Long-term survival in acute myelogenous leukemia

LEAD ARTICLE Long-term Survival in Acute Myelogenous Leukemia: A Second Follow-up of the Fourth International Workshop on Chromosomes in Leukemia G. J. Swansbury, S. D. Lawler, G. Alimena, D. Arthur, R. Berger, H. Van Den Berghe, C. D. Bloomfield, A. de la Chappelle, G. Dewald, O. M. Garson, A. Hagemeijer, F. Mitelman, J. D. Rowley, and M. Sakurai

ABSTRACT: Patients with acute myeloid leukemia (AML, equivalent to acute non-lymphoblastic leukemia [ANLL]) who were studied at the Fourth and Sixth International Workshops on Chromosomes in Leukemia and who have long survival have been re-assessed to identify factors which m a y be associated with good prognosis in AML. In a long-term survivor (LTS) group, there were more cases than expected in each age decade below 50, more cases than expected with FAB type M3, and fewer cases than expected of secondary leukemia. Of the distribution of chromosome abnormalities, t(15;17), t(8;21), and inv/del(16) were over-represented, and -5, -7, and rearrangements of 11q were under-represented. Multivariate analysis of all patients showed that age group, cytogenstic classification, FAB type, and sex all had independent, significant effects on survival. A new observation from o very small subgroup of patients was that deletion of 7q without concurrent abnormality of chromosome 5 appeared to be associated with a good prognosis.

INTRODUCTION

The Fourth International Workshop on Chromosomes in Leukemia, held in Chicago in September 1982 [1], involved a prospective, multi-center study of 716 patients with AML diagnosed between January 1, 1980 and March 31, 1982. A follow-up study was undertaken at the Sixth International

Academic Department of Haematology &Cytogenetics (G. ]. S.), Royal Marsden Hospital, Sutton; Institute of Cancer Research (S. D. L.), London, U.K.;Sezione di Eraatologia {G.A J, Dipartimento di Biopatalogia Urnana, Universita di Rorna, "La Sapienza~ Rome, Italy; Department of Laboratory Medicine & Pathology (D. A.), University o/M/rmesota, Minneapolis, Minnesota; Laboratoire de Cytog6n~tique (R. B.), lnstitut de Recherches sur les Maladies du Sang, Par/s, France; Centre for Human Genetics (I4. V. D. B.), Universitaire Ziekenhuisen, Leaven, Belgium; Department of MedicIne {C. D. B.), Reswell Park Cencer institute, Buffalo, New York;Department of Medical Genetics fA. de la C.), University of Helsink/, Helsink/, I~land; Cytogenetics Laboratory (G. DJ, Mayo Clinic, Rochester, Minnesota; Department of Cytogenetics (0. M. GJ, St. V'mcents Hospital, Victoria,Australia; Department of Cell Biology &Genetics (A. H.), Erasmus University, Rotterdam, Nether/ands; Department of Cl/nical Genetics {E M.I, UniversityHospitol, Land, Sweden; Division of Bialogical Sciences U.D. BJ, Universityof Chicago, Chicago, Illinois; and 8aitarna Cancer Center (M. 8.), Ina, Saltama, Japan. Address reprint request to: G. ]. Swansbury, ~ademic Department of Haernatalogy & Cytogenetics, Royal Marsden Hospital, F Block, Sutton, Surrey SM2 5NG, U.K. Received June 7, 1993; accepted August 31, 1993.

