Abstract

Given their relative rarity, one of the primary diagnostic difficulties in nodal T-cell lymphomas is recognizing their range of histologic patterns. This is complicated by the fact that most mature T-cell lymphomas retain some functional characteristics of nonneoplastic T cells, ie, the capacity to secrete cytokines and costimulate immune cell growth, and, thus, are associated with obscuring nonneoplastic immune cells. Sessions 2 and 3 of the Society for Hematopathology/European Association for Haematopathology Workshop focused on these issues and conditions that may simulate T-cell lymphomas. We summarize salient features of presented cases, including the varied patterns seen in angioimmunoblastic T-cell lymphoma (AITL) and other more poorly characterized morphologic and functional nodal T-cell lymphoma subsets. Many cases illustrated the difficulties distinguishing AITL from peripheral T-cell lymphoma, unspecified, when the neoplasms manifest only some AITL features. The usefulness of separately classifying T-cell lymphomas that demonstrate follicular, perifollicular, or T-zone patterns of infiltration; significance of immunophenotypically distinct subsets that express cytotoxic markers or have features of central memory T cells; diagnostic difficulties posed by B-cell proliferations that accompany T-cell lymphomas; and T-cell lymphoma mimics related to genetic disorders, immune dysregulation, and drug reactions are also discussed.

The World Health Organization classification recognizes only 1 specific type of predominantly nodal T-cell lymphoma, namely angioimmunoblastic T-cell lymphoma (AITL, also commonly abbreviated as AILT) in addition to neoplasms such as anaplastic large cell lymphoma (ALCL) that can manifest at nodal and extranodal sites. The remaining cases of nodal T-cell lymphoma are grouped into the peripheral T-cell lymphoma, unspecified (PTCL-U) category. We discuss the diagnostic boundaries of AITL and whether there is specific evidence to support other morphologic or immunophenotypic subsets of nodal T-cell lymphoma.

What Are the Histologic Patterns Seen in AITL?

Likely owing to the functional nature of the neoplastic T cells, AITL shows the most variability among all T-cell tumors in the density of nonneoplastic cells, including cases with a predominance of B cells, CD4+ or CD8+ nonneoplastic T cells, dendritic cells, histiocytes, and even plasma cells. In fact, in a substantial number of cases, it remains difficult to definitively identify the neoplastic components.

In the Workshop, there were 8 diffuse cases of AITL involving lymph nodes. These cases manifested a diffuse pattern of infiltration Image 1 with some preservation of sinuses and prominent proliferative high endothelial venules (HEVs) (Image 1), often in an arborizing pattern. The background varied from mainly lymphocytic to polymorphous, including eosinophils (extensive in 40% of cases), histiocytes (scattered in about 40% of cases), and plasma cells (usually polyclonal but can be monoclonal and admixed with the lymphoma or manifest as a tumor mass/plasmacytoma).^[1,2] Histiocyte-rich areas, particularly when accompanied by Langhans-type giant cells, can suggest granulomatous disease or be mistaken for T-cell/histiocyte-rich large B-cell lymphoma when accompanied by a large B-cell proliferation (see "B-Cell and Plasma Cell Proliferations in Nodal T-Cell Lymphoma") (case 68).^[3] Only a subset of cases manifests sufficient atypia in small lymphocytes or has sufficient numbers of atypical medium to large cells (often with clear cytoplasm) to provide morphologic evidence to support the diagnosis of lymphoma. Similarly, only a subset of cases has abnormal T-cell antigen loss by flow cytometric or immunohistochemical analysis to support a diagnosis of AITL. Follicular dendritic cell (FDC) markers and germinal center T-cell markers are the most helpful for diagnosis (see "How Helpful Are FDC Markers [eg, CD21 and CD23] in the Diagnosis of AITL?" and "What Is the Role of Germinal Center T-Cell Markers [eg, CD10, bcl-6, and CXCL13] in the Diagnosis of AITL?"). Some cases of AITL will show a prominent population of CD30+ large cells that may raise concern for Hodgkin lymphoma (see "Hodgkin-like Proliferations in AITL and PTCL-U") or ALCL; cases that raise the possibility of ALCL will not show the extensive, strong CD30 expression typical of ALCL and will lack expression of anaplastic lymphoma kinase (ALK).

