Angioimmunoblastic T-Cell Lymphoma

Clinical and Laboratory Features at Diagnosis in 77 Patients

Abstract

We retrospectively analyzed 77 patients with pathologically diagnosed angioimmunoblastic T-cell lymphoma from a single city. There were 43 men and 34 women; the median age was 64.5 years (range, 30-91 yr). Average time between first symptoms of the disease and diagnosis was 3.6 months. At diagnosis, peripheral nodes were present in all but 1 patient, and were generalized in 90% of cases. Constitutional symptoms were reported in 77% of cases and spleen enlargement in 51%. A cutaneous eruption-morbilliform, urticarial, or more polymorphic-was present in 45% of patients; in one-third of them, the eruption occurred after drug administration. Other clinical manifestations included pleuritis (22%); arthralgia or arthritis (17%); ear, nose, and throat involvement (14%); central or peripheral neurologic manifestations (10%); and ascites (5%). Most patients presented with advanced disease at diagnosis (bone marrow involvement in 60% of cases).

The main laboratory abnormalities were elevated lactate dehydrogenase levels (71%), inflammatory syndrome (67%), hypergammaglobulinemia (50%), anemia (51%), and lymphopenia (52%). Auto- or disimmune manifestations were reported in one-third of patients: autoimmune hemolytic anemia was present at diagnosis in 19% of patients and thrombocytopenic purpura in 7%. Documented vasculitis was described in 12% of cases. Clonality was analyzed in lymph nodes in 47 patients: T-cell and B-cell clones were found in 45 (96%) and 20 (45%) patients, respectively. Chromosomal abnormalities were identified in 62% of cases: trisomies 3, 5, 18, 19, additional X chromosome, and deletion of chromosome 7 were the most common abnormalities. The current study underlines the diversity of presenting manifestations of angioimmunoblastic T-cell lymphoma.

Abbreviations: AITL = angioimmunoblastic T-cell lymphoma, AILD = angioimmunoblastic lymphadenopathy with dysproteinemia, CSF = cerebrospinal fluid, EBERs = Epstein-Barr virus-encoded small RNAs, ENT = ear, nose, and throat, HIV = human immunodeficiency virus, Ig = immunoglobulin, IL = interleukin, IPI = International Prognostic Index, LDH = lactate dehydrogenase, PCR = polymerase chain reaction, TCR = T-cell receptor.

INTRODUCTION

Angioimmunoblastic T-cell lymphoma (AITL), previously known as angioimmunoblastic lymphadenopathy with dysproteinemia (AILD)¹⁵, is one of the most common peripheral T-cell lymphomas, accounting for 18% of T-cell lymphomas cases and for 1.2% of all non-Hodgkin lymphomas⁴³. AITL is a systemic disease involving lymph nodes, spleen, and bone marrow. It is characterized by generalized lymphadenopathy, hepatosplenomegaly, fever, and skin rash; anemia, autoimmune features, and polyclonal hypergammaglobulinemia are frequently described^9,37,47.

Recent findings indicate that AITL is derived from centro-follicular T-helper cells^11,18,19. Morphologically, the lymph nodes show partial or total obliteration of the normal architecture by a polymorphic infiltrate of lymphocytes, immunoblasts, eosinophils, histiocytes, and plasma cells and by proliferation of follicular dendritic cells and of high endothelial venules^3,9,13,24,30. Most cases contain a monoclonal T-cell population^{3,30,33,48,53} as well as clonal cytogenetic abnormalities^17,27,31,45. AITL typically follows an aggressive clinical course; despite treatment with polychemotherapy, the prognosis is poor, with a 5-year overall survival of 26%-36% and a median survival of less than 3 years^30,38,49. The diagnosis of AITL may be difficult³, as the morphology and the clinical syndrome can overlap with a wide range of reactive lymphadenopathies, benign lymphoproliferative disorders, and neoplastic conditions such as Hodgkin disease and other non-Hodgkin lymphomas^6,13.

