|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 24, No 15 (May 20), 2006: pp. 2332-2336 © 2006 American Society of Clinical Oncology. DOI: 10.1200/JCO.2005.03.9727 Prognostic Significance of Blasts in the Cerebrospinal Fluid Without Pleiocytosis or a Traumatic Lumbar Puncture in Children With Acute Lymphoblastic Leukemia: Experience of the Dutch Childhood Oncology GroupFrom the University Medical Center St Radboud, Nijmegen; University Medical Center Groningen, Groningen; and the Dutch Childhood Oncology Group, The Hague, the Netherlands Address reprint requests to Siebold S.N. de Graaf, MD, Department of Pediatric Oncology-Hematology, UMC St Radboud, PO Box 9101, 6500 HB Nijmegen, the Netherlands; e-mail: s.degraaf{at}cukz.umcn.nl
PURPOSE: To determine the significance of blasts in the CSF without pleiocytosis and a traumatic lumbar puncture in children with acute lymphoblastic leukemia (ALL). PATIENTS AND METHODS: We retrospectively studied a cohort of 526 patients treated in accordance with the virtually identical Dutch protocols ALL-7 and ALL-8. Patients were classified into five groups: CNS1, no blasts in the CSF cytospin; CNS2, blasts present in the cytospin, but leukocytes less than 5/µL; CNS3, blasts present and leukocytes more than 5/µL. Patients with a traumatic lumbar puncture (TLP; > 10 erythrocytes/mL) were classified as TLP+ (blasts present in the cytospin) or TLP– (no blasts). RESULTS: Median duration of follow-up was 13.2 years (range, 6.9 to 15.5 years). Event-free survival (EFS) was 72.6% (SE, 2.5%) for CNS1 patients (n = 304), 70.3% (SE, 4.7%) for CNS2 patients (n = 111), and 66.7% (SE, 19%) for CNS3 patients (n = 10; no significant difference in EFS between the groups). EFS was 58% (SE, 7.6%) for TLP+ patients (n = 62) and 82% (SE, 5.2%) for TLP– patients (n = 39; P < .01). Cox regression analysis identified TLP+ status as an independent prognostic factor (risk ratio, 3.5; 95% CI, 1.4 to 8.8; P = .007). Cumulative incidence of CNS relapses was 0.05 and 0.07 in CNS1 and CNS2 patients, respectively (not statistically significant). CONCLUSION: In our experience, the presence of a low number of blasts in the CSF without pleiocytosis has no prognostic significance. In contrast, a traumatic lumbar puncture with blasts in the CSF specimen is associated with an inferior outcome.
An early observation in the development of treatment for childhood acute lymphoblastic leukemia (ALL) was that approximately 10% to 35% of patients relapsed in the CNS, even when cytologic examination of the cerebrospinal fluid (CSF) showed no CNS involvement at diagnosis.1-3 The introduction of elective (or so called "prophylactic") CNS-directed treatment, usually cranial irradiation in combination with intrathecal chemotherapy, reduced the incidence of meningeal relapse to less than 5%.4,5 Apparently, at diagnosis leukemic blasts may be present in the CNS at a subclinical level. Current ALL treatment protocols therefore uniformly include some kind of elective CNS treatment, even in children without evidence of CNS involvement at diagnosis. Over the years, the role of radiation therapy in elective CNS treatment has diminished because of its neurotoxicity.6,7 A different category of patients is the group presenting with overt meningeal leukemia. These patients have a significantly worse outcome and need additional CNS treatment.7 A clear definition of overt meningeal leukemia is therefore necessary to identify patients who need additional therapy and to avoid overtreatment in patients who do not need it. At a workshop in Rome in 1985, CNS leukemia was defined as the presence of more than five leukocytes per microliter in the CSF and the presence of morphologic blasts.8 However, questions have been raised regarding the validity of these criteria. Investigators from St Jude Children's Hospital (Memphis, TN) suggested that the presence of even a small number of blasts in the CSF without pleiocytosis in patients with ALL was a significant risk for CNS relapse, and that this group of patients needed additional CNS-directed therapy.9 However, reports about this subject are controversial. While the Pediatric Oncology group confirmed the St Jude results, the Children's Cancer Group and the Berlin-Frankfurt-Münster (BFM) -95 studies failed to show a poorer rate of event-free survival (EFS) for patients with a small number of blasts in the CSF, although these patients tended to have more CNS relapses in the BFM study.10-12 Thus, the prognostic significance of a low number of leukemic blasts in the CSF has remained a subject of investigation. A related issue is the prognostic relevance of a traumatic lumbar puncture at diagnosis. Investigators of the St Jude Children's Hospital were the first to show that a traumatic lumbar puncture with the presence of blasts in the CSF sample negatively affected patient outcome.13 These findings were subsequently confirmed by the BMF group.10 In consideration of all these studies, two issues regarding the analysis of CSF at diagnosis need further elucidation; first, the prognostic relevance of a small number of blasts without pleiocytosis, and second, the consequences of a traumatic lumbar puncture at diagnosis. We retrospectively studied these issues in a cohort of patients who were treated according to Dutch Childhood Leukemia Study Group (DCLSG) protocols ALL-7 and ALL-8 from 1988 to 1997.
