(1999) 13, 335–342  1999 Stockton Press All rights reserved 0887-6924/99 $12.00 http://www.stockton-press.co.uk/leu Intensive high-dose asparaginase consolidation improves survival for pediatric patients with T cell acute lymphoblastic leukemia and advanced stage lymphoblastic : a Pediatric Oncology Group study MD Amylon1, J Shuster2, J Pullen3, C Berard4, MP Link1, M Wharam5, J Katz13,AYu6, J Laver7, Y Ravindranath8, J Kurtzberg9, S Desai10, B Camitta11 and SB Murphy12

1Stanford University, Stanford, CA; 2Pediatric Oncology Group Statistical Office, Gainesville, FL; 3University of Mississippi Medical Center, Jackson, MS; 4St Jude Children’s, Memphis, TN; 5Johns Hopkins University, Baltimore, MD; 6University of California, San Diego, San Diego, CA; 7MUS Carolina, Charleston, SC; 8Children’s Hospital Michigan, Detroit, MI; 9Duke University, Durham, NC, USA; 10Cross Cancer Institute, Edmonton, AB, Canada; 11Midwest Children’s Cancer Center, Milwaukee, WI; 12Children’s Memorial Hospital and Pediatric Oncology Group Operations Office, Chicago, IL; and 13University of Texas, Southwestern, Dallas, TX, USA

This study was designed to test the hypothesis that high-dose blastic lymphoma usually have T lineage disease, and patients asparaginase consolidation therapy improves survival in pedi- with both of these disease entities are at high risk of treatment atric patients with T cell acute lymphoblastic leukemia and 26–34 advanced stage lymphoblastic lymphoma. Five hundred and failure. Initial efforts by the Pediatric Oncology Group fifty-two patients (357 patients with T cell acute lymphoblastic (POG) to improve survival in patients with these aggressive T leukemia (ALL) and 195 patients with advanced stage lym- cell lymphoid malignancies used modifications of the Mem- phoblastic lymphoma) were enrolled in POG study 8704 (T-3). orial Sloan Kettering LSA2L2 regimen. Improved survival was Treatment included rotating combinations of high-dose myelo- accomplished in comparison to historical experience, but the suppressive agents proven to be effective in T overall expectation for cure was still less than 50%.35–38 cell ALL in other POG group-wide or local institutional proto- cols (including , , , The malignant cells in T cell lymphoblastic lymphoma are prednisone, asparaginase, , and indistinguishable from the blasts of T cell ALL. Because of this ). After achieving a complete remission (CR), common biology, and many similar clinical characteristics, patients were randomized to receive or not receive high-dose the POG adopted the strategy in the mid 1980s of treating intensive asparaginase consolidation (25 000 IU/m2) given patients with T cell ALL and advanced stage lymphoblastic weekly for 20 weeks by . Intrathecal lymphoma on a common therapeutic protocol, separate from chemotherapy (, hydrocortisone and cytarabine) was given to prevent CNS disease, and CNS irradiation was patients with B lineage ALL disease. This ‘phenotype-specific’ used only for patients with leukemia and an initial WBC of treatment approach was designed to exploit the unique bio- Ͼ50 000/␮l or patients with active CNS disease at diagnosis. logic features of each diagnostic category in the hope that a CR was achieved in 96% of patients. The high-dose asparagin- superior outcome could be obtained in each subgroup. This ase regimen was significantly superior to the control regimen study was designed to test the hypothesis that high dose for both the leukemia and lymphoma subgroups. Four-year L-asparaginase consolidation improves survival for patients continuous complete remission rate (CCR) for the leukemia patients was 68% (s.e. 4%) with asparaginase as compared to with these T cell lymphoid malignancies. 55% (s.e. 4%) without. For the lymphoma patients, 4-year CCR was 78% (s.e. 5%) with asparaginase and 64% (s.e. 6%) in the controls. The overall one-sided logrank test had a P value Ͻ0.001 favoring asparaginase, while corresponding values Patients and methods were P = 0.002 for ALL and P = 0.048 lymphoblastic lymphoma. Toxicities were tolerable, but there were 18 failures due to sec- Patient selection ondary malignancies (16 with non-lymphocytic leukemia or myelodysplasia). Neither WBC at diagnosis (leukemia patients) nor lymphoma stage were major prognostic factors. We con- Patients with T cell ALL (confirmed by reference laboratories) clude that when added to a backbone of effective rotating were eligible if their age was at least 1 and not more than agents, repeated doses of asparaginase during early treatment 21 years, and if no prior therapy had been given except for improve the outcome for patients with T cell leukemia and emergency radiation (XRT) to the mediastinum (for respiratory advanced stage lymphoblastic lymphoma. Keywords: leukemia; lymphoma; pediatric; asparaginase; T cell embarrassment) or less than 1 week of vincristine and predni- sone. Patients with diffuse lymphoblastic lymphoma (confirmed by central pathology review) were eligible if they were not more than 21 years old, had Murphy stage III or Introduction IV disease,39 and had received no prior therapy except for emergency XRT to the mediastinum or less than 7 days of Lymphoid malignancies can be classified by cell surface