Workshop on Chromosomes in Leukemia, held in London in May 1987 [2], with follow-up being extended to December 31, 1986. The purpose of the present study was to extend the follow-up, which is of special interest in view of the large number of patients, their multi-center origin, and the amount of cytogenetic data available. At the time of the Fourth Workshop, 462 patients had already died and 254 ware alive at last follow-up. At the Sixth Workshop, one of the 716 Fourth-Workshop patients was excluded because of inadequate cytogenetic data. Of the 715 cases included, 598 patients were known to have died, including six whose deaths were nat directly due to their leukemia. Twenty-three were alive after bone-marrow transplant (BMT, two autografls, 17 allogrdts, and four unspecified) and for analyses of survival were censored on the day of their BMT. However, of those still alive without BMT, some had very short follow-up: 39 patients at less than 1 year. This was due to some patients being lost to further foUow-up, and some contributors to the Fourth Workshop being unable to participate in the Sixth Workshop. Fifty-one patients ware identified who were alive with at least four years' follow-up. As described in the Sixth Workshop report [2], these patients had a younger median age than the rest of the group studied, with a smaller proportion having secondary AML and a greater proportion having t(15;17). During 1991, participants in the Fourth and Sixth Workshops ware requested to undertake a second follow-up of longsurviving patients. 1

© 1994 Elsevier Science Inc. 655 Avenue of the ~mericas, New York, NY 10010

Cancer Genet Cytogenet 73:1-7 (1994)

0165-4608/94/$07.00

Table 1 ID

Clinical a n d c y t o g e n e t i c details of t h e l o n g - t e r m s u r v i v o r g r o u p

Age/Sex

FAB

1/2°

BMT

D/A

Days

R

N

24 33 41 45 57 96 105 121 123 126 145 160 164 182 200 203 205 217 219 221 229 259 265 278 280 286 288 293 332 337

42/M 30/M 14/F 5/F 63/F 55/M 53/M 1/M 32/M 50/F 49/M 63/F 52/F 45/F 43/M 38/M 30/F 31/F 26/M 11/F 29/M 9/M 14/F 55/F 14/M 57/F 46/F 67/M 36/F 52/F

4 1 2 2 2 1 4 4 2 4 2 1 1 2 2 3 3 3 2 2 6 2 2 5 3 2 1 1 4 2

1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

-

A D A D D A A A D A A D A D A A D A A A A A D A A D A A D D

3588 3190 2030 1654 1728 3339 4236 3983

< < <

2 29 2 16 4 17 1 15 30 0 20 0 13 4 1 3 3 12 8 9 2 NE 7 0 5 0 11 0 0 29 O 26 0 25 6 6 21 0 11 0 21 O 8 40 2 31 7 23 2 32 45 0 2 46 34 0 14 0 0 13

