S�ezary syndrome: A study of 176 patients atMayo Clinic

Agnieszka W. Kubica, MD,d Mark D. P. Davis, MD,a Amy L. Weaver, MS,b Jill M. Killian, BS,b

and Mark R. Pittelkow, MDa,c

Rochester, Minnesota

From

St

M

M

Fund

Conf

Acce

Background: S�ezary syndrome (SS), a leukemic variant of cutaneous T-cell lymphoma, is characterized byerythroderma and by atypical lymphocytes (S�ezary cells) in peripheral blood. Although numerous studieshave examined the range of disease in cutaneous T-cell lymphoma, a relative paucity of data exists todescribe the long-term outcome of patients with SS.

Objective: We sought to study long-term survival and prognostic factors of patients with SS.

Methods: A retrospective chart review was conducted to identify patients with SS seen at Mayo Clinic from1976 to 2010. Cox proportional hazards regression models, adjusted for age, were fit to evaluate factorsassociated with overall survival.

Results: In total, 176 patients were identified with a clinicopathologic diagnosis of SS. Overall survival was86.1% and 42.3% at 1 and 5 years, respectively, after diagnosis (median survival, 4.0 years). After adjustmentfor age, potential predictors of worse survival included lactate dehydrogenase level at presentation (hazardratio [HR] 1.71; 95% confidence interval [CI] 1.18-2.47 per doubling), prior diagnosis of mycosis fungoides(HR 2.68; 95% CI 1.44-4.98), and the presence of T-cell receptor gene rearrangements in skin (HR 2.59; 95%CI 1.38-4.87) and in blood (HR 2.05; 95% CI 1.00-4.21).

Limitations: This study is retrospective and represents a single academic center population.

Conclusions: To our knowledge, this research evaluated the largest population of patients with SS studiedto date. It shows that overall survival continues to be poor, with a median survival of 4.0 years afterdiagnosis. ( J Am Acad Dermatol 2012;67:1189-99.)

Key words: cutaneous T-cell lymphoma; outcomes assessment; prognosis; S�ezary syndrome; survival.

Abbreviations used:

CI: confidence intervalCTCL: cutaneous T-cell lymphomaEORTC: European Organization for Research and

Treatment of CancerHR: hazard ratioIQR: interquartile rangeISCL: International Society for Cutaneous

LymphomasLDH: lactate dehydrogenaseMF: mycosis fungoidesSS: S�ezary syndromeTCR: T-cell receptor

S�ezary syndrome (SS) is a leukemic variant ofcutaneous T-cell lymphoma (CTCL) that arisesfrom skin-homing lymphocytes. It is charac-

terized by erythroderma, keratoderma, lymphade-nopathy, and the presence of atypical lymphocyteswith cerebriform nuclei (S�ezary cells) in peripheralblood.1-5 Over the past several decades, the diagno-sis of SS has undergone progressive revisions, withthe most recent diagnostic criteria proposed in2007 by the International Society for CutaneousLymphomas (ISCL)/European Organization forResearch and Treatment of Cancer (EORTC) and

the Department of Dermatology,a Division of Biomedical

atistics and Informatics,b Department of Biochemistry and

olecular Biology,c and Mayo Medical School, College of

edicine,d Mayo Clinic.

ing sources: None.

licts of interest: None declared.

pted for publication April 28, 2012.

Reprint requests: Mark D. P. Davis, MD, Department of

Dermatology, Mayo Clinic, 200 First St SW, Rochester, MN

55905. E-mail: davis.mark2@mayo.edu.

Published online May 28, 2012.

0190-9622/$36.00

� 2012 by the American Academy of Dermatology, Inc.

doi:10.1016/j.jaad.2012.04.043

1189

J AM ACAD DERMATOL

DECEMBER 20121190 Kubica et al

including immunophenotyping and molecular crite-ria, in an effort to better distinguish SS from mycosisfungoides (MF) and other erythrodermic CTCLs.6,7

These criteria updated the 1979 National CancerInstitute and MF Cooperative Group TNM(B) stagingclassification system that was previously used todefine MF and SS, and increased emphasis was

CAPSULE SUMMARY

d S�ezary syndrome is an aggressive variantof cutaneous T-cell lymphoma that isassociated with a poor prognosis.

d To our knowledge, this study populationis the largest of patients with S�ezarysyndrome reported in the medicalliterature and reaffirms the poorprognosis of this disease.

d This investigation analyzes prognosticfactors that are specific to this group ofpatients and adds data to the literatureon S�ezary syndrome.

placed on defining the de-gree of peripheral blood in-volvement in establishing adiagnosis.6 Recognizing theclinical challenges in defin-ing SS and separating it fromother erythrodermic CTCLs,the ISCL in 2002 also furtherestablished criteria and labo-ratory evaluation for thesesubtypes of CTCL.7

As a result of both theconstantly evolving under-standing of CTCL and therarity of SS in comparisonwith the relatively more com-mon CTCL variant, MF, stud-ies have frequently groupedmore indolent forms of

CTCL, such as erythrodermic MF or early to laterstages of MF with SS, resulting in incomplete orunstratified data on the actual survival, progression,and outcomes of documented SS.4 Therefore, fewpublished data exist on the long-term outcome anddisease course of patients with SS. The data in themedical literature suggest a poor prognosis for thisaggressive CTCL variant, with reported median over-all survival ranging from 2 to 4 years.5,7-9 Whenextracutaneous involvement of the viscera is identi-fied, median survival is typically 2.5 years or less.10-13

At our tertiary care academic referral institution,the dermatology department evaluates and cares fora large number of patients with SS, which constitutesonly about 2.5% of all CTCLs, according to dataobtained from Surveillance, Epidemiology, and EndResults.14 With the incidence of CTCL increasingdramatically since the 1970s,3 we examined the SSpopulation at our institution over the recent decades,focusing on long-term survival and identifying prog-nostic factors in these patients. This collective,single-institution experience over the past 4 decadesintends to supplement the literature on this more rareand aggressive lymphoma-leukemia and to compareour survival data with the literature at large.

