This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Ribera, J.-M. Articles by Oriol, A. Search for Related Content PubMed Articles by Ribera, J.-M. Articles by Oriol, A.

In Reply

Institut Català d'Oncologia-Hospital Universitari Germans Trias i Pujol, Badalona, Spain

Juan-José Ortega

Hospital Universitari Vall d'Hebron, Barcelona, Spain

Albert Oriol

Institut Català d'Oncologia-Hospital Universitari Germans Trias i Pujol, Badalona, Spain

Two recent trials for children receiving treatment for very high risk (VHR) acute lymphoblastic leukemia (ALL) in first complete remission have compared the outcome of allogeneic stem-cell transplantation (alloSCT) and chemotherapy using genetic random assignment (donor v no donor availability). While the trial conducted within the International Berlin-Frankfurt-Munster Study Group (I-BFM-SG) reported an advantage in 5-year disease-free survival (DFS) for patients with a donor (56.7% v 40.6%; P = .02), the Programa de Estudio y Tratamiento de las Hemopatías Malignas (PETHEMA) ALL-93 trial, published in the Journal of Clinical Oncology, failed to find such differences (DFS of 45% for both groups). Despite such divergence, a careful analysis of subgroups of patients revealed very comparable results between treatments in both trials and provided important insight into the evolving definition of VHR ALL as equivalent to SCT candidates in children.

In January 2007, we reported the results of a trial that failed to prove an advantage for patients with a donor over those without a matched sibling in children with VHR ALL.¹ One year previously, Balduzzi et al published the results of a larger trial with a very similar design, which detected an advantage in event-free survival for patients with a matched relative.² This apparent discrepancy deserves careful analysis.

Differences in the definition of VHR between the two studies were minor. Hyperleukocytosic B-lineage ALL and infant patients were not VHR criteria in the Balduzzi et al study, although they included children younger than 1 year, probably presenting other VHR features (eg, t(4;11)). However, there were important differences in the frequency of VHR subgroups. Very few Philadelphia-positive (Ph'+) ALL patients were included in our trial (9%), particularly since the emergence of alternative imatinib-based trials, while Ph'+ ALL patients represented the 23% of the overall sample of Balduzzi et al. This is particularly relevant since there is an apparent imbalance in the distribution of these patients: 75 were allocated to chemotherapy and eight to SCT, representing 27% and 10% of the patients in the respective arms (², 11.9; P = .001) and these patients fared particularly poorly in the chemotherapy arm (5-year DFS, 25%).

In our study, all patients received the same induction and early consolidation. Induction and early consolidation (if any) are not detailed in the trial conducted by Balduzzi et al, although probably schedules were very similar. In our study, as in most nonrandomized studies, chemotherapeutic approaches presented higher relapse rates than SCT-based consolidations, which were generally counterbalanced by higher nonrelapse mortality with SCT. Noteworthy, the cumulative death in remission in the work by Balduzzi et al was particularly high (9% v 0% and 5%, respectively, in our chemotherapy and autologous SCT [auto-SCT] arms). Forty-three patients (15%) in the chemotherapy arm deviated from the protocol to undergo an unrelated SCT and 14 patients died in complete remission (CR), accounting for 50% of the deaths in CR in this arm. Had the patients been censored before SCT, the results in 5-year DFS of the chemotherapy arm would have increased from 40% to 45%, coinciding with our results for the no donor arm. This approach would make differences among SCT and chemotherapy not significant in the study by Balduzzi et al.

The outcome of our donor arm was inferior to that reported by Balduzzi et al. In their letter, the authors point out two plausible causes: firstly, a longer interval between CR achievement and SCT, and secondly, the use of busulphan and cyclophosphamide as conditioning regimen instead of the preferable cyclophosphamide and total body irradiation (TBI). We were concerned about the first issue, as the three scheduled early consolidation cycles may have delayed SCT in some cases (6 weeks of median delay with respect to the study by Balduzzi et al). However, a time dependent adjustment of the analysis failed to detect a negative effect of time to transplant. Furthermore, none of the outlier patients, in which transplant was exceedingly delayed, relapsed (probably implying that they already presented solid CR before SCT). As for the conditioning regimen, our data also suggest a possible superiority of cyclophosphamide and TBI over busulphan and cyclophosphamide, but TBI could not be used in children younger than age 2. However, only two patients in the donor arm were conditioned with busulphan versus 14 patients allocated to auto-SCT in the no donor arm. Thus, a better balance in conditioning regimens would have improved, if any, the results of the no donor arm. The cumulative incidences of relapse in the SCT arm were 27.3% and 23.5% for Balduzzi et al and Ribera et al respectively, suggesting similar SCT efficacies. A higher pretransplant relapse (17%) and higher transplant-related mortality (17%) accounted for the poorer outcome in our study, suggesting that less intensive or prolonged consolidation may suffice if SCT is to be performed.

