¹Divisions of Pediatric Hematology Oncology, Medical University of South Carolina, Charleston, SC, USA

²Stanford University, Stanford, CA, USA

³University of Massachusetts, Worcester, MA, USA

⁴Duke University, Durham, NC, USA

⁵Midwest Children's Cancer Center, Milwaukee, WI, USA

⁶University of Mississippi, Jackson, MS, USA

⁷Johns Hopkins Hospital, Baltimore, MD, USA

⁸Children's Hospital of Michigan, Detroit, MI, USA

⁹Children's Memorial Hospital, Chicago, IL, USA

¹⁰Pediatric Oncology Group Statistical Office, University of Florida, Gainesville, FL, USA

Correspondence to: Dr J H Laver, Division of Pediatric Hematology Oncology, Medical University of South Carolina, 171 Ashley Avenue, Charleston, SC 29425-3311, USA

contemporary chemotherapy has significantly improved event-free survival among patients with t cell-lineage acute lymphoblastic leukemia (t-all). unlike b-precursor all, most investigators are still using cranial radiation (crt) and are hesitant to rely solely on intrathecal therapy for t-all. in this study we assessed the effects of crt upon event-free survival and central nervous system (cns) relapses in a cohort of children with high risk features of t cell leukemia. in a series of six consecutive studies (1987-1995) patients were non-randomly assigned their cns prophylaxis per individual protocol. these protocols were based on pog 8704 which relied on rotating drug combinations (cytarabine/cyclophosphamide, teniposide/ ara-c, and vincristine/doxorubicin/6-mp/prednisone) post-induction. modifications such as high-dose cytarabine, intermediate-dose methotrexate, and the addition of g-csf, were designed to give higher cns drug levels (decreasing the need for crt), to eliminate epidophyllotoxin (decreasing the risk of secondary leukemia), and to reduce therapy-related neutropenia (pilot studies pog 9086, 9295, 9296, 9297, 9398). all patients included in this analysis qualified for pog high risk criteria, wbc >50 000/mm³ and/or CNS leukemia. Patients without CNS involvement received 16 doses of age-adjusted triple intra-thecal therapy (TIT = hydrocortisone, MTX, and cytarabine) whereas patients with CNS disease received three more doses of TIT during induction and consolidation. Patients who received CRT were treated with 2400 cGy (POG 8704) or 1800 cGy (POG 9086 and 9295). CNS therapy included CRT in 144 patients while the remaining 78 patients received no radiation by original protocol design. There were 155 males and 57 females with a median age of 8.2 years. The median WBC for the CRT+ and CRT- patients were 186 000/mm³ and 200 000/mm³, respectively. CNS involvement at diagnosis was seen in 16% of the CRT+ and 23% of the CRT- groups. The complete continuous remission rate (CCR) was not significantly different for the irradiated vs non-irradiated groups (P = 0.46). The 3-year event-free survival was 65% (s.e. 6%) and 63% (s.e. 4%) for the non-irradiated vs the radiated group. However, the 3-year CNS relapse rate was significantly higher amongst patients who did not receive CRT; 18% (s.e. 5%) vs 7% (s.e. 3%) in the irradiated group (P = 0.012). Our analysis in a non-randomized setting, suggests that CRT did not significantly correlate with event-free survival but omitting it had an adverse effect on the CNS involvement at the time of relapse. Leukemia (2000) 14, 369-373.

childhood acute lymphoblastic leukemia; T cell leukemia; cranial radiation

Introduction

Over the last two decades the Pediatric Oncology Group (POG) has adopted the strategy of treating T cell acute lymphoblastic leukemia on separate treatment protocols from other forms of childhood ALL.¹ This approach resulted in an improved event-free survival (EFS) for this cohort of patients sharing distinct biological properties.²

Unlike B-precursor ALL, prognostic factors for T-ALL are not well defined and cooperative groups have used different criteria when assigning patients on study by risk category.³ A previous POG analysis demonstrated that patients with T-ALL and WBC >50 000/mm³ had the poorest outcome.⁴ Based on these data, the Group defined high risk T-ALL (HR T-ALL) as WBC >50 000/mm³ and/or CNS involvement and stratified patients on its studies accordingly.

