Cytogenetic Analysis of 434 Consecutively Ascertained ...

From www.bloodjournal.org by guest on February 18, 2016. For personal use only.

Cytogenetic Analysis of 434 Consecutively Ascertained Specimens of
Non-Hodgkin’s Lymphoma: Clinical Correlations

By Kenneth Offit, George Wong, Daniel A. Filippa, Yue Tao, and R.S.K. Chaganti

Cytogenetic and histopathologicdata were correlated with althoughmedianfollow-up was only 2 years. Seven patients
clinical parameters from 423 patients with non-Hodgkin’s with t(8;14) LG NHL had an indolent course. Among 104
lymphoma (NHL). Clinical correlations were performed on patients with DLLC and abnormal karyotypes at diagnosis,
subgroups of 149 patients with low-grade lymphoma (LG) breaks at 1q21-23 or more than 4 marker chromosomes was
and 205 patients with diffuse lymphoma with a large cell associated with a shortened median survival. Using these
component (DLLC).Correlationswere made between clinical variableswe constructed a proportional hazards model with
outcome and individual recurring cytogenetic aberrations, a good fit to observed data. Breaks at 6q21-25 predicted a
each of which was noted in >5% of cases belonging to LG decreased probability of achieving remission. Patients with
NHLand DLLC,and derivedmeasuresof karyotypiccomplex- DLLC and breaks at 1q21-23 or 1~32-36had a shorter
ity, comprising modal chromosome number, number of duration of complete remission. Of 41 DLLC studied at
marker chromosomes, and number of translocation break- relapse, the only long-termsurvivors had t(14;18).
points. No correlationswith survival were noted in LG NHL, o 1991by TheAmerican Society of Hematology.

CHROMOSOMAL abnormalities in non-Hodgkm’slym- histopathology, cytogenetic analysis, cell-surface marker analysis,
phoma (NHL) have been associated with histologic and gene rearrangement studies as previously described.” All cases
subtypes,’-’immunophenotype,6and progression or transfor- were histologically classified by the International Working Formu-
mation of tumor grade.” Cytogenetic identification of lation.’’ Karyotypes were classified by the International System for
recurring reciprocal translocations has provided the stimu- Cytogenetic N~menclature.N’~ine cases of mycosis fungoides were
lus for molecular studies of the genetic events that accom- excluded from this analysis. The results of analysis of the cytoge-
pany and may result in the genesis of NHL.” Recently, netic features of these cases and correlations with histology were
several studies correlated translocations and other struc- reported separately.M Correlations between immunophenotype,
tural and numerical chromosomal aberrations with clinical immunogenotype, and clinical outcome were reported separately.”
features, including response to therapy and patient progno- Detailed cytogenetic data on the first 65 tumors included in this
sis.ll.16 Overall, these studies have been limited by small report were part of a previous histologic correlation analysis.’
subsets of patients, mixing of tumors with differing histo- Clinical features of 65 cases of large cell lymphoma with abnormal
logic grades, and inclusion of both pretreatment and karyotype and 37 specimens with normal or failed karyotypic
posttreatment cytogenetic data. analysis were part of a previous report of 18q21 rearrangement in
1ymph0ma.I~
We report the results of cytogenetic, histopathologic, and
clinical correlations derived from a large consecutive series Clinical evaluation. Clinical records were surveyed for all 423
of NHLs ascertained at a single institution in a 5-year cases; data including age at diagnosis, stage, site of involvement,
period. This analysisconfirms the prognostic value of some, serum lactate dehydrogenase (LDH), human immunodeficiency
but not all, of the previously reported cytogenetic markers
in NHL; in addition, we identified several new markers and virus (HIV) exposure, bulk of disease ( >8 cm or one third of the
established new measures of karyotypic complexity that
correlate significantlywith clinical outcome. intrathoracic diameter), initial treatment, response, response dura-
tion, date of relapse, further treatment, survival interval from time
MATERIALS AND METHODS of diagnosis until time of last follow-up or patient death, duration
of complete remission, and survival status, were collected. Patients
Tissue evaluation. Between January 1984 and December 1988, were routinely staged by physical examination, computed tomogra-
434 consecutive specimens from 423 patients with histologically phy (CT)scanning, chest radiograph, bone marrow (BM) evalua-
confirmed NHL were ascertained. Biopsy material was split for tion and, where indicated, examination of cerebro-spinal fluid,
endoscopy, or barium contrast radiography. Complete remission
From the Laboratory of Cancer Genetics, Sloan-Kettering Institute; was defined as complete absence of clinical and radiographic
and the Departments of Pathology (Cytogenetics and Surgical Pathol- evidence of disease for at least 1 month. Remission status was
ogy Services),Medicine (Lymphoma Service) and Biostatistics,Memo- monitored annually after the first year off treatment.
rial Hospital, New York,Ny.
Correlation analysis. Clinical, pathologic, and immunologicdata
Submitted August I , 1990; accepted November 21, 1990. were entered into a computerized database containing coded
Supported by Grant Nos. CA-34775 and CA-20194 from the cytogenetic variables. These variables represented data on all
National Institutes of Health, Bethesda,MD, and the Camel Research chromosomal breaks whose derivation by band could be deter-
Fund. mined, numerical chromosome abnormalities, reciprocal transloca-
Address reprint requests to Kenneth Ofit, MD, Memorial Sloan- tions, and derived measures of karyotypic complexity. The latter
measures comprised modal chromosomal number, number of
Kettering Cancer Center, Box 192, 1275 York Ave, New York, NY breaks, and number of marker chromosomes. Number of breaks
was calculated as the sum of all breaks whose derivation by band
10021. could be identified in a given karyotype; in this analysis, unidenti-
The publication costs of this article were defrayed in part by page fied marker chromosomes were not included. For the enumeration
of marker chromosomes, any clonally evident derived chromo-
charge payment. This article must therefore be hereby marked some, including unidentified markers, was counted as one. In cases
“advertisement” in accordance with 18 U.S.C. section 1734 solely to with multiple clones, the highest modal number observed in an
indicate thisfact. individual clone was used in the analysis. An identical aberration
observed in multiple clones was counted once; the sum of all such
0 I991 by TheAmerican Sociev of Hematology. breaks or marker chromosomes in all clones was recorded. Multi-
0006-4971I9117707-0016$3.0010 ple rearrangements of the same break site were recorded sepa-

1508 Blood, Vol77, No 7 (April l), 1991:pp 1508-1515

From www.bloodjournal.org by guest on February 18, 2016. For personal use only.