We conducted a retrospective study to describe the clinical and laboratory features, especially immunologic features, at diagnosis in a large series of patients with AITL. Because of the rarity of the disease, most clinical studies have been small, and to our knowledge no large clinical series has been reported in the recent literature. Moreover, most series include patients diagnosed in the early 1980s, before the availability of immunophenotyping and molecular tests, and therefore may include patients with reactive lymphadenopathies. Many patients with AITL are first referred to internal medicine departments because of the polymorphism of their initial symptoms or because of autoimmune manifestations. The diagnosis may be difficult, and as a result the mean time lapse between the presenting symptoms of AITL and the diagnosis is wide³⁷. We think that a better knowledge of the disease could lead to decreased time between first symptoms and diagnosis and to earlier and appropriate treatment.

PATIENTS AND METHODS

Case Selection and Histologic Analysis

We searched all files of the departments of hematopathology of the University Hospital Lyon Sud and of the Léon Bérard Center (Lyon, France) for cases diagnosed as AITL between January 1990 and August 2004. Eight patients were excluded from the study: 2 because of the absence of clinical data and 6 others with the diagnosis of marginal zone lymphoma (n = 1), autoimmune disease (n = 2), reactive lymphadenopathy (n = 2), or large T-cell lymphoma (n = 1). Seventy-seven consecutive patients were included in the study. The stopping date was March 1, 2005.

Diagnoses were made on a lymph node biopsy for 76 patients and on both bone marrow and liver biopsies for 1 patient presenting without peripheral lymph node involvement. Diagnoses were established morphologically and immunophenotypically relying on published common criteria^{9,24,25,30,42}. Immunologic studies were performed on paraffin sections and frozen tissues with anti-CD20, anti-CD3, anti-CD4, anti-CD8, anti-CD21, and anti-CD30 antibodies in all cases; expression of CD10 was analyzed in 28 cases. Lymph node architecture was effaced by dispersed and variegated cellular infiltrates of atypical lymphoid cells that expressed T-cell markers and often included characteristic clear cells. These were accompanied by a marked increased in arborizing venules, plasma cells, B-immunoblasts, and small lymphocytes, admixed with histiocytes and eosinophils (Figure 1). Characteristic proliferations of CD21+ follicular dendritic cell networks could be demonstrated outside follicles.

FIGURE 1: Histology of angioimmunoblastic T-cell lymphoma: nodal biopsy showing vascular proliferation and polymorphic infiltrate including small lymphocytes, plasmocytes, and immunoblasts (hematoxylin and eosin staining).

Cytologic Analysis of Peripheral Blood Smears

In 15 cases, peripheral blood smears (automatically prepared and May-Grünwald Giemsa stained by Sysmex SP-100 [Roche diagnostics, Meylan, France]) were evaluated for cytologic features including smear patterns and lymphoid cell characteristics.

In Situ Hybridization for Epstein-Barr Virus-Encoded Small RNAs (EBERs)

In situ hybridization for Epstein-Barr virus-encoded mRNAs was performed in 47 patients, using the Dako Epstein-Barr virus probe kit following the manufacturer's protocol.

Polymerase Chain Reaction (PCR) for T-cell Receptor γ-Chain Gene and Immunoglobulin Heavy Chain Gene

In this retrospective study, T-cell receptor (TCR) gene rearrangement and immunoglobulin (Ig) or B-cell receptor gene rearrangement were not systematically analyzed. Forty-seven patients were tested for TCR gene rearrangement and 44 for B-cell receptor gene rearrangement.

DNA was extracted from tissue using High Pure PCR Template Preparation Kit (Roche, Mannheim, Germany). TCR γ-chain gene rearrangements were studied using a GC-clamp multiple PCR-γ-DGGE procedure as previously described⁵¹, and the Ig gene rearrangements were studied by multiplex PCR using consensus primers from European BIOMED-2 concerted action⁵⁴.

Clinical Evaluation

We retrospectively reviewed the medical charts of these 77 patients. Data on the following parameters were collected at the time of diagnosis: sex, age, past medical history, presenting symptoms, interval between first symptoms and diagnosis, Eastern Cooperative Oncology Group (ECOG) performance status, constitutional symptoms (defined as fever >38.3 °C, night sweats, and/or weight loss >10% within the last 6 mo) and peripheral lymph node involvement. Extranodal manifestations were also noted, including hepatomegaly; splenomegaly; ascites; pleuritis; myalgia, arthralgia, or arthritis; ear, nose, and throat (ENT) involvement; lung involvement (cough, dyspnea); peripheral neuropathy or central nervous system involvement; cutaneous features; pruritus; skeletal involvement; and gastrointestinal manifestations.

