Extranodal NK/T-cell Lymphoma, Nasal Type, Includes Cases of Natural Killer Cell and αβ, γδ, and αβ/γδ T-cell Origin

ORIGINAL ARTICLE

Extranodal NK/T-cell Lymphoma, Nasal Type, Includes Cases of Natural Killer Cell and ab, gd, and ab/gd T-cell Origin: A Comprehensive Clinicopathologic and Phenotypic Study Tawatchai Pongpruttipan, MD,*w Sanya Sukpanichnant, MD,w Thamathorn Assanasen, MD,z Pongsak Wannakrairot, MD,z Paisarn Boonsakan, MD,y Wasana Kanoksil, MD,y Kanita Kayasut, MD,8 Winyou Mitarnun, MD,8 Archrob Khuhapinant, MD, PhD,z Udomsak Bunworasate, MD,# Teeraya Puavilai, MD,** Anan Bedavanija, MD,ww Adriana Garcia-Herrera, MD,zz Elias Campo, MD,zz James R. Cook, MD, PhD,yy John Choi, MD, PhD,88 and Steven H. Swerdlow, MD*

Abstract: Extranodal NK/T-cell lymphoma (ENKTL), nasal type, may be of NK or T-cell origin; however, the proportion of T-ENKTLs and whether they are of ab or gd type remains uncertain. To elucidate the cell of origin and detailed phenotype of ENKTL and assess any clinicopathologic associations, 67 cases of ENKTL from Thailand were investigated, together with 5 gd enteropathy-associated T-cell lymphomas (EATLs) for comparison. In all, 70% of the ENKTL were T-cell receptor (TCR) b,g and, in cases tested, d negative (presumptive NK origin); 5% were TCR gd+, 3% were TCR ab+, 1% were TCR ab/gd+, and 21% were indeterminate. Out of 17 presumptive NK-ENKTLs tested, 3 had clonal TCR rearrangements. All cases were EBV+ and TIA-1+; >85% were positive for CD3, CD2, granzyme B, pSTAT3, and Lsk/MATK; and all were From the *Department of Pathology, Division of Hematopathology, University of Pittsburgh School of Medicine, Pittsburgh; wDepartment of Pathology; zDepartment of Medicine, Division of Hematology; wwDepartment of Otolaryngology, Head and Neck Surgery, Faculty of Medicine Siriraj Hospital, Mahidol University; #Department of Medicine, Division of Hematology; zDepartment of Pathology, Faculty of Medicine, Chulalongkorn University; **Department of Medicine, Division of Hematology; yDepartment of Pathology, Faculty of Medicine Ramathibodi Hospital, Bangkok; 8Department of Pathology, Faculty of Medicine, Prince of Songkla University, Hat-Yai, Songkhla, Thailand; zzHematopathology Unit, Center for Biomedical Diagnosis, Hospital Clinic, University of Barcelona, Barcelona, Spain; yyPathology and Laboratory Medicine Institute, Cleveland Clinic, Cleveland, OH; and 88Department of Pathology, Children’s Hospital of Philadelphia, University of Pennsylvania School of Medicine, Philadelphia, PA. Conflicts of Interest and Source of Funding: The authors have disclosed that they have no significant relationships with, or financial interest in, any commercial companies pertaining to this article. Correspondence: Steven H. Swerdlow, MD, Department of Pathology, Division of Hematopathology, UPMC Presbyterian, 200 Lothrop Street-Room G335, Pittsbugh, PA 15213-2582 (e-mail: swerdlowsh@ upmc.edu). Supplemental Digital Content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Website, www.ajsp.com. Copyright r 2012 by Lippincott Williams & Wilkins

Am J Surg Pathol



Volume 36, Number 4, April 2012

CD16  . Presumptive NK-ENKTLs had significantly more frequent CD56 (83% vs. 33%) and CXCL13 (59% vs. 0%) but less frequent PD-1 (0% vs. 40%) compared with T-ENKTLs. Of the NK-ENKTLs, 38% were Oct-2+ compared with 0% of T-ENKTLs, and 54% were IRF4/MUM1+ compared with 20% of T-ENKTLs. Only ab T-ENKTLs were CD5+. Intestinal ENKTLs were EBV+ and had significantly more frequent CD30, pSTAT3, and IRF4/MUM1 expression but less frequent CD16 compared with gd EATL. Significant adverse prognostic indicators included a primary non-upper aerodigestive tract site, high stage, bone marrow involvement, International Prognostic Index Z2, lack of radiotherapy, Ki67 >40%, and CD25 expression. The upper aerodigestive tract ENKTLs of T-cell origin compared with those of presumptive NK origin showed a trend for better survival. Thus, at least 11% of evaluable ENKTLs are of T-cell origin. Although TENKTLs have phenotypic and some possible clinical differences, they share many similarities with ENKTLs that lack TCR expression and are distinct from intestinal gd EATL. Key Words: extranodal NK/T-cell lymphoma, gd T-cell receptor, natural killer cells, CXCL13, CD25 (Am J Surg Pathol 2012;36:481–499)