347 353 365 375 383 427

27/M 51/M 48/F 66/F 40/F 18/F

4 2 1 4 2 4

1 1 1 1 1 1

A A A A A A

3764 3675 3471 3467

430 466 467 484 488 491 494 500

31/F 11/F 15/F 2/F 27/M

4 1 2 4 4

1 1 1 1 1

A A A A A

26/F

5

1

-

D

48/F 4/M

2 4

1 1

-

A D

3180 2316 2337 2006 1909 1508 3907 3125

504 512 519 540 544 545

19/F 42/F 1/F 23/F 39/F 65/M

2 3

1 1

L -

A A

1761 2016

5

1

-

A

3209

3 1 4

1 1 1

-

A A D

3063 3912 1512

555

9/M

1

1

-

A

3627

<

562

I/M

4

1

Y

A

3172

578

32/F

3

1

-

A

3260

583 585

72/F

4

1

-

D

1844

31/M

4

1

-

D

1591

0

4

589

35/M

3

1

-

A

3589

<

5

0

594

58/M

1 1

-

3949

<

7

O

57/M

5 5

A

605

D

2382

<

0

10

L

-

Y L L L L U

L

Y

L L -

< < <

1772 3752 2555 1715 3905 1531 3819 3795 2204 3653

3773 3471 3342 3522 2051 3661 3676 2799 3824 3781 2090 2208

3377 3221

< < < < < < < <

< < < < < < <

C1

< < < < < <

13 0 15 2 10 0

0 20 0 13 O 15

<

41 35 17 37 19 13 15 16

0 0 0 0 0

14 5

0 29

<

3

56

< <

9 0 11

<

2 10 8 4 5

<

0

6

1

4

< <

0

0 0

Karyotype 46,XY,der(2]t(2;1I)(p25;q23),inv(16)(p13q22) 46,XY,del(7)(q34) [14]/47,idem, + del(7)(q34) [2] 46,XX,t(8;21)(q22;q22) [ 1 5 ] / 4 5 , i d e m , - X [2] 46,XX,t(4;lO)(q31;q22),t(6;9)(p23;q34) 46,XX [Reh 46,XX [28]) 46,XY 46,XY,del(16)(q22) 47,XY, + 8,i(10q),t(11;19)(q23;p13) 46,XY,del(5)(q13q33) 46,XX,inv(16)(p13q22) 46,XY,t(8;21)[q22;q22)/45,idem,-Y 46,XX (Rel: 46,XX [15]) 46,XX 46,XX (Reh 46,XX,del(5)(q15q33) [5]/46,XX [7]) 46,XY,del(7)(q22) [ReI: 47,idem, +8)[12] 46,XY,t(15;17)(q22;q21)

46,XX,der(15)t(15;17),ider[17)t(15;17) 47,XX,+8 46,XY 46,XX 46,XY 45,X, -Y,ins(21;8)(q22.2;q22q21) 46,XX,t(8;21)(q22;q22) [22] 4 5 , i d e m , - X [9] 47,XX, +8 46,XY,t(15;17)(q22;q21) 46,XX 47,XX,+21 46,XY 46,XX (Reh 46,XX [2]) 45,X, -X,t(8;21)(q22;q22) [8]/45,idem,del(9) (qllq21) [5] 46,XY 46,XY,t(8;21)(q22;q22) 46,XX 47,XX, + 22 46,XX 49,XX,t(3;12)(q25;q24),del(7)(q22), + 8, + 17, + 19 46,XX 46,XX 46,XX 46,XX,t(1;12)(q33;q13)c 46,XY 46,XX 46,XX 46,XY (Reh 46,XY,del(2)(q32),inv(16)(p13q22)

[14])

0 0

46,XX 46,XX,t[15;17)(q22;q21) [26]/45,1dem,- 9 [3] 46,XX,der(9)t(1;9) (qll;q34),t(10; 11)(p14;q14) 46,XX,t(15;17)(q22;q21) 46,XX 47,XY, + 22 46,XY 46,XY 46,XX,t(15;17)[q22;q21) 47,XX, + 8 45,X, - Y,t(8;21)(q22;q22) 46,XY 46,XY 45,XY, - 7 (continued)

Cytogenetics and Long-Term Survival in A M L

3

Table 1 Continued ID

Age/Sex FAB 1/2 ° .BMT D/A Days R N

615 617 623 624 625 629 686 718

42/F 25/F 14/M 42/F 27fM 4 6/M 44/F 48/F

3 3 2 2 2 4 4 2

1 1 1 1 1 1 2 2

L L -

A A D A D A D A

-

3447 3396 2686 3885 1581 3523 1866 3530

< < <

< < <

CI

12 0 11 14 7 0 3 12

O 10 O 3 0 25 23 O

Karyotype 46,XX 46,XX,t(15;17][q22;q21) 46,XY 4 6 , X X , - 8 , - 13, + 2 m a r 46,XY (Rel: 45,X ,-Y ,t (8; 21}{q22; q22} [6]} 47,XY,t[4;18){q273;q271], + 8,inv(16}{p13q22) 48,XX, + 6,del(6)(q16), + 9, - 18, + m a r 46,XX

D/A: D = dead; A ffi alive. R: < - Extended follow-up received for this paper. BIVIT: L = allogrefl; U ffi autograR; Y ffi yes, unspecified type; -

= No.

N, normal cells. C1, clonal cells. Rel, relapse study data, given in pertinent cases.