Laboratory Evaluation

Data on the following parameters were collected: complete blood count, biochemical data (including serum sodium and creatinine levels, uricemia, albuminemia, lactate dehydrogenase [LDH] level and liver enzymes), C-reactive protein, Coombs test, serum protein electrophoresis with detection of M-component, autoantibodies, and other immunologic findings when available (cryoglobulinemia, circulating immune complexes, and complement levels).

Serum LDH was considered abnormal when it was higher than the maximal normal value. Autoimmune hemolytic anemia was diagnosed when the following criteria were associated: anemia, laboratory signs of hemolysis (hyperbilirubinemia, low haptoglobin, and/or elevated reticulocytes) and positive Coombs test or presence of cold agglutinins. Immune thrombocytopenic purpura was diagnosed when thrombocytopenia was associated with normal or raised megakaryocyte counts in bone marrow aspirate or biopsy and when there was no other disease or drug possibly responsible for thrombocytopenia (infection, liver disease, hemophagocytic syndrome).

All patients tested negative for human immunodeficiency virus (HIV), hepatitis B virus, and hepatitis C virus infections.

Initial Staging

Results of thoracic, abdominal, and pelvic computerized tomography scan and of bone marrow histologic analysis were noted in all patients. When available, data on cerebrospinal fluid (CSF) analysis and results of skin, liver, lung, or tonsil biopsies were collected. The Ann Arbor classification⁸ was used for initial staging.

The International Prognostic Index (IPI) was determined for each patient using the following parameters: age (≤60 vs. >60 yr), number of involved extranodal sites (≤1 vs. >1 site), LDH value (≤ normal value vs. > normal value), performance status (0 or 1 vs. 2-4), and stage of disease (I or II vs. III or IV)².

Cytogenetic Analysis

Cytogenetic studies were performed in 40 patients. In 35 patients, chromosomal analysis was performed on lymph node biopsy material, and in 5 patients, on bone marrow aspirate.

Cells were cultured as previously reported⁷. Metaphases were R- and G-banded with Giemsa and Wright stains, respectively. Karyotypes were described according to the International System for Human Cytogenetic Nomenclature²³. In 3 patients, additional fluorescent in situ hybridization (FISH) experiments using appropriate probes were performed.

Statistical Analysis

Survival probabilities were determined using the life-table algorithm according to the Kaplan and Meier method. Overall survival was calculated from the date of diagnosis to the date of death or last follow-up evaluation. Progression-free survival was defined from the date of diagnosis to that of any evidence of disease progression (during or after the first treatment), of death from any cause, or of last follow-up evaluation. Major clinical, laboratory, immunologic, and cytogenetic findings obtained at diagnosis were individually analyzed for their impact on overall survival and progression-free survival using the log-rank test.

RESULTS

Demographic Features

Seventy-seven patients, 43 men and 34 women (sex ratio: 1.26) were included in the current study (Table 1). The patients were mostly elderly: the mean age at the time of diagnosis was 64.5 years (range, 30-91 yr), and 64% of patients were over 60 years old (Figure 2). Two patients had a history of Hashimoto thyroiditis; no autoimmune disease was noted in the other patients' past medical history.

FIGURE 2: Age distribution of patients in the current study.

Presenting Manifestations

Most patients presented with a subacute history of constitutional symptoms, often associated with generalized lymphadenopathy, skin rash, and/or hepatosplenomegaly. Atypical presenting symptoms were reported in more than one-third of the patients. The initial manifestations leading to first investigations included autoimmune cytopenia (immunologic purpura, autoimmune hemolytic anemia, or Evans syndrome), vasculitis, peripheral neuropathy, Quincke edema, polyarthritis or polychondritis, toxidermia-like eruption, subcutaneous nodules, thoracic or abdominal pain, dyspnea, dysphagia, isolated fever, confusion, and solitary peripheral adenopathy.

The place of first referring was an internal medicine department for 46% of patients. The time lapse between the first symptoms and the histologic diagnosis ranged from less than 1 month to 36 months, with a median of 3.6 months (Figure 3).