METHODSThe Mayo Clinic Institutional Review Board

approved this study. A data retrieval specialist

performed an electronic search of the medical diag-nosis index to identify patients with SS diagnosed atMayo Clinic, Rochester, MN, between 1976 and 2010.A retrospective chart review was conducted toensure that the cases fulfilled SS diagnostic criteria.Medical records were reviewed for only the patientswho had not denied access to their records for

research purposes, in accor-dancewith Minnesota Statute144.335.

At primary evaluation, thepatients generally under-went a complete physical ex-amination, complete bloodcell count, general chemistrypanel, skin biopsy, and man-ual S�ezary cell count of pe-ripheral blood smears, orflow cytometry of peripheralblood. When lymph node orvisceral involvement wassuspected, patients typicallyunderwent lymph node bi-opsy or fine-needle aspira-tion and other staging (eg,

bone-marrow biopsy; additional imaging, eg, chestradiograph, computed tomography, or positronemission tomography). Given that all patients inthe study had SS, data on imaging were not collectedbecause it is not included in the 2007 ISCL/EORTCcriteria (Table I).

Because the study period spanned more than 3decades and the definition of SS evolved during thistime, the diagnostic criteria in the study reflectedthese changes over time. For the diagnostic criteriaused in this study, SS cases included those with T4([80% of total body surface area involved witherythroderma) and definitive leukemic involvement.In patients whose SS was diagnosed before the mid-1980s, leukemic involvement was defined as abso-lute S�ezary cell counts greater than or equal to1000 cells/�L, following the proposed criteria ofWinkelmann.15-17 However, beginning in the mid-1980s, clinical molecular diagnostic techniques wereimplemented and became incorporated into SS di-agnostic criteria. Therefore, in patients who receiveda diagnosis of SS after the mid-1980s (and in whichmolecular studies were available at diagnosis), leu-kemic involvement consistent with SS included anabsolute S�ezary cell count greater than or equal to1000 cells/�L or molecular evidence of clonality byT-cell receptor (TCR) gene rearrangement (usingSouthern blot or, more recently, polymerase chainreaction), or both, in the blood. In essence, SS, asdefined by the 2007 ISCL/EORTC classification

Table I. TNMB components and definitions formycosis fungoides and S�ezary syndrome based on2007 International Society for CutaneousLymphomas/European Organization for Researchand Treatment of Cancer

Classification Characteristics

T, tumorT1 Patches, papules, and/or plaques

\10% body surface areaT2 Patches and/or plaques $ 10% body

surface areaT3 $ 1 tumor ($ 1 cm in diameter)T4* Erythroderma (erythema $ 80% of

body surface area)N, nodalN0 No clinically abnormal peripheral LNsN1y Clinically abnormal peripheral LNs,

histopathology Dutch grade 2 orNCI LN0-2

N2y Clinically abnormal peripheral LNs,histopathology Dutch grade 2 orNCI LN3

N3 Clinically abnormal peripheral LNs,histopathology Dutch grade 3-4 orNCI LN4; clone positive or negative

NX Clinically abnormal peripheral LNs; nohistopathology provided

B, peripheral bloodB0y Absence of notable blood

involvement: # 5% of peripheralblood lymphocytes are S�ezary cells

B1y Low blood tumor burden:[5% ofperipheral blood lymphocytes areS�ezary cells

B2* High blood tumor burden. Requirespositive clonal rearrangement ofTCR plus one of the following:absolute S�ezary cell count$ 1000/�L; expanded CD41 orCD31 cells with CD4/CD8 ratio$ 10; expanded CD41 cells withabnormal immunophenotype,including loss of CD7 ($ 40%) orCD26 ($ 30%)

M, metastases/visceralorgans

M0 No visceral organ involvementM1 Visceral organ involvement with

pathologic confirmation

From Olsen et al.6 Used with permission.

LN, Lymph node; NCI, National Cancer Institute; TCR, T-cell

receptor.

*Necessary criteria for S�ezary syndrome.yCan be divided into a and b if information on clone with

polymerase chain reaction or Southern blot analysis of TCR is

available. a signifies clone negative; b signifies clone positive.

J AM ACAD DERMATOL

VOLUME 67, NUMBER 6Kubica et al 1191

guidelines, was T4N0-3M0-1B2 (Table I) in the cur-rent study.6 Time of SS diagnosis was defined as thetime of initial evaluation at Mayo Clinic or an outsideinstitution, following the above criteria.

The information extracted from the medicalrecords included date of birth; sex; race; symptomsat presentation; prior diagnosis of MF; lactatedehydrogenase (LDH) at presentation; S�ezary cellcounts; presence of TCR clones in blood, skin,lymph node, or bone marrow, or a combination;date of last follow-up at our institution; and date ofdeath. For patients who no longer correspondedwith or were no longer seen at our institution, vitalstatus and death date information were queriedthrough an institutionally approved fee-basedInternet research and location service (Accurint,LexisNexis Risk Solutions, Burlington, Vt).Therefore, in reference to the end point that isrelevant to this study (overall survival), no patientswere lost to follow-up.

Standard descriptive statistics were used to sum-marize these data. The primary end point was overallsurvival (death by any cause) because we wereunable to evaluate disease-specific survival giventhe lack of data on disease course and cause of deathfor many patients. Duration of follow-up was calcu-lated from the date of SS diagnosis to the date ofdeath by any cause or of last follow-up. Overallsurvival was estimated with the Kaplan-Meiermethod. The expected survival was derived byapplying age-, sex-, and calendar-yearespecificmor-tality from the US white population to the age- andsex-specific person-years of follow-up in our cohort,using the Hakulinen cohort method. A 1-sample logrank test was used to compare the observed andexpected curves.18 Demographic and clinical factorspresent at diagnosis were each evaluated univari-ately for an association with survival (death by anycause) by fitting separate Cox proportional hazardsregression models. In addition, each factor wasevaluated in a Cox model adjusted for age. Thestrength of each association was summarized usingthe hazard ratio (HR) and corresponding 95% con-fidence interval (CI). All calculated P values were 2sided, and P less than .05 was considered statisticallysignificant. Statistical analyses were performed witha software package (SAS, Version 9.2, SAS InstituteInc, Cary, NC).