Whereas for B-precursor ALL, extended intra-thecal therapy can be substituted for cranial irradiation,⁵ most investigators are still hesitant to omit CRT for T-ALL. Published studies addressing the issue of CRT for T-ALL have been non-randomized and used different risk groups.^6,7 A recent study suggested that CRT may not be necessary for T-ALL patients presenting with a WBC <100 000/mm³ whereas in patients with higher WBC, omitting it had an adverse effect on event-free survival.⁶ Others have shown that in T-ALL patients receiving CRT, other risk factors such as FAB L2 morphology can significantly affect outcome.⁷ In another study, which included patients with T-ALL receiving CRT, good in vivo steroid response, correlated with a better outcome.⁸

Using the POG risk criteria we analyzed the effects of CRT in six non-randomized consecutive studies involving a cohort of high risk T-ALL patients. Our data suggest that cranial radiation does not significantly affect the overall event-free survival rate but omission of radiation is associated with a higher rate of CNS involvement at relapse.

Patients and methods

Patients

T cell acute lymphoblastic leukemia was diagnosed by morphologic criteria of the bone marrow (FAB)⁹ and the expression of T cell markers by marrow lymphoblasts¹⁰ (confirmed by POG reference laboratories). Patients between 1 and 21.99 years of age with previously untreated disease were eligible for study after an IRB-approved informed consent was obtained. Patients were enrolled on study between May 1987 and December 1995 and their characteristics are shown in Table 1. All the patients included in this analysis had high risk features with WBC >50 000/mm³ and/or CNS disease. CNS involvement at diagnosis was defined as more than five mononuclear cells/mm³ on a chamber count and presence of blasts in a cytospin preparation.

Response criteria

Complete remission was defined as no physical signs of leukemia, no detectable leukemic blasts in the peripheral blood smear, a bone marrow with active hematopoiesis with <5% blasts, and normal cerebrospinal fluid.

Treatment programs

In a series of six consecutive studies, patients were non-randomly assigned their CNS prophylaxis per individual protocol. All the studies were based on POG 8704 which opened in May 1987 and relied on rotating drug combinations (cytarabine/cyclophosphamide, teniposide/cytarabine, and vincristine/doxorubicin/6-mercaptopurine/prednisone) after a six-drug induction phase. In an attempt to improve outcome and evaluate toxicities, substitutions including higher dose cytarabine (hiDAC) for standard cytarabine, PEG L-asparaginase for native L-asparaginase, intermediate dose methotrexate/6-mercaptopurine (IDM/6MP) for teniposide/ cytarabine, hiDAC for standard cytarabine combined with IDM/6MP for teniposide/cytarabine, and the addition of G-CSF were tested in a series of pilot trials; POG 9086, 9295, 9296, 9297, 9398, respectively. All protocols consisted of induction and consolidation phases followed by 10 cycles of continuation therapy (Table 2). The cut off date for analysis of outcomes was 21 April 1998.

Statistical methods

The primary analysis of this study is the comparison of the CNS remission duration for the two groups (CRT+ vs CRT-). Prior to looking at any data, this variable was selected because of its direct biological intent, namely to prevent CNS relapses. This variable was selected over isolated CNS relapses, as it is consistent with conservative medical practice to ascribe responsibility to the treatment for all questionable cases. In addition, isolated CNS relapses usually are analyzed as if the joint CNS-other site relapses are uninformed censored data, making the presumption that the censoring event is statistically unrelated to the index event. This assumption is clearly inappropriate.

We also analyzed the overall event-free survival. Since all of the treatments had the same induction, and since one study (8704) randomized patients after achieving a remission, we actually compared the treatments from the time of achieving a complete remission (the overall CR rate was 96.7%). Failures include relapses, remission deaths, and second malignancies. This variable is important, but since CNS relapses represent a minority of overall events, and since this study was relatively small, the power to detect real differences in overall events was expected to be lower. The reader is cautioned in several areas. First, this is a non-randomized study, and as such, one cannot definitively conclude cause-effect relationships from these data. They should be interpreted as contributory to a growing body of evidence on the subject of CRT in high risk T-ALL. Second, one should view non-significant results as inconclusive. While zero difference is a plausible value, so are clinically important differences. Finally, as all studies whose primary endpoint is site-specific failures, there is no way to discount the possibility of interaction amongst treatment, other causes of failure and the natural history of disease.

