Multiparameter characterization of L3 leukemia cell populations

LeukemiaResearchVol. 11. No. 1. pp. 73-83. 1987 Printed in Great Britain.

0145-2126/'87$3.00 + .00 ~ 1987PergamonJournals Ltd.

M U L T I P A R A M E T E R C H A R A C T E R I Z A T I O N OF L3 LEUKEMIA CELL POPULATIONS ALEXANDER J. WALLE,* AYAD AL-KATIB, GEORGE Y. WONG, SURESH C. JHANWAR, RAJU S. K. CHAGANTI and BENJAMIN KOZINER Memorial Sloan-Kettering Cancer Center, The Rogosin Institute and The New York Hospital, and Cornell University Medical Center, New York, U.S.A.

(Received 1 July 1985. Revision accepted 7 July 1986) A~tract--Multiparameter analysis of 14 consecutive patients with L3 cell populations (FAB classification) in bone marrow showed consistently high mean cellular RNA content but heterogeneity with respect to percent cells in S- and G2M-phases of the cell cycle, cell surface phenotype, cytogenetic abnormalities, cellular DNA content, and nuclear activities of terminal deoxynucleotidyl transferase. Five of the 14 patients had the characteristic profile of L3 leukemia with high RNA content, a typical t(8;14) or t(8;22) translocation, low Tdt, B-cell surface phenotype, pseudodiploid modal number of chromosomes, high proliferation and normal DNA index. The remaining 9 patients had features thought to be unusual ~.nd they were very heterogeneous. Median survival time from time of diagnosis was 115 days with no patient alive after 474 days. The survival times exhibited excellent correlation with a parametric model assuming exponential distribution. Accordingly, the presence of abnormal DI predicted for decreased, and the presence of t(8;14) or pseudodiploidy was associated with increased survival rates. However, due to the small number of patients in this study, generally applicable conclusions should be drawn only from a collection of larger numbers of cases of this rare type of leukemia.

Key words: Leukemia, FAB classification, Burkitt's lymphoma, terminal deoxynucleotidyl transferase, cell surface markers, cytogenetics, cellular DNA content, flow cytometry.

INTRODUCTION IN ACCORDANCE with the currently used F A B classification of leukemias [1], L3 leukemia is closely related to Burkitt's l y m p h o m a [2] through the similarity of cytomorphological features. Frequently, L3 leukemia has been regarded the leukemic phase of Burkitt's or Burkitt's-like l y m p h o m a [3]. As such, the predominant n u m b e r of L3 leukemias exhibited B-cell phenotype [3]. Despite the morphologic identity of the immature lymphoblasts a study of 5 patients pointed towards

existing phenotypic heterogeneity of L3 cell populations

[4].

Results of our cytogenetic, kinetic, phenotypic and metabolic studies of L3 cell populations of bone marrow, blood and lymph node specimens demonstrate that the uniform morphologic picture is associated with a remarkable heterogeneity of nuclear and cell m e m brane features each of which may carry its own biologic and clinical relevance.

MATERIALS

AND

METHODS

Patients Fourteen consecutive patients (mean age 37-+ 17yr; 3 females, 11 males; representing 2.1% of all cases of acute lymphoblastic leukemia admitted to Memorial Hospital between 1978 and 1983) had the diagnosis of L3 leukemia made based on the presence of L3 blasts in bone marrow aspirates, with or without blood involvement. Four investigators had determined that more than 50% cells in all of the tetrachrome stained smears had typical features of L3 cells (FAB classification [1]) (Fig. 1). Lymphadenopathy diagnosed as diffuse undifferentiated lymphoma of Burkitt's type was present in some patients. Cases 6 and 7 were partially presented previously [4]. Two cases of L3 leukemia with Burkitt's-like lymphoma had evolved from previous nodular poorly dif-

* A. J. W., at the time of the study, was a Clinical Fellow of the American Cancer Society. Abbreviations: AO, acridine orange fiuorochrome; BM, bone marrow; DI, DNA index; EAC, erythrocyte-amboceptor-complement rosettes; FCM, flow cytometry; FITC, fluoroisothiocyanate fluorochrome; F600 and F530, emission wavelengths of AO-RNA and AO-DNA complexes, respectively; GO~l-, S-, G2M-cells, percent of cells in the respective phases of the cell cycle; LN, lymph node; PB, peripheral blood; SD, standard deviation; SM, surface markers; Tdt, terminal deoxynucleotidyl transferase. Correspondence: Dr Alexander J. Walle, The New York Hospital---Cot'nell University Medical Center, The Rogosin Institute, 430 E 71st Street, New York, NY 10021, U.S.A. 73

74

A . J . WALLEet al.

ferentiated lymphoma (cases 13, 14). Extent of disease and clinical details of all cases at time of diagnosis are described in Table 1.