DCLSG protocols ALL-7 and its successor ALL-8 were open for enrollment from 1988 to 1997, enrolling newly diagnosed patients with ALL from birth to 18 years old. These protocols were similar to the protocols BFM-86 and BFM-90, respectively. At diagnosis, CSF samples (not cytocentrifuged) had to be submitted to the Central Laboratory of the DCLSG (currently Dutch Childhood Oncology Group, DCOG) for central review of cytocentrifuge slides by two experienced cytologists. Patients were excluded from this study if no CSF was received for central review. CNS status of the patients was retrospectively defined according conventional morphologic criteria as follows: CNS1, atraumatic lumbar puncture (LP) (  10 erythrocytes/µL) and no blasts present in the CSF specimen at diagnosis; CNS2, atraumatic LP, leukocyte count less than 5/µL but with blasts present in the CSF; CNS3, atraumatic LP, leukocytes  5/µL and blasts in the CSF. Patients with a traumatic lumbar puncture at diagnosis (> 10 erythrocytes/µL) were classified in two groups, namely, TLP+ (traumatic LP with blasts in the CSF sample after centrifugation) and TLP– (traumatic LP with no blasts in the CSF sample). In addition to conventional morphology, terminal deoxynucleotidyl transferase (TdT) staining16,17 of the samples was performed when possible. Results of TdT testing were recorded only and had no effect on treatment allocation. Systemic relapse was defined as more than 25% blasts in the bone marrow. CNS relapse was defined by the presence of blasts and 5 leukocytes/µL in at least two subsequent CSF samples.
Treatment
Statistical analysis.
EFS was defined as the time from diagnosis until the first date of treatment failure of any kind, death during remission, or the development of a second malignancy. For patients alive and event free at latest follow-up (censored observations), EFS was calculated until this latest follow-up. Distribution of EFS and continuous complete remission were calculated using the Kaplan-Meier method and were subsequently compared using the two-sided log-rank test.18 SEs were also calculated with the Kaplan-Meier method.18 The prognostic value of the different groups considering EFS compared with CNS-negative patients was examined by Cox regression analysis.19 Other prognostic factors such as sex and WBC count were used as covariables. Cox regression analysis was used to investigate the combined prognostic value of different parameters significant in the EFS and for survival without CNS relapse. To analyze differences of variables among patients, the Mann-Whitney U test20 was used for continuous variables and the
From July 1988 until January 1997, a total of 570 patients were enrolled onto DCLSG protocols ALL-7 and ALL-8. Forty-four of these patients were excluded from this analysis because no CSF sample was received for central cytologic review. The remaining 526 patients (92% of the total population) were the subjects of this analysis. Patient and disease characteristics and outcome of these patients were not significantly different from the whole group of patients enrolled onto the protocols. Median duration of follow-up was 11.7 years (range, 7.9 to 16.5 years). CNS1 status was found in 304 patients (58%), CNS2 status in 111 patients (21%), and CNS3 status in 10 patients (2%) according to conventional morphologic evaluation. The remaining 101 patients (19%) had a traumatic lumbar puncture at diagnosis and were divided in two groups: TLP+ (n = 62; 12%) and TLP– (n = 39; 7%). All patients were stratified into standard-risk, medium-risk, and high-risk groups according to BFM risk factor (determined by number of blasts in peripheral blood and hepatosplenomegaly), immunophenotype, response to steroids, and cytogenetics. Of the 526 patients included in this study, 192 patients (36%) were stratified into the standard-risk group (SRG), 282 patients (54%) into the medium-risk group (MRG), and 52 patients (10%) into the high-risk group (HRG). Evaluation of the patients' characteristics among the different CNS status groups showed that patients in groups CNS2, CNS3, and TLP+ had significantly more unfavorable characteristics than patients in the CNS1 or TLP–. Compared with CNS1 patients, patients in the CNS2, CNS3, and TLP+ groups had a significantly higher percentage of patients with WBC counts above 50 x 109/L. Distribution of sex and age was equal between risk groups (Table 1). One patient with a traumatic lumbar puncture was treated as being CNS positive.