FIGURE 3: Interval between first symptoms of the disease and diagnosis.

Clinical Manifestations

Clinical features at the time of diagnosis are summarized in Table 1.

TABLE 1: Clinical Features at Diagnosis of 77 Patients With AITL

Constitutional Symptoms

Constitutional symptoms were present in 77% of patients (fever: 57%, night sweats: 44%, and weight loss: 22%). Despite these symptoms, patients were ambulatory (performance status = 0 or 1) in 60% of cases. Pruritus was reported by 44% of patients.

Lymphadenopathy and Hepatosplenic Involvement

Except for 1 patient with exclusive liver and bone marrow involvement, all patients presented with peripheral lymphadenopathy. In 9% of patients, lymphadenopathy was limited to 1 lymph node group, whereas it was generalized to 2 or more lymph nodes groups in the remaining 91%. Cervical nodes were the most commonly involved sites. In most cases, lymph nodes measured 1-3 cm in diameter. Half of the patients had spleen enlargement, and one-quarter had hepatomegaly.

Cutaneous Involvement

A cutaneous eruption was present initially in 35 patients (45%). Skin lesions were polymorphous. A maculopapular morbilliform rash was the most commonly reported lesion. The other encountered aspects, isolated or in association, were urticarial eruption (n = 10), infiltrated and sometimes necrotic infiltrated purpura (n = 5), noninfiltrated purpura (n = 2), nodules (n = 3), erythroderma with secondary desquamation (n = 3), Quincke edema (n = 1), annular rash (n = 1), severe dermographism (n = 1), and infiltrated plaques of the breast (n = 1). Pruritus was reported by 57% of patients with cutaneous involvement.

In 11 patients (31% of patients with rash), the eruption occurred after drug administration. Incriminated drugs were penicillin A (n = 4), nonsteroidal antiinflammatory drugs (n = 3), and fluoroquinolones, macrolides, allopurinol, and third-generation cephalosporin (n = 1 each). The eruption never regressed with the disruption of the incriminated drug.

Musculoskeletal Features

Joint symptoms were described in 13 patients (17%). Among them, 1 patient presented with nonerosive, symmetric, and seronegative peripheral polyarthritis associated with polychondritis, and 1 with remitting seronegative symmetrical synovitis with pitting edema (RS3PE)-like polyarthritis. Polyarthralgia was reported by the 11 other patients. In 4 patients, articular involvement was associated with myalgia. No patient presented with bone pain; no skeletal involvement was described.

Neurologic Manifestations

Neurologic features were reported at diagnosis in 8 (10%) patients. Several symptoms, isolated or in association, were described, including confusion (n = 3), motor and sensory polyneuritis (n = 2), deafness or partial hearing loss (n = 2), apathy (n = 2), regressive hemiparesia (n = 2), cerebellar syndrome (n = 1), loss of vision (n = 1), aphasia (n = 1), and tinnitus (n = 1). In 2 patients these neurologic symptoms were associated with the presence of tumor cells in CSF, and in 1 patient, with lymphocytic meningoencephalitis.

Other Manifestations

ENT involvement was reported in 11 patients (14%). Among them, 9 patients presented with tonsillar hypertrophy, responsible for upper airway obstruction and dyspnea in 2; 1 patient presented with tonsillitis; and 1 with an oropharyngeal mass.

Cough and/or dyspnea was reported in 21 patients (27%). Uni- or bilateral pleuritis was present at diagnosis in 17 patients (22%). In 1 patient, acute respiratory distress syndrome was the inaugural manifestation of the disease. Ascites was present in 4 patients (5%); no other gastrointestinal symptoms were described.

Laboratory Features

Baseline serum chemistry values and blood cell counts are summarized in Table 2, and immunologic features in Table 3. The most consistent hematologic findings at the time of diagnosis were lymphopenia and anemia, reported in 52% and 51% of patients, respectively. Coombs test was positive in 58% of patients, and cold agglutinins were found in 9% of patients. Nevertheless, only 19% of patients had autoimmune hemolytic anemia. One-third of patients presented with hypereosinophilia and one-third with circulating atypical cells. Thrombocytopenia was present in 20% of patients and was related to immune thrombocytopenic purpura in 5 (7%) patients; 2 of them had Evans syndrome.