anti- prednisone therapy (for respiratory embarrassment). Immuno- gen phenotype as derived from either T cell or B cell lineage. phenotyping was not available for most of the lymphoblastic T cell lymphoblastic malignancies have distinct biochemical, lymphoma patients, and therefore a small number from this immunologic and clinical features which set them apart from group may not have a T lineage phenotype. Informed consent non-T lymphoid malignancies.1–21 Acute lymphoblastic leuke- according to local institutional and federal regulatory guide- mia (ALL) of T cell lineage comprises 10–15% of all cases of lines was required prior to registration. Of the patients childhood ALL.22–25 Children with advanced stage lympho- accrued, 357 had T cell ALL and 195 had advanced stage lymphoblastic lymphoma. Patients were enrolled in the study between May 1987 and Correspondence: MD Amylon, Department of Pediatrics, MC: 5119, Division of Hematology/Oncology, Stanford University Medical January 1992. The cutoff for analysis was October 1995. This Center, Stanford, CA 94305–5119, USA; Fax 650 723 5231 date was the pre-planned selection to obtain 4 year follow- Received 16 June 1998; accepted 5 November 1998 up on nearly all patients. A complete patient accounting Asparaginase in T cell leukemia and lymphoma MD Amylon et al 336 Table 1 Patient accounting Table 3 Treatment regimen

T-ALL Lymphoblastic Induction: Vincristine 1.5 mg/m2 IVP weekly × 5 begin day 1 NHL Prednisone 40 mg/m2/day po Ϭ tid × 28 days Cyclophosphamide 1000 mg/m2 i.v. on day 1 2 Entered 357 195 Doxorubicin 50 mg/m i.v. on day 1 2 Eligible 345 180 Cytarabine 100 mg/m /day IVCI for 5 days begin Attained CRa 331 172 day 22 2 CR never randomizedb 14 5 Cyclophosphamide 600 mg/m i.v. on day 22 2 Randomized 317 167 L-asparaginase 10 000 U/m i.m. on days 27, 29 and 31

a For T-ALL the CR rate was 331/343 (96.5%) with two inevaluable Consolidation: Teniposide 300 mg/m2 i.v. twice weekly × 4 for response, while for lymphoblastic NHL, the CR rate was 172/178 doses begin day 43 (96.6%) with two inevaluable for response. Cytarabine 150 mg/m2 IVP twice weekly × 4 b Randomization was conducted on a callback to the automated doses after teniposide telephone registration system. While reminders were issued, the Vincristine 1.5 mg/m2 IVP weekly × 4 begin day variable time of count recovery following consolidation therapy con- 71 tributed to the loss of these 19 patients to the study question. Prednisone 40 mg/m2/day p.o. Ϭ tid × 28 days begin day 71 Doxorubicin 40 mg/m2 i.v. on day 71 Table 2 Patient demographics CNS Intrathecal triple drugs twice weekly × 7 begin T-ALL Lymphoblastic NHL prophylaxis: day 1, then q 9 weeks throughout maintenance Cranial XRT 2400 cGy begin day 71 for T-ALL Ͼ Control L-ASP Control L-ASP patients with WBC 50K only (n = 157) (n = 160) (n = 83) (n = 84) Maintenance: 9 week cycle repeated 10 times:

Females 33 42 18 28 Cytarabine 150 mg/m2/day IVCI × 3 days begin Males 124 118 65 56 day 1 Caucasian 117 118 65 62 Cyclophosphamide 75 mg/m2 IVP q 12 h × 6 African–American 19 21 10 12 doses begin day 1 Hispanic 13 13 5 5 Other 7 8 3 4 Vincristine 2 mg/m2 i.v. on day 22 Unknown 1 0 0 1 Doxorubicin 30 mg/m2 i.v. on day 22 Mediastinal mass 100 88 75 67 Prednisone 120 mg/m2/day p.o. ÷ tid × 5 days a Age 8.4 8.1 10.4 10.9 begin day 22 (5.5–11.6) (5.2–12.3) (5.7–15.0) (6.2–14.4) 6MP 225 mg/m2/day p.o. Ϭ tid × 5 days begin a WBC at Dx 66 82 day 22 (19–152) (20–220) Teniposide 300 mg/m2 i.v. q 3 days × 2 doses aMedian and quartiles. begin day 43 Cytarabine 150 mg/m2/dose IVP × 2 doses following teniposide appears in Table 1. Of the 552 patients enrolled, 27 failed Randomization: L-asparaginase 25 000 U/m2 i.m. weekly × 20 to meet the eligibility requirements (primarily because central doses begin day 99 review failed to confirm T cell ALL in leukemia patients or lymphoblastic histology in lymphoma patients), 22 did not achieve a documented complete remission as required for post CR randomization, and an additional 19 patients were system was accomplished with triple intrathecal medications lost for failure to call the automated telephone registration sys- (methotrexate, cytosine arabinoside and hydrocortisone) given tem for randomization. This report is based primarily on the seven times during induction and consolidation and once with 484 patients (317 T-ALL and 167 advanced lymphoblastic each continuation cycle. Additionally, leukemia patients with lymphoma) randomized after achieving a CR. Patient demo- an initial WBC greater than 50 000/␮l received 2400 cGy of graphics by treatment and diagnosis are presented in Table 2. cranial irradiation during consolidation. By current standards, this is a high dose, which was chosen because it was the stan- dard dose for POG protocols when this study was written in Treatment the mid 1980s. T cell patients are generally older than B pre- cursor patients, and thus would be more resistant to the toxic The treatment regimen (Table 3) was designed by merging into effects of cranial XRT. Leukemia patients with initial WBC less an intensive rotating backbone several different drug combi- than 50 000/␮l had been shown to be at decreased risk for nations extracted from