CYTOGENETICSOF NHL: CLINICAL CORRELATIONS 1509

rately. Triploid or tetraploid clones were reflected in the measure Patients were managed with a variety of treatment strate-
of modal number; structural abnormalities (if present) were gies: 15 patients received doxorubicin-containing combina-
measured as number of breaks or marker chromosomes. For tion chemotherapy, 12 patients received alkylating agents
analysis of correlations between kalyotype and histologic subtype, with or without radiation therapy, 9 patients were treated
the method of inference from proportions based on a chi-square by a Memorial Hospital protocol using “intermediate
analysis of two-way tables was used?’ Means were compared with intensity” combination chemotherapy? 6 patients received
two-samplet tests. radiotherapy only, 6 patients received no therapy, 2 patients
had surgery only, and 3 patients had biologic response
Survival analysis was performed on subsets of patients with modifiers or other therapies. Of the 53 patients studied
differing cytogenetic characteristics using the method of Kaplan before treatment, 47 were alive at a median follow-up of 2
and Meier.” Cytogeneticsubsets were composed of cases with any years.
breakpoint or numerical chromosomal abnormality observed with
greater than 5% incidence in this series, cases with a recurring When the analysis was restricted to patients studied at
reciprocal translocation, or cases partitioned according to mea- the time of diagnosis there were no correlations between
sures of karyotypic complexity. Actuarial survival curves were survival duration and type of treatment, presence or ab-
compared using the log-rank test. Detailed survival analysis was sence of normal metaphases, modal number, number of
limited to patients with cytogenetics sampled at time of diagnosis translocation breaks, or number of marker chromosomes.
and two histologic subsets: (a) low-grade (LG) NHL, consistingof Univariate analysis was performed on survival of subsets of
the Working Formulation diagnoses of small lymphocytic, small patients whose tumors demonstrated individual recurring
cleaved cell follicular, and mixed follicular; and (b) diffuse lym-
phoma with a large cell component (DLLC), including diffuse cytogenetic aberrations each observed in >5% of LG NHL.
mixed, large cleaved, large noncleaved, and immunoblastic sub-
types. These groups were analyzed separately because these There were no differences between subsets of patients with
subgroups of NHL have been shown to have different natural or without t(14;18) (29 patients). Seven patients had LG
histories and responses to therapy. Multivariate analysis was tumors with t(8;14); the median age of five of these patients
restricted to the larger subset of DLLC patients. The Cox propor- whose tumors were studied before treatment was 59.4 years
tional hazards regression modelUwas used to identify a subset of as compared with 55.7 for the non-t(8;14) LG NHL. There
variables that had significant impact on survival. Diagnostic haz- was no difference in survival of the t(8;14) subgroup as
ards plots were performed whenever feasible to ascertain the compared with patients without this translocation; all five
appropriateness of the proportional hazards assumption. Multiple patientswere alive at median follow-upof 14.6months from
logistic regression was used in the multivariate analysis of factors diagnosis. One patient (patient 59) with stage IV mixed
prognostic for achieving a complete response. follicular lymphoma and t(8;14) was alive at 36 months,
never having received treatment.
RESULTS
Univariate analysis of survival of subsets of cases with
Correlation between cytogenetic status and clinical charac- other chromosomal aberrations observed in more than 5%
teristics. Of 423 patients from whom specimens for cytoge- of LG NHL did not show any significant correlations; this
netic analysis were obtained, the median age at diagnosis analysis was limited by the small numbers of pretreatment
was 54 years (mean 51.4 years, range 6 to 82 years). There samples and the excellent overall survival of the entire
were 231 males and 192females in the series. There was no group. Fourteen patients demonstrated either trisomy 7,
difference in age, stage, or LDH between patients with trisomy 12, or breaks at lp32-36; the median survival of
abnormal karyotypes as compared with patients with only these patients was 56 months from diagnosis, as compared
normal metaphases or those with failed cytogenetic analy- with a median not reached for patients with abnormal
ses. karyotypes without these abnormalities. The survival curves
for these groups were significantly different ( P = .028) (Fig
Lymphomas of LG histology. Of 149 specimens of LG
NHL, 86 demonstrated an abnormal karyotype. The cytoge- 1)-
netic features of these cases were reported separately.2oThe Thirty-three specimens derived from 31 patients with LG
most common cytogenetic aberrations were, in decreasing
NHL were studied at the time of relapse; 22 were alive at a
order of frequency, chromosomes with translocation breaks median follow-up of 74 months from diagnosis and 15
at 14q32(66.3% of cases), 18q21(55.8%), 1q21-23(17.4%), months from cytogenetic analysis. There were no correla-
1p32-36 (14%), 6q21-25 (12.8%), trisomy 3 (12.8%), tri- tions between survival from time of relapse and subsets of
somy 18 (9.3%), 8q24 (8.1%), trisomy 12 (8.1%), trisomy 7 patients with individual recurring aberrations, each ob-
(7.0%), breaks at 1Oq22-24 (7.0%), trisomy 21 (5.8%), served in more than 5% of LG NHL.
trisomy 11 (5.8%), and trisomy 5 (5.8% of cases). Other
aberrations were noted in less than 5% of cases. Of the DLLC. Of 285 specimens of intermediate-grade and
reciprocal translocations, t(14;18)(q32;q21)was noted in 47 high-grade NHL ascertained, 205 (71.9%) met the criteria
(54.6%) specimens, t(8;14)(q24;q32) in 7 (8.1%) speci- of DLLC; of these, 145demonstrated abnormal karyotypes.
mens, and t(11;14)(q13;q32) in 3 (3.5%) specimens. The cytogenetic features of these cases were reported
separately.” The most frequently observed aberrations in
Cytogenetic analysis was performed before treatment on this group were chromosomes with breaks at 14q32 (28.6%
99 specimens, of which 53 specimens, derived from 53 of cases), 18q21 (17.2%), 1q21-23 (16.6%), trisomy 7
patients, showed abnormal karyotypes. The median age at (16.6%), 8q24 (15.2%), 6q21-25 (14.5%), trisomy 12
diagnosis of the 53 cases was 59 years (mean 56 years, range (13.8%), breaks at 1p32-36 (11.7%), trisomy 5 (9.7%),
23 to 81years). Of these, 28 patientswere stage IV, 11were trisomy 3 (9%), trisomy 18 (9%), breaks at lp22 (7.6%),
stage 111, 7 were stage 11, and 7 were stage I at diagnosis.

