 patients with HSTCL. Among the 18 who received an allogeneic graft, the 3 year PFS was 48% compared to only 29% for those receiving autologous stem cells.²⁰ 

One potential mimic of HSTCL is a variant of large granular lymphocytic leukemia (LGL) with an often aggressive behavior.²¹  Although LGL typically expresses an alpha/beta T cell receptor (TCR), there may be clinical overlap between these entities and it is important to be as certain as possible as to the diagnosis of HSTCL when applying such aggressive approaches to initial therapy.

As above, HSTCL is too rare and the literature too sparse to be dogmatic about a standard approach. However, it does appear that standard CHOP alone rarely, if ever, results in a durable reemission. As many of these patients are young, a dose-intensified approach is usually feasible and at least in our experience, regimens such as ICE or IVAC for 3 or 4 cycles followed by HSCT have resulted in the majority of our patients surviving. We have long-term survivors who were consolidated with either allogeneic or autologous stem cells, but based on the transplant literature; we do prefer allogeneic transplantation when donor cells are available.

Extranodal NK/T-cell lymphoma, nasal type (ENKTL) is another uncommon subtype of T-cell lymphoma. ENKTL occurs worldwide, with a strong geographic predilection for Asian populations. In the ITCP, ENKTL represents 4%-5% of T-cell lymphomas in North America and Europe but 22% in Asia. ENKTLs are almost exclusively extranodal with the initial presentation most frequently in the nose or nasopharynx and occasionally the paranasal sinuses, tonsil, Waldeyer's ring, and oropharynx. ENKTL may also involve the skin, salivary glands, testis, and gastrointestinal tract either as primary sites or as dissemination from a nasal primary. Local lymph nodes may also be involved. With imaging, such as PET/CT, or careful otolaryngologic evaluation with biopsies, many of the non-nasal presentations are found to be associated with occult nasal primaries.²²  Of note, up to 3% of cases can be associated with hemophagocytosis.²³  Predominantly in patients from South America, it has been reported that ENKTL may be preceded by a diagnosis of hydroa vacciniforme-like lymphoma.²⁴ 

The histopathology shows angiocentric malignant lymphoid cells with vascular invasion and tissue necrosis. The malignant cells express CD2 but may be negative for other T-cell markers. They typically express CD16, CD56, cytoplasmic CD3ε, and may or may not demonstrate clonal rearrangements in the T-cell receptor genes.²⁵  Neoplastic cells are invariably infected by Epstein-Barr virus (EBV) which is detected most reliably by in situ hybridization (ISH) for EBV-encoded RNA (EBER), and many pathologists will not make this diagnosis in an EBV negative tumor.²⁶  EBV DNA PCR measured in plasma has been found to correlate with tumor burden and serial EBV PCR monitoring is useful for assessing responses and screening for disease recurrence.²⁷ 

Like many of the lymphomas discussed here, commonly used chemotherapy regimens, such as CHOP, are less effective in ENKTL and these lymphomas require a disease specific treatment approach. For patients with stage I or II disease with extranodal extension disease confined to or primarily in the nasopharynx, radiation therapy is an essential part of curative therapy. In older studies of CHOP in combination with radiation therapy, CHOP did not appear to improve PFS or OS when compared to radiation alone.^28,29  Some centers, particularly outside the United States, recommend radiation therapy alone for early stage disease.³⁰  However, in studies of radiation therapy alone, systemic relapse rates are as high as 25%-40%.^31,32  With the advent of newer, more effective, disease specific chemotherapy regimens, it now appears that combined modality therapy with concurrent or sequential chemoradiation provides the best outcomes for localized NK/T-cell lymphoma. With regards to the preferred chemotherapy regimen in patients with localized ENKTL, there are multiple options with no standard of care. Long-term follow-up of patients treated with DeVIC (dexamethasone, etoposide, ifosfamide, and carboplatin) in combination with radiation therapy has shown an ORR of ∼80% and 5 year OS of 70%.^33,34  Similarly, VIPD (etoposide, ifosfamide, cisplatin, dexamethasone) has been studied with concurrent radiation therapy with 100% ORR and 85% OS and PFS at 3 years.³⁵  L-asparaginase was found to have significant single agent activity in relapsed NK/T-cell lymphomas with an ORR of 82%,³⁶  and has now been incorporated into multiple other regimens.^37,38  One intensive L-asparaginase containing regimen, SMILE (dexamethasone, methotrexate, ifosfamide, l-asparaginase, and etoposide), in combination with radiation demonstrated an 82% ORR with a complete response (CR) rate of 78%.³⁹  Although radiation is essential for those with localized disease, the optimal sequencing of radiation therapy with chemotherapy remains undefined as some advocate concurrent chemoradiation, while others suggest consolidation with radiation therapy or even radiation followed by chemotherapy.