Cytofluorometry of immunofluorescence Heparinized venous blood (n = 1) and BM (n = 13) cell suspensions were subjected to Ficoll-Hypaque (Lymphoprep, Nyegaard & Co., Oslo, Norway) gradient centrifugation according to Boyum [5]. Four LN cell suspensions were studied. Immunofluorescent staining of these mononuclear cell suspensions was routinely performed as described [6]. Cells were analysed for FITC fluorescence using a Cytofluorograph System 30-L (Ortho Instruments, Westwood, Mass.) or by immunofluorescence microscopy (cases 6,7) using a Leitz Ortholux II microscope equipped with vertical fluorescence epiillumination. SM determinations included HLA-DR, polyvalent immunoglobulin, lgM and lgG heavy, and kappa and lambda light chains. Clonal excess of surface light chain bearing cells was calculated [7] for patients 12-14.

Immunological and biochemical tests Sheep red blood cell rosette determination after incubation at 37°C for 1 h was carried out as reported [8]. EAC rosettes were prepared as described [9]. Tdt was determined biochemically [10].

Cytofluorometry of A O fluorescence The mononuclear cell suspensions (PB, n = 1; BM, n = 12), obtained through Ficoll-Hypaque gradient centrifugation (see above), and one LN cell suspension were stained with AO as described [11], and excited with blue light at 488nm of an argon ion laser beam. Orthochromatic red fluorescence at 630-670nm (F600) is proportional to RNA content under the experimental conditions. Fluorescence intensities of both wavelengths were simultaneously measured in a flow cytometer (FC200, Ortho) for 5-10 x 103 cells. Evaluation of the data stored in list-mode on disks, namely definitions and calculations of various portions of cells in the GI-, S- and G2Mphases of the cell cycle, was performed as previously published [11, 12]. The DNA index of G1 cell populations of each individual sample was calculated by dividing the mean channel number of F530 by the mean channel of F530 of the control lymphocytes admixed to each individual sample. Measuring mixtures at high photomultiplier tube voltages enabled us to detect DNA deviations from the diploid DNA content of control lymphocyte populations of 5%, or more. The F600 index of the G1 cell populations of each sample and normal control BM aspirate specimen was calculated by dividing its mean channel of F600 by the mean channel of F600 of the control lymphocytes measured under identical conditions [16].

Cytogenetic methods For cytogenetic studies 12 BM aspirates were cultured for 24 h in RPMI 1640 medium with fetal bovine serum and antibiotics. Chromosome preparations were made following conventional methods using 0.075 M potassium chloride as the hypotonic solution and 3 : 1 methanol : acetic acid as a fixative. Karyotypic analysis was performed on Q-banded preparations. Two of the samples yielded no analyzable metaphases. Two patients had no cytogenetic studies done.

Statistical methods There were 8 variables potentially useful for predicting survival. These included abnormal DI, aneuploidy, pseudodiploidy, translocations t(8;14), translocations other than t(8; 14), all translocations, null-cell and B-cell phenotypes. The outcome of each variable was dichotomized as presence or absence of the specific parameter. Since this study included only 14 patients, multivariate analysis on all of those variables was not feasible. Instead, each variable was studied univariately for its prognostic importance, i.e. whether the presence or absence of the variable had an impact on survival. For each type of variable, it was empirically determined that the survival times of the patients followed an exponential distribution. This is a simple parametric model which says that the proportion of patients surviving until time t, s(t), drops exponentially with t, or, s(t)= exp(-t/p.), where I~ is the unknown population mean survival time. The parametric assumption was verified in each case by plotting In g(t) against t, where g(t) is the observed s(t) in the patient sample. Plotting In g(t) against t results in a straight line. In the present very small sample study it is crucial that such a simple sensitive parametric model be used instead of the usual non-parametric procedure to study survival. For each of the variables, a twosided F-test at the exact significance level of 0.05 was performed to determine whether there was a significant difference between the two categories. RESULTS

Cellular D N A and D N A content Cellular D N A and R N A content are listed for each individual specimen in Table 2. Nine out of 14 patients had cell populations with distinctly abnormal DI. O f these, two patients had more than one abnormal DI. Cells with abnormal DI and L3 blasts comprised an average of 7 4 - 17.8% (1 S.D.) and 73--. 18.9%, respectively, of the cell populations in PB and BM samples of nine patients (correlation coefficient r = 0.59). A m o n g the L3 populations with abnormal DI there was no correlation between the degree of abnormality and the percentage of proliferating cells (r = 0.49). O n e patient with a cell suspension of an involved LN had no cells with abnormal DI but had 25% SG2M cells and 80% cells with high F600 index. The F600 index of all L3 cell populations was significantly higher than that of normal BM cell suspensions (2.3 vs 1.2, respectively; one-sided Student's t-test: p < 0.001).