Overall, 10-year EFS for this cohort was 71% (SE, 2.1%). The 10-year EFS estimates for patients in each CNS status group were as follows: CNS1 patients 72.6% (SE, 2.5%), CNS2 patients 70.3% (SE, 4.7%; not significantly different; Fig 1). EFS for the 10 patients in the CNS3 group was 67.7% (SE, 19.0%). CNS3 patients were not considered further in this analysis because the low number of patients in this group precluded meaningful comparisons and overt meningeal leukemia was not the focus of this study. The 10-year EFS estimate for TLP– patients was 82% (SE, 5.2%) and for TLP+ patients 58% (SE, 7.6%; P < .01; Fig 2). Prognosis of patients with CNS2 status and TLP– status was not significantly different compared with patients with CNS1 status. In contrast, TLP+ patients had a significantly inferior outcome than patients with CNS1 status. Cox regression analysis, including covariables such as age as a categoric variable, WBC count, treatment, and risk group, showed that a traumatic lumbar puncture with the presence of blasts at diagnosis remained a prognostic risk factor (risk ratio, 3.5; 95% CI, 1.4 to 8.8; P = .007).
To evaluate whether the reduced EFS of TLP+ patients was due to CNS relapses, distribution of relapses according to CNS status group were studied (Table 2). Cumulative incidence (CuIn) of CNS relapses after 10 years, either isolated incidents or combined with relapse in the bone marrow, was 0.05 and 0.07 in CNS1 and CNS 2 patients, respectively (not significant). In comparison with CNS1 patients, TLP+ patients tended to relapse more frequently in the CNS (CuIn, 0.08 v 0.05; P = .236) and in the bone marrow (CuIn 0.294 v 0.205; P = .07). The total number of relapses at any location was significantly higher in TLP+ patients than in patients with CNS1 status (CuIn, 0.255 v 0.375; P = .025).
As mentioned in the Patients and Methods section, TdT staining of CSF samples was performed whenever possible. TdT staining was performed in 494 patients. When morphologic CNS1 and CNS2 patients (n = 401) are reclassified on the basis of TdT positivity instead of morphology, 296 patients (60%) had TdT-CNS1 status and 105 patients (21%) TdT-CNS2 status. The 10-year EFS estimates according to a TdT-based classification were 73.8% (SE, 2.5%) for TdT-CNS1 patients and 68% (SE, 5.1%) for TdT-CNS2 patients (P = .169). Thus, analyses based on TdT staining did not change the conclusion that CNS2 status has no significant effect on prognosis. Unfortunately, there were too many missing data to allow a meaningful analysis of TdT results in patients that had a traumatic lumbar puncture at diagnosis.
