567

Mantle Cell LymphomaPresenting Features, Response to Therapy, and Prognostic Factors

BACKGROUND. The goal of this study was to analyze the presenting features, naturalFrancesc Bosch, M.D.1

history, and prognostic factors in 59 patients with well characterized mantle cellArmando Lopez-Guillermo, M.D.1

lymphoma (MCL).Elıas Campo, M.D.2

METHODS. Cases were classified as nodular or diffuse and as typical or blasticJosep M. Ribera, M.D.3

variants. Age, performance status (PS), histologic variants, mitotic index (MI), he-Eulogio Conde, M.D.4

matologic parameters, tumor extension data, and International Prognostic IndexMiguel Angel Piris, M.D.5

(IPI) were recorded and evaluated for prognosis.Teresa VallespıB , M.D.6

RESULTS. The median age of the patients was 63 years (range, 39–83 years), andSoledad Woessner, M.D.7

the male to female ratio was 3:1. Fifty-three patients had typical histology (3 nodu-Emilio Montserrat, M.D.1

lar and 50 diffuse), and 6 had the blastic variant. Approximately 95% of patients

presented with advanced stage disease (Ann Arbor Stage III–IV). Leukemic expres-1 Department of Hematology, University of Bar-celona, Barcelona, Spain. sion was observed in 58%. Complete and partial response rates were 19% and 46%,

respectively. Parameters associated with lower response rate were Stage IV, high/2 Department of Pathology, Hospital ClıBnic, Uni-intermediate or high risk IPI, and increased lactate dehydrogenase (LDH) level. Inversity of Barcelona, Barcelona, Spain.the logistic regression analysis, high LDH level and Stage IV disease were associated3 Department of Hematology, Hospital ‘‘Ger-independently with lower response rate. Median survival was 49 months. Parame-mans Trias i Pujol,’’ Badalona, Spain.ters associated with a short survival were: poor PS, splenomegaly, B-symptoms,

4 Department of Hematology, Hospital ‘‘Marques MI ú 2.5, leukocyte count ú 10 1 109/L, high LDH level, blastic variant, and high/de Valdecilla,’’ Santander, Spain.

intermediate or high risk IPI. In the Cox proportional hazards regression model,5 Department of Pathology, Hospital ‘‘Virgen de only poor PS (relative risk [RR] Å 3.3; P Å 0.002), splenomegaly (RR Å 2.8; P Åla Salud,’’ Toledo, Spain. 0.007), and MI ú 2.5 (RR Å 2.4; P Å 0.012) were associated with short survival.

CONCLUSIONS. In this series, patients with MCL presented with advanced stage6 Laboratory of Hematology, Hospital Generalde la Vall d’Hebron, Barcelona, Spain. and extranodal involvement. Only a minority of patients achieved a complete

response. The median survival was 4 years, with PS, splenomegaly, and MI being7 Laboratory of Hematology, Hospital Central dethe most important factors predicting survival. These results show clearly thatl’Alianca, Barcelona, Spain.more effective therapies for MCL are needed. Cancer 1998;82:567–75.

q 1998 American Cancer Society.

KEYWORDS: lymphoma, mantle cell, prognosis, therapy, cyclin D1.

Mantle cell lymphoma (MCL) is a recently redefined category ofmalignant lymphoma that represents 5–10% of all lymphoma

cases. It comprises the lymphomas formerly known as intermediateSupported in part by Grants 98/0202, 97/0626, according to Berard et al1 and the centrocytic lymphomas in the Kieland 96/0236 from Fondo de Investigaciones classification,2 whereas in the Working Formulation the majority ofSanitarias de la Seguridad Social, Spanish Min-

cases were included within the category of diffuse small cell malignantistry of Sanidad y Consumo and beca de Xarxalymphoma.3 Histologically, MCL usually shows a diffuse growth pat-Tematica 1997XT00023.tern; in addition, other morphologic variants, including a nodular

Address for reprints: Emilio Montserrat, M.D., pattern and the so-called blastic (or blastoid) variant, can be observedDepartment of Hematology, Hospital ClıBnic, Vil- occasionally.4–7

larroel 170, 08036-Barcelona, Spain.Patients with MCL usually are elderly and the disease predomi-

nates in males; patients present with advanced stage and frequentReceived May 12, 1997; revision received July25, 1997; accepted July 25, 1997. extranodal involvement (e.g., gastrointestinal tract, Waldeyer’s ring,

q 1998 American Cancer Society

/ 7ba8$$0740 01-16-98 13:40:56 cana W: Cancer

568 CANCER February 1, 1998 / Volume 82 / Number 3

lung, and pleura).7,8 The mantle cells have a distinctive izations [4 patients]). Posttherapy restaging includedrepetition of previously abnormal tests and/or biop-immunophenotype, with expression of pan B-cell

markers (CD19, CD20, and CD22) and the T-cell sies.Complete response (CR) was defined as the totalmarker CD5; conversely, these cells do not express

CD23 and CD10 antigens.5,9 In the cytogenetic analy- disappearance of tumor masses and disease-relatedsymptoms as well as the normalization of the initialsis, MCL is characterized by the t(11;14)(q13;q32) that

juxtaposes the bcl-1 locus on chromosome 11 to an abnormal tests for at least 1 month. Partial response(PR) was considered when tumor masses or organ in-immunoglobulin enhancer located on chromosome