Previous POG group-wide or local CNS failure in the POG T-2 study,22,38 and did not receive institution in-house protocols for T cell malignancies. Each radiation therapy unless they had CNS leukemia at diagnosis. combination chosen had been well documented to have All patients who had CNS disease at the time of initial diag- activity against newly diagnosed or relapsed T cell ALL or nosis received two additional doses of intrathecal medication lymphoblastic lymphoma.7,40–42 After an intensive induction during induction and consolidation and then received cranio- and consolidation, continuation consisted of 10 repetitions of spinal irradiation. All patients received three ‘standard’ doses a 9 week cycle with three different drug combinations. Total of L-asparaginase (Elspar; Merck, West Point, PA, USA) during duration of therapy was planned to be 2 years from the time induction (10 000 IU/m2), and those who achieved CR were of initiation of treatment. Prophylaxis of the central nervous randomized to receive or not to receive ‘high-dose’ L-asparag- Asparaginase in T cell leukemia and lymphoma MD Amylon et al 337 inase (25 000 IU/m2) i.m. weekly for 20 doses beginning con- treatment and diagnosis. As expected, the major site of relapse currently with the initiation of the continuation phase of ther- was the bone marrow, which was involved with 70/87 (36/45) apy. Patients who developed to E. coli L-asparaginase relapses in leukemia patients (control (asparaginase)), and in were switched to Erwinia asparaginase at the same dose. 16/31 (5/17) relapses for lymphoma patients. There was an Patients and/or their legal guardians signed informed consent excess of second malignancies (primarily acute non-lympho- according to FDA and local institutional guidelines prior to cytic leukemia or myelodysplastic syndrome) on the asparagi- the initiation of therapy. nase arm (15 vs three on the control arm). Remission deaths were rare (seven total, three on the control arm and four with asparaginase). Statistical considerations Treatment comparisons for continuous complete remission overall (most important) and within subgroups of special inter- The major endpoint for all comparisons is the duration of est are presented in Table 5. Power within the smaller sub- complete continuous remission (CCR): the time from ran- groups is very limited. Despite the problem of increased domization (which occurred after achievement of CR) to the second malignancy, the asparaginase arm had a significant earliest of relapse, second malignancy, death from any cause, advantage in CCR (P Ͻ 0.001 overall, P = 0.002 in T cell ALL, or last clinical contact. Failures included relapses, second and P = 0.048 in advanced stage lymphoblastic lymphoma). malignancies and deaths. Comparisons were conducted by The P value for second malignancy was 0.009 favoring the the logrank test, stratified where indicated. Complete continu- control arm. ous remission curves were constructed by the method of The prognostic significance of white cell count at diagnosis Kaplan and Meier,43 with standard errors of Peto and co-work- for the leukemia patients is further explored in Table 6. Over- ers.44 Comparisons of treatments for CCR are one-sided, as all, white cell count is only marginally significant in this study, designated in the protocol, as the management decision for a and there is no significant predictive value of white count at non-significant difference or for a significantly different result diagnosis in the experimental arm. WBC retains predictive sig- favoring control over L-asparaginase would be identical: drop- nificance in the control arm. Patients with extremely high ping L-asparaginase consolidation from future protocols. Prog- white cell counts (Ͼ300 000/␮l) fared surprisingly well on the nostic factors and time to second malignancy were analyzed asparaginase arm, with 4 year CCR of 58%. by two-sided logrank tests. To impute an approximate 4 year A prognostic significance for stage (stage III vs stage IV) event-free survival from 4 year CCR, the reader should multi- among the lymphoma patients was not observed in this study. ply the 4 year CCR by the CR rate (96.5%). The standard errors On the control arm, there were 26 failures (24.2 expected) are closely approximated by the standard error for CCR. among the 62 stage III patients as compared to eight failures Since the study was designed for overall outcome and not (9.8 expected) among the 21 stage IV patients (P = 0.50). On for outcome within subsets, the overall result should be the asparaginase arm, there were 14 failures (16.9 expected) inferred as holding in all subsets unless a significant difference among the 59 stage III patients as compared to 10 failures (7.1 in the opposite direction is observed. expected) among the 25 stage IV patients (P = 0.20). Overall, there were 40 failures (41.1 expected) among the 121 stage III patients as compared to 18 failures (16.9 expected) among Results the 46 stage IV patients (P = 0.75). No significant difference in outcome was seen. Induction therapy CNS involvement at diagnosis did not significantly correlate with prognosis in patients with leukemia or lymphoma. There As can be seen in Table 1, complete remission rates for both was no significant difference in either overall CCR or CNS T cell ALL and advanced stage lymphoblastic lymphoma were CCR when patients with CNS disease at diagnosis were com- over 96%. Induction failures were evenly distributed to early pared to those who were CNS negative. Age at diagnosis, sex, death (11) and resistant disease (11). presence of a mediastinal mass, cytogenetic abnormalities, and ethnicity likewise had no significant correlation with outcome. Continuous complete remission