Cell surface marker studies Cell SM analysis revealed (Table 3) a predominance of cells with Null-cell phenotype (no rosette formation with sheep erythrocytes, no surface immunoglobulins) in 5 out of 14 patients. Nine of the 14 patients had L3 cell populations with B-cell phenotype. Five of the 9 had monoclonal surface IgM, three were IgM negative, and two out of three examined for IgG did have monoclonal surface IgG. Monoclonality was also observed for light chain expression in all patients with B-cell phenotype: six had lambda and three had kappa light chains. Flow cytometric analysis of SM profiles detected clonal excess even if only small numbers of cells were positive c o m p a r e d to the reference m e a s u r e m e n t (cases

Heterogeneity of L3 leukemia

eL

,v

+

I

I

I

+1

+

I

+

+

75

I

I + +

.¢ t~

.#-

<

z

r~

•

e~

"~ c ~

z~z

z" ~'=~ ~

'.~"-"Z

"~

_

0

(1~1-~

~

_

~¢~-c~ ~,,I

---

.~ .~-

~

~ "~"

~

. - ~

~

t,~ ~

•

"~

-

¢.~

< t-, < r~ <

P_. z ,-I

¢o )1~

~,J .~ ~a

e~ <

.,-, 0-----"

~

~

~

~

~

~1

~

~X

o:1

.o

o.o_ ~ . ~ ~

A. J. WALLE et

76

al.

TABLE 2. A C R I D I N E ORANGE FLOW CYTOMETRY DATA, AND L3 CELLS (BLASTSJ IN SMEARS OF BONE MARROW ASPIRATES STAINED WITH T E T R A C H R O M E COMPARE PERCENrI'AGES OF S G 2M PHASE CELLS AND DNA AND F 6 0 0 INDICES OF PATIENTS' SAMPLES WITH NORMAL CONTROLS

Sample

Marrow blasts (%)

Cells with abnormal DI (ck)

SG2M cells (%)

DNA index

F600 index

1

BM

95

2

BM

52

3 4 5 6 7 8 9 10

BM BM BM BM PB BM BM LN BM BM BM BM PB BM

90 73 90 92 50 82 90 -91 90 61 55 0 0

76 18 62 10 86 65 97 94 84 0 0 0 (1 0 21 55 (I 0

4 11 7 9 21 34 26 25 35 25 52 25 30 22 24 35 0 13

1.56 2.15 1.15 1.40 2.24 1.11 1.13 1.13 1.09 1.011 1.011 1.00 I.(X) 1.00 1.25 11.95 1.00 1.00

3.3 3.7 1.8 2.1 4.9 1.8 2.2 1.6 2.2 1.8 1.9 1.7 1.9 1.8 1.7 1.9 1.0 1.3

Patient

11

12 13 14 Control* Control*

(4)

(0.2)

* Mean values (--+1 S.D.) of 12 control specimens each.

12-14, T a b l e 3). C o n c u r r e n t SM s t u d i e s in L N cell s u s p e n s i o n s in s o m e p a t i e n t s largely c o n f i r m e d the p h e n o t y p i c s u r f a c e c h a r a c t e r i s t i c s o f cell p o p u l a t i o n s taken from BM of the same patient.

t h r e e o u t o f five Null-cell cases. T d t m e a s u r e d in m o s t of the B-cell c a s e s ( T a b l e 3) e x h i b i t e d activities well b e l o w n o r m a l levels, o r n o n - m e a s u r a b l e activities.

