1684
Good Steroid Response In Vivo Predicts
A Favorable Outcome in Children with
T-cell Acute Lymphoblastic Leukemia
Maurizio A r i d , M.D.,* Giuseppe Basso, M.D.,tlI Franco Mandelli, M.D.,$
Carmelo Rizzari, M.D.,§ Roberto Colella, M.D., 11 Elena Barisone, M.D.,T
Luigi Zanesco, M.D.,t Roberto Rondelli, M.D.,# Andrea Pession, M.D.,#
and Giuseppe Masera, M.D.,§ for the Associazione ltaliana Ematologia
Oncologia Pediatrica (AIEOP)
Background. Improved outcome of children with
acute lymphoblastic leukemia (ALL) who received inten-
sive chemotherapy was observed by the Italian Associa-
tion for Pediatric Hematology Oncology (AIEOP).To ver-
From the Department of *Pediatrics, University of Pavia; De-
partment of tpediatrics, University of Padova; Department of LjPedi-
atrics, University of Milan, Monza, Department of 11 Pediatrics, Uni-
versity of Bari; Department of Wediatrics, University of Torino; De-
partment of #Pediatrics, University of Bologna; Department of
$Hematology, University of Rome; Italy.
Supported in part by C.N.R. Grant No. 93.02200.PF39, and
AIRC.
The authors thank Dr. M. G. Valsecchi (Institute of Biometry and
Medical Statistics, University of Milan, Italy) for her valuable contri-
bution.
Participants in the AIEOP studies: Torino C1. Pediatrica, E. Ma-
don; Genova I.G. Gaslini, L. Massimo; Milano C1. Fediatrica I, V. Car-
nelli; Milano-Monza C1. Pediatrica, G. Masera; Pavia C1. Pediatrica,
F. Severi; Bergamo Ematologia, T. Barbui; Brescia C1. Pediatrica, A. G.
Ugazio; Padova C1. Pediatrica, L. Zanesco; Trieste C1. Pediatrica, P.
Tamaro; Parma C1. Pediatnca, G. Izzi; Modena C1. Pediatrica, F. Mas-
solo; Bologna C1. Pediatrica, G. Paolucci; Firenze C1. Pediatrica, C.
Guazzelli; Siena C1. Pediatrica, A. Fois; Pisa C1. Pediatrica, P. Mac-
chia; Perugia C1. Fediatrica, A. Amici; Ancona C1. Pediatrica, G. V.
Coppa; Ancona Div. Pediatrica, G. Caramia; Pescara Div. Ematologia,
G. Torlontano; Roma Catt. Ematologia, F. Mandelli; Roma C1. Pediat-
rica, G. Digilio; Roma Bambin Gesh, A. Donfrancesco; Napoli C1. Pe-
diatrica I, S. Cutillo; Napoli C1. Fediatrica 11, S. Auricchio; Napoli Pau-
silipon, V. Poggi and M. Cosco; Napoli SS. Annunziata, F. Tancredi;
Napoli Pediatria XXIX, P. Antonelli; Bari C1. Pediatrica I, F. Schettini;
Ban C1. Fediatrica 11, A. Ceci; Catanzaro Div. Ematologia, S. Magro;
Palermo C1. Pediatrica, M. LoCurto; Catania C1. Fediatrica, L. Schil-
ir6; Cagliari C1. Pediatrica, P. F. Biddau; Milano Niguarda, A. Nicolini;
Verona C1. Fediatrica, D. Gaburro; S. Giovanni Rotondo Div. Emato-
logia, M. Carotenuto; Reggio Calabria Div. Ematologia, F. Nobile;
Cremona Sez. Ematologia C. T., A. Porcellini.
Address for reprints: Dr. Maurizio Aric6, Clinica Fediatrica, Uni-
versita, IRCCS Policlinico S. Matteo, 1-27100 Pavia, Italy.
Received July 11, 1994; revisions received October 6 , 1994, and
December 14,1994; accepted December 14,1994.
ify if this improved outcome was also extended to T-cell
acute lymphoblastic leukemia (T-ALL) patients after in-
troduction of intensive chemotherapy, treatment re-
sponses of 2011 patients, including 184 with T-ALL
treated over the decade 1982-1991, were analyzed. Fur-
ther, the potential use of the association of presenting
clinical and biologic features with treatment outcome to
determine risk factors that might be useful for planning
risk-directed therapeutic studies was analyzed.
Methods. Of the 2011 children consecutively entered
on the three sequential AIEOP trials ('82, '87, and '88),
1528 (76%) had successful immunologic studies of the
bone marrow blasts. One hundred eighty-four (12%) had
T-ALL. During these studies, four consecutive high-risk
ALL treatment protocols (i-e.,8303,8503,8703, and 8803)
were used. Because the treatment schedule in protocols
8503 and 8703 were almost identical, those patients were
grouped together for the purpose of survival analysis.
The 137 boys and 47 girls ranged in age from 16 months to
15.5 years (median, 7.8 years) at diagnosis, and 38% had
a mediastinal mass. The rates of treatment response [i.e.
complete remission (CR) and event free survival (EFS)
rates] were compared for patients with T-ALL or B-cell
progenitor ALL, overall and by individual study. Present-
ing features and early response to steroid monotherapy
were also tested as possible prognostic factors.
Results. Overall, the CR rate was 94%, and the 7-year
survival (SE) was 51.9% (4.2). T-ALL patients had a signifi-
cantly worse outcome than B-lineage ALL patients [7-year
EFS 40.4% (5.2) vs. 61.7% (1.7), P < 0.001). Progressive im-
provement in EFS for T-ALL patients treated during 1decade
was observed, with 7-year EFS (SE)of 23.2% (8.3) for protocol
8303,J-year EFS of 39.5% (7.1) for combined protocols 8503-
8703, and 54.6% (7.1) for study '88, respectively. The analysis
of prognostic factors for T-ALL patients showed that poor in
vivo steroid response was the most unfavorable prognostic
factor, followed by leukocyte level count > 50 X 109/1(P =
AIEOP Studies for Childhood T-ALL/Aricd et al.
0.04).The EFS for patients with T-ALL and good steroid re-
sponse [63%(3.0)]was comparable with that of unselected B-
lineage ALL patients.
Conclusions. EFS for patient with T-ALL has steadily
increased in consecutive AIEOP ALL trials. However T-ALL
patients still have a significantly worse EFS compared with