To provide support for a presumptive diagnosis of lymphoma, many cases of AITL will require gene rearrangement studies. However, as many as 20% of cases will not show a T-cell clone when studied by DNA-based polymerase chain reaction primers on paraffin-embedded tissue samples using the primers developed by the BIOMED 2 initiative (InVivoScribe Technologies, San Diego, CA).^[4] A double-labeling study showed that the inability to identify a clone in cases of AITL correlates with the extent of the host response, ie, cases that lacked a clone had a lower percentage of proliferating CD4+ lymphoma cells and a more exuberant host response.^[5] In addition, about a third of cases will manifest clonal immunoglobulin gene rearrangement.^[4] Some but not all of these cases with clonal immunoglobulin gene rearrangements will have an associated Epstein-Barr virus (EBV)+ or EBV– B-cell proliferation (see "How Often Are These B-Cell Proliferations Associated With EBV?"). In contrast with the consistent cytogenetic abnormalities seen in most cells in other malignancies, AITL shows a high frequency of unrelated clones and single cell aberrations with completely different karyotypes.^[6] Nevertheless, the constellation of clinical, histologic, immunophenotypic, and molecular findings will generally allow a firm diagnosis of AITL. Although double-labeling studies suggest that most if not all AITL cases derive from CD4+ T cells,^[5] only a subset of cases will show an abnormal CD4+ large cell population. In contrast, many cases will show an admixture of CD4+ and CD8+ small cells that makes definitive identification of the tumor population difficult without more sophisticated studies (eg, polymerase chain reaction–based T-cell receptor [TCR]β spectrotyping followed by studies with β-chain variable region segment-specific antibodies).^[7]

Although AITL essentially always involves lymph nodes, it can manifest at a number of extranodal sites. This can occasionally present difficulties in diagnosis, particularly with skin involvement by AITL, which is often subtle and associated with clinical symptoms such as drug hypersensitivity or a full body rash that may suggest Sézary syndrome. Case 188 presented such a scenario. A 63-year-old man had a 7-month history of pruritic violaceous papules over the posterior part of the neck, back, and upper arms that comprised lymphoma but simulated an allergic reaction; a computed tomography scan revealed bilateral axillary and inguinal lymphadenopathy, and a lymph node biopsy specimen showed involvement by typical AITL.

How Helpful Are FDC Markers (eg, CD21 and CD23) in the Diagnosis of AITL?

An extensive proliferation of FDCs outside of follicles that surrounds the proliferating HEVs Image 2 provides strong support for the diagnosis of AITL.^[8,9] This finding is most often present in lymph nodes but may be seen in extra-nodal sites such as skin (case 188). However, in some bona fide cases of AITL, this extrafollicular proliferation of FDCs involved only a portion of the lymph node (eg, case 29); other cases in the Workshop were merely suspected to represent AITL because the extrafollicular proliferation of FDCs was only focal (eg, cases 152 and 158). One case that did not show this FDC pattern (case 135) was acknowledged to represent AITL. In addition, in some cases of AITL, 1 FDC marker (CD21) may show this diagnostic pattern, whereas another (CD23) will not (case 68). Another case (case 159) comprising the early phase of AITL had a normal pattern of CD23 labeling.

At present, the most sensitive marker for extrafollicular FDCs in the setting of AITL seems to be CD21.^[10] In practice, we look for an abnormal FDC pattern in each case that we suspect AITL, but when an abnormal pattern is only present focally, we generally request molecular studies for diagnosis. If T-cell lymphoma is supported by molecular findings in such a case, a diagnosis of PTCL with features suggestive of AITL can be issued, rather than an unequivocal diagnosis of AITL. Of course, the expression of germinal center T-cell markers (see next section) in such a problematic case may obviate the need for molecular studies and clarify the type of T-cell lymphoma.

A more subtle pattern of proliferating FDCs that can be seen in AITL, particularly in early follicular cases, is the presence of irregular, ragged proliferations at the margins of the follicles Image 3, in contrast with the normal well-developed and well-circumscribed FDC meshwork of normal secondary follicles. Sometimes these abnormal proliferations take the form of "snouts" or "sprouts."^[11]

What Is the Role of Germinal Center T-Cell Markers (eg, CD10, bcl-6, and CXCL13) in the Diagnosis of AITL?

The expression of germinal center T-cell markers, such as CD10, bcl-6, or CXCL13, generally will provide support for an AITL diagnosis.^[11-14] However, as there are rare T-cell lymphomas that are distinct from AITL but seem to be derived from germinal center type T cells (see "Are the Follicular or Perifollicular Variants of T-Cell Lymphoma Distinct From AITL?"), the correct histologic picture of AITL is required, usually with the abnormal FDC proliferation. Not all otherwise typical examples of AITL express CD10 (about 60%-90%, possibly dependent on the presence or absence of an accompanying B-cell proliferation^[11,15]). Among the 8 bona fide cases in the Workshop, 5 showed typical expression of CD10, 1 showed weak and focal staining, 1 demonstrated labeling that was regarded as equivocal, and 1 case lacked CD10 labeling. Most important, only a subset of presumptive lymphoma cells typically expresses this marker. Support for an AITL diagnosis is readily obtained when scattered large lymphoma cells in the paracortex express CD10 Image 4; support is less certain when there is only a mild or focal increase in labeled cells that are not large or abnormal. One must make certain that the paracortical labeling for CD10 is in lymphoid cells and not in granulocytes, monocytes/histiocytes, or stromal cells. Of course, it is not practical to microdissect the CD10-labeled cells and show that they possess the same T-cell gene rearrangement to support the diagnosis of AITL, as done by Attygalle et al.^[11]

