|
|
||
 |
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Originally published as JCO Early Release 10.1200/JCO.2006.09.6396 on March 12 2007 © 2007 American Society of Clinical Oncology. Reclassification of 300 Primary Cutaneous B-Cell Lymphomas According to the New WHO–EORTC Classification for Cutaneous Lymphomas: Comparison With Previous Classifications and Identification of Prognostic Markers
From the Departments of Dermatology and Pathology, Leiden University Medical Center, Leiden; Department of Pathology, Medisch Spectrum Twente, Enschede; Department of Pathology, University Medical Center Nijmegen, Nijmegen; Department of Pathology, University Medical Center Utrecht, Utrecht; Department of Pathology, University Medical Center Groningen, Groningen; Department of Pathology, Erasmus Medical Center, Rotterdam; Department of Pathology, Vrije Universiteit Medical Center, Amsterdam, the Netherlands; and the Department of Dermatology, University Hospital Gent, Gent, Belgium Address reprint requests to Nancy Senff, MD, Department of Dermatology, B1-Q, Leiden University Medical Center, PO Box 9600, 2300 RC Leiden, the Netherlands; e-mail: N.J.Senff{at}LUMC.nl
Purpose: In the new WHO–European Organisation for Research and Treatment of Cancer (WHO-EORTC) classification for cutaneous lymphomas three major groups of primary cutaneous B-cell lymphoma (CBCL) are distinguished: primary cutaneous marginal zone B-cell lymphoma (PCMZL) and primary cutaneous follicle center lymphoma (PCFCL) with a good prognosis, and primary cutaneous large B-cell lymphoma, leg type (PCLBCL-LT), with an intermediate-level prognosis. This study aimed to assess the clinical significance of the new classification compared with previous classification schemes (EORTC 1997; WHO 2001) and to define prognostic factors within the newly defined categories. Patients and Methods: In the present study clinical data and histologic sections of 300 patients with CBCL, formerly classified according to the EORTC classification, were reviewed and reclassified according to the WHO and the new WHO-EORTC classification schemes. Results: After reclassification, the study comprised 71 patients with PCMZL, 171 patients with PCFCL, and 58 patients with PCLBCL-LT, showing 5-year disease-specific survivals of 98%, 95%, and 50%, respectively. When compared with the EORTC and WHO schemes, 5.3% and 36.3% of patients with CBCL were reclassified into another prognostic category. Multivariate analysis of PCFCL revealed localization on the leg and expression of FOXP1 as independent parameters associated with a poor prognosis. Expression of Bcl-2 or MUM-1 had no significant effect on survival in this group. In PCLBCL-LT, no independent prognostic parameters were found. Conclusion: These results emphasize the clinical significance of the WHO-EORTC classification, but suggest that within the group of PCFCL, distinction should be made between lymphomas presenting on the legs and lymphomas presenting at other sites.
In recent years, there has been considerable debate regarding the classification and terminology of the group of primary cutaneous B-cell lymphomas (CBCL).1-4 This controversy, which resulted from differences between the European Organisation for Research and Treatment of Cancer (EORTC) and WHO schemes in the classification of CBCL, may have come to an end with the recent introduction of the new WHO-EORTC classification for cutaneous lymphomas.5-8 In this new classification, three main types of CBCL are recognized: primary cutaneous marginal zone B-cell lymphoma (PCMZL), primary cutaneous follicle center lymphoma (PCFCL), and primary cutaneous large B-cell lymphoma, leg type (PCLBCL-LT). In addition, it contains a category called primary cutaneous large B-cell lymphoma, other (PCLBCL-other) for diffuse large B-cell lymphomas, which do not fit in the group of PCLBCL-LT or the group of PCFCL with a diffuse infiltration of large centrocytes. PCFCL is defined as a tumor of neoplastic follicle center cells, which may show a follicular, a follicular and diffuse, or diffuse growth pattern, and which is predominantly composed of generally large centrocytes (large cleaved cells). In most instances, the tumor cells do not express Bcl-2 and MUM-1. PCLBCL-LT is defined by a predominance or confluent sheets of medium-sized to large B cells with round nuclei resembling centroblasts and/or immunoblasts, which generally show strong expression of Bcl-2 and MUM-1. These definitions imply that differentiation between PCFCL and PCLBCL-LT is determined above all by morphologic criteria, namely the absence or presence of confluent sheets of centroblasts and/or immunoblasts (designated previously as cleaved cell v round cell morphology9), and no longer by site (non-leg v leg) as in the EORTC classification.6,10,11 This implies that a proportion of patients classified as primary cutaneous follicle center cell lymphoma (PCFCCL) or primary cutaneous large B-cell