Terminal deoxynucleotidyl transferase activity

In m o s t c a s e s , c y t o g e n e t i c s t u d i e s p r o v i d e d e v i d e n c e o f the e x i s t e n c e o f t r a n s l o c a t i o n s in v a r i o u s p e r c e n t a g e s

B i o c h e m i c a l activity o f T d t s h o w e d e l e v a t e d levels in

Cytogenetic studies

TABLE 3. CELl. SURFACE MARKER AND Tdt DATA

Patient

Phenotype

1 2 3 4 5 6

Null Null Null Null Null B,IgG,Kappa

7 8

B,IgM,Lambda B,IgM,Lambda

9 10

B,lgM,Lambda B,IgM,Lambda

11 12 13 14

B,IgM,Kappa B,--,Kappa* B,--,Lambda* B,IgG,Lambda*

Sample

Tdt (<0.1(IU/ 10s cells)

BM BM BM BM BM BM LN PB BM LN BM BM LN BM BM BM BM LN

19.60 3.15 QNS >3.00 NMA ---NMA -NMA 0.01 -NMA -0.03 0.01 --

SRBCR (%) 1.0 3.0 3.(I 16.5 4.5 6.0 10.(1 3.(I 22.0 160 4.0 13.(I 12.0 3.0 -15.5 6.5 14.0

EACR (%)

HLA-DR (v/b)

1.5 1.0 0.(} -1.5 9.0 9.0 3.5 0.0 26.5 -. 4.(I 6.0 -0.0 4.0 0.0

56.0 32.0 18.1 15.0 56.0 ---77.1/ -94.0 . --9.7 20.3 9.8 --

.

Polyval. lg (%)

lgM (%)

IgG (%)

Kappa (%)

Lambda (%)

0.0 4.5 6.4 2.(I 39.0 ---54.(1 42.0 39.7

0.0 3.0 8.2 1.5 0.0 1.0 1.(I 85.0 63.0 48.0 49.5

-----80.0 -0.0 ----

88.0 90.0 -1.7 0.4 3.0

----6.2 86.(I

-0.0 9.6 -2.5 79.2 79.2 0.0 6.0 2.0 3.3 2.6 0.0 88.11 1.8 3.9 4.6 0.0

-2.(1 9.5 -2.0 2.0 2.0 88.0 ,12.0 ,12.0 29.0 28.5 84.0 3.0 11.2 2.2 1.3 87.0

.

. 82.0 87.0 -5.9 5.8 89.(I

* B-cell monoclonality of surface immunoglobulin light chains determined by clonal excess calculation. ( - - ) , no data available. QNS, quantity of cells not sufficient to perform test. NMA, not measureable amount of activity. SRBCR. sheep red blood cell rosettes; E A C R , erythrocyte amboceptor complement rosettes.

jm

N"

.q

111

FIG. 1. P h o t o m i c r o g r a p h s ( t e t r a c h r o m e , 6 5 0 x ) of b o n e mar. row a s p i r a t e s of p a t i e n t s A = 3, B = 4, C = 13, D = 6, E = 1 F = 7 , G = 5 , H = 2 , I = 14 of the Tables. 77

Heterogeneity of L3 leukemia

79

+

+

+

E ~

+

~2 +

++

~

+

~.~ +

+

+~

_~

I +

~

T~

T _~

T~-

T +

+ + un"

~m ¢M¢..

~

¥

+7

T~

+

u,~'+

-.~

~d

++

1'+

+

o

+

< ,-]

+

+++~_~

+

t~ ?=~4+

~

~

1

+t ++ I.,-.

~++

,<

I ~-~-.

× 7 ¥ ×t"- -~~

,,,,.

,<

[.-

z q~

z

zz

~

~

~D

z~

0

< z

A. J. WALLEet al.

80

of the metaphases (Table 4) of the chromosome preparations, besides various numerical chromosome aberrations. All numerical aberrations were confirmed by FCM-determined abnormal DI (Table 5) except patient 13 who had 21% cells with abnormal DI (Table 2) but a pseudodiploid karyotype in 89% of evaluable metaphases. Three patients had t(8;14), three patients had t(8;22) and 2 patients had t(14;18). Patients 1 and 3 had no translocations detected in the analyzable metaphases.

Prognostic factors According to our simple parametric model (see Methods) the survival data of all patients as well as that of patients included in groups of certain variables (see Methods) fit exponential curves in an excellent manner. It was determined that the presence of one of these variables, namely abnormal DI and/or pseudodiploid karyotype had an impact on survival at significance level 0.05 (two-sided F-test). Three patients with t(8;14) survived for 371--474 days whereas five patients with other translocations survived for only 32-180 days after diagnosis of L3 disease was made. DISCUSSION All 14 patients had the BM aspirate diagnosis of L3 leukemia made when they were admitted to Memorial Hospital. Some of them had also blood involvement (Table 1). L3 cell populations were also found in lymphatic tissues (Table 1). Two patients had previously been treated for nodular poorly differentiated lymphoma, and they presented now with conversion into L3 disease. From the time of diagnosis till death the clinical course of all patients had a median length of 115 days which is similar to previous reports on L3 leukemia regardless of the mode of treatment [4, 1315]. No treatment protocol-related differences in survival were noted; excellent correlation of survival times

TABLE

Patient 1

2 5 3 4 13

14 9 10 8 11 12 6 7

5.

with an exponential curve was established. All data in this study pertain to the time of first diagnosis of L3 disease in each patient.