14.10,11 The PRAD-1/CCND1/cyclin D1 gene is the bcl- filtration decreased by at least 50% along with the dis-appearance of disease-related symptoms. Patients not1-related gene and its expression is up-regulated in cell

lines or lymphoproliferative disorders with the t(11;14) included in these categories and early deaths wereconsidered as nonresponders. The median follow-uptranslocation. The cyclin D1 gene, which encodes

cyclin D1, initially was identified on chromosome for the surviving patients was 29 months.11q13 in patients with parathyroid adenoma.12,13 Thisgene product plays an important role in the cell cycle

Histopathologycontrol at the G1-S transition, most likely by inter-

The specimens studied were: lymph nodes (55 pa-acting with the retinoblastoma gene protein.14 Re-

tients), spleen (5 patients), gastrointestinal biopsiescently, we and other authors demonstrated that the

(10 patients), and tonsils (5 patients). In two cases,cyclin D1 gene overexpression is a highly specific and

different tissues were analyzed (lymph nodes, spleen,sensitive molecular marker of MCL.13,15,16

and tonsil). Each biopsy was reviewed by two of usThe unique histologic and biologic features of

(E.C. and M.A.P.) and the cases were classified ac-MCL have renewed interest in this lymphoma subtype.

cording to the currently accepted histologic crite-Thus, in 1994 the European Lymphoma Task Force

ria.5,6,17 In all cases the mitotic index (MI) was calcu-defined the diagnostic criteria of MCL17; conversely,

lated and expressed as mitoses per high-power fieldthis subtype of malignant lymphoma has been in-

(mitoses/HPF). Immunophenotypic studies were per-cluded as a distinct entity in the Revised European-

formed in all the cases using immunohistochemistryAmerican Classification of Lymphoid Neoplasms.18

on frozen or fixed paraffin embedded tissue sectionsNevertheless, the presenting features and outcome of

and/or flow cytometry on cell suspensions. Althoughpatients with MCL only have been reported in few

the number of antibodies used varied from case tostudies. Moreover, in some previous reports, patients

case, the immunophenotyping included immunoglob-were not diagnosed according to the currently ac-

ulin light and heavy chains, and B-cell (CD19, CD20,cepted criteria, biologic and histologic data were not

CD22, CD23, CD45RA, CD10, and FMC7) and T-cellincluded, and/or prognostic factors were not assessed

markers (CD2, CD3, CD5, CD7, CD43, and CD45R0);properly.6,8,9,19–23

follicular dendritic cells were identified using CD35The objectives of this study were to describe the

and/or R4/23.presenting features and to analyze the natural historyand prognostic factors in a series of 59 patients withMCL. Cytogenetic and Molecular Analysis

Cytogenetic analysis was performed in 21 patientswith leukemic expression. Twenty of these patientsPATIENTS AND METHODS

Patients had a typical MCL and 1 had a blastic variant. Tumoralcells isolated from peripheral blood were analyzed us-Fifty-nine patients (median age, 63 years, 74% male)

diagnosed with MCL between December 1983 and No- ing G banding after 72-hour phytohemagglutinin andlipopolysaccharide from Escherichia coli-stimulatedvember 1995 in 7 Spanish institutions were included

in the study after a centralized review of the histologic cultures. The karyotypes were established accordingto the International System for Human Cytogeneticdiagnosis.

Staging maneuvers at diagnosis in most patients Nomenclature.bcl-1 rearrangement was studied by Southern blotincluded bone marrow biopsy (54 patients), and tho-

racic and abdominal computed tomograpy (CT) scans analysis using a 2.3-kilobase Sac I fragment for theMTC probe and/or by polymerase chain reaction ac-(all patients). Liver biopsy was performed in 12 pa-

tients. Biopsies of extranodal sites were performed cording to a previous described technique.15,24 mRNA-cyclin D1 gene overexpression was assessed by meanswhenever its involvement was suspected (tonsil [5 pa-

tients], gastrointestinal tract [10 patients], pleural/pul- of Northern blot analysis in 28 cases according to pre-viously described methods.15monary [10 patients], skin [2 patients], and other local-

/ 7ba8$$0740 01-16-98 13:40:56 cana W: Cancer

Prognostic Factors in Mantle Cell Lymphoma/Bosch et al. 569

TABLE 1Parameters EvaluatedMain Initial Characteristics of 59 Patients with Mantle CellIn each patient the following initial data were recordedLymphomaand evaluated for prognosis: 1) clinical data: age, gen-

der, performance status (PS) according to the Eastern AgeCooperative Oncology Group (ECOG) scale, and pres- Median (yrs) 63

Age ú60 yrs 45%ence of B-symptoms (fever, night sweats, weight loss);Male/female ratio 3:12) histologic data: histologic variants (typical or blasticHistologic subtypeand nodular or diffuse) and the MI; 3) hematologic

Typical 53and biochemical parameters: hemoglobin, leukocyte Blastic variant 6count, leukemic expression (presence of atypical lym- CD5 antigen (/) 96%