From www.bloodjournal.org by guest on February 18, 2016. For personal use only.

1510 OFFIT ET AL

o-l survival of this cohort, however, was no different from that
of other DLLC patients (Table 1). Fourteen cases of DLLC
:: demonstrated t(14;18)(q32;q21) in biopsies obtained be-
6- fore cytotoxic therapy. The mean age of the t(14;18) DLLC
patients was 11.3 years older than for the non-t(14;18)
a> DLLC; there was no difference in mean LDH or proportion
Ian stage IV, although there was a trend for decreased survival
of the t(14;18) subset of patients (P = .07) (Table 1).Two
P2 .4- p =.03 patients had a history of transformation of histologic grade.
la- One had a previous history of chronic lymphocytic leuke-
mia, and another had a history of follicular small cleaved
p2 .2- lymphoma. Neither patient had received cytotoxic chemo-
n m1P32-36 Break, trisomy 7 or trisomy 12 present therapy at the time of cytogenetic analysis, both showed a
t( 14;18) with multiple additional abnormalities, and both
Other karyotypic abnormalities died within 2 months of transformation.

0 I I1 Among the individual recurring chromosomal aberra-
0 12 24 3 6 48 60 72 tions observed in more than 5% of DLLC, significant
MONTHS FROM DIAGNOSIS correlations with survivalwere noted for subsets of patients
with breaks on the long arms of chromosomes 1 or 6.
Fig 1. Actuarial survival from time of diagnosis of 14 patientswith Eighteen patients whose tumors demonstrated a transloca-
LG lymphoma and abnormal karyotypes at diagnosis with trisomy 7, tion break at 1q21-23and 12patients with breaks at 6q21-25
trisomy 12, or breaks at lp32-36 as compared with 39 patients had median survivals of less than 1 year; their survival
without these abnormalities. curves are shown in Figs 2 and 3 (P = <.001). The
shortened survival of patients with DLLC demonstrating
breaks at 7q32 (7.0%), trisomy 11 (6.2%), and aberrations breaks at 1q21-23 or 6q21-25 remained significant in the
of 17p,q, or cen (6.2% of cases). The most frequent smaller subset of 41 patients receiving “third-generation’’
recurring translocations were t( 14;18) (17.2% of cases),
t(8;14) (14.5%), and t(3;22)(q27;qll) (4.1% of cases). combination chemotherapy (P < .001, P < .001, respec-

Abnormal karyotypes were detected at the time of tively). Partitioning of the survival analysis according to
diagnosis in 106 specimens derived from 104 patients; the
median age of the 104 patients was 55 years (mean 53.3 Table 1. Survival of Clinicaland Cytogenetic Subsets of DLLC
years, range 12 to 76 years), 5 were stage IE, 26 were stage PatientsWith KaryotypicAnalysis Performedat Time of Diagnosis
11, 28 were stage 111, and 42 were stage IV. The median
LDH at the time of diagnosis for the 76 patients on whom Median
data were availablewas 272 (mean 441, range 102 to 3,240).
Of the five stage IE patients, all had solitary extranodal Mean Mean CR Survival P
lesions that were resected and treated with adjuvant chemo-
therapy. Thirty-two patients received treatment with first- Subgroup n Age LDH (%) (mo) Value’
generation combination chemotherapy (NHL-5;’ CHOP,
or BACOP), 21 with second-generation chemotherapy Clinical
(NHL-7,” m-BACOD], and 41 with third-generation chemo-
therapy (MACOP-B, NHL-9Y L17/2OZ8]8; received radio- All DLLC 104 53.3 441 63 NYR -
therapy plus nonanthracycline-containing chemotherapy,
and 2 died before treatment could be given (one death was stage IV 42 49.3 697t 50 13.2 .01
a suicide). The median survival of the 104 patients was not
reached; actuarial survival for the group was 53% at 5 years. LDH >500 19 53.8 - 3 7 t 10.1 .02