patients with B-lineage ALL. Patients with T-ALLand a poor
in vivo response to steroid monotherapy represent a particu-
larly high risk treatment group. Cancer 1995;75:1684-93.
Key words: T-ALL, acute lymphoblastic leukemia, chil-
dren, steroid response.
Approximately 15% of newly diagnosed children and ado-
lescents with acute lymphoblastic leukemia (ALL) have T-
cell acute lymphoblastic leukemia (T-ALL). Patients with
T-ALL are often older males who present with mediastinal
mass and high leukocyte counts. Although the prognosis
of children with ALL has improved dramatically over the
past quarter century, most studies show that patients with
T-ALL continue to fare worse than those with B-progenitor
ALL, with cure rates in the 45 to 50% range.'-3 Whether
the poorer treatment outcome of children with T-ALL is
due to phenotype or to certain cluucal risk features associ-
ated with phenotype [i.e., hgher leukocyte level at the di-
agnosis, older age, hgher incidence of leukemia-lym-
phoma syndrome, more frequent central nervous system
(CNS) dlsease at presentation, and proportionally more
black patients] remains controversial.More importantly,re-
sults of some recent small studies of intense therapy suggest
that the prognostic importance of T-ALL may dunirush or
disappear in the context of intense therapy.'
To assess the therapeutic effectiveness of increas-
ingly intense therapeutic protocols that were used over
the past decade by the Italian Association for Pediatric
Hematology Oncology (AIEOP) for treatment of chil-
dren with T- or high risk B-progenitor ALL, we ana-
lyzed the treatment responses of 201 l patients, includ-
ing 184 with T-ALL. The results for patients with T-ALL
form the basis for this report. We show an improved
outcome with use of increasingly intense chemotherapy
over this time period, and demonstrate that favorable
early steroid response and low leukocyte level at diag-
nosis are significantly associated with favorable out-
come in the context of these therapy protocols.
Patients and Methods
From July 1982 to March 1992, 2011 eligible children
with newly diagnosed ALL were consecutively enrolled
in three protocols of the AIEOP (i.e., AIEOP studies '82,
'87, and '88). Study '82 was open for accrual of patients
starting in July 1982. Because poor results were noted at
1685
an early interim analysis, the high risk arm of this study
(protocol 8303) was closed in February 1985 and proto-
col 8503 was substituted. Study '87 was started from
February 1987, and the high risk arm (8703)maintained
the same treatment schedule used in 8503. From Febru-
ary 1988, Study '88 was started in eight major centers
and subsequently (November 1990) extended to all cen-
ters. Eligible children were older than 1 year (in studies
'82 and '87 only) but younger than 15 years of age at
diagnosis with a morphologc diagnosis of ALL (see be-
low). Of the 201 1 children enrolled, 1528 (76%)had suc-
cessful immunologic marker studies to determine major
immunophenotype (i.e., T- or B-progenitor ALL) and
184 (12%)of these had T-ALL. We examined the clinical
features and outcomes of the patient groups with and
without successful immunologic marker stules and
were unable to identify significant differences in the pre-
senting features (i.e., age distribution, proportion with a
mediastinal mass or CNS disease at presentation, and
proportion of patients with leukocyte count > 50 X lo9/
1) or treatment outcome (i.e., EFS at 5 years). Therefore,
we assumed that the exclusion of the patients without
successful immunologic marker studies did not bias the
analysis and that the study population focused upon in
this report (i.e., those with T-ALL or documented B-pro-
genitor ALL) is representative of the entire patient popu-
lation. However, we recognize that a proportion of chil-
dren with T-ALL will not be identified using the E-roset-
ting technique that was employed during the early phase
of the studies (1982-1984, n = 26).4
Patients with T-ALL were treated with high risk regi-
mens of the AIEOP studes only (Table l).5-7 In brief, all
patients received induction therapy comprised of vincris-
tine, predrusone (or dexamethasone),daunorubicin, aspar-
aginase, and intrathecal methotrexate; protocol 8303 pa-
tients received addltional intermediate dose methotrexate
and cytarabine, whereas those in study '88 received addi-
tional cyclophosphamide, cytarabine, and mercaptopurine.
Consolidation therapy comprised repeated infusions of in-
termediate (500 mg/mz) or high (5 g/m') dose methotrex-
ate or cytarabine, plus oral mercaptopurine and intrathecal
methotrexate.Reinduction therapy was similar to induction
therapy except that it was less intense. Patients in the last
three studies also received cranial irradiation. Continuation
therapy consisted of mercaptopurine and methotrexate,
with vincristine and prednisone pulses, in all but study '88.
The duration of treatment was 15 months in 8303, 22
months in 8503-8703 and 24 months in 8803. Based on
reports of the BFM study group indicating prognostic im-
portance of in vivo response to early steroid monotherapy,'
we examined the early in vivo response (i.e., lymphoblast
count at day 7 of monotherapy) to glucocorticoid in 103
patients treated in protocols 8703 and 8803, where such
1686 CANCER April 1,1995, Volume 75, No. 7