RESULTS

Further survival data were supplied for 45 of the previously reported 51 cases, the other six being lost to follow-up. Eight of the 51 have died, two of these deaths occurring 8 years after diagnosis. Fifteen other patient from the Fourth Workshop have since had extended follow-up and are included in this paper [3]. This makes a total of 66 patients who are identified here as forming a "long-term survivor~group (LTS}, whose clinical and cytogenetic data [updated in accordance with ISCN, 1991 [4]) are shown in Table 1. Of the other 649 cases, 10.5% were alive when last seen at less than 4 years' follow-up {7% with less than I year); some of these may, of course, have also survived for more than 4 years. A survival curve for 711 cases {four had no documented follow-up) of the Fourth Workshop cases, including the extended follow-up data, is shown in Figure 1. The longest survivor is still alive and well at over 11 years after diagnosis. Of the 23 patients known to have received BMT, 14 (60%} survived for more than 4 years and are included in the LTS Figure 1

group. Seventeen LTS patients had experienced a relapse, including one patient in relapse at 9 years 10 months after diagnosis, who subsequently died of liver failure due to hepatitis C. Using the survival data available in this series, the probability of surviving a further year has been calculated for each year already survived (Table 2}. After 5 years a plateau is reached such that each year patients have appm×imately 94% probability of surviving a further year without recurrence of their leukemia. Analyses of survival, using the BMDP statistical package {University of California), were performed comparing age groups, FAB type, and cytogenetic findings, to examine the effect of longer follow-up on these characteristics, previously shown to have significance [2]. The results are summarized in Table 3 and survival curves are shown for age group {Fig. 2), clone comple×ity classification [2] [Fig. 3), and for selected clonal abnorm~litias (Fig. 4). The longer follow-up has served to emphasize the significance of the different effects of these factors on prognosis. An observation not made previously became apparent

Overall s u r v i v a l for 711 p a t i e n t s in t h i s series.

II~."

Table 2 Probability of surviving a further year

gO

Number of Patients 80

Start Year

7O

4O 3O Z0 |

10

i

I.||1

ms|

im

ms//.

J

0 1 2 3 4 5 6 7 8 9

Entering

Dying

No further follow-up

Probability (%)

711 a 232 117 76 66 56 48 42 40 32

433 b 100 34 10 9 4 3 2 2 O

46 15 7 0 1 4 3 0 6 32

36.8 55.6 69.6 86.8 86.3 92.6 93.6 95.4 94.7 100.0

a Excludes four patients with no survival data. Sm.vlm[! (t~t,.s)

J - ffi atrviv0r

b Includes 61 patients who died before starting treatment.

4

G. J. Swansbury et al.

Table 3

Summ.ry of results of analyses of survival

Cases Dead Expected 392 319

339 257

312.37 283.63

Alive

Sex

53 62

MMe Fema~

Cases Dead Expected Alive

Genemlized Savage(Mantel-Cox) statistic 4.798 l d f p < 0.03. Generalized Wilcoxon (Bmslow} statist~ 3,862 l d f p < 0.05. Cases Dead 656 55

550 46

Expected Alive 561.04 34.96

106 9

Leukemia type

Primary Secondary

Generalized Savage CMantel-Cox) statistic 3.726 1 df p < 0.06. Generalized Wilcoxon (Breslow} statistic 4.403 1 df p < 0.04, Cases Dead Expected 58 44 74

87 93 118 116 98 21

46 32

Alive

56.66

12

< 10

58

12 16

10-19 20-29

73 69 102 110 87 19

89.83 92.81 105.97 69.56 46.61 8.14

14 24 16 6 11 2

30-39 40-49 50-59 60-69 70-79 80-

Generalized Savage (Mantal-Cox} statistic 101.104 8 df p < 0.0001. Generalized Wilcoxon (Breslow) ~ t i s t i c 89.538 8 df p < 0.O001.