Cellular RNA, D N A and proliferation fractions In correspondence with morphologic criteria of L3 disease [11, all L3 cell populations exhibited total cellular RNA content which was uniformly higher than that of normal cell populations (Table 2). However, each L3 cell population exhibited a high degree of heterogeneity in terms of RNA content per cell [16]. Since L1 and L2 type acute lymphoblastic leukemia have significantly lower mean cellular R N A content [17] than L3 cell populations, A O FCM is apt to identify a type of lymphoblastic leukemia which carries a particularly unfavorable prognosis compared to L1 and L2 types [14, 18.19]. Despite the uniformly high mean R N A content of L3 cell populations, they do display wide heterogeneity of cellular DNA content. Nine out of 14 patients had abnormal DI (ranging from 0.95 to 2.24), with two patients having more than one distinct abnormal DI. All of the DI abnormalities were confirmed by aneuploid karyotypes (Tables 2, 4, 5) except case 13. The abnormal clone in case 13 may have escaped cytogenetic detection, while the 25% increase in cellular D N A of G0/1 cells was easily detected by A O FCM. Based on our parametric model, statistical evaluation suggested that patients with abnormal DI have significantly shorter survival times (Table 6). In childhood acute lymphoblastic leukemia [20], as well as in lymphoma [21,22] abnormal DI was found to herald grave prognosis. Yet, the degree of aneuploidy and abnormality of DI were not conclusively related with the proliferation fractions of the abnormal cell populations in our study. Twelve out of 14 patients had L3 populations with high SG2M cell fractions. Commonly, high proliferation has been reported a characteristic feature of Burkitt's lymphoma cell populations

SYNOPSIS OF FEATURES CHARACTERIZING HETEROGENEITY OF BURKITr'S LYMPHOMA/L3 CELL POPULATIONS

Morphology

F600 index

SG2M (%)

Translocation

Tdt

Cell surface phenotype

Modal No. of chromosomes

DNA index

1.3 1.3 I_3 1_3 1_3 1_3 1_3 L3 L3 1.3 L3 1.3 L3 L3

High High High High High High High High High High High tligh High High

4/11" 7/9* 26 21 34 24 35 52 25 25 30 22 25 35

No trans. NAM NAM No trans. T(8;22), 14q+.14q+ T(14; 18) T(14-18) T(8;22) T(8;22) T(8:14) T(8:14) T(8:14) ND ND

High High Low ND High l.ow Low Low Low Low Low Low Low ND

Null Null Null Null Null B. Lambda B, Lambda B. Lambda B, Lambda B. Lambda B. Kappa B, Kappa B, Kappa B. Lambda

Hyperdiploid NAM NAM Hyperdiploid Hyperdiploid Pseudodiploid Hypodiploid Pseudodiploid Pseudodiploid Pseudodiploid Pseudodiploid Pseudodiploid ND ND

Increased Increased Increased Increased Increased Increased Decreased Normal Normal Normal Normal Normal Increased Increased

ND, not done. * Indicates the separately evaluated SG,M % of the two aneuploid cell populations present in these samples. NAM, no analysable metaphases.