lymphoma of the leg (PCLBCL-leg) in the EORTC classification will be assigned to another prognostic category, which could have important therapeutic consequences. Recent studies suggest that this concerns 10% to 15% of PCFCCL and PCLBCL-leg, but exact data are not available.12,13 These studies also illustrate that the definition of the group of PCLBCL-other leaves room for different interpretations. In this study, clinical data and histologic sections of 300 patients with CBCL who were included in the registry of the Dutch Cutaneous Lymphoma Group and originally classified by the EORTC classification were reviewed and reclassified by the new WHO-EORTC classification as well as by the WHO classification. The aims of this study were to (1) assess the clinical significance of this new classification; (2) determine the percentage of CBCL patients assigned to a different prognostic category when compared with the EORTC and WHO schemes; (3) define more precisely the clinicopathologic features and prognostic factors of the newly defined WHO-EORTC categories; and (4) establish whether PCLBCL-LT that did not express Bcl-2 should be assigned to the group of PCLBCL-other, as previously suggested.12
Patient Selection Between 1985 and 2005, 320 patients with CBCL, including 72 patients with PCMZL/primary cutaneous immunocytomas (PCI), 185 patients with PCFCCL, and 63 patients with PCLBCL-leg, had been included in the registry of the Dutch Cutaneous Lymphoma Group. All patients had been classified by an expert panel of dermatologists and pathologists at the time of diagnosis using the criteria of the EORTC classification. In all patients, the presence of extracutaneous disease at the time of diagnosis had been excluded by standard staging procedures, including computed tomography of thorax and abdomen, bone marrow cytology, and histology. Follow-up data had been collected yearly for each patient.
Histologic Review Bcl-2, Bcl-6, and CD10 staining was considered positive if more than 50% of the neoplastic B cells showed an unequivocal positive staining.12,16,17 For CD10, only a membranous staining pattern on the tumor cells was considered positive. For expression of MUM-1, nuclear staining of more than 30% of the neoplastic B cells was considered positive.18,19 Consistent with prior studies,12,20 FOXP1 was scored into three groups: positive (> 90% strongly positive cells), weak (20% to 90% weakly to moderately positive cells), and negative (< 20% weakly stained cells). From the initial group of 320 patients, 10 were excluded because of incomplete clinical details or lack of follow-up. Another 10 patients were excluded because histologic sections were either not available or were of insufficient quality to allow proper reclassification. The final study group comprised 300 patients with CBCL, including 71 patients with PCMZL, 169 patients with PCFCCL, and 60 patients with PCLBCL-leg (Table 1). These 300 patients were reclassified as PCMZL, PCFCL, or PCLBCL-LT per the WHO-EORTC classification, and as extranodal marginal zone B-cell lymphoma (EMZL), cutaneous follicle center lymphoma (cFCL), or diffuse large B-cell lymphoma (DLBCL), using the WHO classification.
Statistical Analysis All statistical calculations were performed using SPSS version 12.0.1 (SPSS Inc, Chicago, IL). Overall survival (OS) was calculated from the date of diagnosis until the patient's death or date of last follow-up. Disease-specific survival (DSS) was calculated from the date of diagnosis until death as a result of lymphoma or date of last follow-up. Survival curves were estimated by the method of Kaplan and Meier and statistical comparison between curves was done by log-rank testing. Prognostic factors within the different entities were evaluated by univariate and multivariate analysis with OS and DSS as end points, and P values below .05 were considered significant. Parameters included for univariate analysis were: age (  70 v > 70 years), sex, localization of skin lesions (leg v non-leg), extent (solitary v localized v multifocal), growth pattern (only in patients with PCFCL: follicular v follicular-diffuse v diffuse), Bcl-2 (positive v negative), MUM-1 (positive v negative), FOXP1 (positive v weak v negative). Multivariate analysis was performed with the use of a Cox proportional hazards regression model with a stepwise selection of the significant variables. Multivariate analysis was stratified for age. All other parameters included in the univariate analysis were introduced in the multivariate analysis and all parameters were categoric.
Comparison of the EORTC, WHO, and the WHO-EORTC Classification Schemes This is the first study that allows direct comparison of the clinical significance of the EORTC,6 WHO,5 and WHO-EORTC7,8 classification schemes for the group of CBCL. The number of patients who were reclassified into another prognostic category are listed in Table 1 (EORTC v WHO-EORTC) and Table 2 (WHO v WHO-EORTC).