356 118 237

278 102 216

365.70 105.97 124.34

78 16 21

102 243 65 162 95 32 12

80 209 52 130 84 30 11

88.21 217.13 58.68 141.47 62.87 21.01 6.63

< 50 50-59 60+

Alive

FAB type

22 34 13 32 11 2 1

M1 M2 M3 and 3v M4 M5 M6 Other

Generalized Savage (Mantel-Cox) statistic 16.776 6 d f p
Alive

315 281

250 243

287.68 227.09

65 38

115

103

81,22

12

250 184 106 56

287.68 163.60 117.42 27.30

65 24 21 5

Clone complexity N o clone found

Simple Complex

Very complex

Generalized Savage (Mantel-Cox} statistic 39.625 3 df p < 0.0001. Generalized Wilcoxon {Breslow) statistic 35.742 3 df p < 0.0001. Cases Dead Expected Alive

Selected abnormality

315 15 27

250 13 24

287.68

65

7.87

2

- 5

13.57

3

5q-

24

23

13.84

1

- 7

11

7

13.30

4

7q -

36 45 44 29 26 13 126

32 38 36 25 25 9 114

30.99 40.48 48.15 20.31 9.49 16.09 94.23

4

+8 t[15;17} t[8;21] Abnormalities of l l q Other abnormalities of 5/7 inv/del(16) All other abnormalities

7

8 4 1 4 12

N o abnormality found

Generalized Savage {Mantel-Cox} statistic65.012 11 df p < 0.0001. Generalized Wilcoxon (Breslow} statistic47.262 11 df p < 0.0001.

Age group

Generalized Savage~Mantel-Cox}statistic 93.619 2 d f p < 0.0001. Generalized Wflcoxon (Bmslow)statistic 82.599 2 d f p < 0.0001. Cases Dead Expected

315 208 127 61

Age group

51.34 74.85

Cases Dead Expected Alive

Table 3 Continued

during this study: four patients with a deletion of part of the long arm of chromosome 7 as the sole abnormality or with abnormalities not involving chromosome 5 tended to have a longer survival than patients with any other abnormalities of chromosome 7 or chromosome 5 {Fig. 5). Deletion of 7q is usually associated with secondary leukemia but in this series all but two of the patients with this abnormality had apparently primary disease.

Figure 2

Effect of age-group classificationon survival.

.\ m

Presence/absence of clone

---

'~

(n= 355) age ( , : 116) a l e

(50

No clonal abnormality found Mixture of clonal a n d normal cells All clonal cells

G e n e r a l i z e d Savage 0Vlantel-Cox] statistic 11.980 2 df p < 0.0025.

Generalized Wilcoxon {Breslow} statistic 12.291 2 df p < 0.0021.

zo

~

"-~.~.

10

~ - - ~ ,

. ~

~ "~"L.

o (continued)

,i ,i Ii ,i ,; ,; • Lreleal(Umrs)

,

, ..................

....

~

. . . . . . . . . .

,J,

_,

,.L--..,,~

,; --~= sereitMa"

11, b.

Cytogenetics a n d Long-Term Survival i n AML

5

~0

Ca)~I- (n:11) --- (5) -? (a~/) ..... (c) ~1- (n~:4) -.- (d) -5 (.:15) ---(e) abn(5~?)(n:Z6)

9O i

81) L.., ?0

?9

50 ~

~

(a) - - H m ~ l (b) --- $i@le

(n: 315) (n: 269)

(c) ..... Cmplex (n: 127) (d) - - - LkmjC~lex (n: 51)

50 40 30 ZO 10

405050 l L~--Ill ]

J

30 ' ',-k. '

~'--

(a)

e i (all

0 Suruiv~l (ge~s)

~- :

Figure 3

surviuc~

Survival (ge~mO

Effect of clone complegcity on survival.