E

xtranodal natural killer/T-cell lymphoma (ENKTL), nasal type, is defined in the 2008 WHO classification as a “predominantly extranodal lymphoma characterized by vascular damage and destruction, prominent necrosis, cytotoxic phenotype and association with Epstein-Barr virus (EBV).”14 It is designated as an “NK/T-cell” lymphoma because, although most of the cases are believed to be of “true” natural killer cell (NK) origin, “some” of the cases are known to demonstrate a cytotoxic T-cell phenotype.14 The actual proportion of cases of T-cell origin is not well established and, even when documented to be of putative T-cell origin, whether the T cells are of ab or gd type is often unknown. Furthermore, little www.ajsp.com |

481

Pongpruttipan et al

attention has been focused on phenotypic variations among the more numerous NK cell cases and their clinical implications, even though NK cells, similar to other lymphoid cells, are phenotypically heterogeneous.36,78,103,104 Some of the phenotypic heterogeneity such as expression of CD16 is well recognized,36 but little attention has been paid to the expression of many other antigens, such as the B-cell–associated transcription factor Oct-2 or the T follicular helper-associated marker CXCL13.78,104 ENKTL is generally considered to be a very aggressive lymphoma; however, some patients have more indolent or therapy-responsive disease.6,14,22 Although there are some well-established mostly clinical prognostic factors,6,14,56 it remains unknown whether any pathologic subsets of ENKTL should be distinguished for clinical or therapeutic purposes and whether any of the antigens that are only expressed by subsets of these cases have prognostic implications. Greater attention has been paid in recent years to the relationship of varied T-cell subsets to specific T-cell neoplasms. For example, it has been suggested that many of the nonhepatic gd T-cell lymphomas, in addition to those involving the skin, might be included under the umbrella term of mucocutaneous gd T-cell lymphoma.3,34 Whether any of the ENKTLs, another neoplasm classically of the innate immune system that frequently involves mucosal or cutaneous sites, should also be considered for inclusion in this potential category is uncertain. This is an even more important question given the recent finding that NK cell neoplasms of varied types have molecular features similar to a group of nonhepatosplenic gd T-cell lymphomas.45 In contrast, other recent studies have suggested that these nonhepatosplenic gd T-cell lymphomas are heterogeneous.32,101 Finally, although much of our knowledge about ENKTL is from Asian countries, a large detailed series of Thai cases has not been published. To address these issues, the clinicopathologic features of 67 cases of ENKTL from Thailand were investigated, with particular attention paid to the expression of T-cell receptor (TCR) proteins and the detailed phenotype of the neoplastic cells. In addition, the findings were compared with those of enteropathy-associated T-cell lymphomas of gd type (gd EATL) to address the question as to whether intestinal ENKTL might represent an EBV+ variant of EATL.

MATERIALS AND METHODS This study was approved by the Institutional Review Boards of the University of Pittsburgh School of Medicine and the Faculty of Medicine Siriraj Hospital, Mahidol University.

Case Selection and Clinical Data Collection Sixty-seven cases from Thailand that fulfilled the WHO criteria for ENKTL were selected for review. None of the cases in this study were included in a recently conducted Thai ENKTL study.71 Five cases of EATL

482 | www.ajsp.com

Am J Surg Pathol



Volume 36, Number 4, April 2012

were also included for comparison purpose (3 type II and 1 type I gd EATL from the University of Pittsburgh Medical Center previously reported in abstract form101 and 1 type II EATL from Siriraj Hospital). Clinical data including age, sex, sites of involvement, Ann Arbor stage, International Prognostic Index (IPI), any underlying immunodeficiency states, mode of treatment, duration of follow-up, and patient status at time of last follow-up were collected.

Histopathologic Assessment Hematoxylin and eosin-stained sections of all cases were reviewed by S.H.S. and T.P. The following features were evaluated: tissue necrosis, ulceration with acute inflammation, angioinvasion, thrombosis, epithelial hyperplasia, epitheliotropism, nuclear irregularity (predominantly round vs. irregular), prominent nuclear elongation (present or absent), cytoplasmic granularity, prominent admixed histiocytes, and granulomas. The average number of admixed neutrophils, eosinophils, and plasma cells was estimated (0 to 10, 11 to 25, 26 to 75, and >75/HPF,  400, field number 22). When possible, evaluation was performed away from inflamed ulcer bases. Only limited morphologic evaluation could be performed in cases with extensive necrosis and/or crush artifact. Mitoses were counted in 10 HPFs ( 400, field number 22, fewer fields counted if limited tissue). Cases were divided into 5 groups on the basis of nuclear size: (1) small cell predominant—small cells >90%; (2) mixed small and larger cells—small cells 25% to 90% plus larger cells >10%; (3) medium-sized cell predominant—medium-sized cells >75%; (4) mixed medium-sized and large cells—medium-sized cells 25% to 75% plus large cells >25%; and (5) large cell predominant—large cells >75% (with or without anaplastic cells). Groups 1 and 2 (subtypes with small cell component) and groups 3 to 5 (subtypes without small cell component) were combined for subsequent analyses. Nuclear size was defined as follows: small—similar to that of mantle zone lymphocytes; large—at least twice the size of small lymphocytes; and medium—intermediate between small and large.