Table 1. Treatment Schedule for T-cell Acute Lymphoblastic Leukemia in Three

Subsequent AIEOP Studies
8303 8503-8703 8803
(mg/m q X no. of doses) (mg/m q X no. of doses) (mg/m q X no. of doses)
Induction
Vincristine 1.5/wk X 5 1.5/wk X 4 t 1.5/wk X 4
Prednisone 60 X 28 days 40 X 33 days 60 X 28 days
Daunorubicin 30/wk X 3* 30/wk X 4 40/wk X 4
L- Asparaginase 6000 1U X 9 6000 IU X 9 10,000 IU X 8
Cytarabine IV 150 X 2 75 X 16
Cyclophosphamide - 1000 x 2
Mercaptopurine - 60 X 28 days
Methotrexate IV 500 X 1 - -

Methotrexate IT by age X 3 by age X 2$ by age X 3#

Cytarbine IT 30 X 1
Consolidation
Methotrexate IV - 500 X 3 5000 X 4
Methotrexate IT - by age X 45 by age X 4
Mercaptopurine - 75 X 21 days 25 X 56 days
Cranial RT - 18 Gy -
Cytarabine IV 3000 X 4
L-Asparaginase 6000 X 2
Reinduction
Vincristine 1.5/wk X 3 1.5/wk X 411 1.5/wk X 4
Prednisone 40/day X 14 40 X 14 days -

Dexamethasone - - 10 X 21 days
Daunorubicin 30/wk X 3* 30/wk X 3 30/wk X 4*
Asparaginase - 25,000 IU/wk 10,000 IU x 4
Methotrexate IT - - by age X 2
Cytarabine IV - 75 X 8
Cyclophosphamide 1000 x 1 1000 x 1
Thioguanine - 60 X 14 days
Cranial RT - 18 Gy#
Continuation7
Mercaptopurine 75 X 21 days 75 X 21 days 50/day
Methotrexate - 20/wk 20/wk
Vincristine 2x3 1.5 X 1 -
Prednisone 40 X 14 days 40 X 5 days
Teniposide 165 X 1 -
Cytarabine IV 300 X 3 -

Asparaginase - 25,000 IU/wk

5 2 0 doses
Duration of therapy 15 mo 22 mo 24 mo
IT: intrathecal; IV: intravenous; CNS: central nervous system; MTX: methotrexate.
Age-adlusted doses of IT methotrexate were as follows: tl year 6 mg, > 1 I2 years 8 mg, <2 I 3 years 10 mg, > 3 years
12 mg.
* Doxorubicin,
t 2/wk x 5,
$ By age X 4,
5 Age dose X 6,
/I 2/wk X 3.
1Repeated every 84 days (8303) or every 28 days (8503-8703).
# Cranial irradiation was given at the following doses: age >2 years 18 Gy (24 Gy for patients with CNS leukemia at
diagnosis), age >1 I 2 years 12 Gy (18 Gy if CNS+); for age i l year extended intrathecal MTX during maintenance
was substituted for radiation therapy. Patients with CNS leukemia had additional intrathecal methotrexate on days 7,
14,21, and 28.

therapy preceded induction treatment. Good steroid re- ate) therapy. The proportion of patients with leukemia-
sponse was defmed as 4 0 0 0 lymphoblasts/mm3 after 7 lymphoma syndrome defined as presence of bulky disease
days of steroid (and one injection of intrathecal methotrex- including mediastinal mass and massive lymphoadenopa-
AIEOP Studies for Childhood T-ALL/Aricd et al. 1687

Table 2. Clinical and Biological Features of 184 Patients With T-cell Acute
Lymphoblastic Leukemia
Total 8303 8503 8703 8803