TABLE 2: Laboratory Features at Diagnosis of 77 Patients With AILT

TABLE 3: Immunologic Features at Diagnosis of 77 Patients With AILT

Elevated β2-microglobulin, elevated LDH, and inflammatory syndrome were the 3 most common biochemical findings (82%, 71%, and 67%, respectively). Mean LDH level was 1.8 times the upper limit of the normal value. Serum protein electrophoresis revealed hypergammaglobulinemia in 51% of patients and hypogammaglobulinemia in 10%; circulating monoclonal Ig, mostly of the IgG type, was present in 27% of patients.

Data on autoantibodies, cryoglobulinemia, complement, and circulating immune complexes were not available for all patients. Several autoantibodies could be identified at diagnosis in our patients: antinuclear antibodies, antismooth muscle antibodies, antithyroperoxidase antibodies, antiphospholipid antibodies, and antiENA antibodies. Cryoglobulinemia, mostly type II mixed cryoglobulinemia, was found in one-half of the tested patients. Since the use of these immunologic tests may have varied according to the patient's symptoms and to the place of initial evaluation, our results may not be representative of the whole population, and therefore no percentages were determined.

Lastly, a hemophagocytic syndrome proven by bone marrow aspiration was present in 3 (4%) patients; all of them had multiple organ failure, severe hyperferritinemia, hypertriglyceridemia, and pancytopenia.

Initial Staging

Mediastinal and retroperitoneal lymphadenopathy was present in 72% and 65% of patients, respectively. One patient presented with a bulky disease (nodal mass' greatest dimension >10 cm) and 9 with 1 or more nodes >5 cm; in other cases, lymph nodes measured 1-3 cm in diameter. Bone marrow was involved at diagnosis in 60% of cases. In 6% of patients, the lymphoid infiltrate was associated with myelofibrosis.

Skin biopsy was performed in 16 patients. In 8, skin specimens showed a histologic picture of vasculitis. A perivascular infiltrate of atypical lymphocytes was present in 3 patients. One biopsy revealed typical histologic aspects of acute generalized exanthematous pustulosis, and the last 2 biopsies were normal. A lymphomatous infiltrate was also present in 1 liver biopsy, in 3 ENT biopsies, in the CSF of 4 patients, and in the spleen of a splenectomized patient.

At diagnosis, 2 (3%) and 4 (5%) patients, respectively, were classified as having stage I and II disease, while 21 (27%) were classified as having stage III, and 50 (65%) stage IV. When the patients were grouped by risk factor as defined by the IPI, 13% of them belonged to the low-risk group (IPI = 1), 23% to the low-to-intermediate-risk group (IPI = 2), 22% to the high-to-intermediate-risk group (IPI = 3), and 42% to the high-risk group (IPI = 4 or 5).

Histologic Initial "Misdiagnosis"

For 28 patients (36%), the disease first diagnosed based on histologic samples was not AITL but reactive lymphadenopathy (16%) or another subtype of lymphoma (19%) (Table 4). Among them, transformed marginal zone lymphoma and transformed lymphoplasmacytic lymphoma were the most common. The diagnosis of AITL was established secondarily by reviewing the initial node biopsy because of clinical or laboratory atypical features or at relapse. The median time between the misdiagnosis and the final diagnosis was 2.3 months, but ranged from 0.4 to 29 months.

TABLE 4: Suspected Histologic Diagnosis on First Lymph Node Biopsy

Clonality and EBERs Analysis

Clonality analyses were mostly performed on nodal biopsies. TCRγ gene rearrangement was monoclonal in 39 (83%) of 47 tested cases and oligoclonal in 6 (13%) cases. Eighteen (41%) of the 44 cases tested for IgH gene rearrangement also showed evidence of a monoclonal B-cell population, and 2 (5%) showed evidence of an oligoclonal B-cell population. A monoclonal T-cell population was found in 1 skin biopsy, 1 CSF sample, and 1 spleen.

EBV RNAs were detected by in situ hybridization in 28 (60%) of 47 nodal samples.