Immunophenotypic and EBV-encoded RNA (EBER) In Situ Hybridization (ISH) Studies Three tissue microarray (TMA) blocks were constructed using a single 3 mm core from all cases from which there was sufficient available tissue; tonsil and gd EATL were included for controls. Most immunohistochemical (IHC) stains were performed on the TMA. A limited number of stains were performed on whole-tissue sections. TCRg IHC staining was performed using clone 3.20 (Thermo Fisher Scientific, Rockford, IL) and the previously reported method.77 TCRd IHC staining was performed with antigen retrieval by incubating deparaffinized slides with EnVision FLEX Target Retrieval Solution, low pH (Dako), using a PT Link (Dako), followed by a 90-minute incubation with antibody for the TCRd constant region (clone 5A6.E9, Thermo Fisher Scientific). r

2012 Lippincott Williams & Wilkins

Am J Surg Pathol



Volume 36, Number 4, April 2012

The immunostaining was performed at room temperature using EnVision Mouse LINKER (Dako) for 15 minutes, followed by EnVision FLEX+, Mouse (Dako), for 20 minutes, and 3-30 diamonibenzidine for 10 minutes. Blocking of endogenous peroxidase was carried out for 5 minutes. Only membranous staining was considered positive. Details of clones, dilution, sources, and methods for all other IHC stain are given in Supplemental Digital Content 1, http://links.lww.com/PAS/A108. EBER ISH was also performed.48 All preexisting and new IHC and EBER ISH stains were reviewed by S.H.S. and T.P. All stains except for Ki67 and p53 were interpreted as follows: negative; probably negative—minimal equivocal staining; probably positive—weak or partial indefinite staining; partially positive—definite positivity on 10/HPF (20%, 5%, and 23%, respectively). Four cases (7%) had extensive neutrophil infiltration mimicking an acute inflammatory process, 2 were histiocyte-rich, and a granulomatous reaction was identified in the subcutaneous fat in 1 case. One ab T-ENKTL had 26 to 75 eosinophils/HPF, 4 ENKTLs had >75 plasma cells/HPF, and 3 ENKTLs had 26 to 75 plasma cells/HPF mimicking inflammatory processes. www.ajsp.com |

483

Pongpruttipan et al

Am J Surg Pathol



Volume 36, Number 4, April 2012

TABLE 1. Clinical Features* and Mode of Treatment of ENKTL Subtypes and gd EATL ENKTL Subgroups and cd EATL

ENKTL Defined by TCR Expression

UADT NKENKTL

Cutaneous NKENKTL

Intestinal ENKTL

cd EATL

59 [37-80] 2: 0

47 [34-52] 1.5: 1

73 [63-80] 1.5: 1

Characteristics

Overall ENKTLw

NK

cd Tz

Age, median [range] (y) Sex, M:F ratio

44 [11-83] 4.2: 1

44 [22-80] 5.7: 1

52 [39-55] 1: 2

23 [18-27] 34 43 [22-74] 1: 1 0: 1 5: 1 No. of Cases/Total Cases (%)

(63)

30/47 (64)

3/3 (100)

1/2 (50)

0

30/42 (71)

0

0

0

(25) (4) (3) (3) (2)

12/47 (26) 2/47 (4) 1/47 (2) 2/47 (4) 0

0 0 0 0 0

1/2 (50) 0 0 0 0

0 0 1/1 (100) 0 0

12/42 (29) 0 0 0 0

0 0 0 2/2 (100) 0

0 3/5 (60) 2/5 (40) 0 0

0 0 5/5 (100) 0 0

(45) (38) (18)

14/30 (47) 10/30 (33) 6/30 (20)

2/3 (67) 1/3 (33) 0

1/2 (50) 1/2 (50) 0

0 0 1/1 (100)

14/26 (54) 9/26 (35) 3/26 (22)

0 0 2/2 (100)

0 2/4 (50) 2/4 (50)

1/3 (33) 2/3 (67) 0

(65) (35)

19/30 (63) 11/30 (37)

2/3 (67) 1/3 (33)

2/2 (100) 0

0 1/1 (100)

18/26 (69) 8/26 (31)

0 2/2 (100)

1/4 (25) 3/4 (75)

1/3 (33) 2/3 (67)

(41) (19) (41)

9/27 (33) 7/27 (26) 11/27 (41)

2/3 (67) 0 1/3 (33)

2/2 (100) 0 0

0 0 1/1 (100)

9/25 (36) 7/25 (28) 9/25 (36)

0 0 1/1 (100)

0 0 3/3 (100)

NA NA NA

(33) (52) (14)

9/32 (28) 20/32 (63) 3/32 (9)

2/3 (67) 1/3 (33) 0

1/2 (50) 1/2 (50) 0

0 0 1/1 (100)

9/28 (32) 18/28 (64) 1/28 (4)

0 1/2 (50) 1/2 (50)

0 1/4 (25) 3/4 (75)

0 2/5 (40) 3/5 (60)