CD23 antigen (0) 86%phocytes in peripheral blood), platelet count, and se-t(11;14) bcl-1 rearrangement 10/21rum lactate dehydrogenase (LDH) level; 4) tumor ex-Cyclin D1 gene overexpression 28/28tension data: number of lymph node and extranodalPerformance status (ECOG ¢2) 51%

involved sites, spleen palpable below the left costal B symptoms 25%margin, hepatomegaly, Ann Arbor stage, and bone Bulky diseasea 5%

Splenomegaly 44%marrow infiltration; and 5) the International Prognos-Increased LDH level 40%tic Index (IPI), as defined by the International Non-Leukocyte count ú10 1 109/L 53%Hodgkin’s Lymphoma Prognostic Factors Project24

Leukemic expression 58%considering the following variables: age (° 60 years Ann Arbor Stagevs. ú 60 years), performance status (ECOG 0 or 1 vs. I, II 5%

III 9%2–4), Ann Arbor stage (I–II vs. III–IV), extranodalIV 86%involvement (õ 2 vs. ¢ 2 involved sites), and serum

Extranodal involvement 91%LDH level (normal vs. high). Patients with none or oneBone marrow 81%

unfavorable variable were considered low risk, those GI tract 17%with two variables as low/intermediate risk, those with Pleural 17%

Tonsil 8%three variables as high/intermediate risk, and those¢2 extranodal involved sites 33%with four or five adverse variables as high risk.25

IPILow risk 15%

Statistical Methods Low/intermediate risk 22%Differences among the different subgroups in terms High/intermediate risk 36%

High risk 27%of the parameters evaluated were compared by theFisher’s exact test (two-tailed) and the Student’s t test.

ECOG: Eastern Cooperative Oncology Group; LDH: lactate dehydrogenase; GI: gastrointestinal; IPI:The actuarial survival analysis was performed ac-International Prognostic Index.

cording to the method described by Kaplan and a Bulky disease indicates tumor with a dimension ¢10 cm.Meier26 and the curves were compared by the log ranktest.27 The univariate analysis was performed for eachof the parameters mentioned earlier. All significantprognostic variables in the univariate study, as well histologic variants (50 diffuse, and 3 nodular) and 6

with blastic histology. The MI varied from 0.1 to 9.4,as those with clinical relevance, were considered formultivariate analysis performed by the stepwise Cox with a mean of 1.78 mitoses/HPF. All the cases were

CD10 negative; 55 of 57 cases (96%) expressed the CD5proportional hazards regression model (survival) andlogistic regression (response). P values ° to 0.05 were antigen, whereas 32 of 37 cases (86%) lacked the CD23

antigen. The t(11;14)(q13;q32) or bcl-1 rearrangementconsidered to indicate statistical significance. Datawere analyzed with the BMDP statistical software was demonstrated in 10 of 21 cases studied. All cases

analyzed (28 of 28) overexpressed the cyclin D1 gene.package (BMDP Statistical Software, Inc., Berkeley,CA).28 Overall, there were 35 cases (59%) with cytogenetic or

molecular features of MCL.In 30 patients (51%) the performance status wasRESULTS

Patient Characteristics poor (ECOG ¢ 2), and 15 (25%) had B-symptoms. In28 patients (47%) hepatomegaly was found, and in 26The main initial characteristics of the 59 patients are

detailed in Table 1. The median age was 63 years (44%) the spleen was palpable below the left costalmargin. The great majority of patients presented with(range, 39–83 years); 45% of patients were age ú 60

years, and 74% were males (male/female ratio of 3:1). advanced disease (95% in Stage III–IV). Extranodalinvolvement was found in 54 patients (92%), includingFifty-three patients were diagnosed with typical

/ 7ba8$$0740 01-16-98 13:40:56 cana W: Cancer

570 CANCER February 1, 1998 / Volume 82 / Number 3

bone marrow infiltration in 44 of 54 patients investi- Å 0.017), and high-intermediate or high risk IPI (P Å0.009) (Table 2). No differences were observed ac-gated (81%), liver in 3 of 12 cases (25%), lung and/or

pleura in 10 (17%), gastrointestinal tract in 10 (17%), cording to the treatment administered; the responserate (CR and PR) of patients receiving regimens con-tonsils in 5 (8%), and other sites in 5 (8%) (skin in 2,

and peritoneum, ovary, and parotid in 1 patient each). taining anthracyclines was 61% compared with 66%in those patients who were treated without anthracy-Overall, 33% of patients had ¢ 2 extranodal sites in-

volved. Of note, although no patients had central ner- clines. Conversely, although not statistically signifi-cant, none of the six patients with blastic variantsvous system (CNS) infiltration at diagnosis, 7 patients

(12%) presented with CNS involvement during follow- achieved CR. In the logistic regression, the variableswith predictive value for CR achievement were stageup, between 4 and 65 months from diagnosis. The

actuarial risk for such a complication was 28% (95% (I–III vs. IV) (P Å 0.03) and LDH level (P Å 0.003)(Table 3).confidence interval, 7–49) at 72 months. It is interest-

ing to note that three of six patients with blastic vari- Six patients (66%) in CR and 14 (63%) in PR re-curred or progressed after a median of 32 and 18ants developed CNS infiltration. As per the IPI, 9 pa-

tients (15%) were low risk, 22% were low-intermediate months, respectively. The median of failure free sur-vival was 17 months. The analysis of variables associ-risk, 36% were high-intermediate risk, and 27% were

high risk. ated with recurrence was not possible due to the lim-ited remaining number of cases.Leukemic expression was found in 58% of pa-

tients. The mean leukocyte count was 39.7 1 109/L(range, 4–322 1 109/L), and the mean percentage of Survival