One pattern of CD10 labeling that may be more common in early involvement is the labeling of scattered lymphoid cells in the mantle zones and beyond; ie, the CD10+ cells are not confined to the germinal center but are closer to the germinal center than in other cases Image 5. The abnormal labeling patterns for bcl-6 and CXCL13 are similar to those of CD10. One may see labeling of more cells for bcl-6 than for CD10, but a higher degree of labeling of other T cells may complicate the interpretation of this marker. Furthermore, bcl-6 is less sensitive than CD10 in the experience of some observers. A recent report suggests that CXCL13 Image 6 may be preferable to CD10 or bcl-6 in this diagnostic problem area, but experience with this new marker is still limited.^[14,16]

CD10 and presumably other markers of germinal center T cells are more often expressed in lymph nodes involved by AITL than in other extranodal sites of involvement.^[12] Nevertheless, these markers can be a useful adjunct to diagnosis in extranodal sites (eg, case 188 that manifested in skin) and in some instances may label the neoplastic T cells in the bone marrow.

Is It Possible to Make an Unequivocal Diagnosis of AITL in Its Earliest Stages?

AITL is particularly difficult to identify when it manifests with a pronounced follicular hyperplasia pattern Image 7 (cases 152 and 159).^[11,17] Often there will be partial effacement of the architecture and marked hypervascularity, suggesting the possibility of AITL. Some follicles will show extension of the lymphoma to the edge of the germinal centers Image 8 with an obliteration of follicle mantle zones that can be highlighted by B-cell markers such as CD20, CD79a, and PAX5. The perifollicular sinus has also been shown to be prominent in early stages of AITL, and a role in determining the complex immunoarchitecture of AITL has been proposed.^[18] Sometimes the follicles will show increased histiocytes or decreased tingible body macrophages as a manifestation of partial infiltration by the T-cell lymphoma (also highlighted by B-cell markers). However, increased tingible body macrophages outside of these germinal centers has also been described.^[17] Thus, there are cases of AITL that raise the differential diagnosis with follicular lymphoma; germinal center lymphocyte markers that label the B and T cells in this compartment can be useful in the diagnosis of both (for AITL, see previous section).

What Are the Boundaries of AITL vs PTCL-U?

As discussed, a number of PTCL lymphomas may show some but not all features of AITL; and even when a feature is present, it may not be sufficiently well developed to provide convincing support for a diagnosis of AITL. In addition, there are cases that display all of the histologic and immunophenotypic features of AITL but occur in extranodal locations, or the patients lack the clinical syndrome of AITL. It is difficult to know whether such cases belong in the AITL category or in PTCL-U. There were at least 6 cases in the Workshop that illustrate this problem, and several are summarized herein.

For example, case 152 occurred in a 30-year-old man with striking weight loss and progressive neurologic abnormalities manifested by weakness, spasticity, and profound muscle atrophy resulting in the patient being wheelchair-bound. Some observers favored an early follicular phase of AITL. There were focal collections of extrafollicular FDCs, focal CD10 labeling of extrafollicular lymphocytes, and a clonal TCR gene rearrangement. There was an associated striking histiocytosis with hemophagocytosis. The patient's signs and symptoms responded dramatically to high-dose steroids. Because of the unusual constellation of features in this case, some observers even questioned a malignant diagnosis.

The lymphoma in case 51 showed abnormal FDC proliferation, and the cells expressed CD4, CD10, and bcl-6. However, the cells also expressed CD56 and CD57, and the disease was localized to the tonsil. In case 138, there was a proliferation of extrafollicular FDCs, but the lymphoma was composed of pleomorphic large T cells that showed extensive CD30 expression and anomalous expression of CD20. The lymphoma cells in case 147 were CD4+CD10+ T cells, and HEVs were prominent, but there was no extrafollicular proliferation of FDCs.

Are There Recognizable Subgroups of PTCL-U?

Not surprisingly, the PTCL-U cases presented at the Workshop comprised a heterogeneous collection of lymphomas. Some with certain distinctive features will be discussed under subsequent headings. Some of the others will be discussed briefly here. Case 219 was composed of large T cells that lacked granzyme expression and showed a striking interfollicular pattern of infiltration. Case 245 was a CD8+ lymphoma with a multifocal epithelioid histiocytic reaction (so-called Lennert lymphoma).^[19] Case 17 occurred in a 48-year-old man with a several-year history of apparent porphy-ria cutanea tarda and dermatopathic lymphadenopathy. The subsequent nodal lymphoma was perifollicular in distribution and weakly expressed CD10. Case 40 occurred in a 70-year-old woman with a 9-year history of Hashimoto thyroiditis. The lymphoma in the thyroid mimicked mucosa-associated lymphoid tissue lymphoma, was composed of cells intermediate in size that expressed CD4, and was accompanied by bone marrow and peripheral blood involvement.

Are the Follicular or Perifollicular Variants of T-Cell Lymphoma Distinct From AITL?