Overall CCR curves are presented in Figure 1a, and by disease Toxicity (leukemia vs lymphoma) in Figure 1b. The 4 year CCR rates were as follows: for all patients, control 57.8% (s.e. = 3.4%) Toxicities were significant, but manageable. Allergic reactions and asparaginase 71.3% (s.e. = 3.3%); for T cell ALL patients, were more common on the asparaginase arm with 59/244 control 54.5% (s.e. = 4.1%) and asparaginase 67.9% (24%) patients reporting allergic reactions vs 24/240 (10%) on (s.e. = 4.1%); for lymphoblastic lymphoma patients, control the control arm. Almost all of these allergic reactions were 64.4% (s.e. = 6.0%) and asparaginase 78.0% (s.e. = 5.2%). local urticarial reactions to the asparaginase injections, and We are 95% confident that the advantage of the asparaginase resulted in a switch to the Erwinia asparaginase preparation. consolidation is at least 6.8% overall at 4 years. There were also allergic reactions reported in association with In previous POG T cell ALL studies, WBC subdivided at teniposide doses and antibiotics administered during febrile 50 000/␮l was the most significant prognostic indicator.22,35,38 neutropenic episodes. There were no fatal allergic reactions. CCR curves for T cell ALL patients in this report divided by Neurotoxicity was similar on the two arms with 28 patients WBC cut at 50 000/␮l (the cut point for inclusion of cranial (24 patients) reporting grade 3+ neurotoxicity on the asparagi- radiation therapy) are presented in Figure 2. In the asparagin- nase arm (control arm), respectively. One patient developed ase-treated patients, there is no evidence of a significant fatal leukoencephalopathy, but the other neurologic toxicities association with outcome. were predominantly single seizure episodes following intra- Table 4 provides an accounting of site-specific failures by thecal medications or peripheral neuropathy following vincri- Asparaginase in T cell leukemia and lymphoma MD Amylon et al 338

Figure 1 (a) Comparison of continuous complete remission rates for all randomized patients treated with (ASP) or without (Control) high- dose L-asparaginase consolidation. In the life table: P estimated % failure-free through the end of the interval; s.e., standard error of P: F, number of failures in interval; N, number at risk at start of interval. (P Ͻ 0.001). (b) Comparison of continuous complete remission rates for randomized patients with leukemia (T-ALL) and advanced lymphoblastic lymphoma (NHL) treated with (ASP) or without (Control) high-dose L-asparaginase consolidation. (For T-ALL P = 0.002; for NHL P = 0.048.) Asparaginase in T cell leukemia and lymphoma MD Amylon et al 339

Figure 2 Comparison of continuous complete remission rates for leukemia patients by white blood count (WBC) subdivided at 50 000/␮l treated with (ASP) or without (Control) high-dose L-asparaginase consolidation. (For WBC Ͻ50 000 P = 0.15; for WBC Ͼ50 000 P = 0.002.)

Table 4 Failures Table 5 Treatment comparisons

T-ALL Lymphoblastic Control L-ASP P NHL valueb N Fail Expected N Fail Expected Control L-ASP Control L-ASP Overall 240 124 99.8 244 82 106.2 Ͻ0.001 Patients 157 160 83 84 T-ALL 157 90 72.1 160 58 75.9 0.002 Total failures 90 58 34 24 Lymphoblastic 83 34 27.7 84 24 30.3 0.048 Relapses 87 45 31 17 NHL Marrow only 54 30 16 3 T-ALL(WBC Ͻ50) 66 31 27.2 65 22 25.8 0.15 Marrow + CNS 93 00 T-ALL (WBC 91 59 44.8 95 36 50.2 0.002 Marrow + testicular 3 1 0 0 Ͼ50) Marrow + mediastinum 4 0 0 2 Stage III NHL 62 26 19.2 59 14 20.8 0.015 Marrow + test + CNS 01 00 Stage IV NHL 21 8 8.8 25 10 9.2 0.86 Marrow + med + CNS 01 00 Femalesa 51 22 18.0 70 22 26.0 0.11 Isolated CNS 7 4 2 4 Malesa 189 102 82.1 174 60 79.9 0.002 Isolated testicular 4 0 0 0 Caucasiana 182 90 71.3 180 58 76.7 0.001 Mediastinal only 1 1 8 3 African– 29 14 11.4 33 11 13.6 0.15 Other extramedullary 5 4 5 5 Americana Second malignancy 2 8 1 7 Hispanica 18 11 11.1 18 9 8.9 0.52 Remission deaths 1 4 2 0 Unknown cause 0 1 0 0 aStratified for T-ALL vs lymphoblastic NHL. bOne-sided logrank test. Sixteen of the 18 second malignancies were ANLL or MDS. The two others were non-lymphoblastic . in therapy were most common after the continuous infusion stine administration. Deaths in CR were unusual, accounting cytarabine and cyclophosphamide combination in continu- for only five failures in the leukemia group (three bacterial ation. Other frequently reported toxicities are presented in sepsis, one Pneumocystis carinii pneumonia and one Table 7. There was no significant difference in the reported leukoencephalopathy) and two failures in the lymphoma incidence of neutropenia (36% control vs 38% experimental, group (one accidental fall from the roof of a building and one P = 