There was no difference in survivalbetween patients with Cytogenetic
abnormal karyotypes as compared with patients with nor-
mal karyotypes or those who were cytogenetic failures. t(a;i4) 16 44.2t 398 56 NYR .45
Univariate analysis showed no correlation between survival
and age greater than the median, A v B symptoms, bulk of t(14; 18) 14 62.2t 499 57 12 .07
disease, histologic subtype of DLLC, or generation of 1q21-23 18 54.4 522 44
chemotherapy treatment. Of the 76 patients with LDH data 7.6 < ,0001
at diagnosis, 19 patients with values greater than 500 had a
diminished survival (P = .02); of the 101patients on whom 1p32-36 12 61.9t 194 58 12 .1
staging data were available, those with stage IV disease had
diminished survival (P = .01). 6q21-25 12 58.8 485 3 3 t 7.5 ,0003

Of 16 patients with t(8;14) or variants, the mean age of f 7 19 58.5 392 58 9.3 .07
the t(8;14) patients was 10 years younger than DLLC with
abnormal karyotypes lacking t(8;14); the mean LDH and strl7 5 65.0 648 25 20 .09

>4 Markers 34 53.3 373 4 7 t 9.3 <.001

>4 Breaks 27 57.5 403 48 9.3 .01

The cytogenetic groups are defined by the presence of listed
abnormalities, numbers of marker chromosomes, or chromosome
breaks as defined in the text; 1q21-23, 1p32-36, and 6q21-52 indicate
tumors with rearrangementsinvolvingthese sites appearing as nonran-
dom abnormalities,and strl7 denotestumorswith any clonalrearrange-
ments affecting chromosome 17 (p,q, or cen).

Abbreviations: NYR, not yet reached; LDH, lactate dehydrogenase;

DLLC, diffuse lymphoma with large cell component; CR, complete
remission (survivalin months).

*P value for survival differences by logrank test comparison of
Kaplan-Meier survival curves.

tsignificantly different mean value from subset without this aberra-
tion.

From www.bloodjournal.org by guest on February 18, 2016. For personal use only.

CYTOGENETICSOF NHL: CLINICAL CORRELATIONS 1511

1.0. apy. In none of these patients was the t(1;6) accompanied
by a t(14;18), t(8;14), or t(11;14). In all of these patients, the
+ No Break a t lq2l-23 break on chromosome 1 was at band q21; the break on
0 Break at lq21-23 chromosome 6q could not be unequivocally resolved in one
.8. patient, it was 6q21 in one patient, and 6q25 in another.

(1 Among the other cytogenetic subsets analyzed by univari-
ate analysisthere were no correlations with patient survival;
f.6, there was a trend for diminished survival for 19 patients
with trisomy 7 ( P = .07). The median survival was only 20
a5 months for five patients with structural abnormalities of
chromosome 17.The mere presence or absence of unidenti-
3
u) PC.0001 fied marker chromosomes was not a prognostic factor for
survival ( P = .27). Significant correlations were noted be-
E* tween survival and derived measures of karyotypiccomplex-
ity. The cohorts of 34 patients with more than four marker
t
B chromosomes ( P < .001) and 27 patients with more than 4
breakpoints (P = .01) had decreased survival. There was
2.2
limited coincidence of these measures of chromosomal
L complexity and other cytogenetic aberrations of prognostic
utility; of the 18patients with 1q21-23breaks, 10 (56%) had
Ic more than five markers; of the 12 patients with 6q21-25
breaks, 6 (50%) had more than five markers. Of the five
a II 1i patients with structural abnormalities of chromosome 17,
all had complex karyotypes with more than four marker
13 26 39 52 65 78 chromosomes and four to nine breaks.

MONTHS FROM DIAGNOSIS Cox regression analysis was used to identify a subset of

Fig 2. Actuarial survival from time of diagnosisof 18 patientswith variables which, in combination, significantly predicted
DLLC and breakpointsat 1q21-23 at time of diagnosis as compared survival. However, multivariate analysiswas limited both by
with 86 patientswith abnormalkaryotypeswithout this abnormality. overall sample size and by OUT ability to assess the assump-
tion of proportional hazards in the multivariate setting.
level of LDH did not diminish the prognostic value of the Two variables, number of marker chromosomes greater
than 4 and breaks at 1q21-23,where chosen in this analysis
two cytogenetic variables; 13patients with either a break at because of their individual contribution to prediction in the
bivariate setting (Fig 4). Breaks at 6q21-25 appeared to
1q21-23or 6q21-25 and LDH < 441 had a shorter survival contribute additional prognostic power to the multivariate
model; however, this variable was not included because of
as compared with 44 patients without these aberrations and difficulties in validation of the more complex model. The
other clinical and cytogenetic variables were outweighed by
LDH < 441 ( P = .001). Similarly,of 19 patients with LDH these cytogenetic variables in the multivariate analysis.
> 441, five patients with breaks at l q or 6q had a decreased
With respect to the remission status, univariate analysis
survival (P = .01). There was no difference in the mean age identified one cytogeneticvariable for study in the multivari-
ate model; breaks at 6q21-25 correlated with a diminished
or LDH of patients with breakpoints at 6q21-25 or 1q21-23 probability of achieving complete remission in a stepwise
logistic regression model (P = .005).
as compared with patients without such breaks (Table 1).
Among the 66 patients who achieved complete remission,
In 3 of the 18 patients with 1q21-23breaks, the tumors there was no correlation between duration of the remission
and age, stage, or LDH at diagnosis. Among the cytogenetic
had a reciprocal translocation involving both l q and 6q variables tested, duration of complete remission was also
shorter in nine patients with breaks at 1q21-23 and seven
breaks; the survival of these patients from diagnosis was 6, patients with breaks at 1p32-36; the actuarial curves dif-
8, and 10months, despite intensive combination chemother- fered from those for patients without such aberrations
( P = .02, P = .0002, respectively).
1.01 No Break at 6q21-25
Break at 6q21-25 In 41 specimens derived from 41 patients with DLLC,
!A.8 cytogenetics was studied at the time of relapse. As ex-
pected, the overall outcome for these patients was dismal,
-u P'.0003 with a median survival of only 14 months. Patients with
t(14;18) had a median survival not yet reached from time of
zr .6- posttreatment cytogenetic sampling, as compared with 12.8
months in patients without this translocation (P = .07) (Fig
3

Ez .4-

a

0
0

P .2-

0

04 1

0 13 26 39 52 65 78

MONTHS FROM DIAGNOSIS

Fig 3. Actuarial survival from time of diagnosisof 12 patientswith
D U C and breakpoints at 6q21-25 at time of diagnosis as compared
with 92 patientswith abnormal karyotypes without these abnormali-
ties.