Heterogeneity of L3 leukemia [23, 26]. In leukemia [27] and lymphoma [21-24, 28, 29] high SG2M cell fractions were high-risk factors. In this study, high SG2M cell fractions were more frequently associated with chromosomal translocations than with aneuploidy (Table 5) in that all patients who had one of the described translocations (Table 4) had high proliferation fractions. One of the two patients with no apparent translocation had a remarkably low SG2M cell fraction (Table 5, case 1). Cytogenetic findings in L3 leukemia Few cytogenetic studies have been reported in this disease [13, 15]. A q14+ was, expectedly, the most common finding in these predominantly B-cell acute lymphoblastic leukemias. Chromosome 8 was the most frequently identified donor chromosome in a small series (for review see [30]). Six of those 7 cases were pseudodiploid and one was hypodiploid. Our study includes one hypodiploid case (Table 4). Other frequent karyotypic changes involved trisomies of chromosomes 1 and 7, and changes in chromosome 6 [30, 31]. In another study, 94% of L3 leukemias (n = 21) had the t(8;14) structural abnormality [15]. However, only rarely have attempts been made to correlate these various cytogenetic findings with other cellular features [13] or with clinical response to treatment as in cases of L1 and L2 leukemia [32] or non-lymphoblastic leukemia [33]. A recent report on 330 newly diagnosed patients with the lymphoblastic variety demonstrated the independent clinical significance of chromosomal abnormalities even when age, initial leukocyte count and FAB-type were accounted for [15]. In our small series, the presence of pseudodiploid karyotype predicted for increased survival times (Table 6). Furthermore, the presence of t(8;14) was associated with increased survival times, although statistical validation of this observation was precluded by the small number of patients. Caution must be exercized, however, in considering these differences because of the very small number of cases. The t(14;18) abnormality has not been reported to occur in L3 disease. This finding is linked to the fact that our two patients with this transiocation had converted into L3 disease from previous nodular poorly differentiated lymphoma (see Methods) in which disease t(14;18) is the most common translocation. Despite the presence of this t(14;18) in the L3 populations of these two patients, other biological and clinical features were strikingly similar to the five "classic" L3 cases (Fig. 1, Tables 1, 5).

81

Cell surface markers and Tdt The present evaluation of SM immunofluorescence by FCM is superior to previously performed fluorescence microscopy, in particular since clonal excess can easily be determined [7]. Nine out of 14 cases of L3 disease had B-cell phenotype, six of which expressed monoclonal lambda and three kappa light chains. In the remaining five cases of Null-cell phenotype the possibility has to be entertained that they may have been of pre-B cell type [34, 35], in particular the three Null-cell cases with strong Tdt positivity. The only case of T-cell phenotype of L3 blast cells has been described earlier by this laboratory [4]. Phenotypic classifications by SM and Tdt of lymphoblastic leukemia and lymphoma convey prognostic information [10, 35--40]. The present data does not reveal cell surface phenotype to have an impact on survival rates (Tables 1, 6). The present findmgs (cases 9, 10) confirm the previously reported relationship between expression of immunoglobulin light chains and specific type of translocation [41]. Prognostic factors in L3 leukemia The extraordinarily high levels of nuclear RNA in L3 ceils [16, 42] combined with the high cytoplasmic ribosomal RNA content and the high proliferative activity of L3 cells reflect a high degree of both transcriptional as well as translational activity characteristically associated with aggressive malignant growth, distinguishing L3 measurably from L1 and L2 type leukemia with their superior survival rates. The strictly exponential distribution of survival times of our L3 patients is a new observation. Despite the inherent statistical power of the exponential model, the tentative characterization of these biological parameters as prognostic factors, namely the presence of abnormal DI, t(8;14) and pseudodiploid karyotype (Table 6), must be interpreted with caution owing to the small number of patients in each group of variables.

REFERENCES 1. Bennett J. M., Catovsky D., Daniel M. T., Flandrin G., Galton D. A. G., Gralnick H. R. & Sultan C. (1976) Proposals for the classification of the acute leukemias. Br. J. Haematol. 33, 451. 2. Burkitt D. (1958) A lymphoma involving the jaws in African children. Br. J, Surg. 46, 218. 3. Magrath I. T. & Ziegler J. L. (1979) Bone marrow involve-

TABLE 6. IMPACT OF PROGNOSTIC VARIABLES ON CALCULATED MEDIAN SURVIVAL TIMES BASED ON PARAMETRIC EXPONENTIAL MODEL (DETAILS SEE METHODS)

Variable Abnormal DI Pseudodiploidy

Presence (+) absence ( - )

Sample size

Median survival (days)

S.D.