Case 11 occurred in a 60-year-old woman with a history of rheumatoid arthritis, peptic ulcer disease, fever, weight loss, and diffuse lymphadenopathy. The lymph node biopsy specimen showed a neoplasm with a nodular-follicular growth pattern composed of somewhat atypical small and large CD4+ lymphocytes that also expressed bcl-6 but not CD10. Similar cases of so-called follicular T-cell lymphoma are reported in the literature.^[20,21] Although it was concluded that case 11 represented a rare follicular T-cell lymphoma, the presence of FDC snouts and prominent HEVs raised the alternative possibility of the follicular phase of AITL, albeit with an unusual intrafollicular localization.

Case 105 occurred in a 66-year-old woman with systemic lupus erythematosus, asthma, and cervical lymphadenopathy. Waxing and waning lymphadenopathy in the face of prednisone and hydroxychloroquine therapy for 7 months was followed by a lymph node biopsy that showed follicles replaced or surrounded by medium to large T cells with clear cytoplasm Image 9 that expressed CD4 and CD10 Image 10. In addition, embedded in the proliferation were small B cells that expressed B-cell markers Image 11 and CD23 Image 12, but not CD5. The clonal T cells lacked CD7 expression and showed a complex karyotype; the small B cells were polyclonal (? expanded mantle zones).

Immunophenotypic Variants of T-Cell Lymphoma

Given the wide variation in individual cases and the changes in pattern during the disease course, there is a limit to the usefulness of the pattern of infiltration in classifying nodal T-cell lymphoma. In this section, we consider whether an immunophenotypic marker (or panel of markers) can be used to subclassify PTCL, not otherwise specified into relevant subgroups.

Does Expression of Cytotoxic Markers Characterize a Subset of PTCL, Not Otherwise Specified?

Recent evidence suggests that patients whose nodal lymphomas express cytotoxic markers (TIA-1 or granzyme) and show CD4+, CD8+, or CD4–CD8– immunophenotypes have a worse clinical outcome than those whose T-cell lymphomas lack cytotoxic marker expression.^[22] There were 4 cases with this immunophenotype that showed variability in other features.

Case 9 occurred in a 61-year-old man with a mass involving the base of tongue and hypopharyngeal soft tissue. The lymphoma was composed of small irregular lymphocytes (CD8+, TIA-1+, and granzyme+ with aberrant expression of CD20) that infiltrated in an interfollicular and diffuse pattern and associated with many HEVs.

Case 151 occurred in a 55-year-old woman with B symptoms, generalized lymphadenopathy, splenomegaly, and pancytopenia. The interfollicular infiltrate in the lymph node comprised large cells Image 13 associated with prominent apoptosis Image 14 and many HEVs. The lymphoma cells expressed CD8, TIA-1, granzyme Image 15, and perforin.

Case 187 occurred in an 80-year-old woman with a left groin mass. The lymphoma cells were intermediate to large, occurred in a mixed inflammatory background, and expressed CD4, CD30, TIA-1, and granzyme.

Case 20 occurred in a 53-year-old man with an enlarged inguinal lymph node. The lymphoma cells were small and large, infiltrated in an interfollicular and diffuse pattern, lacked expression of T-subset markers, expressed granzyme, and weakly expressed CD20. The lack of T-subset marker expression and the lack of expression of ab TCR raised the possibility of a gd T-cell lymphoma. This lymphoma was also associated with prominent apoptosis and associated tingible body macrophages.

Can PTCL-U Be Immunophenotypically Typed According to the Central Memory T-Cell Phenotype?

Case 148 was 1 of a series of 8 PTCL-U cases that were studied by double labeling with various T-cell markers in combination with lymphoma-specific TCRb V-region antibodies.^[7] This case was 1 of 5 that expressed CD45RO, CD27, CCR7, and bcl-2. This immunophenotype suggested a differentiation stage corresponding to the central memory subset of antigen-experienced T cells. This stage differs from that of AITL and ALCL. The interfollicular pattern comprising small and large cells corresponds to the homing pattern of the putative normal T-cell counterpart.

What Is the Significance of Expression of CD20, CD15, or Cyclin D1 in T-Cell Lymphoma?

A number of the cases submitted to the Workshop showed unexpected patterns of cell marker expression as described. Five cases of PTCL-U showed expression of CD20 on the lymphoma cells. Such expression has previously been described, eg, in a PTCL composed of small and large cells accompanied by an epithelioid histiocytic reaction.^[23] Blakolmer et al^[24] also described CD20 expression in a PTCL-U involving skin. These authors also described CD79a expression in 4 extranodal T-cell lymphomas, 3 enteropathy-type and 1 nasal NK/T type. CD20- and CD79a-expressing cases also expressed cytotoxic molecules. Of the 4 CD20+ nodal T-cell lymphomas that were tested for cytotoxic molecules, 2 were positive.