From www.bloodjournal.org by guest on February 18, 2016. For personal use only.

1512 OFFIT ET AL

1.o tumors who died, 3 had previous LG lymphomas. All 10 of
the t(14;18) patients sampled at relapse had multiple
.8 additional secondary abnormalities, but small numbers in
cytogenetic subsets precluded correlation with median
=! survival after relapse. No other primary or secondary
cytogenetic abnormality was of prognostic value in the
mma cohort of 41 previously treated patients; homogeneously
0na .6 staining regions (HSRs) were detected in 2 patients, both of
Aa whom died within 1year of their detection.

2 DISCUSSION

L .4 Correlations between nonrandom cytogenetic aberra-

3 tions and histologic and immunophenotypic subsets of
In
0 NHL have been established and confirmed in large series.'-9

sY Recently, several studies identified cytogenetic subsets of
.2
NHL with differing response to therapy, survival, and
4
clinical behavior.''-l6 These studies were limited by small
c
numbers of patients in the cytogenetic sub~ets,'~~m','i*xing
In
together of subsets of patients with different histologies and
W
grades of NHL,'','2,'5,'6inclusion of samples studied after
0
16 32 48 64 80 cytotoxic and lack of breakpoint-specific
MONTHS OF SURVIVAL
chromosomal
Fig 4. Proportional hazards model for overall survival and two
cytogenetic variables: breakpoints at 1q21-23, and greater than four In the current study, sufficient numbers of patients were
marker chromosomes.Curve A representspatientswith four or fewer
markers and no break at 1q21-23; curve B represents patients with available to permit analysis by histologic subgroup. Al-
more than four markers and no break at 1q21-23; curve C represents
patientswith four or fewer markersand a break at 1q21-23; and curve though different therapies have yet to demonstrate a
D represents patients with more than four markers and a break at
lq21-23. significant effect on survival of patients with LG lymphomas

5). Of 10 patients with t(14;18) detected at time of relapse, in single-institution analysis of this cohort in the
6 were alive, 4 of the 6 at more than 1 year from time of
relapse. Of the 4 long-term survivorswith t(14;18) detected current study was limited by different treatment ap-
at relapse, 3 had a documented history of transformation proaches and short median follow-up. The survival of
from a lower histologic grade; the t(14;18)was documented
in the LG tumor in 3 of these patients. The fourth patients showing trisomy 7 or trisomy 12 or breaks at
long-term survivor had an unusual history of recurrent
DLLC for 7 years; the t(14;18) was detected at the time of 1p32-36 was shorter than that of patients without these
the second relapse. Of the 4 patients with t(14;18) bearing
abnormalities; however, the relatively long median survival
t(14; 18) Absent
tm. :,181 Present of this cohort ( >4 years) limits the clinical utility of this

I observation. Other cytogenetic features reported to corre-

u M e with a differing clinical behavior in LG lymphomas,

2' including the presence of normal metaphases, trisomy 2, or

deletions of 2p, did not correlate with clinical outcome;

other aberrations, eg, chromosome 17 abnormalities and

breaks at 13q32,were observed too infrequently to analyze.

Most surprising in the LG lymphomas was the observa-

tion of t(8;14)(q24;q32) in seven patients, five of whom had

not yet been treated. The indolent clinical behavior of the

disease in these patients (one patient remained untreated

>2 years) suggests a biology of disease different from that

of high-grade NHL bearing the identical translocation.

Analysis of the DLLC-containing histologies did confirm

a trend for decreased survival of t(14;18) bearing DLLC,

although not at the level of significance of another series.I3

This translocation, usually associated with follicular NHL,

was also associated with an older age at diagnosis and a

trend for prolonged survival at the time of relapse, suggest-

ing a resemblance to the natural history of follicular

lymph~ma.'C~ontinued follow-up of this cohort will, how-

ever, show continued lymphoma-related mortality as pa-

0 1 tients die of the complications of recurrent disease. Al-
0 12 24 36 48 60 72
though observed in 15% of DLLC tumors, the clinical

MONTHS m M RELAPSE behavior and clinical features at presentation of patients

Fig 5. Actuarial s W a l from time of relapse for 10 patients with with t(8;14)-bearing DLLC were no different from that of
t(14;18)(q34;q21) detected at time of relapse as compared with 31
patientswithout this translocation. other DLLC.

Multivariate analysis showed that two cytogenetic vari-

From www.bloodjournal.org by guest on February 18, 2016. For personal use only.