After a median follow-up of 29 months, 27 patientsatypical lymphocytes was 44% (range, 5–99%). Only18% of patients had ú 50 1 109/L leukocytes. In 25% (46%) had died. The overall median survival of the

series was 49 months (Fig. 1). Disease progression wasof the patients the hemoglobin level was õ 110 g/L orthe platelet count was õ 100 1 109/L. The serum LDH the cause of death in 22 patients; 5 patients died from

infection with active disease. The median survival forlevel was increased in 40% of patients, and the b2-microglobulin level in 64%. Finally, in 12% of patients patients with typical histology was 50 months whereas

it was 18 months for those with blastic variants. In thea monoclonal serum component was observed (im-munoglobulin [Ig]G in 5 patients and IgM in 1 patient). univariate analysis, eight initial variables were associ-

ated with a shorter survival (Table 2): poor PS (ECOGNone of the five IgG patients had surface IgG expres-sion in the tumoral cells, whereas the IgM patient had ¢ 2) (P Å 0.0014) (Fig. 2), presence of B symptoms (P

Å 0.05), splenomegaly (P Å 0.0122), high leukocytethis heavy chain surface expression.According to the histologic subtype, patients with count (ú 10 1 109/L) (P Å 0.0496), increased serum

LDH level (P Å 0.0068), histologic variant (blastic sub-blastic variants presented with more mitoses/HPF (5.7vs. 1 mitoses in mean) (Põ 0.0001), higher LDH levels type) (P Å 0.0072) (Fig. 3), high MI (ú 2.5 mitoses/

HPF) (P Å 0.001) (Fig. 4), and advanced IPI (high/(83% vs. 33%) (PÅ 0.028), and less frequent bone mar-row involvement (50% vs. 84%) (P Å 0.045) than those intermediate and high risk) (P Å 0.05) (Fig. 5). In the

multivariate regression model, 3 variables retainedwith other histologic variants; no significant differ-ences between these two groups were found as per prognostic significance: poor PS (relative risk [RR] Å

3.3; P Å 0.002), splenomegaly (RR Å 2.8; P Å 0.007),the other initial characteristics.and high MI (RR Å 2.4; P Å 0.012) (Table 3).

The median survival for patients treated with com-Response to Therapy and RecurrenceTreatment varied over time and according to the insti- bination chemotherapy including anthracyclines was

46 months, whereas it was 69 months for those whotutions, and was comprised of combination chemo-therapy regimens without anthracyclines in 16 pa- did not receive anthracyclines (P Å 0.029). Patients

who received anthracyclines were younger (mean age,tients, combination chemotherapy including anthra-cyclines in 38 patients, and radiotherapy as unique 61 years [range, 39–78 years] vs. 67 years [range, 51–

83 years]) (P Å 0.03) and had a higher MI (2.4 vs. 0.46treatment in 1 patient; 5 patients were not treated, inthree cases because they died before starting treat- mitoses/HPF) (P Å 0.013) than those treated without

anthracyclines. All patients with blastic variants werement.Forty-eight patients were evaluable for response. treated with anthracyclines. Nevertheless, the type of

treatment was not significant for survival when in-CR, PR, and failure rates were 19%, 46%, and 35%,respectively. Parameters associated with a lower re- cluded in the multivariate regression model.

Response to therapy (considering both CR and PR)sponse rate in the univariate analysis were advancedAnn Arbor stage (P Å 0.017), increased LDH level (P was associated with longer survival (P õ 0.0001), with

/ 7ba8$$0740 01-16-98 13:40:56 cana W: Cancer

Prognostic Factors in Mantle Cell Lymphoma/Bosch et al. 571

TABLE 2Initial Prognostic Factors for Survival and Response to Therapy in the Overall Series Derived from the Univariate Analysis

% survivingNo. % CR P value at 2 years P value

Overall series 59 19 — 69 —Age (yrs)°60 21 22 50ú60 38 17 NS 72 NS

GenderMale 44 16 63Female 15 30 NS 77 NS

HistologyTypical 53 20 73Blastic variant 6 — NS 20 0.0072

Mitotic index°2.5 m/HPF 49 15 72ú2.5 m/HPF 10 17 NS 36 0.0292

Performance statusECOG 0 and 1 29 25 82ECOG ú1 30 12 NS 49 0.0014

B symptomsNo 43 22 70Yes 16 9 NS 46 0.05

SplenomegalyNo 33 25 80Yes 26 10 NS 47 0.0122

Extranodal involvementõ2 sites 40 22 74¢2 sites 19 12 NS 59 NS

Ann Arbor StageI, II, III 8 57 83IV 51 12 0.017 63 NS

Leukocyte countõ10 1 109/L 28 21 78¢10 1 109/L 31 15 NS 51 0.049

Leukemic expressionNo 25 24 73Yes 34 15 NS 62 NS

Increased LDH levelNo 35 29 70Yes 24 — 0.017 58 0.0268

IPI°2 22 39 81ú2 37 7 0.009 37 0.042

TreatmentNo anthracyclines 16 21 85Anthracyclines 38 15 NS 55 0.005

CR: complete response; NS: not significant; M/HPF: mitoses per high-power field; ECOG: Eastern Cooperative Oncology Group; LDH: lactate dehydrogenase; IPI: International Prognostic Index.