Our results suggest that the presence of morphologic blasts in the spinal fluid in combination with a CSF leukocyte count of fewer than five per microliter has no prognostic relevance. These results are further underlined by the fact that no prognostic relevance could be found when only TdT-positive cells were considered to be true leukemic blasts. Consequently, our results do not provide any evidence for the need of administering additional intrathecal therapy to CNS2 patients, in the context of treatment administered in Dutch protocols ALL-7 or ALL-8. In contrast, our data show that a traumatic lumbar puncture with the presence of blasts in the CSF sample is associated with a worse prognosis. Overall survival of patients treated in accordance with the Dutch protocols ALL-7 or ALL-8 was similar to the outcome of the contemporary trials of the BFM study group, although in the Dutch protocols cranial irradiation was restricted to those few patients (2%) who had overt CNS leukemia at diagnosis.10 In contrast with our experience, the BFM group found that CNS2 patients had an almost threefold (10% v 3.5%) higher risk of CNS relapses compared with patients with CNS1 status. Although this increased risk was statistically not significant, they recommended giving additional CNS-directed therapy to these patients. Our data suggest that the importance of CNS2 status is treatment dependent. In the context of our treatment, CNS2 was not associated with an increased risk of relapse. A major difference between the BFM study and our study is the absence of a central review of CSF cytocentrifuge slides in a substantial number of BFM patients. Only 48% of CSF samples at diagnosis were centrally reviewed in the BFM trial.10 In contrast, in the DCLSG protocols, a central review of cytocentrifuge slides was mandatory. Only one institution did not comply with the requirement to submit slides for review, resulting in 44 patients who were ineligible for this study. The difference in central review between the BFM and DCLSG trials may, at least in part, account for the lower incidence of CNS2 status in the BFM experience; 5.1% versus 21% in our cohort. Accurate cytologic assessment of cells in the CSF from patients with ALL requires analytic skills that may not be available at all of the hospitals at which patients were evaluated. Therefore, it is possible that some patients with only a few blasts in the CSF are diagnosed as having CNS1 status in the BFM study. Thus, the relatively small group of patients with CNS2 status in the BFM study may represent a selection bias. To prevent a selection bias and to identify the true leukemic nature of cells, Pui22 suggested that repeat examinations should be performed, preferably with TdT staining or immunophenotyping. In this study, TdT staining was performed in addition to the cytologic assessment of cells. Even when TdT positivity was used to define a blast, no significant difference in outcome was found between CNS1 and CNS2 patients. Our results are in agreement with previous studies10,13 with regard to the adverse prognostic implication of a traumatic lumbar puncture with blasts in the CSF sample,. The outcome of our TLP+ patients (10-year EFS, 59%) is similar to that of the TLP+ patients treated at St Jude Children's Hospital (5-year EFS, 60%); 5-year EFS of TLP+ patients in the BFM trial was 73%. We should ask the question why patients with a traumatic lumbar puncture with blasts in the CSF have an inferior outcome. One hypothesis regards iatrogenic contamination of the CSF by blasts circulating in the peripheral blood. If such contamination is just a matter of bad luck or an unfortunate procedure, we should try to prevent the latter from happening. A logical recommendation would be performing lumbar punctures in newly diagnosed patients by experienced professionals only under optimal circumstances, including adequate sedation or general anesthesia. However, we wonder if some patient categories are more prone to traumatic lumbar punctures than others. We found that patients with a high WBC count were more likely to have TLP+ status, but immunophenotype and age were not associated with a higher incidence of a traumatic lumbar puncture. Another possibility is that the poor outcome of TLP+ patients is due to undertreatment of patients, who were actually patients with CNS involvement (CNS3 status), but were not identified as such at diagnosis due to the presence of erythrocytes in the CSF. Although we have no definite explanation for the inferior outcome in TLP+ patients, we would recommend some kind of additional intensified therapy to these patients. Intensifying systemic chemotherapy and/or administering additional intrathecal therapy may be needed to prevent a higher incidence of relapses in these patients. The efficacy of such additional therapy can only be evaluated in a prospective clinical trial.
The authors indicated no potential conflicts of interest.