CYTOGENETICS OF NHL: CLINICAL CORRELATIONS 1513

ables, one a specific breakpoint at 1q21-23and the other a of chromosomes 5,6, and 18.Because of limited numbers in
measure of karyotypic complexity, significantly predicted histologic subsets, each of these studies mixed together all
survival when used in combination. Another common histologies of NHL. Restricting the analysis to large cell
breakpoint, at 6q21-25,was also a useful prognostic marker lymphomas,we were able to confirm some of these findings.
in the univariate analysis, and, to a much lesser extent, in There was a trend for diminished prognosis among the
the multivariate analysis. In three patients, both break- subgroup with chromosome 17 abnormalities; however,
points were mutually involved in the reciprocal t(1;6), not each of these patients had highly complex karyotypes with
previously described. Although clinicalvariables (eg, serum multiple additional aberrations. This study was unable to
LDH) were also associated with prognosis, these correla- confirm the previously reported correlations between tri-
tions were overshadowed by chromosomal abnormalities in somy 3 or duplications of 3p with a favorable prognosis in
the multivariate analysis. With respect to LDH, however, DLLC and trisomies of chromosomes 2,5, 6, or 18, dupli-
data were not available for all patients in this series; such cations of 2p, or monosomy 7 with poor prognosis in
data might have increased the prognostic value of this DLLC.131'6
variable in the multivariate analysis.
The variable treatment regimens and short median fol-
The prognostic utility of a measure of karyotypiccomplex- low-up in this study suggest the need for large-scale
ity in NHL was suggested in an early study" but has not prospective trials using uniform treatment approaches.
been previously derived based on number of breakpoints or Although such trials may help resolve some of the reported
marker chromosomes. The prognostic subgroup with com- differences in cytogenetic-clinical correlations in DLLC,
plex karyotypes identified in this study may encompass this approach will not necessarily be definitive. First, in
smaller subsets of specific chromosomal aberrations noted DLLC it is unclear that one anthracycline-containingtreat-
in low frequency but indicative of an aggressive growth ment regimen is superior to another.40 In this study,
pattern. One such marker may be on chromosome 17; all restricting the analysis to the subset of patients receiving
five cases of DLLCwith structural abnormalities of chromo- third-generation chemotherapy did not diminish the corre-
some 17 detected at diagnosis in this series were associated lations. Second, significant genetic heterogeneity of NHL
with marked karyotypic complexity.Alternatively, the serial clearly exists. Translocations such as t(14;18) or t(8;14) are
accumulation of a number of low-frequency cytogenetic not completely specific to subtypes of the disease. In
aberrations, including the unidentified clonal markers, may addition, comparison of geographically disparate series of
contribute toward the aggressive clinical behavior of these NHL shows marked cytogenetic heterogeneity; eg, the
tumors. incidence rates of t(8;14) or 6q breaks in low-grade and
intermediate-grade NHL and t(14;18) in intermediate-
This study confirmed a previous correlation of breaks at grade to high-grade NHL vary considerably in comparable
1p32-36 and decreased disease-free survival.L2Breaks in series.5.6,8,9,4WSuch changes as duplications of 2p and 3p,
this region, detected at time of presentation or relapse, abnormalities of 17p, and breaks at 1q21-23were common
have also been correlated with BM involvement in NHL.30 in some series but rare enough to preclude meaningful
With the exception of immunophenotype," no clinical analysis in other series.'331s3T1h6ird, this analysis suggests
prognostic variables currently exist to predict relapse from that accumulation of cytogenetic aberrations may serve as
complete remission in DLLC.31A combination of clinical, an accurate predictor of clinical behavior. Further analysis
immunophenotypic, and cytogenetic prognostic markers of the additive effects of multiple genomic aberrations may
may better define populations of patients with DLLC who resolve apparent contradictions of single aberration-based
would benefit from more intensive treatment with autolo- analyses. Finally, not all morphologically identical struc-
gous transplantation3' or growth factor support. tural chromosomal aberrations have the same biologic
meaning, as illustrated by the Ph' chromosome in acute and
The method of breakpoint analysis used in this series chronic le~kemias.4T~his appears to be true of t(8;14),
showed correlations between specificbreakpoints and clini- perhaps the most intensively analyzed at the molecular
cal endpoints. The biologic mechanisms underlying these level of the genomic aberrations in NHL. The presence of
correlations remain unclear. Although known genes have this translocation in seven patients with histologically and
been mapped to the regions 1q21-23, 6q21-6q25, 1p32-36, clinically indolent NHL in this series suggests the possibility
and several of the other common breakpoints we report, of different types of oncogene dysregulation in these
dysregulation of none of these genes has been demon- tumors. Direct assay of altered or dysregulated gene prod-
strated in human NHL.33"8Dysregulation of MYB, mapped ucts may provide more reproducible and useful prognostic
to the broad region 6q21-25, has been shown in T-cell markers in future studies.
leukemia cell lines.33Although such oncogene dysregula-
tion may result from deletion of regulatory sequences, more ACKNOWLEDGMENT
complex models involvingdeletion of suppressor genes with
or without subsequent mutation at the remaining allele We are indebted to Quanguarig Chen, Prasad Koduru, Alice
have been proposed to account for correlations between Hampton, Michele Simon, Salah Ebrahim, Carolyn Schultz, Philip
genomic aberrations and tumor progression in other malig- Pastalis, and Nasser Parsa, who participated in the karyotyping of
nancie~.~~
the tumors on which this analysis was based. We also thank Drs
Three recent s t u d i e ~ ' * ~re' ~po~r't~ed correlations between
survival of patients with NHL and cytogenetic abnormali- Bayard Clarkson and Carol Portlock for critical reading of the
ties, including 2p breaks, structural abnormalities of chro-
mosome 17,monosomy 7 or 7p abnormalities, and trisomies manuscript.

From www.bloodjournal.org by guest on February 18, 2016. For personal use only.