Primary site Nasal ± other 42/67 UADT Non-nasal UADT 17/67 Large intestine 3/67 Small intestine 2/67 Skin 2/67 Deep soft tissue 1/67 No. other sites of involvement None 18/40 1 sitey 15/40 2-4 sitesy 7/40 Ann Arbor stage Low stage (1 and 2) 26/40 High stage (3 and 4) 14/40 IPI score 0 15/37 1 7/37 2-4 15/37 Modes of treatment Chemoradiotherapy 14/42 Chemotherapy only 22/42 None of above 6/42

ab Tz

ab/cd T

*None of the 40 patients with ENKTL and available information were HIV+. One patient with NK-ENKTL was receiving methotrexate for psoriasis, and the patient with ab/gd T-ENKTL was receiving steroids for systemic lupus erythematosus. wOverall, ENKTL includes ENKTL of NK, T, and NOS type. The ENKTL, NOS (cases in which either TCRb or g stains were not evaluable or could not be performed), is only included in aggregate with the other ENKTLs in this table because it does not represent a unique category. zThe UADT T-ENKTL (includes T-ENKTL except for the non-UADT double-positive case) had a female predominance in contrast to the UADT NK-ENKTL (P = 0.057). yKnown secondary sites of involvement included regional lymph nodes (28%), skin (13%), distant lymph nodes (8%), bone marrow (8%), and/or other organs (3% to 7%). NA indicates not available; NOS, not otherwise specified.

As identified with IHC, 41% of cases had >10% admixed reactive T cells, including the 9% of cases also with >10% admixed reactive B cells. Focal B-cell aggregates were present in 28% of cases. Cases with >10% admixed reactive T cells were associated with a mitosis count of r60 mitoses/10 HPF (16/22 cases with >10% T cells vs. 9/30 cases with r10% T cells had r60 mitoses/10 HPF; P = 0.004).

Phenotypic Features TCR Expression Five groups of ENKTLs were identified on the basis of TCR protein expression. Forty-seven of 67 cases (70%) were TCR silent and designated as presumptive NK type (NK-ENKTL) on the basis of negativity for at least TCRb and g, with TCRd also negative in 33 tested cases (Fig. 2). Three cases (5%) were of gd T-cell type (gd T-ENKTL) on the basis of positivity for both TCRg and d but not TCRb (Fig. 3). Two cases (3%) were of ab T-cell type (ab T-ENKTL) on the basis of positivity for TCRb but not for TCRg or d (Fig. 4). Both cases of ab T-ENKTL had rare larger cells with TCRg expression;

484 | www.ajsp.com

however, they were significantly fewer than the number of EBER+ or bF1+ cells. One (1%) case expressed TCRb, g, and d proteins (ab/gd T-ENKTL) (Fig. 5). Staining for TCRg and TCRd was concordant in all cases and showed similar staining patterns. Fourteen cases (21%) were of not otherwise specified type because either TCRb or g stains were not evaluable or could not be performed.

Other Phenotypic Features All ENKTLs were positive for EBER ISH and TIA-1. Many expressed CD3, CD2, granzyme B, pSTAT3, Lsk/MATK, and to a somewhat lesser extent CD30 and LMP-1 without documented differences between NKENKTL and T-ENKTL, whereas all tested cases were negative for CD16, CD20, EBV nuclear antigen-2 (EBNA2), and ALK (Table 3; Figs. 3–5). CD56 expression was more common in cases of NK origin (83%) than in those of T-cell origin (33%) (P = 0.019). In some cases, although the TMA core was CD56  , staining of the whole section showed CD56 positivity. Expression of other antigens was more variable. Loss of CD7 or CD45 expression was commonly found in both NK-ENKTL r

2012 Lippincott Williams & Wilkins

Am J Surg Pathol



Volume 36, Number 4, April 2012

NK and T Subtypes of Extranodal NK/T-Cell Lymphoma

and T-ENKTL. P53 expression of >10% was associated with high-stage disease (7/7 high-stage vs. 6/14 low-stage cases were p53+; P = 0.018), absence of a small cell component (21/25 cases without vs. 0/6 cases with a small cell component were p53+; P40%, and CD25 expression (Table 5; Fig. 6). Among the UADT ENKTL cases, the cases of NK versus T-cell origin

Molecular Genotypic Studies Clonal PCR-based TCR gene rearrangement was identified in 3 of 6 T-ENKTLs and 3 of 17 NK-

TABLE 2. Pathologic Features of Extranodal ENKTL Subtypes and gd EATL ENKTL Subgroups and cd EATL

ENKTL Defined by TCR Expression

Characteristic

Overall ENKTL*

NK

cd Tw

ab Tw

ab/cd T

UADT NKENKTL

Cutaneous NKENKTLz

Intestinal ENKTL

cd EATLy

No. of Cases/Total Cases (%) Size of lymphoma cells Predominantly small cells Mixed small and larger cells Predominantly medium-sized cells Mixed medium-sized and large cells Predominantly large cells Predominantly irregular nuclei Frequent elongate nuclei Necrosis Ulcer edge with many PMNs Angioinvasion Thrombosis Epithelial hyperplasia Epitheliotropism Admixed reactive B cells >10% Admixed reactive T cells >10% Mitoses/10 HPF [median, range] Mitoses >60/10 HPF

3/62 (5)

2/45 (4)

1/3 (33)

0

0

2/40 (5)

0

0

0

9/62 (15)

4/45 (9)

1/3 (33)