Treatment comparisons were conducted by the one-sided logrank test,¹¹ which compares the differences of the entire remission curves. Since this was a non-randomized study, one-sided post-stratified logrank tests were also conducted to see if the results held up, when adjusted for potential prognostic factors such as age, white count, gender, and CNS involvement at diagnosis. Remission curves were constructed by the method of Kaplan-Meier,¹² with standard errors of Peto and colleagues.¹¹

Results

The study included 222 high risk T-ALL patients, of whom 144 received cranial radiation and 78 did not receive radiation by original protocol design. Although these data were pooled from different POG studies they were all based on the 8704 treatment schema with substitutions which we believe are comparable for the purpose of evaluating the effects of CRT in this cohort of patients. Event-free survival rates for the individual studies were quite similar; P values comparing the EFS for complete continuous remission were 0.55 and 0.21 within the CRT+ and CRT- protocols, respectively.

Figure 1 shows the continuous complete remission rate (CCR). The 3-year CCR rates in the irradiated and non-irradiated groups are 65% (s.e. 6%) and 63% (s.e. 4%), respectively (P = 0.46). If CNS events are censored, the P value is 0.16. Figure 2 depicts the CNS remission status of these patients. It includes all CNS treatment failures and is not restricted to isolated CNS relapses (ie a patient with a CNS relapse in addition to a marrow relapse was included in the CNS failure group). Patients who received CRT and those who did not receive CRT had 3-year CNS remission rates of 93% (s.e. 3%) and 82% (s.e. 5%), respectively (P = 0.012). Thus, although the overall CCR rate does not significantly correlate with radiation the overall number of CNS events is significantly increased by omitting radiation. Assessment of the isolated CNS relapse rates rather than overall CNS events, shows no significant difference between CRT+ and CRT- groups but the data favored CRT (data not shown).

There were 18 and 23 patients with CNS involvement at diagnosis in the CRT- and CRT+ groups, respectively. Eight patients failed in the CRT- group and six in the CRT+ (P = 0.17, two sided). However, due to the small number of events the power of the test is very limited.

We re-analyzed the data, using a post-stratified analysis, stratifying in turn for each of the following: age (subdivided at 10.0 years), gender, white count at diagnosis (subdivided at 100 000/mm³), and CNS involvement at diagnosis. In each case, the overall results held up with significance for CNS remission at P < 0.05, but no significant difference in overall remission duration. Numbers were too small to report other specific subset results.

Discussion

CNS prophylaxis with intensified intrathecal chemotherapy and cranial radiation, has contributed significantly to the increased survival of children with ALL.¹³ This combination although highly effective, has resulted in acute and late effects including increased risk of CNS tumors, as well as impaired intellectual function and neuropsychologic development.^14,15 Attempts to ameliorate such morbidity have led investigators to explore CNS treatments without radiation therapy. While CRT as pre-symptomatic therapy could be eliminated for B-lineage ALL, most investigators are reluctant to omit it for high risk T-lineage ALL.^16,17,18

This manuscript reports a series of POG protocols which incorporated intensified chemotherapy in a group of patients with higher risk T-ALL. Our intent was to test different drug combinations that could improve outcome while minimizing long-term side-effects such as secondary leukemia and late radiation toxicity. Some of these treatment programs included high-dose cytarabine, intermediate-dose methotrexate, and intravenous 6-MP in doses expected to give increased CNS drug levels.^19,20 These strategies might allow omission of CRT. However, the overall number of CNS events was significantly increased by omitting radiation despite the lack of effect on EFS. One possible explanation for these results is that the CNS drug penetration achieved in our protocols may not be adequate to eradicate occult CNS disease. This explanation is supported by an earlier study of standard risk ALL. In that study intermediate-dose methotrexate resulted in a lower risk of both systemic and testicular relapse but was inferior to cranial radiation in protecting against CNS events.²¹