1514 OFFIT ET AL

REFERENCES

1. Kaneko Y, Rowley JD, Variakojis D, Haren JM, Ueshima Y, 18. Non-Hodgkin’s lymphoma pathologic classification project:
Daly K, Kluskens LF: Prognostic implications of karyotype and NCI sponsored study of classifications of non-Hodgkin’s lympho-
morphology in patients with non-Hodgkin’s lymphoma. Int J mas: Summary and description of a working formulation for
Cancer 32:683,1983 clinical usage. Cancer 49:2112,1982

2. Yunis JJ, Oken MM, Theologides A, Howe RB, Kaplan M: 19. Harnden DE, Klinger HP, (eds) ISCN: An International
Recurrent chromosomal defects are found in most patients with System for Human Cytogenetic Nomenclature. Basel, Karger, 1985
non-Hodgkin’s lymphoma. Cancer Genet Cytogenet 13:17,1984
20. Offit K, Jhanwar SC, Ladanyi M, Filippa DA, Chaganti
3. Bloomfield CD, Arthur DC, Frizzera G, Levine EG, Peterson RSK Cytogenetic analysis of 434 consecutively ascertained speci-
BA, Gajl-Peczalska KJ: Nonrandom chromosome abnormalities in mens of non-Hodgkin’s lymphoma: Correlations between recur-
lymphoma. Cancer Res 43:2975,1983 rent aberrations, histology, and exposure to cytotoxic treatment.
Genes Chrom Cancer (in press)
4. Levine EG, Arthur DC, Frizzera G, Peterson BA, Hurd DD,
Bloomfield CD: There are differences in cytogenetic abnormalities 21. Armitage P: Statistical Methods in Medical Research, 4th
among histologic subtypes of the non-Hodgkin’s lymphomas. Blood ed, London, Blackwell Scientific, 1977
66:1414, 1985
22. Kaplan EL, Meier P: Nonparametric estimation from incom-
5. Koduru PRK, Filippa DA, Richardson ME, Jhanwar SC, plete observations. J Am Stat SOC53:457, 1958
Chaganti SR, Koziner B, Clarkson BD, Lieberman PH, Chaganti
RSK Cytogenetic and histologic correlations in malignant lym- 23. Cox DR: Regression models and life tables (with discus-
phoma. Blood 69:97,1987 sion). J R Stat SOCB 34187,1972

6. Levine EG, Arthur DC, Gajl-Peczalska KJ, LeBien TW, 24. Kempin S, Cirrincione C, Myers C: Combined modality
therapy of advanced nodular lymphomas. Proc Am SOCClin Oncol
Peterson BA, Hurd DD, Bloomfield CD: Correlations between 256,1983 (abstr)
immunological phenotype and karyotype in malignant lymphoma.
Cancer Res 46:6481,1986 25. Carratto A, Koziner B, Wong G, Lowenthal D, Chapman D,
Little C, Lee BJ, Straus D, Nisce L, Clarkson BD: Randomized
7. Yunis JJ, Frizzera G, Oken MM, McKenna J, Theoligides A, comparison of CHOP versus Bleo CHOP in NHL. Proc Am SOC
Arnesen M: Multiple recurrent genomic defects in follicular Clin Oncol6:779,1987(abstr)
lymphoma. N Engl J Med 316:79,1987
26. Lowenthal D, Carrato A, Koziner B, Little C, Warrell R,
8. Richardson ME, Quanguang C, Filippa DA, Offit K, Hamp- Young CY, Straus D, Lee B, and Clarkson BD: The NHL-7
ton A, Koduru PRK, Jhanwar SC, Lieberman PH, Clarkson BD, protocol: Alternating non cross resistant chemotherapy for diffuse
Chaganti RSK Intermediate to high grade histology of lymphomas non-Hodgkin’s lymphoma. Proc Am SOCClin Oncol 6:778, 1987
carrying t( 14;18) is associated with additional non-random chromo- (abstr)
some changes. Blood 70444,1987
27. Warrell RP, Straus DJ, Martin M, Kempin SJ, Lee BJ,
9. Armitage JO, Sanger WG, Weisenburger DD, Harrington Koziner B, Clarkson BD: Short term intensive treatment of
DS, Linder J, Bierman PJ, Vose JM, Purtillo DT: Correlation of intermediate-grade non-Hodgkin’s lymphoma using infusional che-
secondary cytogenetic abnormalities with histologic appearance in motherapy. Proc Am SOCClin Oncol8:1054,1989 (abstr)
non-Hodgkin’s lymphomas bearing t(14;18)(q32;q21). J Natl Can-
cer Inst 80576,1988 28. Straus DJ, Oppenberg J, Little C, Filippa DA, Offit K,
Clarkson BD: Small non-cleaved cell lymphoma in adults treated
10. Croce CM, Nowell PM: Molecular basis of human B cell with acute leukemia chemotherapy: Results at 12.5 years. Proc Am
neoplasia. Blood 65:1,1985 SOCClin Oncol7:927, 1988 (abstr)

11. Kristoffersson U, Heim S, Mandahl N, Olsson H, Ranstam J, 29. Young RC, Longo DL, Glatstein E, Ihde DC, Jaffe ES,
Akerman M, Mitelman F Prognostic implication of cytogenetic DeVita VT Jr: The therapy of indolent lymphomas: Watchful
findings in 106 patients with non-Hodgkin lymphoma. Cancer waiting v aggressive combined modality therapy. Semin Hematol
Genet Cytogenet 25:55,1987 25:11,1988