a median survival of 69 months for patients having DISCUSSIONThe term MCL has been proposed to describe a cate-achieved a response compared with 22 months in

those who failed to respond; there was no statistical gory of malignant lymphoma that arises in the follicu-lar mantle zone and has unique histologic and biologicdifference in survival between CR and PR patients.

When response to therapy was included in the multi- characteristics. Immunophenotypically, the neoplasticcells express the CD5 antigen along with a number ofvariate analysis, it became the most important prog-

nostic factor (RR Å 5.2; P Å 0.0053), with only MI pan B-cell markers (CD19, CD20) but not CD23. Incytogenetic study, the t(11;14)(q13;q32) usually isretaining its prognostic value (RR Å 4.2; P Å 0.04).

/ 7ba8$$0740 01-16-98 13:40:56 cana W: Cancer

572 CANCER February 1, 1998 / Volume 82 / Number 3

TABLE 3Main Prognostic Factors Derived from the Multivariate Analysis

Response Survivalb

achievementa

P value P value RR

Poor PS (ECOG ú1) NS 0.002 3.3Splenomegaly NS 0.007 2.8Mitotic index (ú2.5

mitoses/HPF) NS 0.012 2.4Increased LDH level 0.003 NS NSStage (IV) 0.03 NS NS

RR: relative risk; PS: performance status; ECOG: Eastern Cooperative Oncology Group; NS: not signifi-

cant; HPF: high-power field; LDH: lactate dehydrogenase.a Logistic regression.

FIGURE 3. Overall survival curve of patients with mantle cell lymphomab Cox proportional hazards regression model.grouped according to the histologic subtype (typical vs. blastic variants,P Å 0.0072). The median survival of patients with typical histology was50 months vs. 18 months for patients with blastic variants.

FIGURE 1. Overall survival curve of patients with mantle cell lymphoma(59 patients; median survival, 49 months).

FIGURE 4. Overall survival by mitotic index (MI) (expressed as mitoses/high-power field) in the 59 patients with mantle cell lymphoma.

found implicating the cyclin D1 gene in the pathogen-esis of MCL. Moreover, the cyclin-D1 gene characteris-tically is overexpressed at mRNA and protein levels inthe majority of cases. These features have resulted ina renewed interest in MCL. Although the diagnosticcriteria of MCL have been defined,17,18 the clinicalcharacteristics and outcome of patients are not so wellknown. In a number of series patients were not diag-nosed according to previous classifications based onthe current concept of MCL or the biologic features ofMCL were not analyzed. In our series, in addition toa centralized review of the histology, in 10 of 21 casesanalyzed the bcl-1 rearrangement or t(11;14)(q13;q32)FIGURE 2. Overall survival curve of patients with mantle cell lymphoma

according to performance status. ECOG: Eastern Cooperative Oncology was found, which is a proportion similar to that re-ported in other studies29; Conversely, all cases ana-Group.

/ 7ba8$$0740 01-16-98 13:40:56 cana W: Cancer

Prognostic Factors in Mantle Cell Lymphoma/Bosch et al. 573

TABLE 4Comparison of the Current Study with Other Previously Published Series

Current Velders Zucca TeodorovicVariable study et al.22 et al.8 Norton et al.20 Fisher et al.19 et al.31 Pittaluga et al.9 Berger et al.21 Majlis et al.23

No. of patients 59 41 65 66 36 64 55 52 46Median age (yrs) 63 68 64 62 55 58 68 58% ú60 yrs 54Male:female ratio 3:1 1.6:1 2:1 3.7:1 4:1 2.7:1 6.8:1 NR 1.7:1Molecular studies Yes Yes No No No No Yes No YesHistologic distribution Nodular: 5% All diffuse NR Nodular: 36% Nodular: 39% Nodular: Nodular: 11% Nodular: 22% Nodular: 13%

Blastic: 10% Blastic: 15% Blastic: 33% 3% Mantle zone: 26%Poor PS 50% 20% 20% NR NR 6% NR 23% NRSplenomegaly 44% NR 35% 48% NR 33% NR 59% NRBone marrow involvement 75% 80% 58% 80% 53% 60% 66% 82% 69%Leukemic expression 58% NR 20% NR NR NR NR 49% NRGI tract involvement 17% NR 15% 12% 19% NR NR 20% 24%Stage IV 86% 78% 72% 82% NR NR 62% 89% 82% (III & IV)Increased LDH level 40% 55% 30% NR NR NR NR 45% NRCR 19% 32% 51% 9% NR 52% NR 31% a