Our observation of increased CNS events in the non-CRT group without significant difference in EFS raises the possibility of increased non-CNS events in the CRT cohort. Therefore, we re-analyzed the data by censoring all CNS events. Although there was an apparent excess in marrow relapse in the CRT group, there was no significant difference in the non-CNS events between the two groups. We also experienced a relatively high rate of secondary acute myelocytic leukemia (seven cases) most probably due to the intense etoposide treatment.²

Although omission of CRT did not adversely affect EFS, this approach might affect salvage therapy. Currently 40-80% of patients with CNS relapse can be cured with aggressive chemotherapy and delayed cranial or cranial-spinal irradiation.²² The majority of these patients had no radiation as their primary treatment. Patients with CNS relapse who have been irradiated might not be able to receive adequate doses of radiation at the time of salvage. In addition they may be at higher risk of late effects following radiation. However, the increased number of CNS events in the CRT- group might decrease the effectiveness of bone marrow transplantation as salvage therapy.²³

A recent report indicated that CRT may be required for T-ALL patients with a WBC count of over 100 000/mm^3,6 to improve EFS. The apparent discrepancy between this study and ours could not be entirely explained by different cut off points defining risk groups. We used a cut off point of 50 000/mm³ based on a previous study showing that CRT was not needed for patients with WBC <50 000/mm³.²⁴ Re-analyzing our current data using a WBC cut off of 100 000/mm³ did not change our conclusions.

Event-free survival in other reports of T-ALL treatment seems to demonstrate a slightly better outcome as compared to ours.^18,25,26 However, comparing these series raises significant statistical problems,²⁷ especially in the context of the higher risk characteristics of our patients. Regardless, our results raise the question as to whether CRT should be included in therapeutic regimens for T-ALL. Due to the diversity of systemic treatment strategies now employed for T-ALL and the very large patient population required to study the CRT question properly, this question might best be asked as an international 'Prospective Meta Analysis'.^28,29 The cranial radiation question could be grafted in a factorial fashion on to individual group trials for T cell patients, even if in some groups they are part of larger trials. Such an effort would require the combined resources of the North American, European and other cooperative groups.

Acknowledgements

This work was supported in part by grants: CA-69177, CA-33603, CA-15525, CA-30969, CA-15989, CA-29139, and CA-39139 from the National Cancer Institute.

1 Falletta JM, Shuster JJ, Crist WM, Pullen DJ, Borowitz MJ, Wharam M, Patterson R, Foreman E, Vietti TJ. Different patterns of relapse associated with three intensive treatment regimens for pediatric E-rosette positive T cell leukemia: a Pediatric Oncology Group study. Leukemia 1992; 6: 541-546, MEDLINE

2 Amylon MD, Shuster J, Pullen J, Berard C, Link MP, Wharam M, Katz J, Yu A, Laver J, Ravindranath Y, Kurtzberg J, Desai S, Camitta B, Murphy SB. Intensive high dose asparaginase consolidation improves survival for pediatric patients with T cell acute lymphoblastic leukemia and advanced stage lymphoblastic lymphoma: Pediatric Oncology Group study 8704. Leukemia 1999; 13: 335-342, MEDLINE

3 Smith M, Arthur D, Camitta B, Carroll AJ, Crist W, Gaynon P, Gelber R, Heerema N, Korn EL, Link M, Murphy S, Pui CH, Pullen J, Reaman G, Sallan SE, Sather H, Shuster J, Simon R, Trigg M, Tubergen D, Uckun F, Ungerleider R. Uniform approach to risk classification and treatment assignment for children with acute lymphoblastic leukemia. J Clin Oncol 1996; 14: 18-24, MEDLINE

4 Shuster JJ, Falletta JM, Pullen DJ, Crist WM, Humphrey GB, Dowell BL, Wharam MD, Borowitz M. Prognostic factors in childhood T cell acute lymphoblastic leukemia: a Pediatric Oncology Group study. Blood 1990; 75: 166-173, MEDLINE