All 71 patients with PCMZL/PCI in the EORTC classification were classified with EMZL per the WHO classification and with PCMZL using the criteria of the WHO-EORTC classification (Tables 1 and 2). Of the 169 patients classified as PCFCCL using the EORTC classification scheme, 62 patients were presenting a follicular (four patients; 2.4%) or a follicular and diffuse (58 patients; 34.3%) growth pattern, and were classified as cFCL, whereas 107 patients (63.3%) presenting a diffuse population of large centrocytes were classified as DLBCL using the WHO classification (Table 2). Using the WHO-EORTC classification, the large majority of PCFCCL patients (162 of 169 patients; 95.9%) were reclassified as PCFCL. The remaining seven patients with PCFCCL who were presenting a predominance or confluent sheets of cells with a centroblast- or immunoblasts-like morphology were reclassified as PCLBCL-LT (see Appendix Fig A1, online only). All 60 patients classified as PCLBCL-leg by the EORTC scheme were reclassified as DLBCL using the WHO classification. Fifty-one of 60 patients showed diffuse infiltrates of centroblasts and/or immunoblasts and were therefore classified as PCLBCL-LT according to the WHO-EORTC classification (Table 1). Nine (15.0%) of 60 patients showed diffuse infiltrates predominantly composed of large cleaved cells, in the absence of sheets or clusters of centroblasts and/or immunoblasts, and were therefore classified as PCFCL (see Appendix Fig A2, online only). The 5-year OS for the three main types of CBCL according to the EORTC, WHO, and WHO-EORTC classifications are presented in Figure 1.
Clinical Features and Prognostic Parameters of the WHO-EORTC Groups After reclassification, the study group contained 71 patients with PCMZL, 171 patients with PCFCL, and 58 patients with PCLBCL-LT. The main clinical characteristics and follow-up data are presented in Table 3. In the following sections, relevant clinicopathologic and immunophenotypical data and their prognostic significance are presented.
PCMZL Follow-up data of the 71 patients with PCMZL showed development of extracutaneous disease in six (8.5%) of 71 patients, and overall and disease-specific 5-year survival rates of 94% and 98%, respectively (Table 3). The clinicopathologic details of 50 of these 71 patients have been reported recently and will, therefore, be discussed no further.21
PCFCL Histologically, a follicular, a follicular and diffuse, or diffuse growth pattern was observed in 2.3% (four of 171), 33.9% (58 of 171), and 63.7% (109 of 171) of patients. Expression of Bcl-2 by more than 50% of the neoplastic B cells was observed in 16 (10.8%) of 148 patients and expression of MUM-1 by more than 30% of the neoplastic B cells in 13 (10.2%) of 128 patients (Table 4). Strong nuclear expression of FOXP1 was observed in only four (3.8%) of 104 patients. Bcl-6 and CD10 were expressed by 122 (92.4%) of 132 patients and nine (6.0%) of 141 patients, respectively. CD10 expression was found in seven (14.6%) of 48 patients with PCFCL with a follicular or follicular and diffuse growth pattern, and in only two (2.2%) of 93 patients with PCFCL with a diffuse growth pattern. After a median follow-up period of 62 months (range, 2 to 336 months), 18 of 171 patients developed extracutaneous disease and nine patients died as a result of lymphoma; four of whom due to CNS involvement. The overall and disease-specific 5-year survival ratess were 87% and 95%, respectively.
Univariate analysis of OS on the total group of patients with PCFCL (n = 171) showed that the following variables were related to a shorter survival: age (P < .0001), female sex (P = .028), localization on the leg(s) (P < .0001), extent of skin lesions (P = .045), and expression of FOXP1 (P = .002). Growth pattern, expression of Bcl-2 or MUM-1 had no significant effect on OS. In multivariate analysis, only localization on the leg(s) (P = .001) and expression of FOXP1 (P = .033) remained significant independent factors associated with a poor prognosis. Both weak and positive expression of FOXP1 was associated with inferior prognosis compared with the negative cases. Patients with PCFCL presenting with skin lesions on the leg(s) developed extracutaneous disease in five (46%) of 11 patients and had a 5-year OS and DSS of 22% and 41%, respectively, whereas patients with PCFCL who presented at sites other than the leg developed extracutaneous disease in 13 (8%) of 160 patients and showed 5-year OS and DSS of 92% and 98%, respectively (Fig 2). Multivariate analysis of OS in only those patients with PCFCL with a diffuse infiltrate of large centrocytes (n = 109) revealed localization on the leg as the only independent parameter significantly related with a poor survival (P = .006).