Duration of CR (mos) 17 25 44 10 6%b 21 NR 14 a

Median survival (mos) 49 32 42 36 8%b 45 32 52 a

NR: not reported; PS: performance status; GI: gastrointetinal; LDH: lactate dehydrogenase; CR: complete response.a Variable according to histologic subtype.b At 10 years.

cases. Involvement of the gastrointestinal tract wasobserved in 17%, which is similar to the rate de-scribed in other series8,9,19 – 23,30,31 (Table 4), althoughbecause gastrointestinal studies were not con-ducted routinely in all patients this could be anunderestimation. A significant proportion (12%) ofpatients developed CNS infiltration during thecourse of the disease, which is in agreement with aprevious report from our group.32 In contrast toother extranodal infiltration, CNS involvement gen-erally appears as a late event in the history of MCLpatients. Finally, in the current study a significantproportion of patients (58%) showed peripheralblood involvement; this high proportion can be dueto the fact that some of these patients were referred

FIGURE 5. Overall survival curve of 59 patients with mantle cell to us with a diagnosis of chronic lymphocytic leuke-lymphoma grouped according to the International Prognostic Index (IPI) mia.(low and low/intermediate risk vs. high/intermediate and high risk). With regard to the histologic distribution, the

proportion of different infiltration patterns is vari-able according to the series analyzed. In our series,

lyzed (n Å 28) overexpressed the cyclin D1 gene. Thus, only 5% of the cases had a nodular pattern andapproximately 60% of the cases from this series had a 10% of the cases corresponded to blastic variants.cytogenetic or molecular marker of MCL, supporting Differences in diagnostic criteria and patterns ofthe histologic diagnosis. referral could explain the different proportion of

The clinical characteristics found in the cur- nodular cases observed in the literature (from 3 –rent series were similar to those reported pre- 39%) (Table 4). For example, some of our cases wereviously8,9,19 – 23,30,31 (Table 4): a median age of approx- referred as ‘‘atypical’’ chronic lymphoid leukemias,imately 65 years, male predominance, advanced with the final diagnosis being MCL with leukemicstage, and frequent extranodal involvement, in- expression, a situation in which either diffuse or

blastic variants predominate. In addition, as including bone marrow infiltration in the majority of

/ 7ba8$$0740 01-16-98 13:40:56 cana W: Cancer

574 CANCER February 1, 1998 / Volume 82 / Number 3

other series,9,30 no cases of mantle zone variant due to the fact that most patients were consideredhigh/intermediate or high risk.were observed. This is of interest, because such

mantle zone patients apparently would have a more In most series, including the current one, MCLpatients have been treated very heterogeneously. In-indolent outcome.23,33 The histologic and cytologic

distinction between diffuse and blastic variants has dependent of the regimen used, CRs are rare (approxi-mately 10–40% of the cases) (Table 4). The role ofprognostic implications because blastic variant

cases have a worse behavior than those with typical anthracycline-containing regimens in MCL therapy isunclear. Thus, no differences were found in a random-histology (Fig. 2). In fact, none of the patients with

blastic variants from our series achieved CR and ized study comparing the COP (cyclophosphamide,vincristine, and prednisone) regimen versus the CHOPtheir median survival was only 18 months. The rea-

sons for the unfavorable outcome of patients with (cyclophosphamide, doxorubicin, vincristine, andprednisone) regimen in patients with centrocyticblastic variants currently are under study. In this

sense, the differences in levels of cyclin D1 gene lymphoma.38 In this study, the type of treatment didnot reach independent significant value in the multi-expression did not correlate with the pathologic

characteristics of MCL and survival.15,22,34 This indi- variate analysis. Finally, taking into account the poorprognosis of patients with MCL with current therapies,cates that other molecular alterations such as p53

gene mutations might be implicated in those cases the role of new and more aggressive strategies in themanagement of this type of lymphoma, including stemwith a more aggressive outcome.35,36 In this series,

two cases classified as blastic variants had p53 gene cell transplants, warrant investigation.39,40

mutations,35 three had deletion and lack of expres-sion of the p16/INK4a gene, and in the remaining case REFERENCESthere was a lack of protein and mRNA p21/Waf1 gene 1. Berard CW, Dorfman RF. Histopathology of malignantexpression without p53 mutation.37 These biologic lymphomas. Clin Haematol 1974;3:39–43.

2. Lennert K. Centrocytic lymphoma. In: Lennert K, editor. His-features that were not found in any of the 18 exam-topathology of non-Hodgkin’s lymphoma (based on the Kielined typical variants may explain, at the molecularclassification). New York, NY: Springer-Verlag, 1981:57–62.level, the different clinical course of blastic vari-

3. Non-Hodgkin’s lymphoma pathologic classification project.ants.37

National Cancer Institute sponsored study of classificationsIn the current study, the overall median survival of non-Hodgkin’s lymphomas: summary and description of

a Working Formulation for clinical usage. Cancer 1982;49:was 49 months with a continuous decline of the sur-2122–35.vival curve (Fig. 1). The median survival of MCL pa-

4. Weisenburger DD, Kim H, Rappaport H. Mantle-zonetients ranges between 2–5 years (Table 4). The multi-lymphoma. A follicular variant of intermediate lymphocytic

variate regression model for individual prognostic fac- lymphoma. Cancer 1982;49:1429–38.tors showed that the MI, along with splenomegaly and 5. Banks PM, Chan J, Cleary ML, Delsol G, De Wolf-Peeters