0.16) or sepsis (1.6% control vs 2.3% experimental, bronchiolitis obliterans). P = 0.083). Thrombocytopenia (2.6% control vs 6.2% experi- Most patients tolerated therapy with minimal dose mental, P Ͻ 0.001), liver function abnormalities (transaminase reductions, although there were frequent delays due to slow elevation 0.9% control vs 2.6% experimental, P Ͻ 0.001; bili- count recovery (G-CSF was not used in this protocol). Delays rubin elevation 0.05% control vs 1.2% experimental, Asparaginase in T cell leukemia and lymphoma MD Amylon et al 340 Table 6 Prognostic significance of WBC survival in the Dana Farber Cancer Institute protocol for treat- ment of patients with T-ALL and advanced stage Control L-ASP lymphoblastic NHL.42,46 In most studies of T cell leukemia, a higher WBC at diag- N Fail Expecteda N Fail Expecteda nosis is an adverse prognostic factor.22,35,38,49 In the current study, with the addition of intensive asparaginase, WBC lost WBC Ͻ10 23 14 13.1 24 11 8.3 its prognostic significance. This may have been due in part to WBC 10–50 43 17 28.3 41 11 15.9 the lesser degree of improvement seen in the patients with a WBC 50–300 69 42 39.2 66 23 23.9 WBC Ͼ300 22 17 9.4 29 13 9.9 low WBC who received asparaginase. The only difference in the therapy between patients with a high or low initial WBC P values: Ͻ50 vs Ͼ50: P = 0.037 (overall), P = 0.008 (control), was the inclusion of cranial XRT for CNS prophylaxis in the P = 0.57 (L-ASP); Ͻ300 vs Ͼ300: P = 0.009 (overall), P = 0.008 high WBC group only. The possible role of XRT as a contribu- (control), P = 0.29 (L-ASP). ting factor in systemic control is of interest because of the aExpected to observed values are compared vertically. differential efficacy of this regimen in patients who received or did not receive cranial XRT in comparison to previous treat- ment regimens. CNS control was acceptable in all subgroups, Table 7 Toxicity comparison with CNS event-free survival greater than 90%. Secondary malignancies, predominantly non-lymphoid leu- Frequency of grade 3 or 4 12345 toxicity reported kemia and myelodysplasia, accounted for a significant subset of the treatment failures. These failures may be related to the Neutrophil count 125 42 41 667 783 use of the epipodophyllotoxin VM-26 in the treatment regi- Platelet count 42 10 9 49 129 men. The time to development of the secondary malignancies Bacterial sepsis 35 3 7 29 48 (14 months to 4 years from original diagnosis) and the occur- Liver transaminases 7 3 7 17 53 rence of cytogenetic abnormalities at chromosome band Total bilirubin 0 0 0 1 24 11q23 in seven patients suggests topoisomerase II inhibitor- Serum amylase 1 0 1 2 17 related leukemogenesis.50–54 have been impli- cated in a similar pattern of leukemogenesis, and might also 1, induction (519 patients); 2, consolidation (control arm) (240 have contributed to these failures. The failures in our study, patients); 3, consolidation (asparaginase arm) (244 patients); 4, maintenance (control arm) (1865 cycles); 5, maintenance however, are not suggestive of alkylator-induced secondary (asparaginase arm) (2066 cycles). malignancies which generally have a longer latent period and cytogenetic abnormalities of chromosome 5 or 7.55–57 The use of CNS irradiation had no detectable impact on the develop- P Ͻ 0.001) and (amylase elevation 0.10% control ment of secondary malignancies. vs 0.82% experimental, P Ͻ 0.001) were more commonly Of interest, the development of second malignancy as a reported on the asparaginase arm. There was no reported cause of failure was significantly higher in the experimental grade 3 or 4 or thrombotic events. As previously arm (P = 0.009, two-sided, see Table 4). This was not a pro- mentioned, second malignancies also occurred more fre- spective hypothesis, however, and should be interpreted with quently on the intensive asparaginase regimen,45 but nonethe- caution. Additional study of this phenomenon is ongoing, and less the outcome is superior. will be the subject of a future report. The mechanism of the apparent contribution of the asparag- inase consolidation to secondary leukemia is not understood, Discussion but may involve an interaction with the epipodophyllotoxin VM-26, perhaps by limiting the ability of cells to repair DNA T cell lymphoblastic malignancies share a distinct clinical, damage by inhibiting synthesis of repair . The second biochemical, histochemical, cytogenetic and surface antigen malignancies were all counted as failures in the analysis, how- phenotype which distinguishes the blast cells from B precursor ever, and the experimental arm still had a superior outcome. lymphoid malignancies.1–21 Traditionally, the presence of a T Thus the apparent advantage of the high-dose asparaginase cell phenotype has been a poor prognostic feature in acute consolidation outweighed the risk of second malignancy in lymphoblastic leukemia in children, and patients with this study. Limiting the use of epipodophyllotoxin in future advanced stage lymphoblastic lymphoma have also been at treatment regimens may decrease the risk of second malig- high risk for relapse.26–34 However, intensified therapy has nancy acute non-lymphocytic leukemia, and further improve improved the