RESULTSA total of 487 patients with possible SS were

identified on a preliminary search over 35 years ofpatient records. Among these patients, 311 wereexcluded because another final diagnosis wasidentified (eg, erythrodermic MF, generalized

Table II. Demographic and clinical features anddiagnostic criteria for 176 patients with S�ezarysyndrome

Characteristic Value,* N = 176

Demographic and clinical featuresSex, %Male 110 (62.5)Female 66 (37.5)

Age, yMean (SD) 66.2 (12.7)Range 26.2-89.6

RaceCaucasian 168 (95.4)African American 4 (2.3)Middle Eastern 1 (0.6)Unknown 3 (1.7)

Duration of symptoms beforepresentation, y

Mean (SD) 4.4 (4.4)Median (IQR) 3.0 (1.3-6.0)Range 0.2-32.0

Erythroderma at presentation 176 (100)Duration of erythroderma before

presentation, yData available, No. of patients 81Mean (SD) 1.7 (1.5)Median (IQR) 1.7 (0.6-2.0)Range 0.2-7.0

Other symptoms at presentationy

Adenopathy 100 (56.8)Plaques 39 (22.2)Tumors 4 (2.3)Pruritus 176 (100)Alopecia 29 (16.5)Keratoderma 30 (17.1)Lichenification 13 (7.4)Patches 1 (0.6)

Prior MF diagnosis 12 (6.8)LDH level at presentationData available, No. of patients 81Median, U/L (IQR) 197 (130-275)# 222 U/L 48 (59.3)

Diagnostic criteriaMolecular data available 131 (74.4)TCR clone present in bonemarrow, lymph node, blood,or skin

119 (90.8)

Skin 69/89 (77.5)Lymph node 21/27 (77.8)Blood 104/120 (86.7)Bone marrow 15/28 (53.6)

Lymph node biopsy performed 71 (40.3)

Continued

Table II. Cont’d

Characteristic Value,* N = 176

Lymph node biopsy resultsArchitecture consistent with SSz 37 (52.1)Molecular data consistentwith SS

3 (4.2)

Architecture and molecular dataconsistent with SSz

18 (25.4)

No lymph node involvement 13 (18.3)Bone-marrow biopsy performed 114 (64.8)Bone-marrow biopsy resultsArchitecture consistent with SS 20 (17.5)Molecular data consistentwith SS

6 (5.3)

Architecture and moleculardata consistent with SS

9 (7.9)

Nondiagnostic 1 (0.9)No bone-marrow involvement 78 (68.4)

S�ezary cell count, median, cells/�L(IQR)

1778 (1215-3481)

S�ezary cells, median, % (IQR) 19.0 (11.5-30.0)Results of skin biopsyConsistent with CTCL/SSx 163 (92.6)Nondiagnostic 13 (7.4)

CTCL, Cutaneous T-cell lymphoma; IQR, interquartile range, 25th

and 75th percentiles; LDH, lactate dehydrogenase; MF, mycosis

fungoides; SS, S�ezary syndrome; TCR, T-cell receptor.

*Values are presented as No. (%) unless specified otherwise.yTotal percentage exceeds 100% because some patients had

[1 symptom at presentation.zArchitectural involvement included lymph node effacement

caused by tumor and accumulation of dense S�ezary cell infiltrate

in lymphoid tissue.xMost frequently noted to be atypical epidermal lymphocytic

band.

J AM ACAD DERMATOL

DECEMBER 20121192 Kubica et al

contact dermatitis, reactive erythroderma, graft-versus-host disease, atopic dermatitis), resulting in176 patients who had SS as defined by theestablished criteria. Among these patients, 23 had

S�ezary cell counts that were not recorded (n = 7) orwere recorded to be less than 1000 cells/�L at thetime of diagnosis (n = 16); in these cases, thepatients fulfilled the inclusion criteria for diagnosisof SS on the basis of the other molecular criterialisted in Table I. Table II summarizes the patientdemographic characteristics and clinical features atpresentation for the 176 patients with SS in thisstudy. The mean (range) age at SS diagnosis was66.2 (26.2-89.6) years. Patients presented with var-ious symptoms, with a median duration of 3.0 yearsof dermatologic symptoms before diagnosis (inter-quartile range [IQR] 1.3-6.0 years). Because eryth-roderma is an essential component of thediagnosis, all patients had this finding at presenta-tion. In particular, the average duration of erythro-derma before diagnosis was 1.7 years (IQR 0.6-2.0years). Other symptoms are also listed in Table II.To verify the diagnosis of SS and exclude otherhematologic or dermatologic processes, many pa-tients underwent biopsies of lymph node and bone

Table III. Overall survival after diagnosis of S�ezarysyndrome

Time since

diagnosis, y

No. of patients

at risk, N = 176 Survival, % (95% CI)

1 147 86.1 (81.1-91.4)2 126 75.5 (69.3-82.2)3 104 63.3 (56.4-71.0)4 79 48.6 (41.6-56.8)5 65 42.3 (35.5-50.5)6 43 28.0 (21.8-35.9)7 37 24.1 (18.2-31.7)8 28 19.4 (14.1-26.7)9 23 17.3 (12.2-24.4)10 18 13.5 (9.0-20.3)

CI, Confidence interval.

Fig 1. Overall survival from diagnosis of S�ezary syndrometo 10 years after diagnosis.

J AM ACAD DERMATOL

VOLUME 67, NUMBER 6Kubica et al 1193

marrow, and molecular studies of the lymph nodes,bone marrow, blood, and skin.

Among the 176 patients at follow-up, 150 wereknown to be deceased, with a median time to deathafter SS diagnosis of 3.5 years (IQR 1.7-5.7 years).Among the other 26 patients, the median duration offollow-up was 4.4 years (IQR 1.6-10.1 years). Mediansurvival after SS diagnosis was 4.0 years. As summa-rized in Table III and Fig 1, overall survival was 86.1%and 42.3% at 1 and 5 years, respectively, after diag-nosis. The observed survival was significantly worsethan the expected survival derived from the USCaucasian population for the same sex, age, andcalendar year (P\.001). Of note, themedian survivalwas 5.5 years after onset of erythroderma in the 81patients for whom this clinical finding was recorded.