C, Gatter K, et al. Mantle cell lymphoma: a proposal forpoor PS, retained prognostic importance on survival.unification of morphologic, immunologic, and molecularThe patients with ú 2.5 mitoses/HPF had a mediandata. Am J Surg Pathol 1992;16:637–40.survival of 24 months, whereas those with ° 2.5 mito-

6. Lardelli P, Bookman MA, Sundeen J, Longo DL, Jaffe ES.ses/HPF had a survival of 50 months (Fig. 4). In fact, Lymphocytic lymphoma of intermediate differentiation:all the patients with blastic variants hadú 2.5 mitoses/ morphologic and immunophenotypic spectrum and clinical

correlations. Am J Surg Pathol 1990;14:752–63.HPF. Other authors have reported the prognostic value7. Weisenburger DD, Armitage JO. Mantle cell lymphoma-anof Ki-67 antigen expression.22 In this setting we re-

entity comes of age. Blood 1996;87:4483–94.cently showed a correlation between Ki-67 expression,8. Zucca E, Roggero E, Pinotti G, Pedrinis E, Cappella C, Venco

the proportion of cells in S-phase, and mitotic activity. A, et al. Patterns of survival in mantle cell lymphoma. AnnThe proliferative activity also correlated with the levels Oncol 1995;6:257–62.

9. Pittaluga S, Wlodarska I, Stul MS, Thomas J, Verhoef G, Cas-of expression of the retinoblastoma gene protein andsiman JJ, et al. Mantle cell lymphoma: a clinicopathologicalwere independent of cyclin D1 mRNA levels.34 Withstudy of 55 cases. Histopathology 1995;26:17–24.regard to prognostic factors, Zucca et al8 found that

10. Weisenburger DD, Sanger WG, Armitage JO, Purtilo DT. In-age ú 65 years and increased LDH level correlated termediate lymphocytic lymphoma: immunophenotypicwith a poor prognosis, whereas Norton et al,20 in a and cytogenetic findings. Blood 1987;69:1617–25.

11. Vandenberghe E, De Wolf-Peeters C, Wlodarska I, Stul M,series in which 30% of patients had a nodular growthLouwagie A, Verhoef G, et al. Chromosome 11q rearrange-pattern, found that age ú 70 years, splenomegaly, lowments in B non-Hodgkin’s lymphoma. Br J Haematolalbumin serum level, and advanced stage of disease1992;81:212–7.

were associated with short survival. It is interesting to 12. Motokura T, Bloom T, Kim HG, Juppner H, Ruderman JV,note that in our analysis the IPI did not retain prognos- Kronenberg HM, et al. A novel cyclin encoded BCL1-linked

candidate oncogene. Nature 1991;350:512–5.tic value in the multivariate analysis; this probably is

/ 7ba8$$0740 01-16-98 13:40:56 cana W: Cancer

Prognostic Factors in Mantle Cell Lymphoma/Bosch et al. 575

13. E)2/E in the log-rank test for survival data or tumor incidenceRosenberg CL, Wong E, Petty EM, Bale AE, Tsujimoto Y,data. Biometrics 1973;29:759–84.Harris NL, et al. PRAD1, a candidate bcl1 oncogene: map-

28. Dixon WJ, Brown MB, Engelman L, Hill MA, Jennrich RI.ping and expression in centrocytic lymphoma. Proc NatlDixon WJ, editors. BMDP statistical software. Berkeley, CA:Acad Sci USA 1991;88:9638–42.University of California Press, 1988:719–43.14. Dowdy SF, Hinds PW, Louie K, Redd SI, Arnold A, Weinberg

29. Williams ME, Swerdlow SH, Rosenberg CL, Arnold A. Char-RA. Physical interaction of the retinoblastoma protein withacterization of chromosome 11 translocation breakpoints athuman D cyclins. Cell 1993;73:499–11.the bcl-1 and PRAD1 loci in centrocytic lymphoma. Cancer15. Bosch F, Jares P, Campo E, Lopez-Guillermo A, Piris MA,Res 1992;52:5541s–4s.Villamor N, et al. PRAD1/Cyclin D1 gene overexpression in

30. Zucca E, Fontana S, Roggero E, Pedrinis E, Pampallona S,chronic lymphoproliferative disorders: a highly specificCavalli F. Treatment and prognosis of centrocytic (mantlemarker of mantle cell lymphoma. Blood 1994;84:2726–32.cell) lymphoma: a retrospective analysis of twenty-six pa-16. Rimokh R, Berger F, Delsol G, Charrin C, Bertheas MF,tients treated in one institution. Leuk Lymphoma 1996;13:Ffrench M, et al. Rearrangement and overexpression of the105–10.bcl-1/PRAD-1 gene in intermediate lymphocytic lympho-

31. Teodorovic I, Pittaluga S, Kluin-Nelemans JC, Meerwaldt JH,mas and in the t(11q13)-bearing leukemias. BloodHagenbeek A, van Glabbeke M, et al. Efficacy of four differ-1993;81:3063–7.ent regimens in 64 mantle-cell lymphoma cases: clinico-