outlook for patients with T cell ALL and the overall outcome. advanced stage lymphoblastic lymphoma.35–38 T cell ALL Two statistical points are worth mentioning. First, the strat- patients have fared progressively better with each generation egy of randomizing after achievement of a CR caused the loss of POG protocols, with a 4 year CCR of 65% on the asparagin- of 19 patients (who achieved a CR, but the institution failed ase arm of the current study. Non-POG regimens also have to randomize them) and 22 patients who never achieved a resulted in improved survival for patients with T-ALL and documented CR. On balance, if randomization had occurred advanced stage lymphoblastic NHL.46–48 up front, the outcome of those patients not achieving a CR This study demonstrated a significant improvement in would have been ‘pure noise,’ since they failed before the event-free survival (EFS) for patients who received high-dose divergence of the therapies. From a power perspective, the intensive asparaginase consolidation added to an effective rot- exclusion of 22 uninformative failures is a much greater gain ating backbone. This was particularly striking in the leukemia than the exclusion of 19 CR patients prior to randomization patients with an initial WBC Ͼ50 000/␮l, but also was sig- is a loss. The second issue deals with the fact that we have 2 nificant in the patients with lymphoma. Early intensive aspara- additional years of follow-up beyond the October 1995 cutoff. ginase is felt to be a key element contributing to improved Although an analysis at the October 1997 cutoff was not a Asparaginase in T cell leukemia and lymphoma MD Amylon et al 341 planned one, it is important to note that the overall con- antibody characterization of surface antigens in childhood T-cell clusions on the basis of the most current information are lymphoid malignancies. Blood 1983; 61: 830–837. unchanged. There were only six additional failures in these 2 11 Crist WM, Grossi CE, Pullen DJ, Cooper MD. Immunologic mark- ers in childhood acute lymphocytic leukemia. Semin Oncol 1985; years (two on the control arm and four on the experimental Ͻ 12: 105–121. arm). The overall comparison has P 0.001, within T-ALL 12 Borowitz MJ, Dowell BL, Boyett JM, Falletta JM, Pullen DJ, Crist P = 0.002, and within NHL P = 0.046 all favoring the experi- WM, Humphrey GB, Metzgar RS. Monoclonal antibody definition mental arm. of T-cell acute leukemia. A Pediatric Oncology Group study. The ultimate goal of dividing treatment groups based on Blood 1985; 65: 785–788. lymphoid lineage is that truly lineage-specific therapeutic 13 Borowitz MJ, Dowell BL, Boyett JM, Pullen DJ, Crist WM, Quddus FM, Falletta JM, Metzgar RS. Clinicopathologic aspects of E-rosette interventions will be devised. In the absence of truly T-specific negative T-cell acute lymphocytic leukemia. A Pediatric Oncology therapy, however, the unique biological features of the T cell Group study. J Clin Oncol 1986; 4: 170–177. blasts may result in patterns of responsiveness that differ from 14 Raimondi SC, Behm FG, Robertson PK, Pui CH, Rivera GK, Mur- other phenotypes. The high-dose intensive asparaginase con- phy SB, Williams DL. Cytogenetics of childhood T-cell leukemia. solidation which significantly improved survival in this study Blood 1988; 72: 1560–1566. had no significant impact in a POG study of B precursor 15 Carroll AJ, Crist WM, Link MP, Amylon MD, Pullen DJ, Ragab ALL.58 Therefore, even though more intensive treatment regi- AH, Buchanan GR, Wimmer RS, Vietti TJ. The t(1;14)(p34;q11) is nonrandom and restricted to T-cell acute lymphoblastic leukemia: mens have eliminated to a large extent the prognostic signifi- a Pediatric Oncology Group study. Blood 1990; 76: 1220–1224. cance of the T cell phenotype in childhood, it will remain 16 Bash RO, Crist WM, Shuster JJ, Link MP, Amylon M, Pullen J, important to track the results of treatment interventions by Carroll AJ, Buchanan GR, Smith RG, Baer R. Clinical features and phenotype in future trials. The use of intensive asparaginase outcome of T-cell acute lymphoblastic leukemia in childhood consolidation should be considered in future treatment with respect to alterations at the TAL1 locus. A Pediatric Oncology regimens for T cell lymphoid malignancies. Group study. Blood 1993; 81: 2110–2117. 17 Raimondi SC. Current status of cytogenetic research in childhood acute lymphoblastic leukemia. Blood 1993; 81: 2237–2251. 18 Steinherz PG, Siegel SE, Bleyer WA, Kersey J, Chard R Jr, Coccia Acknowledgements P, Keikin S, Lukens J, Neerhout R, Nesbit M, Miller DR, Reaman G, Sather H, Hammond D. Lymphomatous presentation of child- hood acute lymphoblastic leukemia. Cancer 1991; 68: 751–758. This work was supported in part by the following grants from 19 Vannier JP. Childhood non-Hodgkin’s lymphomas: clinical and the National Cancer Institute; CA-33603, CA-29139, CA- therapeutic aspects. Biomed Pharmacother 1988; 42: 263–271. 69428, CA-31566, CA-28476, CA-28439, CA-69177, CA- 20 Amylon MD, Link MP, Murphy SB. Malignant T-cell and other 29691, CA-15525, CA-32053, CA-30969. non-Hodgkin’s lymphoma of childhood. In: Voute PA, Barrett A (eds). Cancer in Children, 2nd edn. Springer Verlag: New York, 1986, pp 152–163. 