Cytologic Analysis of Peripheral Blood Smears

In all studied cases but 2, cytologic analysis of peripheral blood smears showed a polymorphic lymphoid population. This population could contain a minority of small to large atypical cells, immunoblasts, lymphoplasmacytic cells, large granular lymphocytes, and sometimes plasma cells. Lymphopenia was common. A pronounced rouleaux-formation was present in all cases with hypergammaglobulinemia. Immunocytochemical analysis was not systematically performed; it revealed CD10 positive circulating neoplastic T-cells in 1 case.

Cytogenetic Findings

Clonal chromosomal aberrations were found in 62.5% of analyzed patients (Table 5). Trisomy 5, trisomy 3, trisomy 19, rearrangements involving the location of TCR genes, and an additional X chromosome, alone or in association, were the most frequent cytogenetic abnormalities detected in the current series (Figure 4). Cytogenetically unrelated clones were reported in 2 patients (5%).

TABLE 5: Cytogenetic Findings

FIGURE 4: Example of karyotype. Cell from lymph node of a patient; R-Banding. Karyotype: 48,XY, +X, ?add (1) (q43), +2, add (3) (p23), -4, +5, -6, add (7) (p14), add (9) (p23-24), +add (9), -13, -14, -15, -17, +mar.

Outcome and Prognostic Factors

At the time of the analysis, 46 patients (60%) had died. The actuarial overall survival at 3 years was 49% (confidence interval [CI], 37%-60%), and the progression-free survival at 3 years was 31% (CI, 20%-46%) with a median follow-up of 2.96 years (Figures 5A and 5B).

FIGURE 5: Outcome of the 77 patients with AITL. A. Overall survival; B. progression-free survival.

Factors significantly associated with shorter overall survival were respectively performance status >1 (p = .01), ascites (p = .002), thrombocytopenia (p = .002), hypoalbuminemia (p = .022), renal failure (p = .036), presence of circulating atypical cells (p = .040), hyperuricemia (p = .046), and presence of an additional X chromosome (p = .041). Progression-free survival was negatively influenced by performance status (p = .013), generalized (vs. localized) lymphadenopathy (p = .041), thrombocytopenia (p = .015), and hypoalbuminemia (p = .042). The negative influence of the presence of an additional X chromosome on progression-free survival was of borderline significance (p = 0.061). The outcome of patients with an initial misdiagnosis was not significantly different from that of patients with no misdiagnosis. Finally, overall and progression-free survival were strongly improved in patients obtaining a complete response after the first-line treatment (p = .0001); nevertheless, no plateau was apparent on patients' survival curves (data not shown).

DISCUSSION

AITL is a rare disease⁴³. Therefore, relatively little is known about its clinical and laboratory presentation and about the autoimmune phenomena reported in the disease. To our knowledge, the current series of 77 patients is the largest clinical report of AITL cases in the literature. It differs from the other series on 2 main points. First, we included only patients diagnosed after 1990 who had immunophenotyping and molecular tests for diagnosis, whereas the previous series included patients diagnosed before 1986 and patients eventually found to have reactive lymphadenopathy. Second, the present patients were selected from the files of departments of hematopathology in a single town and not from the files of departments of hematology or oncology as was the case in previous studies. This enabled us to include patients too old or too severe to be referred to hemato-oncologists.

The current series represents a valuable supplement to our knowledge of the presenting manifestations of AITL and emphasizes the polymorphism of the disease. The main findings are the following. AITL is a disease of the elderly: the mean age was 64.5 years in the current series, and no patient was younger than 30 years. Peripheral lymphadenopathy was present in all but 1 patient, and lymph nodes were commonly small (1-3 cm). Thus, the diagnosis should be considered by physicians facing aged patients with lymphadenopathy, especially when it is associated with general symptoms, rash, and/or autoimmune manifestations. A lymph node biopsy should then be performed, even if those nodes are small. As bone marrow was infiltrated in 60% of our cases, we think that bone marrow aspirate or bone marrow biopsy should also be performed when the presenting manifestations are suggestive of the diagnosis. Although most patients presented with polyadenopathy and general symptoms, more than one-third of them had atypical presenting symptoms such as immune cytopenia, neuropathy, polyarthritis, or vasculitis. These symptoms could be suggestive of autoimmune, infectious, or allergic disease, which can explain the frequency of admission in departments of internal medicine and the significant delay often observed between the first symptoms of the disease and the diagnosis (median interval in the current study, 3.6 mo). We think that a better knowledge of the disease could lead to decreased time between first symptoms and diagnosis and to earlier and appropriate treatment.