Despite the 15% difference in 5-year DFS between the donor and no donor arms reported by Balduzzi et al, they fail to prove any difference in overall survival (OS). As inferred from comparisons between DFS and OS curves, 6% to 10% of the patients in both studies treated with chemotherapy may have been salvaged after relapse with SCT, while there is no effective rescue treatment after SCT. This fact would lead to questioning the place of SCT in first CR if chemotherapy is capable of providing similar results as a first-line therapy and SCT may be the only alternative later on. For the same reason, chemotherapy would be preferable to auto-SCT as long as the superiority of auto-SCT is not apparent.

Our study was closed prematurely due to absence of expected differences with further recruitment. Balduzzi et al postulated that our final sample was underpowered to detect differences. The criticism is right a priori but it is not so a posteriori, taking into account the available data. With data for the first 100 patients yielding a 5-year DFS of 45% for both the donor (n = 24) and no donor (n = 76) arms, the probability of one group having a 10% superior outcome over the other was only 4%, thus the interruption of recruitment was fully justified. Further follow-up, up to a median of 6.5 years, corroborated the adequacy of this decision.

In conclusion, a careful analysis of subsets revealed interesting similarities in the results of two different studies with different recruitment times and number of patients. The criteria to consider an ALL patient as having VHR (equivalent to alloSCT candidate in first CR) tended to be even more restrictive. Only 8% and 13% of all patients (in Balduzzi et al and Ribera et al respectively) qualified as VHR and a potential benefit of SCT may apply only to a subset of these patients. Patients with Ph'+ALL or with suboptimal response to initial therapy, probably the only subgroups to be currently considered as VHR, account for 46% of the overall sample in the study by Balduzzi et al. This fact, together with 15% of patients analyzed by intention-to-treat in the chemotherapy arm who deviated from protocol to undergo an unrelated donor SCT with overall bad results, explain per se the difference in outcome between the two arms. Our study probably failed to find differences because of these same reasons: inferior inclusion of Ph'+ALL and a reasonable outcome of slow responders in the chemotherapy arm (5-year DFS, 47%) partially explained by the use of our less stringent criteria of failure or slow response to initial treatment. Ph'+ ALL is unlikely to be merged with other ALLs in current and future trials^3,4 and this is probably true for t(4;11)/MLL-AF4 ALL.⁵ Adequate monitorization and interpretation of minimal residual disease will undoubtedly influence the way we approach patients with ALL,^6,7 particularly those experiencing failure to achieve an optimal response to initial therapy, and may further define the exact time point in which an alloSCT is mandatory. International collaboration is essential to redefine the indications of SCT as the definition of VHR-ALL tends to include a further limited number of children.

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

The author(s) indicated no potential conflicts of interest.

ACKNOWLEDGMENTS

Supported in part by Grants No. 97/1049 and PI 05/1490 from Fondo de Investigaciones Sanitarias, and Grant No. FIJC P-EF/06 from José Carreras International Leukemia Foundation. The authors are writing on behalf of the PETHEMA Group.

REFERENCES

1. Ribera JM, Ortega JJ, Oriol A, et al: Comparison of intensive chemotherapy, allogeneic, or autologous stem-cell transplantation as postremission treatment for children with very high risk acute lymphoblastic leukemia: PETHEMA ALL-93 trial. J Clin Oncol 25:16-23, 2007[Abstract/Free Full Text]

2. Balduzzi A, Valsecchi MG, Uderzo C, et al: Chemotherapy versus allogeneic transplantation for very-high-risk childhood acute lymphoblastic leukaemia in first remission: Comparison by genetic randomisation in an international prospective study. Lancet 366:635-642, 2005[CrossRef][Medline]

3. Aricó M, Valsecchi MG, Camitta B, et al: Outcome of treatment in children with Philadelphia chromosome –positive acute lymphoblastic leukemia. N Engl J Med 342:998-1006, 2000[Abstract/Free Full Text]

4. Sharathkumar A, Saunders EF, Dror Y, et al: Allogeneic bone marrow transplantation vs. chemotherapy for children with Philadelphia chromosome-positive acute lymphoblastic leukaemia. Bone Marrow Transplant 33:39-45, 2004[CrossRef][Medline]

5. Pieters R, Schrappe M, de Lorenzo P, et al: Outcome of infants less than one year of age with acute lymphoblastic leukemia treated with the Interfant-99 protocol. Blood 108:47a, 2006 (abstr)

6. Biondi A, Valsecchi MG, Seriu T, et al: Molecular detection of minimal residual disease is a strong predictive factor of relapse in childhood B-lineage acute lymphoblastic leukaemia with medium risk features: A case-control study of the International BFM Study Group. Leukemia 14:1939-1943, 2000[CrossRef][Medline]

7. Coustan-Smith E, Sancho J, Hancock ML, et al: Clinical importance of minimal residual disease in childhood acute lymphoblastic leukemia. Blood 96:2691-2696, 2000[Abstract/Free Full Text]

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Ribera, J.-M. Articles by Oriol, A. Search for Related Content PubMed Articles by Ribera, J.-M. Articles by Oriol, A.