Age (yr)
1-9 128 16 36 42 34
10-14 56 11 9 18 18
Sex
Males 137 21 30 44 42
Females 47 6 15 16 10
Leukocyte count (lO”/L)
<10 26 4 5 13 4
10-49 46 5 14 17 10
50-99 27 4 7 6 10
2100 85 14 19 24 28
Mediastinal mass
Yes 69 12 19 14 24
No 111 13 25 46 27
NE 4 2 1 0 1
C N S involvement
Yes 9 0 1 0 8
NO 171 27 44 58 42
NE 4 0 0 2 2
CDlO
Positive 32 0 10 12 10
Negative 144 25 33 45 41
NE 8 2 2 3 1
Myeloid antigen
Positive 6 1 1 1 3
Negative 65 6 7 14 38
NE 113 20 37 45 11
Steroid response
Good 69 - - 34 35
Poor 34 - - 20 14
NE
~~ 81 27 45 0 0
CNS: central nervous svstem: NE: not evaluated.

thy at diagnosis was determined. Central nervous system
leukemia at diagnosis was d e h e d as >5 mononuclear
WBC on a chamber count plus blast cells in a cytospin prep-
aration of such spinal fluid.
Morpkologic and Cytockemical Studies
The diagnosis of ALL was based on morphologic and
cytochemical criteria of the French-American-British
working group.’ Thus, by definition, all patients had
less than 3% blast cells positive for myeloperoxidase
and/or aspecific esterase.
Immunopkenotyping
Bone marrow cells were separated on a Ficoll-Hipaque
gradient. After preincubation with normal AB human
serum, cell surface antigens were detected by standard
indirect immunofluorescence assay by fluorescence mi-
croscopy (>400 cells counted) and/or by flow cytome-
try (Profile 11, Coulter or Facscan, Becton Dickinson),
using a panel of commercial monoclonal antibodies in-
cluding the T-cell-associated antigens CDla (OKT6),
CD3 (Leu4), CD4 (OKT4A), CD5 (Leul), CD7 (3A1),
CD8 (OKT8), the B-lineage-associated antigens CD19
(B4), CD20 (Bl), CD24 (OKB2),the myeloid-associated
antigens CD13 (My7), CD14 (My4), CD15 (Leu Ml),
CD33 (My9), CD65 (Vim2), and nonlineage-restricted
CDlO (CALLA; IS), HLA DR, and CD34 (HPCA 1). Re-
sults were considered positive when more than 20% of
the lymphoblasts expressed the antigen (except for cIg
and TdT, where expression of TdT in more than 10%
of lymphoblasts was considered positive).” In the early
phase of the study before the wide availability of mono-
1688 CANCER April 2,2995, Volume 75, No. 7

Table 3. Outcome of Treatment in 184 Children With T-cell Acute Lymphoblastic

Leukemia
Total 8303 8503 8703 8803
No. of patients 184 27 45 60 52
Induction failure 11 0 0 4 7
Resistant 5 0 0 2 3
Induction death 6 0 0 2 4
Responders 173 (95%) 27 45 56 45
Relapse 79 (45%) 20 24 21 14
Isolated
Bone marrow 32 7 11 7 8
CNS 24 6 8 5 5
Testis 4 1 0 3 0
Other 5 0 2 3 0
Combined
Bone marrow CNS + 5 1 3 0 0
Bone marrow testis + 2 0 0 1 1
CNS eye + 2 1 0 1 0
Other 5 4 0 1 0
Adverse events in responders
Death in CCR 4 0 1 1 2
Lost to follow-up 3 0 0 3 0
Off study 9 ix 1 5 3 0
CCR 78 (45%) 6 15 28 29
7-year survival (SE) 51.9 (4.2) 35.7 (9.4) 55.0 (7.0)t 63.6 (7.0)$
7-year EFS (SE) 40.4 (5.2) 23.2 (8.3) 39.5 (7.l)t 54.6 (7.1)$
CNS: central nervous system; SE: standard error; CCR: continuous complete remission; EFS: event free survival.
* See text for details.
t Survival and EFS estimates for 8503 and 8703 groups are presented combined.
$ Five years.

clonal antibodies, only E-rosetting was used to identify

T-ALL (n = 26 cases). A case was considered evaluable
for major immunophenotype if it had successful sIg, E-
rosetting (or later T-associated antigen expression as
specified above), and CDlO studies. T-ALL was diag-
nosed if the lymphoblasts expressed CD3, surface CD7
and TdT. Generally, other T-cell-associated antigens
were coexpressed as well, whereas B-progenitor cell an-
tigens were absent. T-ALL maturational stage was de-
fined as follows: early thymocyte phenotype cCD3+,
v
+
Z
w
60-
50-
TL--

--=., ....................... ............ :...... 40.4 (5.2)

0 -
CD7+, TdT+, CDI- and sCD3-; intermediate stage 6 40- ...........................
cases expressed the same antigens plus CD1; late stage
cases lacked CD1 expression but expressed sCD3 as
well.