Cox proportional hazards models were fit tounivariately evaluate potential predictors of survivalin patients with SS. Advanced age at SS diagnosis,diagnosis of MF before SS, presence of skin clone,presence of blood clone, and higher LDH levels atpresentation were each identified as significantlyassociated with poorer survival (Table IV). In addi-tion, the association between each factor and sur-vival was assessed after adjusting for age (Table V).

The following factors were individually associatedwith survival after adjustment for age: LDH level atpresentation, prior diagnosis of MF, and presence ofTCR gene rearrangements in skin and blood. Amultivariable model was not fit because the datafor each of these factors were not available for allpatients.

DISCUSSIONIn this study, we present the survival data over the

past 35 years on, to our knowledge, the largest groupof patients with SS studied to date. Currently, someheterogeneity exists in the literature regarding sur-vival and prognosis for patients with SS because ofthe discrepancies in diagnostic criteria and theinclusion of patients without SS in many of theanalyses.4 Nevertheless, our survival data on 176patients reinforce the trends found in the literatureand reaffirm several prognostic markers of worsesurvival outcomes in patients with SS.

Over the past several decades, extensive researchadvances in SS and CTCL have generated improveddiagnostic techniques in an effort to more effectivelydiagnose SS and distinguish it from other subsets ofCTCL.1,4 Nevertheless, only limited data exist regard-ing the disease course itself. Longitudinal data arealso limited regarding incorporation of the changesthat occurred over the past decades, such as inclu-sion of new molecular tests and diagnostic criteria,and use of extracorporeal photopheresis, newerretinoids, immunomodulatory agents, monoclonalantibodies, and more targeted agents as treatment.4

Additional confounding factors in establishing clearepidemiologic data on survival include the frequentexclusion of patients with SS from clinical trials onCTCL treatment and the myriad of definitions used asend points in various studies.4 Recently, Olsenet al,19 in collaboration with the ISCL, published aconsensus recommendation for clinical trials in pa-tients with CTCL, to standardize assessments ofsymptoms; skin, lymph node, blood, and visceralorgan involvement; and end points for response. Theimplementation of such uniform criteria will greatlyassist in providing international collaboration forresearch in CTCL, particularly in the subset ofpatients with SS, where a larger, multicenter cohortof patients should allow for more standardized,validated, and accurate future studies.

The group of patients in this study reflected thereported male predilection, with 110 men (62.5%)and 66 women (37.5%); generally, men are affectedtwice as often as women.1,8,13,14,20 The mean age atSS diagnosise66.2 yearsein our study was similar toother reported data on SS,21-23 with a significantincrease in incidence associated with increasing age.

Table IV. Summary of factors evaluated univariately for association with survival after S�ezary syndromediagnosis

Factor

No. of

patients

Survival, % (95% CI)Median

survival, y HR (95% CI)

P

valueAt 1 y At 3 y At 5 y

Age at SS diagnosis, y* 176 1.39 (1.20-1.60) \.001SexMale 110 85.1 (78.7-92.1) 61.1 (52.5-71.2) 39.4 (31.0-50.1) 3.6 1.21 (0.87-1.68) .27Female 66 87.7 (80.1-96.1) 66.8 (56.1-79.5) 47.1 (36.1-61.3) 4.6 Referent

Symptom durationbefore presentation, y

\1.3 43 85.7 (75.8-97.0) 61.9 (48.8-78.5) 49.7 (36.6-67.5) 4.0 Referent .251.3-2.9 36 83.3 (72.0-96.4) 54.9 (40.7-74.0) y 3.3 1.51 (0.94-2.44)3.0-5.9 43 90.2 (81.6-99.8) 74.1 (61.4-89.4) 38.2 (25.3-57.7) 3.9 1.20 (0.75-1.92)$ 6.0 52 86.5 (77.7-96.3) 64.6 (52.7-79.2) 50.3 (38.2-66.3) 5.0 0.99 (0.64-1.52)

Adenopathy atpresentation

Yes 100 85.8 (79.2-93.0) 65.1 (56.2-75.3) 43.6 (34.7-54.9) 4.0 1.02 (0.74-1.41) .92No 76 86.5 (79.0-94.6) 60.9 (50.5-73.4) 40.4 (30.4-53.8) 3.9 Referent

Alopecia atpresentation

Yes 29 79.3 (65.9-95.5) 55.2 (39.7-76.6) y 3.9 1.06 (0.69-1.61) .80No 147 87.5 (82.2-93.1) 65.0 (57.5-73.4) 44.7 (37.0-53.9) 4.1 Referent

Plaques atpresentation

Yes 39 79.5 (67.8-93.2) 61.1 (47.4-78.6) 44.9 (31.5-63.9) 4.0 0.85 (0.58-1.26) .43No 127 88.0 (82.7-93.7) 63.9 (56.1-72.7) 41.5 (33.8-51.1) 4.0 Referent

Systemic symptomsat presentation

Yes 10 y e e 7.4 0.62 (0.28-1.29) .19No 166 86.5 (81.4-91.9) 63.6 (56.5-71.5) 41.4 (34.3-49.9) 4.0 Referent

MF before SSYes 12 y e e 1.5 2.36 (1.27-4.39) .007No 164 87.5 (82.6-92.8) 65.5 (58.5-73.4) 43.6 (36.4-52.3) 4.0 Referent

Lymph node biopsyPositive 58 89.5 (81.9-97.8) 62.2 (50.6-76.4) 47.1 (35.5-62.5) 4.3 1.76 (0.88-3.50) .11Negative 13 100 (NA) 91.7 (77.3-100) y 5.8 Referent

Bone-marrow biopsyPositive 35 85.7 (74.9-98.1) 60.0 (45.8-78.6) 34.3 (21.7-54.2) 3.3 1.39 (0.90-2.14) .14Negative 79 84.3 (76.6-92.9) 62.5 (52.4-74.6) 41.7 (31.8-54.7) 4.1 Referent