+ + -

9 5 6 4

65.2 203.5 177.4 61.9

21.7 91.0 72.4 30.9

82

A . J . WALI.E et al.

ment in Burkitt's lymphomas and its relationship to acute B-cell leukemia. Leukemia Res. 4, 33. 4. Koziner B., Mertelsmann R., Andrecff M., Arlin Z.. Hansen H.. DeHarven E., McKenzie S.. Gee T., Good R. A. & Clarkson B. D. (1980) Heterogeneity of cell lineages in L3 leukemias. Blood 55, 694. 5. Boyum A. (1977) Separation of lymphocytes, lymphocyte subgroups and monocytes: a review. Lymphology 10, 71. 6. Koziner B., Filippa D.. Mertelsmann R., Gupta S., Clarkson B. D., Good R. A. & Siegal F. B. (1977) Characterization of malignant lymphomas in leukemic phase by multiple differentiation markers of mononuclear cells: correlation with clinical features and conventional morphology. Am. J. Med. 63, 556. 7. Ligler F. S., Smith R. G., Ketterman J. R., Hernandez J. A., Hymes J. B., Vitetta E. S., Uhr J. W. & Frenkel E. P. (1980) Detection of tumor cells in the peripheral blood of nonleukemic patients with B-cell lymphoma: analysis of "clonal excess". Blood 55, 792. 8. Borella L. & Sen L. (1975) E receptors on blasts from untreated acute lymphocytic leukemia (ALL). Comparison of temperature dependence of E rosettes formed by normal and leukemic lymphoid cells. J. Immunol. 114, 187. 9. Shevaeh E. M., Herberman R., Frank M. D. & Green I. (1972) Receptors for complement and immunoglobulin on human leukemic cells and human lymphoblastoid cell lines. J. Clin. Invest. 51, 1933. 10. Mertelsmann R., Koziner B., Filippa D. A., Grossbard E., leery G.. Moore M. A. S. & Clarkson B. D. (1979) Clinical significance of Tdt, cell surface markers and CFUC in 279 patients with hematologic neoplasia. In Modern Trends in Human Leukemia III (Neth R., Gallo R. C.. Hofschneider P. H. & Mannwieler K . Eds) p. 131. Springer, New York. 11. Darzynkiewicz Z., Traganos F., Sharpless T. & Melamed M. R. (1976) Lymphocyte stimulation: a rapid multiparameter analysis. Proc. natn. Acad Sci. U.S.A. 73, 2881. 12. Darzynkiewicz Z.. Traganos F. & Melamed M. R (1980) New cell cycle compartments identified by multiparameter flow cytometry. Cytometry !, 98. 13. Berger R. & Bernheim A. (1982) Cytogenetic studies on Burkitt's lymphoma--leukemia. Cancer Genet. Cytogenet. 7, 231. 14. Miller D. R., Leikin S., Albo V.. Sather H. & Hammond D. (1981) Prognostic importance of morphology (FAB classification) in childhood acute lymphoblastic leukemia (ALL), Br. J. Haematol. 48, 199. 15. Third International Workshop on Chromosomes in Leukemia: chromosomal abnormalities and their clinical significance in acute lymphoblastic leukemia (1983). Cancer Res. 43, 868. 16. Walle A. J. (1986) Identification of L3 leukemia and Burkitt's lymphoma cells by flow cytometric quantitation of nuclear DNA and RNA content. Leukemia Res. 10, 303. 17. Walle A. J. (1985) Distribution and content of nuclear and cellular RNA among cell population of acute lymphoblastic and non-lymphoblastic leukemia. Cancer Res. 45, 5193. 18. Brunning R. D., McKenna R. W.. Bloomfield C. D., Coceia P. & GajI-Peczalska K. J. (1977) Bone marrow involvement in Burkitt's lymphoma. Cancer 40, 1771.