Genetic Studies 0
...T-ALL

1 2
0=94

3 4 5 6
YEARS FROM DIAGNOSIS
-
Log-Rank p

7
<
8
,001

Karyotype and flow cytometric testing of DNA content Number of patients ot risk:

were not assessed in this study. 1344 743 163 NON T-ALL
184 77 17 T-ALL
Statistical Analysis Figure 1. Overall event free survival of 1528 children with acute
lymphocytic leukemia (ALL);T-cell ALL (n = 184) versus non-T-cell
Data were collected by patient-oriented and protocol-spe- ALL (n = 1344). Numbers under the graphics indicate the number of
cific forms. All information was stored, controlled, and an- patients at risk at specific points.
AIEOP Studies for Childhood T-ALL/Aricd et al.
alyzed by VENUS, an integrated software facility system
running on an IBM mainframe computer at the Italian In-
teruniversity Computing Centre CINECA. Event free sur-
vival (EFS) and survival curves were estimated according
to Kaplan-Meier. For both curves, the starting point was
date of study entry at dagnosis of ALL. For EFS, induction
failure (resistant disease or death during induction), death
in continuous complete remission, relapse, and dagnosis
of second malignancy in continuous complete remission
were counted as failures. Death from any cause was con-
sidered a failure in calculating survival time. Nine patients
were excluded from analysis at the time of treatment with-
drawal due to bone marrow transplant in first complete
remission (n = 3), treatment refusal (n = 2), toxicity (n =
3), or other unreported reason (n = 1). Six of these patients
were still alive at the time of writing. Follow-up was up-
dated as of April 15, 1994. Univariate and multivariate
analyses were conducted for covariates such as age at d-
agnosis, sex, hemoglobin level, leukocyte level, and plate-
let count, liver and spleen size, presence of mediastinal
mass, leukemia-lymphoma syndrome according to the
Children's Cancer Group definition, expression of T-asso-
ciated and myeloid markers, presence or absence of CNS
leukemia, response to early steroid monotherapy, and
therapy used. Relative hazard rate was computed accord-
ing to the estimate proposed by Pike: (observed l/ex-
pected l)/(observed 2/expected 2)." Fischer's exact test
was used to compare complete remission rate between
different subgroups.'2
Results
Of the 1528 cases analyzed, 184 (12%) were identified
as T-ALL. The percentage of T-ALL was 12% in study
1982, 10% in study 1987, and 13% in study 1988. The
median observation times were 55 months (range, 1 to
99 months) for the entire patient cohort; 74 months
(range, 20-84 months) for patients in protocol 8303, 79
months (range, 67-99 month) for protocol 8503, 45
months (range, 1-64 months) for protocol 8703, 54
months (range, 1-99 months) when 8503 and 8703
were combined, and 54 months (range, 15-68 months)
for protocol 8803.
Presenting Fea tu yes
The presenting features of the 184 patients with T-ALL
are described in Table 2. The ages at diagnosis of the
137 boys and 47 girls ranged from 16 months to 15
years (median, 7.8 years). Only 2 children were younger
than 18 months; 21 were younger than 3 years of age.
1689
Clinical Outcome
The complete remission rate for patients with T-ALL
was 94% overall. Details of treatment results for pa-
tients, overall and within each treatment protocol, are
described in Table 3. The overall 7-year survival and
EFS rates of patients with T-ALL were 51.9 (4.2) and
40.4 (5.2) respectively, compared with 72.4 (1.4) and
6 1.7 (1.7), respectively, for patients with B-lineage ALL
treated on the same studies (P < 0.001) (Fig. 1).T-ALL
patients did worse than B-lineage ALL patients in all
age groups (P < 0.001). The EFS rates of T-ALL patients
according to treatment studies were as follows: 23.2%
(8.3) for protocol 8303, 36.9% (8.2) in protocol 8503,
36.5% (15.6) for 8703, with 39.5 (7.1) for combined
8503 and 8703 groups, and 54.6% (7.1) for protocol
8803 respectively (Fig. 2).
Prognostic Factors
High WBC count was the only presenting feature asso-
ciated with a n increased rate of treatment failure (Table
4, Fig. 3). The outcome of children with T-ALL was not
substantially affected by expression of any of the
different T-cell-associated antigens (CD1, CD2-ERFC,
CD3, CD4, CD5, CD7, CD8). For early response to ste-
roid monotherapy, three groups of patients could be de-
fined: (1) patients who had fewer than 1000 lympho-
blasts/mm3 before and after 7 days of steroid [33.0%;
EFS 67.8% (8.0)], (2) patients who had greater than
1000 lymphoblasts/mm3 before but fewer than 1000
after 7 days of steroids [32.1%; EFS 60.3% (8.3)], and
(3) patients who had greater than 1000 lymphoblasts/
mm3 after 7 days of steroid [34.9%; EFS 43.5% (8.9)].
Overall, good responders to steroids (n = 69) had a sig-
nificantly better prognosis than poor responders (n =
34). Their 5-year EFS was 63.3% (3.0) vs. 43.5y0,'-~re-
spectively (P = 0.036). Within the above (2) group (i.e.,
the patients who showed a favorable response to ste-
roids), the presenting leukocyte count (less or greater
than 50,000 WBC/mm3) was not an important prog-
nostic feature [EFS 72.9% (13.5) versus 53.3% (10.3),
respectively; P = 0.17, not significant].
Pattern of Failures
The induction failure rate was 6% overall. The in-
creased rate of induction failure observed in the 8803
group was due to toxicity (hemorrhage n = 1, infection
n = 3) more than to resistant disease. Leukemia relapse
was the most common cause for treatment failure. The
pattern of relapses and their distribution among differ-
ent protocols is described in Table 3. No significant as-
*
1690 CANCER April 1,1995, Volume 75, No. 7