17. Zucca E, Stein H, Coiffier B. European Lymphoma Taskpathologic comparison with 498 other non-Hodgkin’s

Force (ELTF): report of the Workshop on mantle celllymphoma subtypes. European Organization for the Re-

lymphoma (MCL). Ann Oncol 1994;5:507–11.search and Treatment of Cancer Lymphoma Cooperative

18. Harris NL, Jaffe ES, Stein H, Banks PM, Chan JK, Cleary ML, Group. J Clin Oncol 1995;13:2819–26.et al. A revised European-American classification of 32. Montserrat E, Bosch F, Lopez-Guillermo A, Graus F, Terollymphoid neoplasms: a proposal from the International MJ, Campo E, et al. CNS involvement in mantle-cellLymphoma Study Group. Blood 1994;84:1361–92. lymphoma. J Clin Oncol 1996;14:941–4.

19. Fisher RI, Dahlberg S, Nathwani BN, Banks PM, Miller TP, 33. Garcia-Conde J, Cabanillas F. Mantle cell lymphoma: a newGrogan TM. A clinical analysis of two indolent lymphoma lymphoproliferative entity with definite histopathologicalentities: mantle cell lymphoma and marginal zone patterns, clinical characteristics and prognostic factors, andlymphoma (including the mucosa-associated lymphoid tis- an investigational therapeutic approach. Ann Oncol 1995;6:sue and monocytoid B-cell subcategories): a Southwest On- 305–6.cology Group study. Blood 1995;85:1075–82. 34. Jares P, Campo E, Pinyol M, Bosch F, Miquel R, Fernandez

20. Norton AJ, Matthews J, Pappa V, Shamash J, Love S, Rohati- PL, et al. Expression of retinoblastoma gene product (pRb) inner AZ, et al. Mantle cell lymphoma: natural history defined mantle cell lymphomas. Correlation with Cyclin D1 (PRAD1/in a serially biopsied population over a 20-year period. Ann CCND1) mRNA levels and proliferative activity. Am J PatholOncol 1995;6:249–56. 1996;148:1591–600.

35. Hernandez L, Fest T, Cazorla M, Teruya-Feldstein J, Bosch21. Berger F, Felman P, Sonet A, Salles G, Bastion Y, Bryon PA, etF, Peinado MA, et al p53 gene mutations and protein overex-al. Nonfollicular small B-cell lymphomas: a heterogeneouspression are associated with aggressive variants of mantlegroup of patients with distinct clinical features and out-cell lymphomas. Blood 1996;87:3351–9.come. Blood 1994;83:2829–35.

36. Louie DC, Offit K, Jaslow R, Parsa NZ, Murty VV, Schluger22. Velders GA, Kluin-Nelemans JC, De Boer CJ, Hermans J,A, et al p53 overexpression as a marker of poor prognosisNoordijk EM, Schuuring E, et al. Mantle-cell lymphoma: ain mantle cell lymphomas with t(11;14)(q13;q32). Bloodpopulation-based clinical study. J Clin Oncol 1996;14:1269–1995;86:2892–9.74.

37. Pinyol M, Hernandez L, Cazorla M, Balbin M, Jares P, Fer-23. Majlis A, Pugh WC, Rodriguez MA, Benedict WF, Cabanillasnandez PL, et al. Deletions and loss of expression of p16/F. Mantle cell lymphoma: correlation of clinical outcomeINK4a and p21/Waf1 genes are associated with aggressiveand biologic features with three histologic variants. J Clinvariants of mantle cell lymphomas. Blood 1997;89:272–80.Oncol 1997;15:1664–71.

38. Meusers P, Engelhard M, Bartels H, Binder T, Fulle HH,24. Pinyol M, Campo E, Nadal A, Terol MJ, Jares P, Nayach I,Gorg K, et al. Multicentre randomized therapeutic trial for

et al. Detection of the bcl-1 rearrangement at the majoradvanced centrocytic lymphoma. Anthracycline does not

translocation cluster in frozen and paraffin-embedded tis-improve the prognosis. Hematol Oncol 1988;7:365–80.

sues of mantle cell lymphomas by polymerase chain reac- 39. Stewart DA, Vose JM, Weisenburger DD, Anderson JR, Rubytion. Am J Clin Pathol 1996;105:532–7. EI, Bast MA, et al. The role of high-dose therapy and autolo-

25. The International non-Hodgkin’s lymphoma prognostic fac- gous hematopoietic stem cell transplantation for mantle celltors project: a predictive model for aggressive non-Hodg- lymphoma. Ann Oncol 1995;6:263–6.kin’s lymphoma. N Engl J Med 1993;329:987–94. 40. Romaguera J, Khouri I, Kantarjian H, Terol-Castera MJ,

26. Kaplan GL, Meier P. Non-parametric estimation from in- RodrıB guez MA, Hagemeister FB, et al. HCVAD/MTX-ARAC:complete observations. J Am Stat Assoc 1958;53:457–81. a new effective regimen for diffuse and nodular mantle

27. Peto R, Pike MC. Conservatism of the approximation (O- lymphoma (MCL). Blood 1996;88:568a.

/ 7ba8$$0740 01-16-98 13:40:56 cana W: Cancer