1. Lampert F: Results of Memphis-study VII ("Pinkel-Therapy") in 659 children with acute lymphoblastic leukemia: Cooperative therapeutic study of the German working group on childhood leukemia with participation of 42 hospitals. Klin Pädiatr 190:57-64, 1978[Medline] 2. van Eys J, Berry D, Crist W, et al: A comparison of two regimens for high-risk acute lymphocytic leukemia in childhood: A Pediatric Oncology Group Study. Cancer 63:23-29, 1989[CrossRef][Medline] 3. Kawashima K, Nagura E, Yamada K, et al: Leukemia relapse in long-term survivors of acute leukemia. Cancer 56:88-94, 1985[CrossRef][Medline] 4. Pui CH, Sandlund JT, Pei D, et al: Improved outcome for children with acute lymphoblastic leukemia: Results of Total Therapy Study XIIIB at St Jude Children's Research Hospital. Blood 104:2690-2696, 2004 5. Reiter A, Schrappe M, Ludwig WD, et al: Chemotherapy in 998 unselected childhood acute lymphoblastic leukemia patients: Results and conclusions of the multicenter trial ALL-BFM 86. Blood 84:3122-3133, 1994 6. Tubergen DG, Gilchrist GS, O'Brien RT, et al: Prevention of CNS disease in intermediate-risk acute lymphoblastic leukemia: Comparison of cranial radiation and intrathecal methotrexate and the importance of systemic therapy—A Childrens Cancer Group report. J Clin Oncol 11:520-526, 1993 7. Clarke M, Gaynon P, Hann I, et al: CNS-directed therapy for childhood acute lymphoblastic leukemia: Childhood ALL Collaborative Group overview of 43 randomized trials. J Clin Oncol 21:1798-1809, 2003 8. Mastrangelo R, Poplack D, Bleyer A, et al: Report and recommendations of the Rome workshop concerning poor-prognosis acute lymphoblastic leukemia in children: Biologic bases for staging, stratification, and treatment. Med Pediatr Oncol 14:191-194, 1986[Medline] 9. Mahmoud HH, Rivera GK, Hancock ML, et al: Low leukocyte counts with blast cells in cerebrospinal fluid of children with newly diagnosed acute lymphoblastic leukemia. N Engl J Med 329:314-319, 1993 10. Burger B, Zimmermann M, Mann G, et al: Diagnostic cerebrospinal fluid examination in children with acute lymphoblastic leukemia: Significance of low leukocyte counts with blasts or traumatic lumbar puncture. J Clin Oncol 21:184-188, 2003 11. Gilchrist GS, Tubergen DG, Sather HN, et al: Low numbers of CSF blasts at diagnosis do not predict for the development of CNS leukemia in children with intermediate-risk acute lymphoblastic leukemia: A Childrens Cancer Group report. J Clin Oncol 12:2594-2600, 1994 12. Tubergen DG, Cullen JW, Boyett JM, et al: Blasts in CSF with a normal cell count do not justify alteration of therapy for acute lymphoblastic leukemia in remission: A Childrens Cancer Group study. J Clin Oncol 12:273-278, 1994[Abstract] 13. Gajjar A, Harrison PL, Sandlund JT, et al: Traumatic lumbar puncture at diagnosis adversely affects outcome in childhood acute lymphoblastic leukemia. Blood 96:3381-3384, 2000 14. Kamps WA, Bokkerink JP, Hahlen K, et al: Intensive treatment of children with acute lymphoblastic leukemia according to ALL-BFM-86 without cranial radiotherapy: Results of Dutch Childhood Leukemia Study Group Protocol ALL-7 (1988-1991). Blood 94:1226-1236, 1999 15. Kamps WA, Bokkerink JP, Hakvoort-Cammel FG, et al: BFM-oriented treatment for children with acute lymphoblastic leukemia without cranial irradiation and treatment reduction for standard risk patients: Results of DCLSG protocol ALL-8 Leukemia 16:1099-1111, 2002 16. Bradstock KF, Papageorgiou ES, Janossy G: Diagnosis of meningeal involvement in patients with acute lymphoblastic leukemia: Immunofluorescence for terminal transferase. Cancer 47:2478-2481, 1981[Medline] 17. van Wering ER, Veerman AJ, van der Linden-Schrever BE: Diagnosis of meningeal involvement in childhood acute lymphoblastic leukemia: Cytomorphology and TdT. Eur J Haematol 40:250-255, 1988 18. Kaplan EL, Meier P: Nonparametric estimation from incomplete observation. J Am Stat Assoc 53:457-481, 1958[CrossRef] 19. Cox DR, McCullagh P: Some aspects of analysis of covariance. Biometrics 38:541-546, 1982[CrossRef][Medline] 20. Rosner B, Grove D: Use of the Mann-Whitney U-test for clustered data. Stat Med 18:1387-1400, 1999[Medline] 21. Kalbfleisch JD, Prentice RL: The Statistical Analysis of Failure Time Data. Wiley, New York, NY, 1980 22. Pui CH: Toward optimal central nervous system-directed treatment in childhood acute lymphoblastic leukemia. J Clin Oncol 21:179-181, 2003 Submitted August 30, 2005; accepted February 24, 2006. This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2006 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
2 test for categorized variables. Cumulative incidence curves were calculated using NCSS statistical analysis software (Kaysville, UT).