Table 4. Distribution and Influence of Presenting Features on Treatment

Outcome in T-cell Acute Lymphoblastic Leukemia
No. of No. of Relative
oatients % failures EFS % (SE) risk P value
Sex
Male 137 74 73 35.7 (6.8) 1.19 0.31
Female 47 21 50.0 (7.8)
Age (yr)
210 56 30 33 35.3 (7.6) 1.24 0.14
<10 128 61 43.0 (6.5)
Leukocytes
250 109 59 64 30.7 (6.9) 1.51 0.003
<50 72 28 56.1 (6.7)
CNS
Yes 9 4.9 4 53.3 (17.3) 0.87 0.52
No 171 87 40.5 (5.3)
Mediastinal mass
Yes 69 37 38 42.4 (6.1) 1.17 0.45
No 111 52 36.7 (9.7)
Hemoglobin
210 98 53 56 34.4 (7.1) 1.3 0.16
<10 86 38 47.8 (7.2)
Steroid response
Poor 34 33 18 43.5 (8.9) 1.52 0.036
Good 69 24 63.3 (3.0)
CDlO
Positive 32 18 14 45.1 (12.5) 0.82 0.46
Negative 144 76 39.7 (5.6)
Myeloid antigen
Positive 6 8 2 0 0.57 0.32
Negative 65 38 41.2 (6.1)
Treatment protocol
8303 27 15 20 23.2 (8.3)” a.b 1.33 0.13*
1.70
a:d 1.69

8503 45 24 25 36.9 (8.2)b b‘1.28

bd1.27
8703 60 33 26 36.5 (15.6)’ ‘:d0.99
8803 52 28 23 54.6 (7.1)d
EFS: event free survival; SE: standard error; CNS: central nervous system
* Overall.

sociation was observed between the frequency of fail- Discussion

ures, overall or by site and the thymocyte maturation
stage (P, not significant). The control of CNS disease
was suboptimal in 8303, but thereafter it markedly im-
proved. There were four testicular relapses (3.1%). No
second malignancy was reported.
Therapy Associated Toxicity
There were eight toxic deaths (four during induction
and four during remission) due to infection (n = 7) or
hemorrhage (n = 1).
The incidence of T-ALL observed in our studies was
comparable with that of other European studies13al-
though it was slightly lower compared with that of
other studies from the US. This may depend on the
lower cut-off age for eligibility in our study (i.e., 15
years) compared to most US studies, which usually
include 10-15% Black people, who have a considera-
bly higher incidence of T-ALL.14Otherwise, our pa-
tient population appeared to be similar to that of
other studies in terms of sex ratio, proportion with
AIEOP Studies for Childhood T-ALLIAricd et al. 1691

- PROTOCOL 88 OK23

54.6 (7.1) .....................................

...... ....................................... ............
i 45.3 (4.1)
......... I.

404 17 ' :
>..
39.5 (7.1)
..........................
L- L _
L -- -- - - - -
--I 30.7 (6.9)
,,.. T-ALL/<50K O= 28

- NON T-ALL/>50K 0=101

10
-- T-ALL/>JOK O= 64
10 ° ~
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9
YEARS FROM DIAGNOSIS
YEARS FROM DIAGNOSIS Number of patients at risk:
Number of patients at risk 1128 661 139 NON T-ALL/<50K
72 36 9 T-ALL/t50K
27 6 4 PROTOCOL82 210 80 24 NON T-ALL/>50K
105 43 13 PROTOCOLS 85-87 109 40 8 T-ALL/>50K
52 28 0 PROTOCOL88
Figure 3. Impact of high leukocyte count and/or T phenotype on the
Figure 2. Event free survival of 184 children with T-cell acute event free survival of children with acute lymphocytic leukemia.
lymphocytic leukemia treated according to three different treatment Number under the graphics indicate number of patients at risk at
schedules between 1982 and 1991 (AIEOP protocols 8303,8503- specific points.
8703, 8803). Numbers under the graphics indicate the number of
patients at risk at specific points.

tor.22The changing effectiveness of therapy over the

hepatosplenomegaly, distribution of WBC count at study period, together with the limited number of pa-
the diagnosis, and proportion with a mediastinal tients, may have blurred the relevance of other puta-
mass,2.13-15 tive prognostic factors, including the expression of
The introduction of intense multiagent chemo- surface antigens as CDlO (CALLA),23CD3 without
therapy, together with improved supportive care, has CD1, or myeloid antigen(s), which all were not sig-
increased the cure rate of children and adolescents nificant in our ~ e r i e s . ~ ~ - * ~
with T-ALL in three large sequential AIEOP studies The prognostic relevance of steroid response for
conducted over the last decade. The latest study '88 childhood ALL had been first reported by the BFM
featuring high dose methotrexate and the BFM con- group several years ago and was recently con-
solidation scheme including cyclophosphamide, cy- The results of our study suggest that favor-
tarabine and antimetabolite, had the best result. This
is in keeping with good results in T-ALL obtained by
the BFM in the recently reported ALL 86 study.16It is 1001:

not yet clear what element in this combination may

lead to the significant increase in EFS, and the sug-
gested role played by high dose methotrexate in in-
creasing the EFS in a proportion of T-ALL cases re- ........
mains i n t r i g ~ i n g . Nevertheless,
'~ about 45% T-ALL 50- .,,,~
43.5 (8.9)
patients in our series suffered an adverse event, usu- 0 -
<,,.
.......................................................................................
i
6
a
40-
-
ally leukemia relapse, within a few years after achiev-
ing complete remission. The salvage rate for patients
with T-ALL who experience a relapse is only approx-
imately l o % , regardless of treatment program. Al-
though B-lineage ALL subsets have been recently as-
sociated with different prognosis based on karyo-
0 1 2 3 4
YEARS FROM DIAGNOSIS
---- Log-Rank p < .05
7
5
- 1
6

Number of patients at risk

type"," and DNA content" analysis, this has not
69 42 19 STEROID RESPONDERS
been reported for T-lineage ALL.21Thus the analysis 34 13 !2 STEROID NON RESPONDERS
of potential prognostic factors in large series of T-ALL Figure 4. Event free survival of 103 T-cell acute lymphocytic
remains of interest.14z15 Leukocyte count exceeding 50 leukemia children with good (n = 69) or poor (n = 34) early in vivo
X 109/1 was an independent adverse prognostic fac- steroid response, treated in the AIEOP 8703 or 8803 protocols.
1692 CANCER April 1, 2995, Volume 75, No. 7
able early response in vivo to steroid monotherapy
identified a subset of T-ALL patients with a remark-
ably favorable outcome in the range of 70%. Whether
this may reflect different susceptibility to steroid-in-
duced apoptosis of the leukemic blasts remains to be
elucidated.
We conclude that the prognosis for children with
T-ALL has remarkably improved over the past decade
in Italy, largely due to the introduction of intense che-
motherapy together with improved supportive care.
Early in vivo response to steroid monotherapy identi-
fies T-ALL patients with a remarkably good outcome
due to lower risk of leukemic relapse. Based on these
results, steroid responder T-ALL patients are now el-
igible for the intermediate risk group of current
AIEOP ’91 study, designed to explore the impact of
randomized addition of high dose L-asparaginase to
a BFM-based intensive chemotherapy. Patients who
fail to respond to steroid in vivo are at increased risk
of leukemia relapse, regardless of their leukocyte
count. Thus, they are enrolled in the high risk group,
which receives the BFM block-therapy, consisting of
cyclic administration of high dose polychemotherapy
“blocks,” formerly used by BFM for relapsed pa-
tient~.~~
References
1. Clavell LA, Gelber RD, Cohen HJ, Hitchcock-Bryan 5,Cassady
JR, Tarbell NJ, et al. Four agent induction and intensive aspara-
ginase therapy for treatment of childhood acute lymphoblastic
leukemia. N Eiigl I M e d 1986;315:657-63.
2. Crist WM, Shuster JJ, Falletta J, Pullen DJ, Berard CW, Vietti
TJ, et al. Clinical features and outcome in childhood T-cell leu-
kemia-lymphoma according to stage of thymocyte differen-
tiation: a Pediatric Oncology Group study. Blood 1988; 72:
1891-7.
3. Crist WM, Shuster JJ, Look T, Falletta J, Pullen DJ, Berard CW,
et al. Current results of studies of immunophenotype-, age-, and
leukocyte-based therapy for children with acute lymphoblastic
leukemia. Leukemia 1992;6:162-6.
4. Dowel1 BL, Borowitz M, Boyett JM, Pullen J, Crist WM, Quddus
FF, et al. Immunologic and clinico-pathologic features of com-
mon acute lymphoblastic leukemia antigen-positive childhood
T-cell leukemia: a Pediatric Oncology Group Study. Cniirer
1987; 59:2020-6.
5. Pession A, DeRosa M, Vecchi V, Masera G, Pillon M, Comelli A,
et al. L-Asparaginase: tolerability and effectiveness in childhood
high-risk acute lymphoblastic leukemia [abstract]. Proc Am SOC
Cliiz Oiicol 1990.
6 . Vecchi V, Aric6 M, Basso G, Ceci A, Madon E, Mandelli F,
et al. Risk-directed therapy for childhood acute lymphoblas-
tic leukemia: the AIEOP’82 studies. Cancer 1993; 72:
2517-24.
7. Vecchi V, Pession A, Rosati D, Schilir; G, Aric6 M, Arrighini A,
et al. Childhood acute lymphoblastic leukemia: results of
AIEOP’87 protocol generation [abstract]. Proceeds of the XXIV
SIOP Meeting. M e d Pediatr Oilcol 1992; 20:390.