In the setting of patients with disseminated and advanced stage disease, combination chemotherapy remains the standard treatment.²⁵  Here again it appears that L-asparaginase-containing regimens have particular efficacy. Studies of SMILE demonstrate an ORR of 25%-80%,^39,40  whereas L-asparaginase combinations with methotrexate and dexamethasone have an ORR of 78%.⁴¹  Despite the improved efficacy of L-asparaginase based regimens, long-term remissions are rare in patients with advanced stage disease following combination chemotherapy. However, retrospective analyses suggest a benefit to consolidation with ASCT in CR.^42,43  Similarly for allogenic transplant, retrospective analyses show that in patients who are able to achieve a CR, consolidation with allogeneic transplant is a potentially curative option with 5 year OS reported as high as 55%.⁴⁴ 

For NK/T-cell lymphomas, it is clear that a disease specific approach is required. Our approach has been to use a short course of a modification of SMILE followed by radiation. With this approach, most patients with early stage disease received 2 cycles with a CR rate of 88%. This high rate of local disease control has allowed our radiation oncologists to treat the nasopharynx with lower doses of radiation (45 cGy) then the higher doses used with radiotherapy alone. Although this remains a quite aggressive chemotherapy regimen, in our small experience (n = 10), the outcomes have been promising with a PFS at 3 years of 100%. However, we believe that any of the newer combined modality approaches, such as DeVIC, VIPD, or more recent L-asparaginase/gemcitabine combinations with the inclusion of radiotherapy, are appropriate choices. Radiation alone remains an option for patients with localized disease; however, the rates of progression outside the field from some of the larger studies and the high doses of radiation used make us favor, in general, a combined modality approach. The optimal treatment for advanced stage disease truly remains to be defined. We favor combination chemotherapy with one of the regimens above. In our experience, SMILE has been effective at inducing remissions but almost all relapse. The data is too limited to recommend a standard approach; however, we do consider consolidation with HSCT for those patients achieving adequate disease control.

Subcutaneous panniculitis-like T-cell lymphoma is a rare primary cutaneous T-cell lymphoma (CTCL) that makes up <1% of T-cell lymphomas.⁴⁵  SPTCL presents with subcutaneous nodules that may regress and remit.^46-48  The lesions consist of atypical lymphoid cells expressing CD8 and CD56 that rim individual adipocytes with associated reactive histiocytes often associated with coagulation necrosis. Histologically, the cells express an alpha/beta phenotype. The gamma/delta expressing phenotype of this disease is now classified as cutaneous gamma/delta T-cell lymphoma and tends to carry a more aggressive clinical course.^12,48 

Therapy for SPTCL remains controversial. Our recent understanding that SPTCL usually follows an indolent course and carries a favorable prognosis, particularly when hemophagocytic lymphohistiocytosis is not present, has led to approaches more commonly used for other CTCLs. Historically, responses to combination chemotherapy have been reported but are usually of short duration and CRs are rare.^49,50  There are reports of successful allogeneic stem cell transplantation^51-53  and prolonged remissions with combination chemotherapy followed by an autologous stem cell transplant.⁵⁴  However, the rarity of this disease and the infrequent need for aggressive therapy hampers our understanding of the utility of such approaches. Importantly, as with other indolent lymphomas, relapses often have not correlated with shortened survival. Milder approaches include single-agent bexarotene, which has significant clinical activity with an ORR of 82% including CRs.⁵⁵  Other systemic agents as used for CTCL, such as oral methotrexate and histone deacetylase inhibitors have activity. Anecdotal responses to glucocorticoids, interferon-α, zidovudine, and cyclosporine also have been reported.^56-60  The use of denileukin diftitox in two patients has been reported with evidence of activity.⁶¹ 

In contrast to SPTCL alpha/beta, cutaneous gamma/delta T-cell lymphoma usually presents with ulcerating skin lesions and is more frequently associated with hemophagocytic lymphohistiocytosis. In the largest multicenter retrospective series of this disorder, the median survival was 31 months but others have cited median survivals of ∼1 year.^62,63  More recently, as our ability to perform gamma staining by immunohistochemistry has become more reliable, it has been observed that there may be a less aggressive form of cutaneous gamma/delta T-cell lymphoma that has histologic features of SPTCL.^64-66 

Given the poor prognosis, aggressiveness, and poor responsiveness to conventional-dose treatments, we and others prefer to manage most patients with aggressive cutaneous gamma/delta lymphomas patients similarly to other cytotoxic T-cell lymphomas (eg, HSTCL).

It has been our approach to treat those with SPTCL or clinically indolent cutaneous gamma/delta T-cell lymphoma with therapies that may be less aggressive but more tolerable for long periods of time. This strategy mimics what we often do for CTCL as opposed to how we approach those with aggressive systemic PTCL. Therefore, we prefer initial therapy with either bexarotene or oral methotrexate. At times, observation alone may be appropriate. We consider combination chemotherapy, often adding HSCT consolidation with curative intent, in physically fit patients who continue to progress on these and other milder therapies, show a more aggressive disease course, and/or develop HLH.

Rare diseases are always a challenge. Randomized trials are nonexistent and our management decisions must be extrapolated and applied from very imperfect data. However, patients frequently need treatment when they develop a disease, not when the field generates high-level evidence. Although a strict algorithm cannot be written, through the literature that exists and clinical experience themes and principles do emerge that, when coupled with clinical judgment, allow reasonable and logical decisions. Although targeted or immunotherapy is not standardly available for these rare diseases, better understanding of the biology of these rare lymphomas may lead to the application of these new classes of drugs.

Steven Horwitz, Department of Medicine, Memorial Sloan Kettering Cancer Center, 1233 York Ave, New York, NY 10065; Phone: 212-639-3045; Fax: 646-422-2164; e-mail: horwitzs@mskcc.org.