21 Crist WM, Shuster JJ, Falletta J, Pullen DJ, Berard CW, Vietti TJ, References Alvarado CS, Roper MA, Prasthofer E, Grossi CE. Clinical features and outcome in childhood T-cell leukemia–lymphoma according 1 Tsukimoto I, Wong KY, Lampkin BC. Surface markers and prog- to stage of thymocyte differentiation: a Pediatric Oncology Group nostic factors in acute lymphoblastic leukemia. New Engl J Med study. Blood 1988; 72: 1891–1897. 1975; 292: 828–832. 22 Shuster JJ, Falletta JM, Pullen DJ, Crist WM, Humphrey GB, Dow- 2 Murphy SB, Aur RJ, Simone JV, George S, Mauer AM. Pretreatment ell BL, Wharam MD, Borowitz M. Prognostic factors in childhood cytokinetic studies in 94 children with acute leukemia: relation- T-cell acute lymphoblastic leukemia: a Pediatric Oncology Group ship to other variables at diagnosis and to outcome of standard study. Blood 1990; 75: 166–173. treatment. Blood 1977; 49: 683–691. 23 Borowitz M. Immunologic markers in childhood acute lympho- 3 Gupta S, Good RA. Markers of human lymphocyte subpopulations blastic leukemia. Hematol Oncol Clin North Am 1990; 4: 743– in primary immunodeficiency and lymphoproliferative disorders. 752. Semin Hematol 1980; 17: 1–29. 24 Crist W, Shuster J, Look T, Borowitz M, Behm F, Bowman P, Fran- 4 Frei E III, Schnabel FM Jr, Goldin A. Comparative chemotherapy kel L, Pullen J, Krance R, Steuber P, Camitta B, Amylon M, Link of AKR lymphoma and human hematological neoplasia. Cancer M, Land V. Current results of studies of immunophenotype-, age-, Res 1974; 34: 184–193. and leukocyte-based therapy for children with acute lymphoblas- 5 Ohnuma T, Arkin H, Minowada J, Holland JF. Differential chemo- tic leukemia. Leukemia 1992; 6: 162–167. therapeutic susceptibility of human T-lymphocytes and B-lympho- 25 Pui CH, Behm FG, Crist WM. Clinical and biological relevance cytes in culture. J Natl Cancer Inst 1978; 60: 749–752. of immunologic marker studies in childhood acute lymphoblastic 6 Verhoef V, Fridland A. Metabolic basis of arabinonucleoside sel- leukemia. Blood 1993; 82: 343–362. ectivity for human leukemic T- and B-lymphoblasts. Cancer Res 26 Sen L, Borella L. Clinical importance of lymphoblasts with T mark- 1985; 45: 3646–3650. ers in childhood acute leukemia. New Engl J Med 1975; 292: 7 Lauer SJ, Pinkel D, Buchanan GR, Sartain P, Cornet JM, Krance R, 828–832. Borella LD, Casper JT, Kun LE, Hoffman RG, Camitta BM. Cytosine 27 Coccia PF, Kersey JH, Gajil-Peczalska KJ, Krivit W, Nesbit ME. arabinoside/cyclophosphamide pulses during continuation ther- Prognostic significance of surface marker analysis in childhood apy for childhood acute lymphoblastic leukemia: potential selec- non-Hodgkin’s lymphoproliferative malignancies. Am J Hematol tive effect in T-cell leukemia. Cancer 1987; 60: 2366–2371. 1976; 1: 405–417. 8 Quddus FF, Leventhal BG, Boyett JM, Pullen DJ, Crist WM, Boro- 28 Belpomme D, Mathe´ G. Clinical significance and prognostic value witz MJ. Glucocorticoid receptors in immunological subtypes of of the T-B immunological classification of the human primary childhood acute lymphoblastic leukemia cells: a Pediatric acute lymphoid . Lancet 1977; 1: 555–558. Oncology Group study. Cancer Res 1985; 45: 6482–6486. 29 Heideman RL, Falletta JM, Mukhopadhayay MS, Fernbach DJ. 9 Pieters R, Kaspers GJ, Van Wering ER, Huismans DR, Loonan AH, Lymphocytic leukemia in children: prognostic significance of Hahlen K, Veerman AJ. Cellular drug resistance profiles that might clinical and laboratory findings at time of diagnosis. J Pediatr explain the prognostic value of immunophenotype and age in 1978; 92: 540–545. childhood acute lymphoblastic leukemia. Leukemia 1993; 7: 30 Sallan SE, Ritz J, Pesandro J, Gelber R, O’Brien C, Hitchcock S, 392–397. Coral F, Schlossman SF. Cell surface antigens: prognostic impli- 10 Roper M, Crist WM, Metzgar R, Ragab AH, Smith S, Starling K, cations in childhood acute lymphoblastic leukemia. Blood 1980; Pullen J, Leventhal B, Bartolucci AA, Cooper MD. Monoclonal 55: 395–402. Asparaginase in T cell leukemia and lymphoma MD Amylon et al 342 31 Greaves MF, Janossy G, Peto J, Kay H. Immunologically defined Mantel N, McPherson K, Peto J, Smith PG. Design and analysis subclasses of acute lymphoblastic leukemia in children: their of randomized clinical trials requiring prolonged observation of relationship to presentation features and prognosis. Br J Haematol each patient. Br J Cancer 1977; 35: 1–39. 1981; 48: 179–197. 45 Amylon MD, Carroll AJ, Link MP, Katz JA, Shuster JJ. Second 32 Pullen DJ, Boyett JM, Crist WM, Falletta JM, Roper M, Dowell B, malignancies in children treated with teniposide (VM-26) for T- Van Eys J, Jackson JF, Humphrey GB, Metzgar RS, Cooper MD. cell lymphoid malignancy: a role for asparaginase? (A Pediatric Pediatric Oncology Group utilization of immunologic markers in Oncology Group study). Blood 1992; 80 (Suppl. 1): 206a. the designation of acute lymphocytic leukemia subgroups: influ- 46 Schorin MA, Blattner S, Gelber RD, Tarbell NJ, Donnelly M, ence on treatment response. Ann NY Acad Sci 1984; 428: 26–48. Dalton V, Cohen HJ, Sallan SE. Treatment of childhood acute 33 Crist WM, Grossi CE, Pullen DJ, Cooper MD. Immunologic mark- lymphoblastic leukemia: results of a Dana-Farber Cancer ers in childhood acute lymphocytic leukemia. Semin Oncol 1985; Institute/Children’s Hospital acute lymphoblastic leukemia con- 12: 105–121. sortium protocol 85–01. J Clin Oncol 1994; 12: 740–747. 