(a) The m e a n ages at diagnosis ware 35 years (LTS) a n d 47 years (others), and the distributions ware significantly different (Student's t-test, data not shown). A two-way contingency table analysis of the distribution is shown (Table 5). Each age-decade below age 50 had more cases than expected i n the LTS group a n d each age-decade above age 60 had fewer (p < 0.0001). Co) There was a slight excess of females i n the LTS group (p < 0.032 uncorrected, < 0.043 with Yates correction). Although this contingency table analysis is not significant, a multivariate analysis (using BMDP) of the whole series (Table 6) showed that sex was a significant factor after allowance for age group, chromosome classification, a n d FAB type. (c) FAB types did not differ markedly between LTS a n d

l~ure 4

Effectof presence of specific abnormalities on survival.

i~." 90

(n: 15) (n: Z?) (n: 24) --- ?q(.= 11) -.-t(15:17) (n: 45) --tCO;Z1) (n: 14) ~-~

---7 ..... ~ -

~]

80 ~

t-q I ~1

----abn(11q) --.- inv(15)

I

o

Other patients

8 9 8 11 14 10 5 1 0

50 36 66 76 80 108 113 99 21

0-9 years age 10-19 20-29 30-39 40-49 50-59 60-69 70-79 8O

28 38

365 284

Male Female

10 21 11 18 5 1 0

92 223 54 145 92 31 12

M1 FAB type M2 M3 M4 M5 M6 Other

63 3

596 53

Primary leukemia Secondary leukemia

I

tk ~',~-'-~-r~

[_~

.4-,,,

-

'

.... '

, ,

-----------~.~--. . . . . . .~ . . . . . . . ~ abn(llq)

S u m m a r y of clinical features

LTS group

Table 5

-

h

10

Table 4

(n: Zg) (n: 13)

5O '

--~ = StwvJ~

Figure 5 Effectof presence of selected abnormalities of chromosomes 5 and 7 on survival (p < 0.005).

The following comparisons of clinical a n d cytogenetic data ware made between the 66 LTS patients and the other 649 patients (Table 4):

70

J(a)

t(O:Z1)

Total

Two-way analysis of age group LTS

Others

Total

50 33.1

308 324.9

358

Age 0-49 years Expected

10 10.9

108 107.1

118

Age 50-59 years Expected

6 22.1 66

233 216.9 649

239

Age 60 and greater Expected

X2 = 22.5 with 2 dr, p < 0.0001.

715

6

G . J . S w a n s b u r y et ell. Table 6

Results of multivariate analyses of variance

(a) Comparison of age group, NN-AN-AA classification, FAB type, sex, and primary/secondary status Step

Log

Improvement

No.

df

likelihood

0

0

-3467.989

1 2 3 4

1 2 3 4

-3436.811 -3430.804 -3427.179 -3425.149

Global

X2

p value

X2

62.355 12.015 7.250 4.060

0.000 0.001 0.007 0.044

60.472 72.762 80.919 84.954

p value ~ ~ < ~

0.001 0.001 0.001 0.001

Age NN-AN-AAclassificafion FAB Sex

(b) Comparison of age group, clone complexity classification, FAB type, sex, and primAry/secondary status Step No. 0 1 2 3 4

Log

Improvement

clf

likelihood

O 1

-3467.989 -3436.811

2 3 4

-3428.726 -3425.505 -3423.298

Global

~2

p value

X2

62.355 16.171 6.442 4.414

0.000 0.000 0.011 0.036

60.472 77.862 65.182 89.338

p value ~ < < ~

0.001 0.001 0.001 0.001

Age Comp~xityclassificaUon FAB Sex

(c) Comparison of age group, modified Chicago classification, sex, and primary/secondary status Log

Improvement

Step No.

df

0

0

-3467.989

1 2 3

1 2 3

-3436.811 -3434.127 -3432.573

likelihood

X2

p value

X2

p value

62.355 5.369 3.108

0.000 0.020 0.078

60.472 65.562 69.269

< 0.001 < 0.001 < 0.001

others e ~ e p t for an e0~cess of M3 in the LTS group (p = 0.021 comparing M3 w i t h all other FAB types combined). This corr e s p o n d e d to a n e~cess of LTS patients w i t h t(15;17) (Table 7), the abnormRlity characteristic of M3. (d) In t h r e e patients (4.6%) in the LTS group the leukemia was classified as being secondary; one h a d received carb i m a z o l e a n d 1-131 2 years p r e v i o u s l y for thyrotoxicosis; the other two h a d both received radiation a n d c h e m o t h e r a p y 5 years previously, one for ovarian cancer a n d the other fur