Skin clonePresent 69 86.8 (79.0-95.2) 62.9 (52.4-75.6) 35.6 (25.8-49.2) 3.7 2.62 (1.41-4.85) .002Absent 20 90.0 (64.3-99.6) 80.0 (64.3-99.6) 69.6 (52.0-93.2) 6.4 Referent

Blood clonePresent 104 86.4 (80.0-93.3) 64.3 (55.5-74.4) 43.7 (34.9-54.7) 4.0 2.00 (1.00-4.02) .05Absent 16 86.7 (71.1-100) y e 5.2 Referent

Lymph node clonePresent 21 90.5 (78.8-100) 70.7 (53.5-93.6) y 4.3 1.58 (0.53-4.72) .41Absent 6 y e e 7.7 Referent

Log2 (S�ezary count)* 169 1.01 (0.88-1.16) .90Log2 (% S�ezary cells)* 170 0.93 (0.79-1.11) .43Log2 (LDH level atpresentation)*

81 1.80 (1.26-2.59) .001

CI, Confidence interval; HR, hazard ratio; LDH, lactate dehydrogenase; MF, mycosis fungoides; NA, not applicable; SS, S�ezary syndrome.

*HR is per 10-y increase in age and per doubling in values of factors evaluated after applying log2 transformation.yKaplan-Meier estimates are not reported when there are\10 patients left at risk.

J AM ACAD DERMATOL

DECEMBER 20121194 Kubica et al

Table V. Factors associated with increased risk ofdeath after S�ezary syndrome diagnosis and afteradjustment for age at diagnosis

Factor

No. of patients

with data

available

Adjusted

HR (95% CI)*

P

value

Prior mycosisfungoides diagnosis(vs none)

176 2.68 (1.44-4.98) .002

Positive lymph nodebiopsy (vs negative)

71 1.85 (0.93-3.68) .08

Positive bone-marrowbiopsy (vs negative)

114 1.31 (0.85-2.02) .22

Present skin clone(vs absent)

89 2.59 (1.38-4.87) .003

Present blood clone(vs absent)

120 2.05 (1.00-4.21) .05

Log2 (LDH level atpresentation)y

81 1.71 (1.18-2.47) .005

CI, Confidence interval; HR, hazard ratio; LDH, lactate

dehydrogenase.

*Each factor was assessed in separate Cox proportional hazards

regression model that also included age.yHR is per doubling in LDH level.

J AM ACAD DERMATOL

VOLUME 67, NUMBER 6Kubica et al 1195

Accordingly, our study also confirmed prior findingsthat worse prognosis is associated with increasingage (HR 1.39; 95% CI 1.20-1.60; P\.001).10,12,24 It isimportant to note that many data on age, sex, andrace, as with other factors and epidemiologic char-acteristics in patients with SS, are not clearly differ-entiated from CTCL studies as a whole.

Although data on lymph node involvement werenot uniformly available for all patients, 71 patientsdid undergo lymph node biopsy, of which 58(81.7%) showed involvement with SS. Similarly,Kim et al12 evaluated nodal involvement in patientswith blood involvement and found that 26 of 28patients in stage T4B1 (92.9%) had lymph nodeinvolvement. We did not find a statistically significantcorrelation with worse prognosis in patients withlymph node involvement. However, the study ofpatients with MF and SS by Agar et al8 did notedecreases in median survival, overall survival, anddisease-specific survival of patients with advancinglymph node stage (P\ .001). Similar findings wereobserved in other studies.9,25 In addition, somestudies more specifically suggest worse prognosisin cases where lymph nodes had TCR clonepresent.26,27

The presence of a peripheral blood TCR clone,especially when correlated to the skin, has beenshown to be associated with a considerably worseprognosis compared with patients who do not haveclonal involvement, and this finding holds true in

patients with MF and those with SS.8,9,28 Our studyconfirms these findings, with an HR of 2.05 (95% CI1.00-4.21) after adjustment for age. Furthermore, theproportion of S�ezary cells present in the blood,acting as a marker of the degree of peripheral bloodinvolvement, has been correlated with worse prog-nosis. Specifically, an increased disease-specificdeath rate was demonstrated by Scarisbrick et al9

with cutoffs for hematologic stage of 1000 cells/�Land 10,000 cells/�L. Yet, our study did not find thiscorrelation to be statistically significant.

A TCR clone found in the skin also portendedworse prognosis in our patients, with an HR of 2.59(95% CI 1.38-4.87) after adjustment for age.However, this test is nonspecific because only 69 ofthe 89 patients with molecular data present in theskin (77.5%) revealed a clone. Studies have shownthat the presence of identical cutaneous and bloodTCR clones is an independent prognostic factor fordisease progression of SS, indicating that skin biop-sies can indeed be helpful when confirming a clonepresent in the blood.22,29 In contrast to MF, epider-motropism is frequently absent in cutaneous biopsyspecimens of patients with SS.30 Currently, the pres-ence of a cutaneous TCR clone is not a part of themandatory diagnostic criteria of SS, although it hasbeen proposed as part of the laboratory diagnosis forearly MF.31

Of the 131 patients with molecular collectionsperformed (74.4%), a total of 119 (90.8%) had evi-dence of clonal involvement by a T-cell malignancy.More specifically, 120 patients had molecular analy-sis of blood specimens, and 104 (86.7%) of theseanalyses revealed clones. These data reaffirm theaccuracy and utility of molecular tests, which havemarkedly advanced the sensitivity to diagnose ma-lignant lymphoproliferative disorders. Nevertheless,a small percentage of patients with SS do not show aTCR clone by Southern blot or polymerase chainreaction analysis. In a study comparing variousmethods for assessing peripheral involvement bySS, 10 of 11 patients (90.9%) had a TCR clone.32

Another study used similar methods and found a TCRclone in 12 of 17 patients (71%).33 These findingsfurther support clinical laboratory recommendationsthat no single test uniformly establishes the diagnosisof SS.33 As with all diagnostic techniques, false-negative and false-positive results must be consid-ered; therefore, incorporation of all available clinical,pathologic, and immunohistochemical findings re-mains crucial when diagnosing SS.