In a review of CD15, 71 (21.5%) of 331 T-cell lymphomas were identified that expressed CD15.^[25] Case 149 is an example of this phenomenon that occurred in a 31-year-old man with an enlarged inguinal lymph node. The biopsy showed a T-cell lymphoma composed of medium to large lymphoid cells with irregular folded to lobulated nuclei, 1 or 2 nucleoli, and pale cytoplasm accompanied by histiocytes. Although the cells showed extensive labeling for CD30 and CD15 and lacked expression of CD45, raising the possibility of Hodgkin lymphoma, the cells also expressed CD3, CD4, and CD43. Rearrangements of the TCRB and TCRG genes were also identified.

Yatabe et al^[26] reported 1 of 152 T/NK/null-cell lymphomas expressing cyclin D1; this case was described as an ALCL of null immunophenotype that expressed granzyme and lacked ALK expression. Case 120 is an example of this phenomenon in a 65-year-old man with bulky retroperitoneal lymphadenopathy. At surgery, a jejunal mass was resected and a mesenteric lymph node biopsy was performed. In the jejunum, the lymphoma was composed of atypical small lymphocytes; in the lymph node, the cells were of intermediate size with occasional large cells. The neoplastic CD4+ T cells also partially expressed CD30 but not ALK and lacked cytotoxic markers. Fluorescence in situ hybridization did not identify t(11;14) nor evidence of ALK translocation but did identify 3 to 4 copies of 2p23.

B-Cell and Plasma Cell Proliferations in Nodal T-Cell Lymphoma

Proliferations of B cells, particularly large B cells, may complicate AITL and PTCL-U.^[27-29] There were 12 cases in this Workshop. By using 25% large B cells as a criterion for this phenomenon, a French study found an incidence of 18% in AITL.^[29] Without using a firm percentage cutoff, a recent study at Stanford, Stanford, CA, found an incidence of this phenomenon of 21% in AITL and 19% in PTCL-U.^[4] In a given patient, these B-cell and plasma cell proliferations may be more pronounced in some areas of a biopsy specimen than in others or at different sites at the same or different times. The combined B- and T-cell proliferations may persist in staging specimens and/or recurrences, but there are exceptions (eg, case 29 after thalidomide treatment). Although a CD20 stain often highlights this phenomenon and may suggest the diagnosis of B-cell lymphoma, this phenomenon will be even more prevalent if a marker that additionally labels plasmacytoid lymphocytes and plasma cells, such as CD79a, is used. Moreover, κ and λ immunostains may highlight more large B cells than CD20 in these cases, presumably because of a plasmacytoid component. In situ hybridization for κ and λ RNA generally labels only the plasma cell component in such cases. Recently, monoclonal plasmacytoid lymphocyte and plasma cell proliferations in the setting of AITL and PTCL-U have been described.^[2] Posttransplant lymphoproliferative disorders of the T-cell type also may be accompanied by B-cell proliferations (eg, case 177).

What Are the Criteria for Large B-Cell Lymphoma or Plasmacytoma in the Setting of Peripheral T-Cell Lymphoma?

This question generated considerable discussion at the Workshop. There was general agreement that sheets of monoclonal large B cells can be regarded as diffuse large B-cell lymphoma in this setting. Less than sheet-like growths of monoclonal large B cells or sheet-like growths of polyclo-nal B cells shown by immunohistochemical analysis might be better regarded as atypical B-cell proliferations of uncertain clinical significance rather than B-cell lymphoma. Similarly, monoclonal plasmacytoid lymphocyte or plasma cell atypical proliferations that are admixed with T-cell lymphoma cells might be regarded as proliferations, whereas monoclonal plasma cell proliferations that form tumor masses can be regarded as plasma cell neoplasms or plasmacytomas. Molecular findings should not be used to determine whether a large B-cell proliferation fulfills criteria for lymphoma because a significant percentage of AITL and PTCL-U cases will have a B-cell clone in the absence of a recognizable B-cell proliferation (30% in AITL and 33% in PTCL-U) as shown in a recent study.^[4] Approximately 50% of the B-cell proliferations in the setting of AITL and PTCL-U will carry a clone as shown by molecular studies, a much higher percentage than shown by immunophenotyping. Many of these findings parallel those that are seen with B-cell proliferations in the posttransplantation or immuno-deficiency settings.

What about the cases in the Workshop? Two cases were regarded as having large B-cell lymphoma, 7 were regarded as having a B-cell proliferation, 1 was regarded as being questionable for lymphoma, and 1 raised the differential diagnosis with large B-cell lymphoma and Hodgkin lymphoma. Two cases showed a proliferation of small B cells that expressed CD23 but not CD5. One also should keep in mind that a small B-cell proliferation in the setting of T-cell lymphoma may represent involvement by chronic lymphocytic leukemia/small lymphocytic lymphoma (eg, case 113).

How Often Are These B-Cell Proliferations Associated With EBV?