1/2 (50)

0

4/40 (10)

0

0

0

23/62 (37)

18/45 (40)

0

1/2 (50)

0

17/40 (43)

0

1/5 (20)

14/62 (23)

10/45 (22) 1/3 (33)

0

0

10/40 (25)

0

1/5 (20)

0

13/62 (21)

11/45 (24)

0

1/1 (100)

7/40 (18)

2/2 (100)

3/5 (60)

0

45/62 (73)

34/45 (76) 2/3 (67)

1/2 (50)

1/1 (100) 33/40 (83)

1/2 (50)

2/5 (40)

0/5 (0)

45/60 (75) 53/67 (79) 42/62 (68)

32/43 (74) 3/3 (100) 39/47 (83) 2/3 (67) 31/45 (69) 2/3 (67)

1/2 (50) 1/1 (100) 31/38 (82) 2/2 (100) 1/1 (100) 35/42 (83) 2/2 (100) 1/1 (100) 28/40 (40)

0/2 (0) 1/2 (50) 1/2 (50)

2/5 (40) 5/5 (100) 4/5 (80)

0/5 (0) 5/5 (100) 5/5 (100)

30/67 4/61 5/55 23/55 5/58

22/47 4/45 4/39 17/39 3/43

2/2 0/2 0/2 2/2 1/2

2/2 0/2 0/1 1/1 0/1

4/5 0/5 0/5 0/5 0/5

(45) (7) (9) (42) (9)

(47) (9) (10) (44) (7)

0

1/3 0/3 0/3 0/3 0/3

(33) (0) (0) (0) (0)

(100) (0) (0) (100) (50)

23/56 (41)

13/42 (31) 1/1 (50)

[66, 2-248]

[68, 2-248] [25, 17-42] [69, 8-103]

29/56 (52)

23/41 (56) 1/2 (50)

1/1 0/1 0/1 0/1 0/1

(100) 17/42 (41) (0) 4/40 (10) (0) 4/35 (11) (0) 16/35 (46) (0) 3/39 (8)

2/2 (100) 0/1 (0)

0/3 (0)

[87]

13/38 (34) [68, 2-248]

1/1 (100) 18/36 (50)

(100) (0) (0) (100) (0)

0/1 (0)

(80) (0) (0) (0) (0)

5/5 (100)

5/5 1/5 0/5 19/46 0/5

(100) (20) (0) (41) (0)

1/5 (20)

0/5 (0)

[73, 65-81]

[77, 35-155]

[92, 67-148]

2/2 (100)

4/5 (80)

4/4 (100)

*Overall, ENKTL includes ENKTL of NK, T, and NOS type. wWhen compared with UADT NK-ENKTL, UADT T-ENKTL (includes T-ENKTL except for the non-UADT double-positive case) had more cases with a small cell component (P = 0.042). zWhen compared with UADT NK-ENKTL, the cutaneous NK-ENKTL had less frequent elongate nuclei (P = 0.046). yWhen compared with intestinal ENKTL, gd EATL more often had predominantly medium-sized cell morphology (P = 0.048); when gd EATL were compared with the gd T-ENKTL, they had elongate nuclei less frequently (P = 0.018) and >60 mitoses/10 HPF more frequently (P = 0.029). NOS indicates not otherwise specified.

r

2012 Lippincott Williams & Wilkins

www.ajsp.com |

485

Pongpruttipan et al

Am J Surg Pathol



Volume 36, Number 4, April 2012

FIGURE 1. ENKTL, histopathology. A, Note the dense lymphoid infiltrate with a large area of geographic necrosis. B, Prominent angioinvasion of a muscular artery is present. C, Although the neoplastic cells are predominantly small, they have irregular and elongate nuclear contours. D, The neoplastic cells are predominantly medium-sized but still irregular and elongate. E, Note the prominent mitotic figures in this case with predominantly large cells. F, Some of the neoplastic cells demonstrate coarse eosinophilic granules in their abundant pale cytoplasm (all cases of presumptive NK type except for C, which is of gd type, hematoxylin and eosin stains).

showed a trend for worse survival (P = 0.092). Significantly better survival was observed in patients who received combined chemoradiation therapy, compared

486 | www.ajsp.com

with those who received only chemotherapy. All patients receiving radiation had low-stage disease (and none had CD25 expression), but the significance was retained when r

2012 Lippincott Williams & Wilkins

Am J Surg Pathol



Volume 36, Number 4, April 2012

NK and T Subtypes of Extranodal NK/T-Cell Lymphoma

FIGURE 2. NK-ENKTL, phenotypic features. A, The EBER ISH stain for EBV demonstrates numerous cells with nuclear positivity, as well as cytoplasmic staining in cells undergoing mitosis. B, The neoplastic cells are positive for CD2, but (C) CD3 and (D) CD5 are only present on some presumptive normal T cells. E, CD56, (F) CD30, (G) CXCL13, and (H) IRF4/MUM1 are all positive. The CXCL13 shows typical punctate staining. (A), EBER ISH; (B–H), immunoperoxidase stains, all with hematoxylin counterstain.