The primary purpose of our study was to estimatethe overall survival of patients with SS. Availableinformation indicates that the natural disease courseof SS is more aggressive than MF, and the prognosis

Table VI. Major survival studies on cutaneous T-cell lymphoma with identified subsets of patients who have S�e ary syndrome

Study Journal Location

Total No.

of patients

Patients

included

No. of patients

with SS

Survival,

median (overa )

Patients alive

at 1 y, %

Patients alive

at 5 y, %

Patients alive

at 10 y, %

Kubica et al (currentstudy), 2012

Minnesota 176 SS 176 4.0 y (48 mo) 86.1 OS 42.3 OS NA

Agar et al,8 2010 J Clin Oncol United Kingdom 1502 MF and SS 104 3.13 y (37.56 mo NA 26 OS;31 DSS

12 OS;15 DSS

Vidulich et al,36 2009* Int J Dermatol Texas 124 ErythrodermicCTCL

101: 79 H3,22 H4

5.4 y for H3;2.4 y for H4

NA NA NA

Arulogun et al,10 2008 Blood Australia 297 MF and SS 31 11 y (132 mo) NA 60.8 OS 54.7 OSWillemze et al,5 2005y Blood The Netherlands,

Austria1905 CTCL 52 NA NA 24 DSS NA

Foulc et al,22 2003y Br J Dermatol France 28 SS 28 5.4 y (64.55 mo DSS NA NA NAKim et al,12 2003 Arch Dermatol California 525 MF and SS 35 3.0 y (36 mo) NA 20 NAMarti et al,20 2003 Leuk Lymphoma Spain 29 SS 29 4.0 y (48 mo) NA Mean

actuarialrisk, 38at 5 y

NA

Diamandidouet al,11 1999

J Am AcadDermatol

Texas 115 MF and SS 12 2.5 y (30 mo) NA 40 NA

Bernengo et al,21 1998 Ann Oncol Italy 62 SS 62 2.6 y (31 mo) NA 33.50 NASausville et al,25 1988 Ann Intern Med Maryland 152 MF and SS 52 3.3 y (40 mo) NA NA NASchechter et al,34 1987 Blood Maryland 160 CTCL 60 3.5 y (42 mo) NA 22 NA

CTCL, Cutaneous T-cell lymphoma; DSS, disease-specific survival; MF, mycosis fungoides; NA, not available; OS, overall survival; SS, S�ezary sy drome.

*Study divided patients with SS into 2 groups: H3 ($ 1000-# 10,000 S�ezary cells/�L) and H4 ($ 10,000 S�ezary cells/�L).yReported only DSS. J

AM

ACADD

ERM

ATOL

DECEM

BER20

121196

Kubica

etal

z

ll

)

)

n

J AM ACAD DERMATOL

VOLUME 67, NUMBER 6Kubica et al 1197

for these patients is typically poor.12 The medianoverall survival ranges from 2.5 to 4.0years.8,11,12,20,21,25,34 Table VI represents major sur-vival studies on CTCL that have identified subsets ofpatients with SS. Although only a select group ofthese studies were exclusively of SS, we believed thatincluding all major publications that clearly definedthe pool of patients with SS would be important,given the rarity of this diagnosis and the paucity ofdata on patients with SS only. Of note, in the 2008study by Arulogun et al,10 themedian overall survivalwas surprisingly high, at 11 years; it is not under-stoodwhy this finding was so high. Also, Foulc et al22

studied a group of patients with SS and found amedian survival of 5.4 years, but this finding wasdisease specific.

The largest, most recent study from the UnitedKingdom cutaneous lymphoma database evaluated104 patients with SS and revealed that overall sur-vival at 5 years is 26%, with a median survival of 3.13years after diagnosis.8 Similarly, our study of 176patients found a median survival of 4 years, whichparallels the findings of many prior studies and thusprovides further affirmation of the general unifor-mity of staging and treatment response that can bemore broadly applied for patients with SS cared for atvarious institutions over several decades. Of note,the median survival after onset of erythroderma was5.5 years. Disease-specific mortality could not becalculated because of the inability to gather data ontiming of disease progression and cause of death.

Our study also addressed factors at the time ofdiagnosis that correlated to increased risk of death(Table V). After adjusting for age, LDH level atpresentation was a potential predictor of poorersurvival after a diagnosis of SS. Other studies havesimilarly noted this finding.2,10,24 Interestingly, 59.3%of patients in whom LDH levels were available hadLDH levels within the reference range at diagnosis.Another prognostic factor was the presence of TCRgene rearrangements in skin and blood. In theliterature, the prognostic factors associated withworse outcomes include advanced age, enlargementof peripheral lymph nodes, increased leukemic bur-den in the blood, increased LDH levels, low per-centage of CD81 cells in lymph nodes, and large-celltransformation.2,3,8,10,24,35 Because of a lack of dataon the disease course of many patients, we were notable to assess for certain prognostic factors, such aslarge-cell transformation. Frequently, this develop-ment occurs later in the disease course rather thanbeing present at time of diagnosis.

The number of circulating S�ezary cells did notappear to be a prognostic factor in our study. Thisresult has been noted in other publications as

well.21,22 Yet, elevated levels of S�ezary cells, andnotably those exceeding 10,000 cells/�L, in studieshave identified a subset of patients with a worseprognosis.9,36 These findings could call into questionthe role of absolute S�ezary cell counting, particularlyof the 1000 S�ezary cell cutoff, in prognosis of SS.