In the limited literature on this topic, approximately 70% of B-cell proliferations in the setting of AITL or PTCL-U are positive for EBV, usually shown by in situ hybridization for EBV small-encoded RNA (EBER).^[27,29-31] In the Workshop, 9 (69%) of 13 cases showed labeling for EBER. Although combined small and large B-cell proliferations may show EBER, predominant small B-cell proliferations seldom show EBER. For example, the CD23+ small B-cell proliferation in perifollicular T-cell lymphoma (case 105) did not show EBER. However, the CD23+ small B-cell proliferation in case 51 was associated with increased EBER-labeled cells (? EBER-labeled subset of CD23+ B cells or the expected "baseline" increase that is seen in examples of AITL).^[30] Although monoclonal plasma cell proliferations in this setting generally do not show EBER, rare cases also may be EBV+.^[2,28] Bona fide lymphomas that arise in this setting have been reported to show latency types 2 and 3.^[28]

In AITL, What Can Thalidomide Treatment Tell Us About Patterns of Disease and Associated B-Cell Proliferations?

Case 29 occurred in a 78-year-old woman with a 6-month history of generalized lymphadenopathy and weight loss. A lymph node biopsy confirmed AITL. When the patient refused standard treatment options, she agreed to be treated with thalidomide as a single agent.^[16] The initial biopsy specimen showed typical features of AITL but accompanied by a marked EBV+ large B-cell proliferation. A biopsy specimen after 8 weeks of thalidomide treatment showed lower grade cytology of the lymphoma cells that were now localized to follicles rather than present diffusely, an altered vascular architecture (less arborized), and the virtual disappearance of the EBV+ large B cells. A complete remission occurred after 18 months. It has been suggested that the broad immunomodulatory properties of thalidomide might be advantageous in AITL in which immune deregulation rather than aggressive tumor growth may be the major clinical problem. Clinical studies have shown the efficacy of another immunomodulatory agent, cyclosporine.^[32]

Hodgkin-like Proliferations in AITL and PTCL-U

Three submitted cases that raised the possibility of Hodgkin lymphoma are described briefly. Case 217 occurred in a 71-year-old man with generalized lymphadenopathy. His lymph node biopsy specimen showed features of AITL, including an expanded and disrupted pattern of CD23+ FDCs. In addition, there were clusters of large lymphoma cells with clear cytoplasm, some of which surrounded Hodgkin-type cells Image 16. The clear cells expressed T-cell markers and CD10 Image 17 but no Hodgkin-associated markers, whereas the Hodgkin-type cells lacked T-cell markers and CD10 but expressed CD15, CD30 (Image 17), fascin, and CD20 (weak). The Hodgkin-type cells did not label for EBER.

Case 227 occurred in a 51-year-old man with cervical lymphadenopathy. The biopsy specimen showed features suggestive but not diagnostic of AITL, ie, focal extrafollicular FDCs but no definite CD10 expression on lymphoma cells. A clonal TCR gene rearrangement was identified. The Hodgkin-type cells expressed CD15, CD20, CD30, and EBV latent membrane protein type 1. Such cases can easily be confused with Hodgkin lymphoma.^[33]

Case 24 occurred in a 65-year-old man with right femoral lymphadenopathy that was sampled. A month later, a mediastinal lymph node was removed and the histologic findings were similar. Both lymph nodes showed a vaguely nodular process composed mainly of slightly irregular lymphocytes admixed with atypical large lymphocytes as well as rare Reed-Sternberg–like cells. A PTCL-U diagnosis was supported by a lack of surface CD3 and CD7 on the small cell population, whereas the large cells expressed CD15, CD30, CD4, and PAX5 (partial) but not CD20. The cells did not label for EBER. Prior reports have described this rare occurrence.^[33,34] These reports have not described the subsequent development of Hodgkin lymphoma.

The background in these Workshop cases ranged from mildly atypical with or without immunophenotypic abnormalities to clearly malignant, whereas the Hodgkin population showed CD15 and CD30 with or without CD20. Each of the 3 cases described expressed at least 1 B-lineage marker, more in keeping with Hodgkin lymphoma than Hodgkin-like cells in a T-cell lymphoma. Some but not all such cases are associated with EBV.

Reactive, Preneoplastic, and Neoplastic Disorders That May Simulate PTCL-U

A subset of cases in Session 3 was dedicated to the theme of reactive or preneoplastic processes mimicking T-cell lymphoma. Illustrative cases can be grouped into atypical proliferations in the setting of immune dysfunction and pathologic mimics of T-cell lymphomas. All of these cases highlight the overarching theme of atypical proliferations that more or less resemble T-cell lymphoma owing to a combination of loss or disturbance of architecture, mild cellular atypia with morphologic or immunophenotypic findings suggesting a T-cell process, and ancillary studies such as clonal TCR gene rearrangement. Knowledge of the features of these specific entities and their pitfalls, familiarity with the clinical setting (particularly drug exposures and history of immunodeficiency), and strict adherence to diagnostic features for T-cell lymphomas are required to avoid an erroneous diagnosis of T-cell lymphoma.

What Are the Features of T-Cell Lymphoproliferative Disorders That Arise in the Setting of Immune Dysfunction?