r

2012 Lippincott Williams & Wilkins

www.ajsp.com |

487

Pongpruttipan et al

Am J Surg Pathol



Volume 36, Number 4, April 2012

FIGURE 3. gd T-ENKTL, histopathologic and phenotypic features. A, The dense lymphoid infiltrate with a large area of geographic necrosis is indistinguishable from an NK-ENKTL. B, Cytologically, this case with mixed medium-sized and large cells also has irregular and elongate nuclei. C, The EBER ISH stain for EBV shows numerous positive nuclei. D, The neoplastic cells are also positive for CD2 and (E) CD56 but (F) are negative for CD5. G, Although somewhat difficult to interpret, at least a significant proportion of the atypical cells expressed TCRg and (H) TCRd. (A, B) hematoxylin and eosin stains; (C) EBER ISH with hematoxylin counterstain, (D–H) immunoperoxidase with hematoxylin counterstain).

488 | www.ajsp.com

r

2012 Lippincott Williams & Wilkins

Am J Surg Pathol



Volume 36, Number 4, April 2012

NK and T Subtypes of Extranodal NK/T-Cell Lymphoma

FIGURE 4. ab T-ENKTL with clonal TCRg gene rearrangement, phenotypic features. A, There are many scattered EBV+ neoplastic cells, which are also positive for (B) CD3, (C) CD5, and (D) CD4, but negative for (E) CD8 and (F) CD56. G. There are many scattered bF1+ cells, including some larger cells, but (H) there are only infrequent TCRg+ cells. (A) EBER ISH; (B–H) immunoperoxidase with hematoxylin counterstain.

r

2012 Lippincott Williams & Wilkins

www.ajsp.com |

489

Pongpruttipan et al

Am J Surg Pathol



Volume 36, Number 4, April 2012

FIGURE 5. ab/gd T-ENKTL, intestinal: histopathologic and phenotypic features. A, The infiltrative mass involves only a very short segment of bowel and yet is transmural and deeply ulcerated. B, There is prominent angioinvasion, which was most frequent around the ulcer edge. C, The neoplastic cells are EBV, (D) TCRg, (E) TCRb, and (F) CLA+. (A, B) Hematoxylin and eosin stains; (C) EBER ISH; (D–F) immunoperoxidase, all with hematoxylin counterstain.

490 | www.ajsp.com

r

2012 Lippincott Williams & Wilkins

Am J Surg Pathol



Volume 36, Number 4, April 2012

NK and T Subtypes of Extranodal NK/T-Cell Lymphoma

TABLE 3. Phenotypic Features of ENKTL Subtypes and gd EATL ENKTL Subgroups and cd EATL

ENKTL Defined by TCR Expression

Characteristic

Overall ENKTL*

NKw

cd Tz

ab Ty

ab/cd T

UADT NKENKTL8

Cutaneous NKENKTL

Intestinal ENKTLz

cd EATL#

No. of Cases/Total Cases (%) Pan T-cell antigens CD2 32/36 CD3 65/67 CD5 3/66 CD7 15/33 T-/NK-cell antigen subset CD4 1/60 CD8 6/63 CD16 0/31 CD56 44/56 TIA-1 65/65 Granzyme B 51/54 CXCL13 17/33 PD1 2/30 CLA 12/28 EBV-associated markers EBER ISH 67/67 LMP1 18/32 EBNA2 0/35 Miscellaneous CD45 18/29 CD45RA 0/34 CD30 41/57 ALK 0/42 CD25 11/31 pSTAT3 32/33 Lsk (MATK) 31/35 IRF4/MUM1 17/35 Oct-2 8/26 CD20 0/67 Ki67 >40% 13/25 p53 >10% 21/31

(89) (97) (5) (45)

24/28 45/47 0/47 13/27

(86) (96) (0) (48)

3/3 3/3 0/2 1/3

(100) (100) (0) (33)

1/1 (100) 2/2 (100) 2/2 (100) ND

1/1 1/1 0/1 0/1

(100) (100) (0) (0)

21/24 40/42 0/42 10/23

(88) (95) (0) (44)

1/1 2/2 0/2 0/1

(100) (100) (0) (0)

3/4 5/5 0/5 3/4

(75) (100) (0) (75)

3/5 5/5 0/5 5/5

(60) (100) (0) (100)

(2) (10) (0) (79) (100) (94) (52) (7) (43)

0/47 4/47 0/25 35/42 47/47 39/42 16/27 0/23 10/21

(0) (9) (0) (83) (100) (93) (59) (0) (48)

0/3 0/3 0/3 1/3 3/3 2/2 0/3 1/3 0/3

(0) (0) (0) (33) (100) (100) (0) (33) (0)

1/2 1/2 0/1 0/2 2/2 1/1 0/1 0/1 0/1

0/1 1/1 0/1 1/1 1/1 1/1 0/1 1/1 1/1

(0) (100) (0) (100) (100) (100) (0) (100) (100)

0/42 3/42 0/22 32/38 42/42 35/38 14/24 0/20 9/18

(0) (7) (0) (84) (100) (92) (58) (0) (50)

0/2 (0) 1/2 (50) ND 2/2 (100) 2/2 (100) 1/1 (100) ND ND ND

0/5 1/5 0/4 3/4 5/5 4/4 2/4 1/4 2/2

(0) (20) (0) (75) (100) (100) (50) (25) (50)