Of note, an indicator of worse prognosis inpatients with SS in our investigation was the exis-tence or diagnosis of prior MF (HR 2.68; 95% CI 1.44-4.98; P = .002). In our study, 12 patients had prior MF,although the types of MF were not further charac-terized according to stage or subtype. This findingwas also reported in 11.3% of patients with SS in anItalian SS study,21 although overall, this phenome-non is not well documented in the literature.7

Typically, less than 10% of patients with MF progressto more advanced disease.37 Nevertheless, in pa-tients with MF who have development of cutaneoustumors or generalized erythroderma, survival de-creases from a median of 11 years to a median of 3years and 4.5 years, respectively.37 These values aremore consistent with the prognosis of patients withSS. Although CTCL represents a spectrum of disease,with various stages based on TNMB classifications,there is little crossover between MF and SS inprevious reports. It is possible that patients with SSand prior MF have an atypical, more aggressivevariant of CTCL, perhaps with large-cell, CD301

expression or inactivation of the cyclin-dependentkinase inhibitor 2A-2B locus that is associated withinferior outcomes. Alternatively, these survival dataof patients with prior MF could represent an artifactof statistical bias. Further studies are needed toelucidate the clinical course of these patients.37,38

LimitationsLimitations of this study include its retrospective

design. In addition, the patient population repre-sents a single academic center. Therefore, resultsmay not be generalizable to other institutions orpopulations. Because this study spanned almost 4decades, it is important to note that the patientpopulation, diagnostic techniques, and treatmentsevolved over this time. Some results were limitedbecause of missing data and inconsistencies in theextent of laboratory evaluation, imaging, and stag-ing, along with shortcomings inherent in retrospec-tive chart reviews.

CONCLUSIONBecause of the aggressive nature of SS compared

with other subsets of CTCL, such as erythrodermicMF or early-stage MF, it is imperative to stratify andanalyze this distinctive patient population sepa-rately. However, because of the relative rarity of

J AM ACAD DERMATOL

DECEMBER 20121198 Kubica et al

this disease, the continual evolution of diagnosticcriteria to document skin and blood involvement bySS, and the occasional overlap in the spectrum ofCTCL, it is difficult not only to identify, but alsorobustly analyze, the group of patients with SS. Withthe long-standing experience and large volume ofpatients with SS seen at our institution, we were ableto assemble a well-defined SS cohort on the basis ofcharacteristic clinical features and blood involve-ment by CTCL. We found that this group of patientshas a poor prognosis, with a median survival of 4.0years after SS diagnosis. Further studies need toidentify improved diagnostic techniques and molec-ular prognostic markers, and assess the impact ofnovel therapeutics to improve the outcomes ofpatients with SS.

REFERENCES

1. Hwang ST, Janik JE, Jaffe ES, Wilson WH. Mycosis fungoides

and S�ezary syndrome. Lancet 2008;371:945-57.

2. Kim EJ, Lin J, Junkins-Hopkins JM, Vittorio CC, Rook AH.

Mycosis fungoides and S�ezary syndrome: an update. Curr

Oncol Rep 2006;8:376-86.

3. Lansigan F, Choi J, Foss FM. Cutaneous T-cell lymphoma.

Hematol Oncol Clin North Am 2008;22:979-96.

4. Olsen EA, Rook AH, Zic J, Kim Y, Porcu P, Querfeld C, et al.

S�ezary syndrome: immunopathogenesis, literature review of

therapeutic options, and recommendations for therapy by the

United States Cutaneous Lymphoma Consortium (USCLC).

J Am Acad Dermatol 2011;64:352-404.

5. Willemze R, Jaffe ES, Burg G, Cerroni L, Berti E, Swerdlow SH,

et al. WHO-EORTC classification for cutaneous lymphomas.

Blood 2005;105:3768-85.

6. Olsen E, Vonderheid E, Pimpinelli N, Willemze R, Kim Y,

Knobler R, et al. Revisions to the staging and classification of

mycosis fungoides and S�ezary syndrome: a proposal of the

International Society for Cutaneous Lymphomas (ISCL) and

the cutaneous lymphoma task force of the European Organi-

zation of Research and Treatment of Cancer (EORTC). Blood

2007;110:1713-22.

7. Vonderheid EC, Bernengo MG, Burg G, Duvic M, Heald P,

Laroche L, et al. Update on erythrodermic cutaneous T-cell

lymphoma: report of the International Society for Cutaneous

Lymphomas. J Am Acad Dermatol 2002;46:95-106.

8. Agar NS, Wedgeworth E, Crichton S, Mitchell TJ, Cox M,

Ferreira S, et al. Survival outcomes and prognostic factors in

mycosis fungoides/S�ezary syndrome: validation of the revised

International Society for Cutaneous Lymphomas/European

Organization for Research and Treatment of Cancer staging

proposal. J Clin Oncol 2010;28:4730-9.

9. Scarisbrick JJ, Whittaker S, Evans AV, Fraser-Andrews EA, Child

FJ, Dean A, et al. Prognostic significance of tumor burden in

the blood of patients with erythrodermic primary cutaneous

T-cell lymphoma. Blood 2001;97:624-30.

10. Arulogun SO, Prince HM, Ng J, Lade S, Ryan GF, Blewitt O, et al.

Long-term outcomes of patients with advanced-stage cuta-

neous T-cell lymphoma and large cell transformation. Blood

2008;112:3082-7.

11. Diamandidou E, Colome M, Fayad L, Duvic M, Kurzrock R.

Prognostic factor analysis in mycosis fungoides/S�ezary syn-

drome. J Am Acad Dermatol 1999;40:914-24.

12. Kim YH, Liu HL, Mraz-Gernhard S, Varghese A, Hoppe RT.

Long-term outcome of 525 patients with mycosis fungoides

and S�ezary syndrome: clinical prognostic factors and risk for

disease progression. Arch Dermatol 2003;139:857-66.

13. Zackheim HS, Amin S, Kashani-Sabet M, McMillan A. Prognosis

in cutaneous T-cell lymphoma by skin stage: long-term

survival in 489 patients. J Am Acad Dermatol 1999;40:418-25.

14. Criscione VD, Weinstock MA. Incidence of cutaneous T-cell

lymphoma in the United States, 1973-2002. Arch Dermatol

2007;143:854-9.

15. Winkelmann RK. Clinical studies of T-cell erythroderma in the

S�ezary syndrome. Mayo Clin Proc 1974;49:519-25.

16. Duncan SC, Winkelmann RK. Circulating S�ezary cells in hospi-

talized dermatology patients. Br J Dermatol 1978;99:171-8.