The current study also underlines the polymorphism of cutaneous manifestations, analyzed in detail in 35 patients. To our knowledge the largest previous series of AITL patients with cutaneous involvement was reported by Martel et al³⁵ and included only 10 patients; most other cases are described in isolated case reports. Our study confirmed that skin lesions have no specific clinical and histologic patterns^34,35,46. Most patients presented with a generalized morbilliform maculopapular rash that could mimic a viral rash or drug hypersensitivity. Nevertheless, we encountered numerous other clinical features, such as urticarial lesions (much more frequent in the current study than in the literature), infiltrative purpura, nodules, or erythroderma, that may also alert physicians to perform a skin biopsy. In the current study, as in previous series, the onset of the disease coincided with drug exposure in nearly one-third of the patients with rash^16,35,53. Antibiotics (amoxicillin) and nonsteroidal antiinflammatory drugs were the drugs more often involved, but eruptions were described under a wide range of drugs. The triggering place of drugs in this disease is well described, but the mechanisms directing this phenomenon are poorly studied and understood. By analogy to the immunologic phenomena described in the DRESS syndrome (Drug Rash with Eosinophilia and Systemic Symptoms), we suppose that a large variety of cytokines and inflammatory cells are involved in these pseudoallergic manifestations. In AITL, the immune system is substantially activated and an array of cytokines, such as interferon gamma, interleukin (IL)-6, IL-12, and tumor necrosis factor-α, are up-regulated and therefore elevated in the sera of patients^14,36,50,57. Because of this immune system dysregulation, we suppose that small amounts of antigens are sufficient to initiate an inflammatory reaction. By analogy to what is observed during HIV infection, we can postulate that the T immune deficiency associated with AITL may also favor these systemic manifestations^40,41. In some patients, we observed an interval, up to 6 months long, between drug-related rash and occurrence of AITL. This emphasizes the need for follow-up of patients with toxidermia, especially when the rash doesn't regress after stopping the incriminated drugs.

We compared our findings with those of the recent literature^37,38,47,49 (Table 6). The results of the current study are in agreement with those reports. Nevertheless, our patients are older than patients in the previous studies; this can be explained by the differences in patient recruitment. Furthermore, ascites and hepatomegaly were less frequently observed among our patients than among the cases in the literature, whereas pruritus, circulating monoclonal immunoglobulin, and positive Coombs test were most commonly observed.

TABLE 6: Features of Patients With AITL, Present and Previous Reports

Our results underline the difficulty of histologic diagnosis in AITL⁹. In the present series, the disease first diagnosed was not AITL but reactive lymphadenopathy or another subtype of lymphoma in 36% of patients. The difficulty in establishing the diagnosis of AITL has been reported in the literature, especially in the early phase of the disease³, as histologic presentation may be suggestive of autoimmune or infectious disease or drug hypersensitivity^6,9,13. These difficulties probably contribute to the delay observed between the first symptoms of the disease and the diagnosis. Although this delay may cause difficulties for patients and physicians, it does not appear to influence patient outcome. CD10 is a phenotypic marker that specifically identifies the tumor cells in 90% of AITL and distinguishes AITL from other peripheral T-cell lymphomas^3-5; this marker should greatly assist in the diagnosis of this disorder¹¹. The expanded CD21+ follicular dendritic cells meshwork, sometimes subtle, is helpful too in making the diagnosis^26,32. The current study confirmed the high frequency of monoclonal or oligoclonal T-cell population in AITL, as well as the high incidence of detection of EBERs in this disease and the frequency of clonal chromosomal aberrations^1,29,56. We identified many patients with trisomy 18 or 19 or with rearrangement involving the loci of TCR genes; these abnormalities were less frequent in the literature series, in which trisomy 3, 5, and additional X chromosome were the most common cytogenetic abnormalities^28,45. Searching for T and B clonality, cytogenetic abnormalities, and Epstein-Barr virus expression seems useful for the diagnosis of AITL, which we think emphasizes the mandatory exchange between clinicians and pathologists. Even if the precise diagnosis of AITL is achieved after biopsy and histologic examination of 1 of the enlarged nodes, our results suggest that the careful cytologic analysis of peripheral blood smears may also provide valuable help in diagnosis before more invasive investigations. Recent data from the literature showed that identification of circulating CD10 neoplastic T cells by multicolor flow cytometry may also contribute to diagnosis⁵. In some cases however, repeated lymph node biopsies may be required to reach the correct diagnosis.