5 Pullen J, Boyett J, Shuster J, Crist W, Land V, Frankel L, Iyer R, Backstrom L, Van Eys J, Harris M, Ravindranath Y, Sullivan M. Extended triple intrathecal chemotherapy trial for prevention of CNS relapse in good-risk and poor-risk patients with B-progenitor acute lymphoblastic leukemia: a Pediatric Oncology Group study. J Clin Oncol 1993; 11: 839-849, MEDLINE

6 Conter V, Schrappe M, Arico' A, Reiter A, Rizzari C, Dordelmann M, Valsecchi MG, Zimmermann M, Ludwig WD, Basso G, Masera G, Riehm H. Role of cranial radiotherapy for childhood T cell acute lymphoblastic leukemia with high WBC count and good response to prednisone. J Clin Oncol 1997; 15: 2786-2791, MEDLINE

7 Pui C-H, Behm FG, Singh B, Schell MJ, Williams DL, Rivera GK, Kalwinsky DK, Sandlund JT, Crist WM, Raimondi SC. Heterogeneity of presenting features and their relation to treatment outcome in 120 children with T cell acute lymphoblastic leukemia. Blood 1990; 75: 174-178, MEDLINE

8 Arico M, Basso G, Mandelli F, Rizzari C, Colella R, Barisone E, Zanesco L, Rondelli R, Pession A, Masera G. Good steroid response in vivo predicts a favorable outcome in children with T cell acute lymphoblastic leukemia. Cancer 1995; 75: 1684-1693, MEDLINE

9 Bennett JM, Catovsky D, Daniel MT, Flandrin G, Galton DA, Gralnick HR, Sultan C. Proposals for the classification of the acute leukemias. Br J Haematol 1976; 33: 451-458, MEDLINE

10 Pui C-H, Behm FG, Crist WM. Clinical and biologic relevance of immunologic marker studies in childhood acute lymphoblastic leukemia. Blood 1993; 82: 343-362, MEDLINE

11 Peto R, Pike MC, Armitage P, Breslow NE, Cox DR, Howard SV, Mantel N, McPherson K, Peto J, Smith PG. Design and analysis of randomized clinical trials requiring prolonged observation of each patient. II. Analysis and examples. Br J Cancer 1977; 35: 1-39, MEDLINE

12 Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958; 35: 457-481,

13 Rivera GK, Pinkel D, Simone JV, Hancock ML, Crist WM. Treatment of acute lymphoblastic leukemia. 30 years' experience at St Jude Children's Research Hospital. New Engl J Med 1993; 329: 1289-1295, MEDLINE

14 Rimm IJ, Li FC, Tarbell NJ, Winston DR, Sallan SE. Brain tumors after cranial irradiation for childhood acute lymphoblastic leukemia. A 13-year experience from the Dana-Farber Cancer Institute and The Children's Hospital. Cancer 1987; 59: 1506-1508, MEDLINE

15 Brouwers P, Poplack D. Memory and learning sequelae in long-term survivors of acute lymphoblastic leukemia: association with attention deficits. Am J Ped Hematol/Oncol 1990; 12: 174-181,

16 Nachman J, Sather HN, Cherlow JM, Sensel MG, Gaynon PS, Lukens JN, Wolff L, Trigg ME. Response of children with high-risk acute lymphoblastic leukemia treated with and without cranial irradiation: a report from the Children's Cancer Group. J Clin Oncol 1998; 16: 920-930, MEDLINE

17 Pui C-H, Mahmoud HH, Rivera GK, Hancock ML, Sandlund JT, Behm FG, Head DR, Relling MV, Ribeiro PC, Rubnitz JE, Kun LE, Evans WE. Early intensification of intrathecal chemotherapy virtually eliminates central nervous system relapse in children with acute lymphoblastic leukemia. Blood 1999; 92: 411-415,