PCLBCL-LT This group contained 58 patients, including seven patients previously classified as PCFCCL using the EORTC classification. Consistently, most patients presented with skin lesions on one (n = 40) or both legs (n = 6) or with lesions on the leg(s) and other skin site(s) (n = 5); seven patients had presented with skin lesions restricted to sites other than the leg. Characteristically, the neoplastic B cells strongly expressed Bcl-2 (44 [89.8%] of 49 patients), MUM-1 (36 [90.0%] of 40 patients), and FOXP1 (26 [81.3%] of 32 cases). Bcl-6 and CD10 were expressed in 30 (68.2%) of 44 and 0 (0%) of 47 patients, respectively (Table 4). After a median follow-up of 26 months (range, 2 to 276 months), 27 of 58 patients had developed extracutaneous disease and 26 of 58 patients had died as a result of lymphoma. The overall and disease-specific 5-year survival rates were 37% and 50%, respectively. In a univariate analysis of OS, only age (P = .041) and the extent of skin lesions (P = .023) were associated with a poor prognosis. The 5-year OS for patients presenting with a solitary tumor was 70% compared with 27% and 0% for patients presenting with localized or multifocal disease, respectively (Appendix Fig A3, online only). Sex, localization on the leg, expression of Bcl-2, MUM-1, and FOXP1 had no prognostic significance in this group. In multivariate analysis, no independent prognostic factors were found. Importantly, no differences in survival were found between PCLBCL-LT presenting on the leg(s) and PCLBCL-LT presenting at sites other than the leg (Fig 2).
In this study, histologic sections of 300 Dutch patients with CBCL, formerly classified using the EORTC classification,6 were reviewed and classified according to the WHO5 and the new WHO-EORTC classification schemes.7,8 Using the criteria of the WHO-EORTC classification, PCMZL and PCFCL showed 5-year DSS of 98% and 95%, respectively, whereas PCLBCL-LT had a 5-year DSS of 50%. These results are in agreement with the results of recent studies and confirm that the new WHO-EORTC classification adequately distinguishes between CBCL with an indolent and CBCL with more aggressive clinical behavior.3,12,13 Comparison between the three classification schemes showed no differences other than semantic ones in the classification of PCMZL and PCFCL with a follicular or follicular and diffuse growth pattern. However, with respect to CBCL with a diffuse population of large neoplastic B cells, 16 (9.6%) of 167 and 109 (65.3%) of 167 patients who were classified according to the EORTC and WHO classification schemes, respectively, were reclassified into another prognostic category. Seven patients originally classified as PCFCCL using the EORTC classification were reclassified as PCLBCL-LT, because of the presence of confluent sheets of cells with a centroblast- or immunoblast-like morphology. Conversely, nine patients originally classified as PCLBCL-leg showed diffuse infiltrates of large cleaved cells and were therefore reclassified as PCFCL. Of the 167 patients classified as DLBCL using the WHO classification, 109 (65%) were reclassified as PCFCL using the WHO-EORTC classification scheme (Table 2), which implies that these patients should be treated primarily with radiotherapy rather than with multiagent chemotherapy.22-27 Taken together, these results illustrate that the WHO-EORTC classification is a major step forward, and may contribute to more appropriate treatment in patients with CBCL. A second goal of the present study was to define prognostic parameters within the redefined categories PCFCL and PCLBCL-LT. Both in the total group of PCFCL and in the group of PCFCL with a diffuse infiltrate of large cleaved cells, the site of clinical presentation seemed to be the most important independent prognostic factor. Consistent with previous studies, PCFCL presenting on the leg had a much worse prognosis than PCFCL presenting at other sites, but was similar to PCLBCL-LT (Fig 2).9,12 We therefore suggest that within the group of PCFCL, further distinction should be made on the basis of site to prevent undertreatment of patients presenting with skin lesions on the leg (Appendix Fig A4). Expression of Bcl-2 and MUM-1 were no independent prognostic parameters, neither in the total group of PCFCL, nor in PCFCL with a diffuse infiltrate of large cleaved cells. In the group of PCLBCL-LT, the extent of skin lesions at the time of diagnosis and age were the only prognostic parameters. Comparison of Bcl-2–positive and