19. Palmer M. K., Harm I. M., Jones P. M. & Evans D. I. K. (1980) A score at diagnosis for predicting length of remission in childhood acute lymphoblastic leukemia. Br. J. Cancer 42, 841. 2(I. Suarez C.. Miller D. R., Steinherz P. & Andreeff M. (1981) Flow cytometry DNA and RNA determination in 107 cases of childhood acute lymphoblastic leukemia: correlation with FAB classilication. Proc. Am. Ass. Cancer Re~. 22 (Abstr. 770a). 21. Diamond L. W. & Braylan R. C. (1980) Flow analysis of DNA content and cell size in non-Hodgkin's lymphoma. Cancer Res. 40, 703. 22. Diamond L. W., Nathwani B. N. & Rappaport H. (1982) Flow cytometry in the diagnosis and classification of malignant [ymphoma and leukemia. Cancer 50, 1122. 23. Braylan R. C., Diamond L. W., Powell M. L. & HartyGolder B. (1980) Percentage of cells in the S-phase of the cell cycle in human lymphoma determined by flow cytometry: correlation with labeling index and patient survival. Cytometry 1, 171. 24. Braylan R. C., Fowlkes B. J., Jaffe E. S.. Sanders S. K., Berard C. & Herman C. (1978) Cell volumes and DNA distributions of normal and neoplastic human lymphoid cells. Cancer 41, 2(11. 25. Cooper E. H., Frank G. L. & Wright D. H. (1966) Cell proliferation in Burkitt's tumors. Eur. J. Cancer 2, 377. 26. lversen O., Iversen U., Ziegler J. L. & Bluming A. Z. (1974) Cell kinetics in Burkitt's lymphomas. Eur. J. Cancer I0, 155. 27. Dosik G. M., Barlogie B.. Smith T., Gehan E. A., Keating M. J., McCredie R. B. & Freireich E. J. (1980) Pretreatment flow cytometry of DNA content in adult acute leukemia. Blood 55, 734. 28. Barlogie B.. Johnston D. A., Smallwood L., Raber M. N., Maddox A. M., Latreille J., Swartzendruber D. E. & Drewinko B. (1982) Prognostic implications of ploidy and proliferative activity in human solid tumors. Cancer Genet. C)'togenet. 6, 17. 29. Shackney S. E., Skramstad K. S., Cunningham R. E., Dugas D. J., Lincoln T. L. & Lukes R. J. (1980) Dual parameter flow cytometry studies in human lymphomas. J. ('lin. Invest. 66, 1281. 30. Sandberg A. A., (1981) Chromosome changes in the lymphomas. Hum. Pathol. 12, 531. 31. Sandberg A. A. (1980) The Chromosomes in Human Cancer and Leukemia, p. 377. Elsevier North-Holland, New York. 32. Williams D. L., Tsiatis A., Brodeur G. M., Look A. T., Bowman W. P., Kalvinsky D. K., Rivera G. & Dahl G. V. (1982) Prognostic importance of chromosome number in 136 untreated children with acute lymphoblastic leukemia. Blood 60, 864. 33. Golomb H. M., Vardiman J. W., Rowley J. D., Testa J. R. & Mintz U. (1978) Correlation of clinical findings with quinacrine-banded chromosomes in 90 adults with acute non-lymphocytic leukemia. New Engl. J. Med. 229, 613. 34. Cossman J.. Chusid T. M., Fisher R. I., Magrath I.. Bollum F. & Jaffc E. S. (1983) Diversity of immunological phcnotypes of lymphoblastic lymphoma. Cancer Res. 43, 4486. 35. Vogler L. B., Crist W. M., Sarrif A. M., Pullen D. J.,

Heterogeneity of L3 leukemia Bartolucci A. A., Ialetta F. M., Dowell B., Humphrey G. B., Blackstock R., Eys F. V., Metzgar R. S. & Cooper M. D. (1981) An analysis of clinical and laboratory features of acute lymphocytic leukemias with emphasis on 35 children with pre-B leukemia. Blood 58, 135. 36. Bloomfield C. D., Kersey J. H., Brunning R. D. & GajlPeczalska K. J. (1977) Prognostic significance of lymphocytic surface markers in adult non-Hodgkin's lymphoma. Cancer Treat. Rep. 61,963. 37. Greaves M. F., Janossy G., Peto T. & Kay H. (1981) Immunologically defined subclasses of acute lymphoblastic leukemia in children: Their relationship to presentation features and prognosis. Br. J. Haematol. 48, 179. 38. Marks S. M.. Baltimore D. & McCaffrey R. (1978) Terminal transferasc as a predictor of initial responsiveness to vincristine and prednisone in blastic chronic myelogenous leukemia. New Engl. J. Med. 298, 812.

83

39. Thiel E., Rodt H., Huhn D., Netzel B., Grosse-Wilde H., Ganeshaguru K. & Thierfelder S. (1980) Multimarker classification of acute lymphoblastic leukemia: evidence for further subgroups and evaluation of their clinical significance. Blood 56, 759. 40. Tsukimoto I., Wong K. Y. & Lampkin B. C. (1976) Surface markers and prognostic factors in acute lymphoblastic leukemia. New Engl. J. Med. 294, 245. 41. Lenoir G. M., Preud'Homme J. L., Bernheim A. & Berger R. (1982) Correlation between immunoglobulin light chain expression and variant translocation in Burkitt's lymphoma. Nature, Lond. 298, 474. 42. Walle A. J., Darzynkiewicz Z., McKenzie S. & Higgins P. (1983) Characterization by nuclear and cellular RNA content of blood and marrow cells of patients with Burkitttype leukemia (FAB-L3). Proc. Am. Ass. Cancer Res. 24 (abstr. 32).