0/5 2/5 3/5 4/5 5/5 2/4 0/5 1/5 0/5

(0) (40) (60) (80) (100) (50) (0) (20) (0)

(100) (56) (0)

47/47 (100) 13/25 (54) 0/28 (0)

3/3 (100) 2/3 (67) 0/3 (0)

2/2 (100) 1/2 (50) 0/1 (0)

1/1 (100) 1/1 (100) 0/1 (0)

42/42 (100) 11/21 (52) 0/25 (0)

2/2 (100) ND ND

5/5 (100) 3/4 (75) 0/4 (0)

0/5 (0) 0/5 (0) 0/5 (0)

(62) (0) (72) (0) (35) (97) (87) (49) (31) (0) (52) (68)

14/23 0/27 32/43 0/33 9/25 26/27 26/28 15/28 8/21 0/47 12/19 18/25

1/3 0/3 1/3 0/3 0/3 3/3 3/3 0/3 0/3 0/3 0/3 1/3

ND 0/1 (0) 1/1 (100) 0/1 (0) 1/1 (100) 1/1 (100) 0/1 (0) 0/1 (0) 0/1 (0) 0/2 (0) 0/1 (0) 0/1 (0)

1/ 1 (100) ND 1/1 (100) 0/1 (0) 1/1 (100) 1/1 (100) 1/1 (100) 1/1 (100) 0/1 (0) 0/1 (0) 1/1 (100) 1/1 (100)

13/20 0/24 27/38 0/28 8/22 23/24 22/24 12/24 8/19 0/42 10/18 15/22

ND 0/1 (0) 2/2 (100) 0/2 (0) ND ND 1/1 (100) 1/1 (100) ND 0/2 (0) ND ND

2/4 0/2 5/5 0/4 2/4 4/4 4/4 3/4 0/3 0/5 2/2 4/4

5/5 2/5 0/5 0/5 0/5 1/5 5/5 0/5 1/5 0/5 4/5 5/5

(61) (0) (74) (0) (36) (96) (93) (54) (38) (0) (63) (72)

(33) (0) (33) (0) (0) (100) (100) (0) (0) (0) (0) (33)

(50) (50) (0) (0) (100) (100) (0) (0) (0)

(65) (0) (71) (0) (36) (96) (92) (50) (42) (0) (56) (68)

(50) (0) (100) (0) (50) (100) (100) (75) (0) (0) (100) (100)

(100) (40) (0) (0) (0) (20) (100)** (0) (20) (0) (80) (100)

*Overall, ENKTL includes ENKTL of NK, T, and NOS type. wThe NK-ENKTL, when compared with T-ENKTL, had more frequent CXCL13 expression (P = 0.043) but less frequent PD1 expression (P = 0.026). zThe gd T-ENKTL (all UADT), when compared with UADT NK-ENKTL, had less frequent CD56 expression (P = 0.092). yThe ab T-ENKTL (all UADT), when compared with the UADT NK-ENKTL, had more frequent CD5 expression (P = 0.001), more frequent CD4 expression (P = 0.045), and less frequent CD56 expression (P = 0.036). 8The UADT T-ENKTL (only the ab/gd T-cell case excluded) had less frequent CD56 expression compared with UADT NK-ENKTL (P = 0.008). zThe intestinal ENKTL, when compared with the gd EATL, had more frequent expression of CD30 (P = 0.008), pSTAT3 (P = 0.048), and IRF4/MUM1 (P = 0.048) but less frequent CD16 expression (P = 0.048). #The gd EATL, when compared with UADT NK-ENKTL, had more frequent expression of CD7 (P = 0.044), CD16 (P = 0.003), and CD45RA (P = 0.025) but less frequent expression of CXCL13 (P = 0.042), CD30 (P = 0.005), and pSTAT3 (P = 0.001). There are also trends for more frequent CD8 expression (P = 0.081) but less IRF4/MUM1 expression (P = 0.059) in the gd EATL. **Of the type II EATLs, 2 had strong Lsk expression in˜ 90% of cells, and 2 had weaker staining in 50% to 60%. The type I EATL had only approximately 20% weakly Lsk+ cells. NOS indicates not otherwise specified.

only low-stage cases and NK-UADT cases were analyzed. Stage, bone marrow involvement, IPI score, and CD25 expression retained their significance in the UADT NK group.