Table 7

I n c i d e n c e of selected c h r o m o s o m e abnormalities LTS Group

Total

0 1 Ia 3 0 7 7a 2 4a 9 32 66

Other Patients 15 26 23 8 26 38 37 30 9 153 284 649

Global

-5 - 7 5q 7q- 5/5q- and t(15;17)

7/7q -

t(8;21)

abn(11q) inv/del(16) Other clonal abnormality No abnormality detected

a Also one case discovered at relapse, not found at disgnosis.

Age Sex Chicago classification

Hodgkin's disease. Of the other patients, 53 (8.3%) h a d seco n d a r y leukemia. (e) The LTS group h a d a deficit of patients w i t h abnormalities of chromosomes 5 and/or 7, and ane~cess of patients w i t h t(8;21), t(15;17), and, although the n u m b e r s were very small, w i t h inv(16). A contingency table Chi-square analysis of the distribution of these cytogenetic classes gave a p < 0.01 b u t the smRll n u m b e r s in several of the LTS group categories make it difficult to calculate an esttmsW of the true probability. In three patients indicated in Table Z a clone was discovered at relapse w h i c h h a d not b e e n found at diagnosis. In the case w i t h t(8;21), o n l y seven cells h a d been available for analysis at diagnosis, and in the case w i t h del(Sq) o n l y 11 cells, so it is possible that the clone was originally present b u t undetected. In the case w i t h inv(16), 41 cells h a d been analyzed. These cases were not i n c l u d e d among the c h r o m o s o m e a b n o r m a l i t y categories for the analyses of survival in this paper. (f) The b r o a d e r c h r o m o s o m e classifications (NN-AN-AA, a n d the c o m p l e x i t y class pre'Hons]y described [2]) d i d not show any significant difference in distribution or survival between the LTS group a n d the other p a t i e 1 ~ . However, the multivariate AnAlysis of the w h o l e series (Table 6) showed that these classifications were second o n l y to age group for significant effect. In this series it appears that ~ n t s with a p p a r e n t l y n o r m a l karyotype hays a slightly better prognosis and those with very complex ksryetypes a dlsH~ctly worse

Cytogenetics and Long-Term Survival in AML prognosis than all other patients. Some classes of the Chicago classification are, of course, strongly associated with certain FAB types, so FAB type was excluded from the multivariate analysis of variance which included this clone classification. Similarly, after determining that age, chromosome classification, FAB type, and sex had the greatest effect on survival, the multivariate analysis did not identify primary/secondary status as being significant, probably because of the previously reported association between primary/secondary status and chromosome classification [2], i.e., that patients with secondary leukemia are more likely to have complex clones.

DISCUSSION More than a decade after the patients in this study were being diagnosed, AML is still a disease with a poor prognosis. Improved chemotherapy regimens have resulted in more patients achieving remission and surviving relapse, and increasing use of BMT to rescue patients following ablative therapy offers a hope of prolonged survival in some patients, as illustrated by our data. Ultimately, however, with all current treatments available, most patients will succumb to their disease [5] and it has been questioned whether chemotherapy will ever provide a generally effective cure, however refined or intensive it becomes. While other approaches are being investigated, there remains interest in determinln~ why some patients do experience good survival, even after what might be considered, by current standards, to have been suboptimal therapy. That worthwhile responses to treatment do occur may be illustrated by recording that at least three of the fvmMe LTS group patients in this series are known to have subsequently had successful pregnancies. The patients in the LTS group described are vary heterogeneous in terms of presentation characteristics, and of the features studied there is none that would at diagnosis have indicated their good prognosis. However, the greater proportion aged less than 50 years at diagnosis and the higher incidence of M3 and of chromosome abnormalities t(15;17], t(8;21), and inv/del(16) serve to confirm the prognostic value of these factors. A study by Marosi et al. [6] of 125 adult patients with de novo AML found that t(8;21), t(15;17}, and inv(16) correlated with good prognosis and abnormalities of 11q correlated with poor prognosis. However, Keating et al. [7], in a study of 384 patients with AML, while concurring with the good prognostic effect of t(8;21) and inv(16), found that t05;17) was associated with only an intermediate prognosis. A similar conclusion was drawn by Fenanx et al. [8] in another large series (283 patients), who further showed that the initially good response of patients with t(8;21) was undermined by a tendency to relapse, resulting in a lower