18 Uckun FM, Sensel MG, Sun L, Steinherz PG, Trigg ME, Heerema NA, Sather HN, Reaman GH, Gaynon PS. Biology and treatment of childhood T-lineage acute lymphoblastic leukemia. Blood 1998; 91: 735-746, MEDLINE

19 Early AP, Preisler HD, Slocum H, Rustum YM. A pilot study of high-dose 1--D-arabinofuranosylcytosine for acute leukemia and refractory lymphoma: clinical response and pharmacology. Cancer Res 1982; 42: 1587-1594, MEDLINE

20 Shapiro WR, Young DG, Mehta BM. Methotrexate distribution in cerebrospinal fluid after intravenous, ventricular, and lumbar injections. New Engl J Med 1975; 293: 161-166, MEDLINE

21 Freeman AI, Weinberg V, Brecher ML, Jones B, Glicksman AS, Sinks LF, Weil M, Pleuss H, Hananian J, Burgert EO, Gilchrist GS, Necheles T, Harris M, Kung F, Patterson RB, Maurer H, Leventhal B, Chevalier L, Forman E, Holland JF. Comparison of intermediate-dose methotrexate with cranial irradiation for the post-induction treatment of acute lymphocytic leukemia in children. New Engl J Med 1983; 308: 477-484, MEDLINE

22 Ritchey AK, Pollock BH, Lauer SJ, Andejeski Y, Barredo J, Buchanan GR. Improved survival of children with isolated CNS relapse of acute lymphoblastic leukemia. A Pediatric Oncology Group study. J Clin Oncol 1999; 17: 3745-3752, MEDLINE

23 Thompson CB, Sanders JE, Flournoy N, Buckner CD, Thomas ED. The risks of central nervous system relapse and leukoencephalopathy in patients receiving marrow transplants for acute leukemia. Blood 1986; 67: 195-199, MEDLINE

24 Pullen DJ, Sullivan MP, Falletta JM, Boyett JM, Humphrey GB, Starling KA, Land VJ, Dyment PG, Vats T, Duncan MH. 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: 1159-1168, MEDLINE

25 Nachman JB, Sather HN, Sensel MG, Trigg ME, Cherlow JM, Lukens JN, Wolff L, Uckun FM, Ganynon PS. Augmented post-induction therapy for children with high-risk acute lymphoblastic leukemia and a slow response to initial therapy. New Engl J Med 1998; 338: 1663-1671, MEDLINE

26 Reiter A, Schrappe M, Ludwig WD, Hiddenmann W, Sauter S, Henze G, Zimmerman M, Lampert F, Havers W, Niethammer D et al. Chemotherapy in 998 unselected childhood acute lymphoblastic leukemia patients. Results and conclusions of the multicenter trial ALL-BFM 86. Blood 1994; 84: 3122-3133, MEDLINE

27 Shuster JJ. Publishing clinical trials and dealing with external comparisons. Control Clin Trials 1998; 19: 269-270, MEDLINE

28 Shuster JJ, Gieser PW. Meta-analysis and prospective meta-analysis in childhood leukemia: clinical research. Ann Oncol 1996; 7: 1009-1014, MEDLINE

29 Valsecchi MG, Masera G. A new challenge in clinical research in childhood ALL: the prospective meta-analysis strategy for inter-group collaboration. Ann Oncol 1996; 7: 1005-1008, MEDLINE

Figure 1  Comparison of the complete continuous remission for patients with high risk T-ALL who received cranial radiation (dotted line) or not (solid line). The 3-year event-free survival for high risk T-ALL was 65% (s.e. 6%) and 63% (s.e. 4%) for the non-irradiated vs the radiated group, respectively (P = 0.46). Numbers on curve represent number of patients at risk at time shown.

Figure 2  Comparison of CNS remission rates for patients who received cranial radiation (dotted line) vs patients who did not (solid line). The 3-year CNS remission rates were 82% (s.e. 5%) and 93% (s.e. 3%) for CRT- and CRT+ groups, respectively (P = 0.012). Numbers on curve represent number of patients at risk at time shown.

Table 1  Patient characteristics

Table 2  POG treatment programs for T-ALL (1987-1995)