UADT ENKTL: Comparison of Those With and Without TCR Expression Patients with T-ENKTL showed a trend for a lower M:F ratio (1.5:1 vs. 5:1, P = 0.057). The T-ENKTL also included a higher proportion of cases with a small cell component (3/5 vs. 6/40 UADT NK-ENKTLs; P = 0.047) (Table 2) and showed a trend for a Ki67 r

2012 Lippincott Williams & Wilkins

proliferative index of r40% (4/4 vs. 8/18 UADT NKENKTLs; P = 0.096). CD56 expression was more common in NK cases (84%) compared with T-ENKTL (20%, P = 0.008), gd T-ENKTL (33%, P = 0.091), and ab T-ENKTL (0%, P = 0.036) cases (Table 3). Both ab T-ENKTL cases were CD56  /CD5+, with one CD4+ and the other CD8+. All other cases were CD5  and CD4  . CXCL13 (58%), IRF4/MUM1 (50%), CLA (50%), and Oct-2 (42%) were present in a subset of NK-ENKTLs but not in any T-ENKTLs (differences not significant; P = 0.098 for CXCL13). www.ajsp.com |

491

Pongpruttipan et al

Am J Surg Pathol



Volume 36, Number 4, April 2012

TABLE 4. PCR-based TCR Gene Rearrangement Analysis of ENKTL of NK-cell and T-cell Types No. of Cases/Total Cases (%) PCR-based TCR Gene Rearrangement Analysis Overall positive cases TCRg gene TCRb gene* Suboptimal DNAw

NK

cd T

ab T

ab/cd T

3/17 (18%) 2/17 1/10 5/11 (45%)

1/3 (33%) 1/3 0/3 2/3 (67%)

1/2 (50%) 1/2 0/1 2/2 (100%)

1/1 (100%) 1/1 ND 1/1 (100%)

*TCRb gene rearrangement analysis was performed only in TCRg gene rearrangement-negative cases, except for the 1 TCRg-rearranged gdTCR+ case. wQuality of DNA was evaluated only in cases studied with BIOMED-2 primers. Cases with suboptimal DNA quality had internal control housekeeping gene amplicon size ladders of 200 bp. Cases with size ladders 10% of admixed reactive T cells were more likely to have a lower mitotic count and showed a trend for a lower Ki67. Rare cases showed coarse granules in the hematoxylin and eosin-stained sections with finer granules in some others. Although it is well known that granules are seen in ENKTL in Wright-Giemsastained smears, it has been suggested that the lack of granules in histologic sections of an NK-derived proliferation is a feature that helps distinguish ENKTL from “lymphomatoid gastropathy,”86 an indolent NK-cell–derived proliferation similar to “NK-cell enteropathy.”61 Most prior studies have evaluated the cell of origin in ENKTL strictly using PCR-based TCR gene rearrangement analysis,6,52,64 even though TCR gene rearrangements can be present in normal B cells and NK cells, in reactive proliferations, and in B-cell neoplasms.33,37,53,70,88,93,94 The proportion of ENKTLs reported to be of T-cell origin on the basis of PCR-based studies ranges from 0% to 38%, with about 8% based on a Southern blot analysis study.6,27,40,52,59,64,66,82 A gene expression profiling study reports that 11% of ENKTLs are of gd T-cell type.45 Very little information exists on TCR expression in ENKTL, although a recent series reports 0 of 74 ENKTLs with either g or b TCR protein expression, even though there were 18% CD5+ cases and almost 10% had a TCR-g rearrangement.40 In contrast, 11% of the evaluable ENKTLs in this study expressed one or more TCRs, defining at least this subset as being of T-cell origin. They included 3 of gd type, 2 of ab type, and 1 expressed g, d, and b TCR proteins. Whether this latter case represents aberrant coexpression of TCR proteins or spurious staining is uncertain, but ab/gd double-positive T-cell lymphomas (TCL) have been reported, including r

2012 Lippincott Williams & Wilkins

Am J Surg Pathol



Volume 36, Number 4, April 2012

4 TCLs of uncertain type, 3 cutaneous cytotoxic neoplasms with high-grade cytology, and 3 EBV  type II EATLs.13,69,97 The rapidly fatal ab/gd case reported here was an EBV+, CD56+, CD8+, CD5  , CD7  perforated ileal neoplasm with adenopathy but no skin disease. Although TCR expression is better evidence of a T-cell phenotype compared with PCR studies, because TCRsilent TCLs are also well known,11,18,32,35 the possibility

NK and T Subtypes of Extranodal NK/T-Cell Lymphoma

that some of our presumptive NK cases represent CD5  “receptor silent” T-cell neoplasms must be seriously considered and could potentially mask differences between the T and NK groups. It is of interest that one of the presumptive NK-ENKTLs with a clonal TCR rearrangement had a b but not g rearrangement, which is a very infrequent occurrence in neoplastic T cells.92 The limited number of TCR-silent, TCR-gene–rearranged

TABLE 5. Clinicopathologic Correlations With Overall Survival Factors

Total Cases

Overall cases Site of ENKTL Nonintestinal Intestinal Site of ENKTL UADT Non-UADT Stage Low (I and II) High (III and IV) BM involvement Present Absent IPI score 0 1 2-4 Treatment (all stages) Chemoradiotherapy Chemotherapy only None of the above Treatment in stages I and II disease Chemoradiotherapy Chemotherapy only Morphologic subtypes With small cell component Without small cell component IRF4/MUM1 expression Positive Negative CD56 expression Positive Negative CD25 expression Positive (Z10% cells) Negative Ki67 >40% r40% Admixed reactive T cells >10% r10% UADT ENKTL NK origin T origin, overall ab T-cell origin gd T-cell origin Intestinal lymphoma Intestinal ENKTL gd EATL

6-mo Survival (%)

1-yr Survival (%)

Median Survival (mo)

P (Log Rank)

57

47

35

5

—

52 5

52 0

40 0

7 0