8. Riehm H, Reiter A, Schrappe M, Berthold F, Dopfer R, Gerein V,
et al. Corticosteroid-dependent reduction of leukocyte count in,
blood as a prognostic factor in acute lymphoblastic leukemia in
childhood (therapy study ALL-BFM 83). Klin Padiatr 1987; 199:
151-60.
9. Bennett JM, Catovsky D, Daniel M-T, Flandrin G, Galton DAG,
Gralnick HR, et al. Proposals for the classification of the acute
leukemias. French-American-British Cooperative group. Br J
Hneiiiatol 1976; 33:45 1-8.
10. Basso G, Rondelli R, Putti MC, Cantu Rajnoldi A, Granchi D,
Cocito MG, et al. Incidence and prognostic significance of im-
munophenotypic subgroups in childhood acute lymphoblastic
leukemia: the experience of the AIEOP cooperative study. Recent
Results Catzcer Res 1993; 131:297-307.
11. Pike M. Contribution to the discussion on the paper by F. Peto
and J. Peto: asymptotically efficient rank invariant test proce-
dures.JRStat S o c A 1972; 135:201-3.
12. Armitage P. Statistical methods in medical research. New York:
Blackwell Scientific Publications, 1971.
13. Ludwig R, Teichmann JV, Sperling C, Komische B, Ritter J, Reiter
A, et al. Incidence, clinical markers and prognostic significance
of immunologic subtypes of acute lymphoblastic leukemia in
children: experiences of the ALL-BFM 83 and 86 studies. Klin
Padintr 1990;202:243-52.
14. Pui CH, Behm F, Singh B, Schell M, Williams D, Rivera G, et al.
Heterogeneity of presenting features and their relation to treat-
ment outcome in 120 children with T-cell acute lymphoblastic
leukemia. Blood 1990;75:174-9.
15. Shuster JJ, Falletta J, Pullen DJ, Crist WM, Humphrey GB, Dow-
ell BL, et al. Prognostic factors in childhood T-cell acute lympho-
blastic leukemia: a Pediatric Oncology Group study. Blood
1988; 75: 166-73.
16. Reiter A, Schrappe M, Ludwig R, Hiddemann W, Sauter S,
Henze G, et al. Chemotherapy in 998 unselected child-
hood acute lymphoblastic leukemia patients: results and
conclusions of the multicenter trial ALL-BFM 86. Blood
1994;84:3122-33.
17. Bettoni C, Reiter A, Schrappe M, Ludwig WD, Beck J, Graf N, et
al. Improved event-free-survival for of childhood T-cell acute
lymphoblastic leukemia (T-ALL)after introduction of high dose
methotrexate in multicenter trial ALL-BFM 86 [abstract]. Med
Pediatr Oncol 1992;20:385.
18. Pui CH, Crist WM, Look T. Biology and clinical significance of
cytogenetic abnormalities in childhood acute lymphoblastic leu-
kemia. Blood 1990; 76:1449-63.
19. Raimondi SC. Current status of cytogenetic research in child-
hood acute lymphoblastic leukemia. Blood 1993;81:2237-51.
20. Trueworthy R, Shuster J, Look T, Crist W, Borowitz M, Carroll
A, et al. Ploidy of lymphoblasts is the strongest predictor of treat-
ment outcome in B-progenitor cell acute lymphoblastic leukemia
of childhood: a Pediatric Oncology Group Study. Clin Oncol
1992; 10:606-13.
21. Pui CH, Behm FG, Crist WM. Clinical and biologic relevance
of immunologic marker study in childhood acute lymphoblastic
leukemia. Blood 1993;82:343-62.
22. Pullen DJ, Sullivan MP, Falletta JM, Boyett JM, Humphrey GB,
Starling KA, et al. Modified LSA2-L2 treatment in 53 children
with E-rosette positive T-cell leukemia: results and prognostic
factors (a Pediatric Oncology Group Study). Blood 1982;60:
1159-69.
23. Pui CH, Rivera G, 1 lancock M, Raimondi 5,Sandlund JT, Mah-
moud HH, et al. Clinical significance of CDlO expression in
AIEOP Studies for Childhood T-ALLIAricd et al.
childhood acute lymphoblastic leukemia. Leukemia 1993; 7:35-
40.
24. Sobol RE, Mick R, Royston I, Davey FR, Ellison RR, Newman
R, et al. Clinical importance of myeloid antigen expression in
adult acute lymphoblastic leukemia. N Engl I Med 1987;316:
1111-7.
25. Wiersma SR, Ortega J, Sobel E, Weinberg K1. Clinical importance
of myeloid antigen expression in acute lymphoblastic leukemia
of childhood. N Etlgl ] M e d 1991;324:800-8.
1693
26. Pui CH, Schell MJ, Raimondi S, Head D, Rivera GH, Crist
WM, et al. Myeloid antigen expression in childhood acute
lymphoblastic leukemia. N Engl ] Med 1991;325:1378-
9.
27. Henze G, Fengler R, Hartmann R, Kornhuber B, Janka-Schaub
G, Niethammer D, et al. Six-year experience with a comprehen-
sive approach to the treatment of recurrent childhood acute lym-
phoblastic leukemia (ALL-REZ BFM 85): a relapse study of the
BFM Group. Blood 1991;78:1166-72.