Autoimmune Lymphoproliferative Syndrome. Case 18 was an example of autoimmune lymphoproliferative syndrome (ALPS) in a 2-year-old boy. ALPS is a lymphoproliferative syndrome due to defects in the apoptosis pathway, including mutations in CD95, CD95L, and caspases 8 and 10. ALPS is characterized by parafollicular and interfollicular expansion in lymph nodes with interspersed reactive follicles.^[35,36] Marked splenomegaly is common, and patients are at increased risk of developing lymphoma. Patients have been reported to be at greater risk of developing non-Hodgkin (B- and T-cell) and Hodgkin lymphoma (partic ularly nodular lymphocyte predominant Hodgkin lymphoma), 15 times and approximately 50 times, respectively. The lymphomas occur in adulthood most commonly but can occur in childhood.^[37]

The histopathologic features of ALPS are illustrated in case 18 Image 18. Diagnostic difficulties may occur on examination of the cytologic features of the interfollicular areas. Blast-like cells of varying size can be seen with many mitoses, suggesting a neoplastic proliferation. The cells in ALPS are T cells with a particular immunophenotype (CD4–/CD8–CD45RO–CD45RA+).^[38] TIA-1 and CD57 are usually expressed. These cells are also seen in the spleen, expanding the red and white pulp. Plasma cells are increased but are polytypic. Preservation of overall nodal architectural integrity, the immunophenotype, and lack of a dominant lymphoid clone by molecular methods help distinguish this peculiar proliferation from overt malignancy.

Common Variable Immunodeficiency. Case 58 represented an atypical lymphoid proliferation in the setting of common variable immunodeficiency (CVID). Patients with CVID are known to be at risk for follicular hyperplasia, atypical lymphoid hyperplasia, and lymphomas. The hyperplasias occur more frequently than lymphomas. The lymphomas are usu ally non-Hodgkin lymphomas such as diffuse large B-cell lymphoma; however, other lymphomas, including Hodgkin lymphomas and T-cell lymphomas, may occur.^[39-42] Lately, mucosa-associated lymphoid tissue lymphomas have also been recognized and may have been included as atypical hyperplasias in older studies.^[43,44] The risk of lymphoma may be increased as much as 30-fold,^[45] but more recent studies quote a slightly lower figure of 12- to 15-fold.^[46] The submitted retroperitoneal biopsy specimen showed an abnormal proliferation of small irregular lymphocytes with sclerosis Image 19 and Image 20 that occurred in the setting of a monoclonal T-cell population in peripheral blood and bone marrow. This case also highlights the practice of the expert panel to refrain from establishing the diagnosis of lymphoma without unequivocal histologic evidence of lymphoma, despite demonstration of a clone and involvement of more than 1 site, in the setting of CVID.

Chronic Active EBV. Case 96 was an unusual case of chronic active EBV infection. This case could be considered a pathogen-specific type of immune dysfunction because the patient seemed to have a specific inability to clear an EBV infection that most other people handle without adverse effects. Cases of EBV+ T-cell proliferations in chronic active EBV infection have been described previously.^[47-49] The patient was otherwise healthy but experienced several episodes of pancytopenia associated with splenomegaly. The patient subsequently died of progressive cytopenias and hepatic failure. Persistently elevated EBV titers were demonstrated during life, and pathologic features included a transient clonal T-cell proliferation involving bone marrow and spleen and hemophagocytosis Image 21, Image 22, Image 23, and Image 24. The lymphoproliferative disorder was rather subtle, and clinical suspicion of chronic active EBV infection assisted in interpreting the pathologic features. Despite the lack of overt lymphomatous features at various times during the disease, the final outcome was fatal.

EBV+ T-cell lymphomas can occur as a fulminant lymphoproliferative disorder in the setting of primary acute infection and are most commonly seen in patients of Asian descent, particularly in children.^[50-52] Of note, recent studies suggest mutations in the perforin gene as a specific mechanism for susceptibility to chronic active EBV infection and hemophagocytosis.^[53]

What Are Other Pathologic Mimics of T-Cell Lymphoma?

Phenytoin-Associated Lymphoproliferative Disorder. Case 61 highlights the importance of clinical history. This patient had a phenytoin-associated clonal T-cell lymphoproliferative disorder. It was characterized by a diffuse but heterogeneous proliferation of T cells, including immunoblasts Image 25. Immunostaining demonstrated that the atypical lymphocytes had a T-cell immunophenotype Image 26, and gene rearrangement studies showed the presence of a monoclonal T-cell population. In patients taking antiepileptic drugs such as phenytoin (but not limited to it), "pseudolymphomatous" proliferations are known to develop in sites such as lymph node and skin. The latter lesions may mimic mycosis fungoides.^[54,55] Review of lymph node cases seen at the Armed Forces Institute of Pathology showed that most proliferations are immunoblastic hyperplasias, with a minority meeting criteria of non-Hodgkin or Hodgkin lymphoma.^[56] Other studies suggest that gene rearrangement studies are useful in evaluating these lesions.^[57] The regression of disease in this case after cessation of phenytoin demonstrates another case in which mono-clonality does not necessarily equate with malignancy.