34 Pui CH, Behm FG, Singh B, Schell MJ, Williams DL, Rivera GK, 47 Reiter A, Schrappe M, Ludwig WD, Hiddemann W, Sauter S, Kalwinsky DK, Sandlund JT, Crist WM, Raimondi SC. Heterogen- Henze G, Zimmermann M, Lampert F, Havers W, Niethammer D, eity of presenting features and their relation to treatment outcome Odenwald E, Ritter J, Mann G, Welte K, Gadner H, Riehm H. in 120 children with T-cell acute lymphoblastic leukemia. Blood Chemotherapy in 998 unselected childhood acute lymphoblastic 1990; 75: 174–179. leukemia patients, results and conclusions of the multicenter trial 35 Pullen DJ, Sullivan MP, Falletta JM, Boyett JM, Humphrey GB, ALL-BFM 86. Blood 1994; 84: 3122–3133. Starling KA, Land VJ, Dyment PG, Vats T, Duncan MH. Modified 48 Tubergen DG, Krailo MD, Meadows AT, Rosenstock J, Kadin M, LSA2-L2 treatment in 53 children with E-rosette positive T-cell leu- Morse M, King D, Steinherz PG, Kersey JH, Comparison of treat- kemia: results and prognostic factors (a Pediatric Oncology Group ment regimens for pediatric lymphoblastic non-Hodgkin’s lym- study). Blood 1982; 60: 1159–1168. phoma: a Childrens Cancer Group study. J Clin Oncol 1995; 13: 36 Sullivan MP, Boyett J, Pullen J, Crist W, Doering EJ, Trueworthy 1368–1376. R, Hvizdala E, Ruymann F, Steuber CP. Pediatric Oncology Group 49 Pui CH, Crist WM. Biology and treatment of acute lymphoblastic experience with modified LSA2-L2 therapy in 107 children with leukemia. J Pediatr 1994; 124: 491–503. non-Hodgkin’s lymphoma (Burkitt’s lymphoma excluded). Cancer 50 Albain KS, LeBeau MM, Ullirsch R, Schumacher H. Implication 1985; 55: 323–336. of prior treatment with drug combinations including inhibitors of 37 Hvizdala EV, Berard C, Callihan T, Falletta J, Sabio H, Shuster JJ, topoisomerase II in therapy related monocytic leukemia with a 9; Sullivan M, Wharam MD. Lymphoblastic lymphoma in children – 11 translocation. Genes Chromosomes Cancer 1990; 2: 53–58. a randomized trial comparing LSA2-L2 with the ACOP+ thera- 51 Whitlock JA, Greer JP, Lukens JN. Epipodophyllotoxin-related leu- peutic regimen: a Pediatric Oncology Group study. J Clin Oncol kemia. Identification of a new subset of secondary leukemia. Can- 1988; 6: 26–33. cer 1991; 68: 600–604. 38 Falletta JM, Shuster JJ, Crist WM, Pullen DJ, Borowitz MJ, Wharam 52 Pui CH, Ribeiro RC, Hancock ML, Rivera GK, Evans WE, Rai- M, Patterson R, Foreman E, Vietti TJ. Different patterns of relapse mondi SC, Head DR, Behm FG, Mahmoud MH, Sandlund JT, Crist associated with three intensive treatment regimens for pediatric E- WM. Acute myeloid leukemia in children treated with epipodo- rosette positive T-cell leukemia: a Pediatric Oncology Group phyllotoxins for acute lymphoblastic leukemia. New Engl J Med study. Leukemia 1992; 6: 541–546. 1991; 325: 1682–1687. 39 Murphy SB. Classification, staging and end results of treatment of 53 Ratain MJ, Rowley JD. Therapy-related acute myeloid leukemia childhood non-Hodgkin’s lymphomas: dissimilarities from lym- secondary to inhibitors of topoisomerase II: from the bedside to phomas in adults. Semin Oncol 1980; 7: 332–339. the target genes. Ann Oncol 1992; 3: 107–111. 40 Steuber CP, Levy GJ, Nix WL, Shepherd DA, Starling KA, Fernbach 54 Hunger SP, Tkachuk DC, Amylon MD, Link MP, Carroll AJ, Wel- DJ. Use of L-asparaginase and cytosine arabinoside for refractory born JL, Cleary ML. HRX involvement in de novo and secondary acute lymphocytic leukemia with particular reference to T-cell leukemias with diverse chromosome 11q23 abnormalities. Blood leukemia. Med Pediatr Oncol 1978; 5: 33–38. 1993; 81: 3197–3203. 41 Dahl GV, Rivera G, Pui CH, Mirro J Jr, Ochs J, Kalwinsky DK, 55 Grunwald HW, Rosner F. Acute myeloid leukemia following treat- Abromowitch M, Look AT, Murphy SB. A novel treatment of child- ment of Hodgkin’s disease. Cancer 1982; 50: 676–683. hood lymphoblastic non-Hodgkin’s lymphoma: early and intermit- 56 Pedersen-Bjergaard J, Larsen SO. Incidence of acute non-lympho- tent use of teniposide plus cytarabine. Blood 1985; 66: 1110– cytic leukemia, preleukemia, and acute myeloproliferative syn- 1114. drome up to 10 years after treatment of Hodgkin’s disease. New 42 Clavell LA, Gelber RD, Cohen HJ, Hitchcock-Bryan S, Cassady JR, Engl J Med 1982; 307: 965–971. Tarbell NJ, Blattner SR, Tantravaki R, Leavitt P, Sallan SE. Four 57 Tucker MA, Coleman CN, Cox RS, Varghese A, Rosenberg SA. agent induction and intensive asparaginase therapy for treatment Risk of second cancers after treatment for Hodgkin’s disease. New of childhood acute lymphoblastic leukemia. New Engl J Med Engl J Med 1988; 318: 76–81. 1986; 315: 657–663. 58 Harris MB, Shuster JJ, Pullen DJ, Borowitz MJ, Carroll AJ, Behm 43 Kaplan EL, Meier P. Nonparametric estimation from incomplete FG, Land VJ. Consolidation with anti-metabolite based therapy in observations. J Am Stat Assoc 1958; 53: 457–481. standard risk acute lymphocytic leukemia: a Pediatric Oncology 44 Peto R, Pike MC, Armitage P, Breslow NE, Cox DR, Howard SV, Group study. J Clin Oncol 1998; 16: 2840–2847.