Bcl-2–negative PCLBCL-LT showed no significant difference in 5-year DSS (47% v 60%; P = .512) or 5-year OS (38% v 40%; P = .579), which is consistent with the results of Kodama et al.12 Similarly in both studies, no significant differences in 5-year OS and DSS were found between PCLBCL-LT with or without expression of MUM-1 or FOXP1. These data indicate that distinguishing among patients with or without expression of Bcl-2, MUM-1, or FOXP1 is not useful, and that categorization of Bcl-2–negative patients as PCLBCL-other is not justified.12 We therefore propose that this term should only be used for exceptional cases of CBCL that do not fit the criteria of PCFCL or PCLBCL-LT, such as rare cases of intravascular large B-cell lymphoma, T-cell/histiocyte-rich B-cell lymphoma, or plasmablastic lymphoma with only skin lesions at presentation. In the WHO-EORTC classification, differentiation between PCFCL with a diffuse large cell infiltrate and PCLBCL-LT has become more difficult, since it is based primarily on cell morphology (cleaved v round) and no longer on the basis of site (non-leg v leg) as in the EORTC classification. Classification on the basis of morphology may be difficult and is associated with a considerable inter-observer variation.9 In such difficult cases, the presence of a considerable proportion of admixed T cells, the presence of a stromal reaction as well as demonstration of (remnants of) follicular dendritic cell networks by staining with appropriate antibodies (CD35 or CD21) may serve as useful additional criteria suggesting a diagnosis of PCFCL. Moreover, since PCLBCL-LT, unlike PCFCL, characteristically shows strong expression of Bcl-2, MUM-1, and FOXP1, this phenotypic profile might also be a useful adjunct, supporting a diagnosis of PCLBCL-LT. However, since Bcl-2, MUM-1, and to a lesser extent FOXP1 are also expressed by a small minority of PCFCL, these markers cannot be used as a golden standard to differentiate between both conditions (Table 4). In conclusion, the results of this study emphasize the clinical significance of the new WHO-EORTC classification and suggest that it may contribute to better diagnosis and treatment. Reclassification of PCFCCL (EORTC) with a predominant round cell morphology as PCLBCL-LT proved clinically relevant, which implies that such patients should be treated primarily with multiagent chemotherapy. However, reclassification of PCLBCL-leg (EORTC) with a predominant cleaved cell morphology into the group of indolent PCFCL proved less fortunate, since these patients have a much worse prognosis than PCFCL not presenting on the leg and should not be treated routinely with radiotherapy. Consistent with the results of a previous European multicenter study,9 we suggest that in addition to the WHO-EORTC classification, site of presentation also be taken into account to select the most appropriate treatment in patients with PCFCL.
The authors indicated no potential conflicts of interest.
Conception and design: Maarten H. Vermeer, Chris J.L.M. Meijer, Rein Willemze Provision of study materials or patients: Juliette J. Hoefnagel, Patty M. Jansen, Maarten H. Vermeer, Joop van Baarlen, Willeke A. Blokx, Marijke R. Canninga-van Dijk, Marie-Louise Geerts, Konnie M. Hebeda, Philip M. Kluin, King H. Lam, Chris J.L.M. Meijer Collection and assembly of data: Nancy J. Senff, Patty M. Jansen, Joop van Baarlen, Willeke A. Blokx, Marijke R. Canninga-van Dijk, Marie-Louise Geerts, Konnie M. Hebeda, Philip M. Kluin, King H. Lam, Chris J.L.M. Meijer, Rein Willemze Data analysis and interpretation: Nancy J. Senff, Juliette J. Hoefnagel, Patty M. Jansen, Maarten H. Vermeer, Chris J.L.M. Meijer, Rein Willemze Manuscript writing: Nancy J. Senff, Maarten H. Vermeer, Rein Willemze Final approval of manuscript: Nancy J. Senff, Juliette J. Hoefnagel, Patty M. Jansen, Maarten H. Vermeer, Joop van Baarlen, Willeke A. Blokx, Marijke R. Canninga-van Dijk, Marie-Louise Geerts, Konnie M. Hebeda, Philip M. Kluin, King H. Lam, Chris J.L.M. Meijer, Rein Willemze
We thank Enno Dreef for excellent technical assistance and Ron Wolterbeek for providing statistical advice.
published online ahead of print at www.jco.org on March 12, 2007. Supported by Grant No. UL 2006-3705 from the Dutch Cancer Society (N.J.S., M.H.V.). Authors’ disclosures of potential conflicts of interest and author contributions are found at the end of this article.
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2007 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