7 median disease-free interval. A special study of AML relapse undertaken at the Sixth Workshop [9] found that the relapses of AML with t(8;21) were often responsive to treatment unless secondary chromosome abnormalities had appeared. The introduction of ever more intensive therapies for patients with AML carries the risk of increasing the morbidity among those patients who could achieve a cure with gentler regimens. It is therefore becoming more urgent that existing long-term survivors are studied to try to identify laboratory or clinical markers that would have indicated their good response.

REFERENCES

1. Fourth International Workshop on Chromosomes in Leukemia, 1982 (1984): Clinical significance of chromosomal abnormalities in acute nonlymphoblastic leukemia. Cancer Ganet Cytogenet 11:332-350. 2. Arthur DC, Berger R, Golomb HM, Swansbury GJ, Reeves BR, Alimena G, Van Den Berghe H, Bloomfield CD, de la Chapelle A, Dewald GW, Carson OM, Hagemeijer A, Kaneko Y, Mitalman F, Pierre RV, Ruutu T, Sakural M, Lawler SD, Rowley ID (1989): The clinical significance of ksryotype in acute myelogenous leukemia. (Sixth International Workshop on Chromosomes in Leukemia, London, 1987). Cancer Genet Cytogenet 40:203-215. 3. Fourth international Workshop on Chromosomes in Leukemia, 1982 (1984): Overview of association between chromosome pattern and cell morphology, age, sex, and race. Cancer Ganet Cytogenet 11:265:274. 4. ISC~ (1991):Cancer Cytogenetics, supplement to An International System for Human Cytogenetic Nomenclature, F. MitelmRn, ed. S. Karger, Basel 1991. 5. Stone RM, Mayer RJ (1993): Treatment of the newly diagnosed adult with de novo acute myeloid leukemia. Hematol Clin N America 7:47-79. 6. Marosi C, KSller U, Keller-WeberE, Schwarzinger I, Schneider B, J~igerU, Vahls P, Nowotny H, Pirc-Danoewinata H, Stager G, Kreiner G, Wagner B, Lechner K, Lutz D, Bettelheim P, Haas QA (1992): Prognostic impact of karyotype and immunologic phenotype in 125 adult patients with de nero AML. Cancer Gonet Cytogenet 61:14-25. 7. Keating MJ, Cork A, Broach Y, Smith T, Waiters RS, McCredie KB, Trujillo J, Fralreich EJ (1987): ~ward a clinically relevant cytogenetic classification of acute myelogenous leukemia. Leukemia Res 11:119-133. 8. Fenanx P, Preudhomme C, Lal~JL, Morel P, Beuscart R, Banters F (1989): Cytogenetics and their prognostic value in de nero acute myeloid leukaemia: a report on 283 cases. Br J Heematol 73:61-67. 9. Garson OM, Hagemeijer A, Sakurai M, Reeves BR, Swansbury GJ, WilliAm~GJ, Alimena G, Arthur DC, Berger R, de la Chapelle A, DewaldG'W,MitelrnAnF, VanDen BergheH, Lawler SD, Rewley JD (1989): Cytogenetic studies of 103 patients with acute myelogenous leukemia in relapse. (Sixth International Workshop on Chromosomes in Leukemia, London, 1987}.Cancer C,enet Cytogenet 40:187-201.