12. Levine EG, Arthur DC, Frizzera G, Peterson BA, Hurd DA, 30. Offit K, Richardson M, Chen Q, Hampton A, Koduru P,
Bloomfield CD: Cytogenetic abnormalities predict clinical out- Jhanwar S, Filippa D, Leiberman P, Clarkson BD, Chaganti RSK.
come in non-Hodgkin’s lymphoma. Ann Intern Med 108:14,1988 Nonrandom chromosomal aberrations are associated with sites of
tissue involvement in non-Hodgkin’s lymphoma. Cancer Genet
13. Yunis JJ, Mayer MG, Arnesen MA, Aeppli DP, Oken MM, Cytogenet 37:85,1989
Frizzera G: Bcl-2 and other genomic alterations in the prognosis of
large cell lymphoma. N Engl J Med 320:1047,1989 31. Armitage JO: Bone marrow transplantation in the treatment
of patients with lymphoma. Blood 73:1749,1989
14. Offit K, Koduru PRK, Hollis R, Filippa D, Jhanwar SC,
Clarkson BC, Chaganti RSK: 18q21 rearrangement in diffuse large 32. Gulati SC, Shank B, Black P, Yopp J, Koziner B, Straus D,
cell lymphoma: incidence and clinical significance. Br J Haematol Filippa D, Kempin S, Castro-Malespina H, Cunningham I, Berman
72:178, 1989 E, Coleman M, Langleben A, Colvin OM, Fuks Z, O’Reilly R,
Clarkson B: Autologous bone marrow transplantation for patients
15. Cabanillas F, Pathak S, Grant G, Hagemeister F, McLaugh- with poor prognosis lymphoma. J Clin Oncol6:1303, 1988
lin P, Swan F, Rodriguez M, Trujillo J, Cork A, Butler J, Katz R,
Bourne S, Freireich E: Refractoriness to chemotherapy and poor 33. Barletta C, Pier-Giuseppe P, Kenyon LC, Smith SD, Dalla-
survival related to abnormalities of chromosomes 17 and 7 in Favera R: Relationship between the c-myb locus and the 6q-
lymphoma. Am J Med 87:167,1989 chromosomal aberration in leukemias and lymphomas. Science
235:1064,1987
16. Schouten HC, Sanger W, Weisenburger DD, Anderson J,
Armitage JO: Chromosomal abnormalities in untreated patients 34. Morton CC, Diamond A, Lane M, Cooper GM, Leder P:
with non-Hodgkin’s lymphoma: Associations with histology, clini- Mapping of the human B-LYM transforming gene activated in
cal characteristics, and treatment outcome. Blood 75:1841,1990 Burkitt’s lymphoma to chromosome 1. Science 223:173, 1984

17. Offit K, Ladanyi M, Gangi MD, Ebrahim SAD, Filippa D, 35. LeBeau MM, Westbrook CA, Diaz MO, Rowley JD: Evi-
Chaganti RSK Ki-1antigen expression defines a favorable clinical dence for two distinct c-src loci on human chromosomes l and 20.
subset of non-B cell non-Hodgkin’s lymphoma, Leukemia 4:625, Nature 312:70, 1984
1990
36. Tronick SR, Popescu NC, Cheah MSC, Swan DC, Armsbaugh

From www.bloodjournal.org by guest on February 18, 2016. For personal use only.

CYTOGENETICSOF NHL: CLINICAL CORRELATIONS 1515

SC, Lengel CR, DiPaolo JA, Robbins K C Isolation and chromo- an aggressivechemotherapeutic regimen (MACOP-B) for interme-
somal localization of the human fgr protooncogene, a distinct diate and some high grade non-Hodgkin’s lymphomas. J Clin
member of the tyrosone kinase family. Proc Natl Acad Sci 82:6595, Oncol8:94,1990
1985
41. Cabanillas F, Pathak S, Trujillo J, Manning J, Katz R,
37. McBride OW, Kirsh I, Hollis G, Nau M, Battey J, Minna J: McLaughlin P, Valasquez WS, Hagemeister FB, Goodacre A, Cork
Human L-myc proto-oncogene is on chromosome lp32. Eighth A, Butler JJ, Freireich ET: Frequent nonrandom chromosome
International Workshop on Human Gene Mapping. Cytogenet abnormalities in 27 patients with untreated large cell lymphoma
Cell Genet 40694,1985 and immunoblastic lymphoma. Cancer Res 48357,1988

38. Chaganti RSK, Balazs I, Jhanwar SC, Murty W S , Koduru 42. Chaganti RSK, Doucette LA, Offit K, Filippa DA, Allen GJ,
PRK, Grzeschik KH, Stavnezer E: The cellular homologue of the Condon MR, Jhanwar SC, Clarkson BD, Lieberman PH: Specific
transforming gene of SKV avian retrovirus maps to human chromo- translocations in non-Hodgkin’s lymphoma: Incidence, molecular
some region lq22 --$ q24. Cytogenet Cell Genet 43:181,1986 detection, and histological and clinical correlations. Cancer Cells.
7:33,1989
39. Soloman E. Colorectal cancer genes. Science 342412,1990
40. Sciineider AM, Straus DJ, Schluger AE, Lowenthal DA, 43. Butturini A, Gale RP: Oncogenes and leukemia. Leukemia
Koziner B, Lee BL, Wong G, Clarkson BD: Treatment results with 4:138,1990

From www.bloodjournal.org by guest on February 18, 2016. For personal use only.

1991 77: 1508-1515

Cytogenetic analysis of 434 consecutively ascertained specimens of
non- Hodgkin's lymphoma: clinical correlations

K Offit, G Wong, DA Filippa, Y Tao and RS Chaganti
Updated information and services can be found at:
http://www.bloodjournal.org/content/77/7/1508.full.html
Articles on similar topics can be found in the following Blood collections
Information about reproducing this article in parts or in its entirety may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#repub_requests
Information about ordering reprints may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#reprints
Information about subscriptions and ASH membership may be found online at:
http://www.bloodjournal.org/site/subscriptions/index.xhtml

Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly by the American
Society of Hematology, 2021 L St, NW, Suite 900, Washington DC 20036.
Copyright 2011 by The American Society of Hematology; all rights reserved.