To our knowledge, the current study is the first to analyze precisely autoimmune phenomena at diagnosis in patients with AITL; 19% of our patients had autoimmune hemolytic anemia, 7% had immune thrombocytopenic purpura, and 12% had well-documented vasculitis. Other possibly disimmune manifestations such as peripheral neuropathy, polychondritis, or paraneoplastic encephalitis were reported in 6% of patients. Overall, symptomatic autoimmune or disimmune diseases were present at diagnosis in 34% of the patients studied. This percentage is considerably higher than those observed in other lymphoma entities with the exception of splenic marginal zone lymphoma^10,52. For example, in the cohort with non-Hodgkin lymphoma reported by Gronbaek in 1995²⁰, immune cytopenia was described in 4.8% of patients (vs. 25% in the current study). Our results suggest that the association between autoimmune manifestations and AITL is not fortuitous. Nevertheless, the mechanisms directing these autoimmune phenomena are poorly understood. Current data from the literature suggest that autoimmunity in AITL is multifactorial. The process of autoantibody production may first be the result of known defects in the apoptotic pathways such as Fas/FasL⁵⁸. Second, neoplastic T cells may promote antibody production by B cells through excessive cytokine release; the role played by IL-6, which is overexpressed in AITL patients^22,57, in the T- and B-cell interactions observed in AITL may be essential. These abnormalities in the T- and B-cell interaction may also contribute to the generation of a potentially autoreactive monoclonal B-cell population^21,48. Finally, the disturbance of the T cellular immunity associated with AITL (as attested to by the high incidence of fatal infectious complications in this disease and by functional tests^37,41,47) may also promote autoimmunity by favoring persistent antigen stimulation and impairing clearance of immune complexes and control of autoreactive cells. A characteristic feature of AITL is the presence of increased numbers of Epstein-Barr virus-infected cells^1,3,55; we can suppose that this chronic infection of B lymphocytes by Epstein-Barr virus may play a role in the generation of autoimmune phenomena in this disease as in others³⁹, but no data are available up to now. To test some of these pathogenetic hypotheses, we studied the links between autoimmune manifestations and several laboratory features in our patients. We found no significant link between lymphopenia and autoimmune manifestations, between the presence of a monoclonal B-cell population and autoimmunity, between the presence of circulating monoclonal immunoglobulin and autoimmune manifestations, nor between the presence of EBERs and autoimmune features (data not shown). Additional studies are required to analyze the mechanism of autoimmunity in AITL; this could also contribute to a better understanding of the relationship between autoimmunity and lymphoid neoplasia^12,44.

In conclusion, the current study underlines the diversity of presenting manifestations of AITL and emphasizes the distinctive clinical, histologic, immunologic, molecular genetic, and cytogenetic features of this T-cell neoplasm. This disease is also characterized by poor outcome. It significantly differs from other non-Hodgkin lymphomas by the age of patients and by the high incidence of constitutional symptoms, skin rash, systemic manifestations, and autoimmune phenomena.

ACKNOWLEDGMENTS

We thank the following physicians who contributed to the study: Lucile Baseggio, Amine Belhabri, Michel Blanc, Pascale Cony-Makhoul, Bernadette Coron, Philippe Dubourdeau, Pascale Felman, Maya Hacini, Roch Houot, Jérôme Jaubert, Olivier Lejeune, François Maréchal, Serge Martinon, Mauricette Michallet, Jacques Ninet, Michel Pavic, Pierre-Yves Péaud, Jean Roche, Catherine Sebban, Didier Sicard, Jean-Jacques Sotto, Catherine Thieblemont, Catherine Traullé, Denis Vital-Durand, and Nicolas Voirin. We particularly thank Marie-Hélène Delfau-Larue (Hôpital Henri Mondor, Créteil, France) for her contribution to molecular analysis.