17. Buechner SA, Winkelmann RK. S�ezary syndrome: a clinico-

pathologic study of 39 cases. Arch Dermatol 1983;119:979-86.

18. Therneau T, Offord J. Expected survival based on hazard rates

(update). Rochester (MN): Mayo Clinic, Division of Biostatistics

and Informatics; 1999 Feb. 26 p. Technical Report No. 63.

19. Olsen EA, Whittaker S, Kim YH, Duvic M, Prince HM, Lessin SR,

et al. Clinical end points and response criteria in mycosis

fungoides and S�ezary syndrome: a consensus statement of the

International Society for Cutaneous Lymphomas, the United

States Cutaneous Lymphoma Consortium, and the Cutaneous

Lymphoma Task Force of the European Organization for Re-

search and Treatment of Cancer. J Clin Oncol 2011;29:2598-607.

20. Marti RM, Pujol RM, Servitje O, Palou J, Romagosa V, Bordes R,

et al. S�ezary syndrome and related variants of classic cutane-

ous T-cell lymphoma: a descriptive and prognostic clinico-

pathologic study of 29 cases. Leuk Lymphoma 2003;44:59-69.

21. Bernengo MG, Quaglino P, Novelli M, Cappello N, Doveil GC,

Lisa F, et al. Prognostic factors in S�ezary syndrome: a multi-

variate analysis of clinical, hematological and immunological

features. Ann Oncol 1998;9:857-63.

22. Foulc P, N’Guyen JM, Dreno B. Prognostic factors in S�ezary

syndrome: a study of 28 patients. Br J Dermatol 2003;149:

1152-8.

23. Scarisbrick JJ, Child FJ, Evans AV, Fraser-Andrews EA, Spittle M,

Russell-Jones R. Secondary malignant neoplasms in 71 pa-

tients with S�ezary syndrome. Arch Dermatol 1999;135:1381-5.

24. Vonderheid EC, Pena J, Nowell P. S�ezary cell counts in

erythrodermic cutaneous T-cell lymphoma: implications for

prognosis and staging. Leuk Lymphoma 2006;47:1841-56.

25. Sausville EA, Eddy JL, Makuch RW, Fischmann AB, Schechter

GP, Matthews M, et al. Histopathologic staging at initial

diagnosis of mycosis fungoides and the S�ezary syndrome:

definition of three distinctive prognostic groups. Ann Intern

Med 1988;109:372-82.

26. Assaf C, Hummel M, Steinhoff M, Geilen CC, Orawa H, Stein H,

et al. Early TCR-beta and TCR-gamma PCR detection of T-cell

clonality indicates minimal tumor disease in lymph nodes of

cutaneous T-cell lymphoma: diagnostic and prognostic impli-

cations. Blood 2005;105:503-10.

27. Fraser-Andrews EA, Mitchell T, Ferreira S, Seed PT, Russell-Jones

R, Calonje E, et al. Molecular staging of lymph nodes from

60 patients with mycosis fungoides and S�ezary syndrome:

correlation with histopathology and outcome suggests prog-

nostic relevance in mycosis fungoides. Br J Dermatol 2006;155:

756-62.

28. Fraser-Andrews EA, Woolford AJ, Russell-Jones R, Seed PT,

Whittaker SJ. Detection of a peripheral blood T cell clone is an

independent prognostic marker in mycosis fungoides. J Invest

Dermatol 2000;114:117-21.

29. Beylot-Barry M, Sibaud V, Thiebaut R, Vergier B, Beylot C,

Delaunay M, et al. Evidence that an identical T cell clone in

J AM ACAD DERMATOL

VOLUME 67, NUMBER 6Kubica et al 1199

skin and peripheral blood lymphocytes is an independent

prognostic factor in primary cutaneous T cell lymphomas.

J Invest Dermatol 2001;117:920-6.

30. Diwan AH, Prieto VG, Herling M, Duvic M, Jone D. Primary

S�ezary syndrome commonly shows low-grade cytologic atypia

and an absence of epidermotropism. Am J Clin Pathol 2005;

123:510-5.

31. Pimpinelli N, Olsen EA, Santucci M, Vonderheid E, Haeffner AC,

Stevens S, et al. International Society for Cutaneous Lym-

phoma. Defining early mycosis fungoides. J Am Acad Derma-

tol 2005;53:1053-63.

32. Morice WG, Katzmann JA, Pittelkow MR, el-Azhary RA, Gibson

LE, Hanson CA. A comparison of morphologic features, flow

cytometry, TCR-Vbeta analysis, and TCR-PCR in qualitative and

quantitative assessment of peripheral blood involvement by

S�ezary syndrome. Am J Clin Pathol 2006;125:364-74.

33. Klemke CD, Brade J, Weckesser S, Sachse MM, Booken N,

Neumaier M, et al. The diagnosis of S�ezary syndrome on

peripheral blood by flow cytometry requires the use of

multiple markers. Br J Dermatol 2008;159:871-80.

34. Schechter GP, Sausville EA, Fischmann AB, Soehnlen F, Eddy J,

Matthews M, et al. Evaluation of circulating malignant cells

provides prognostic information in cutaneous T cell lym-

phoma. Blood 1987;69:841-9.

35. Prince HM, Whittaker S, Hoppe RT. How I treat mycosis

fungoides and S�ezary syndrome. Blood 2009;114:4337-53.

36. Vidulich KA, Talpur R, Bassett RL, Duvic M. Overall survival in

erythrodermic cutaneous T-cell lymphoma: an analysis of

prognostic factors in a cohort of patients with erythrodermic

cutaneous T-cell lymphoma. Int J Dermatol 2009;48:243-52.

37. Zinzani PL, Ferreri AJ, Cerroni L. Mycosis fungoides. Crit Rev

Oncol Hematol 2008;65:172-82.

38. Laharanne E, Chevret E, Idrissi Y, Gentil C, Longy M, Ferrer J,

et al. CDKN2A-CDKN2B deletion defines an aggressive sub-

set of cutaneous T-cell lymphoma. Mod Pathol 2010;23:

547-58.