Kikuchi-Fujimoto Disease. Kikuchi-Fujimoto disease, an atypical hyperplasia characterized by nonsuppurative "granulomatous" lymphadenitis, can be mistaken for a lymphoma. The mixture of cells includes histiocytes with crescentic nuclei, plasmacytoid monocytes, and immunoblasts with necrosis and apoptosis. However, different phases of the disease can show varying proportions of these cells.^[58-60] Immunoblasts can be increased in the proliferative phase. Case 195 demonstrated a marked proliferation of histiocytic cells with apoptosis associated with immunoblasts Image 27 and Image 28. The relatively monotonous population of cells mimics lymphoma; the recognition of the histiocyte-rich areas with apoptosis is critical for the correct diagnosis. Myeloperoxidase expression by histiocytes may also be helpful in diagnosis.^[61]

Other Immunoblastic Proliferations. Case 123 showed an atypical paracortical immunoblastic reaction in which flow cytometric immunophenotyping identified a heterogeneous T-cell population with "soft" abnormalities related to CD4 and CD8 expression. However, the lack of definitive morphologic, immunophenotypic, and molecular abnormalities precluded the diagnosis of lymphoma. Mesenteric lymphadenitis can be primary or secondary.^[62] In the setting of a volvulus in which the lymph node was sampled, this case likely represents a secondary lymphadenitis due to the localized inflammatory reaction. Primary forms are often of infectious origin, eg, due to organisms such as Salmonella and Yersinia species, although many other organisms can manifest as mesenteric lymphadenitis.^[62,63]

Mastocytosis. Systemic mast cell disease is a neoplastic disorder comprising abnormal mast cells. It is subclassified based on the distribution of disease and the presence or absence of other non–mast cell hematologic malignancies as described in the World Health Organization Classification.^[64] Genetic abnormalities have been described, such as mutations in c-KIT or translocations involving FIP1L1 and PDGFRa; demonstration of such abnormalities may be required for classification in the future.^[65] Involvement in lymph nodes is often in a paracortical distribution, as in case 239 Image 29 and Image 30. Because T-cell lymphomas often involve the paracortical areas, particularly in early involvement, the differential diagnosis of mast cell disease in lymph node will include T-cell lymphoma and other hematologic processes such as myeloid sarcoma and nodal marginal zone B-cell lymphoma with monocytoid B cells. The presence of eosinophils in mast cell disease can accentuate the resemblance to T-cell lymphoma. Immunophenotyping will unequivocally distinguish mast cells that express tryptase and CD117 from T or B cells that express a variety of T- or B-cell antigens. Because abnormal mast cells in mast cell disease may express CD2 and CD25, these markers alone should not be used as evidence of T-cell lineage when this differential diagnosis is being considered.^[66]

The following were contributors or cocontributors of cases to these sessions 2 and 3. Session 2: J. Audouin, P. Banks, P.J. Buckley, G. Bentley, W.R. Burack, P. Cervl, G. Deeb, C. Leith, J. Diebold, A. Dogan, C.H. Dunphy, W.G. Finn, P. Gaulard, L. Germain, J.Z. Gong, A. Hassan, M.J. Hayes, F. H.C. Kreisel, M. Kurrer, A.S. Lagoo, A. Le Tourneau, W.R. Macon, T. Molina, L. Quintanilla-Martinez, M. Piris, E. Ranheim, V. Reddy, E. Remstein, C.W. Ross, D. Roulston, T. Rudiger, B. Schnitzer, L.R. Smith, S.V. Smith, W. Tang, J.A. Thorson, R. Valdez, G.A. VanDerwerker, R. Warnke, and D. Wu.

Session 3: B. Addis, B. Alobeid, A. Bagg, G. Bhagat, J. Burke, F. Craig, J.O. Dailey, D. De Jong, L. De Leval, S.S. Deodhare, K. Fu, J. Garcia, S. Gheith, J.H. Goldberg, L.P. Gordon, T.C. Greiner, J. Herrick, E.D. Hsi, R.E. Hutchison, K. Huynh, E.S. Jaffe, S.P. Jiang, C.M. Jones, N. Karandikar, A.I. Kende, M. Kesler, M. Kim, D.J. Konopka, S.H. Kroft, F. Kuo, P. Lin, Z. Liu, D.P. O'Malley, R.W. McKenna, J. Menarguez, N. Muscato, B. Nelson, P. Peterson, M. Pilichowska, M. Proytcheva, J. Roepke, M. Samoszuk, B. Sawan, E.J. Schlette, S. Schuster, M.A. Thompson, L. Tsao, R. Tubbs, N. Vajpayee, E. Vakiani, J.A. Vergilio, J. Weisberger, D.D. Weisenburger, T. Worgall, C. Yue, and X.F. Zhao.

Dr Warnke: Dept of Pathology, Stanford Medical Center, L235, 300 Pasteur Dr, Stanford, CA 94305-5324.