MINI-REVIEW New Definition of Remission in Childhood Acute

New deﬁnition of remission in childhood acute lymphoblastic leukemia C-H Pui1,2,3 and D Campana1,3

Departments of 1Hematology-Oncology and 2Pathology, St Jude Children’s Research Hospital; and 3The University of Tennessee College of Medicine, Memphis, Tennessee, USA

The extent of clearance of leukemic cells from the blood or patients.5,6 Either method is at least 100-fold more sensitive bone marrow during the early phase of therapy is an inde- than morphologic examination of marrow specimens and con- pendent prognostic factor in acute lymphoblastic leukemia 4 (ALL). Several methods are available to measure the minimal sistently identifies one leukemic cell among 10 or more nor- residual disease (MRD) remaining after initial intensive chemo- mal bone marrow cells. Recent demonstration of the inde- therapy. The most promising are flow cytometric detection of pendent prognostic significance of MRD levels in remission aberrant immunophenotypes and polymerase chain reaction bone marrow samples using these methods7–9 provided the analysis of clonal antigen-receptor gene rearrangements. When impetus to redefine complete remission in patients with ALL.10 applied together, these techniques enable one to monitor MRD However, several critical issues must be addressed before in virtually all cases of ALL. Patients who achieve an ‘immunologic’ or ‘molecular’ remission (ie leukemic involvement of MRD determinations can be routinely considered in clinical ,0.01% of nucleated bone marrow cells at the end of remission decision making. induction therapy) are predicted to have a better clinical outcome than patients whose remission is defined solely by morphologic criteria. In studies to date, patients with MRD at a level Universal monitoring of MRD of 1022 or more at the end of induction have fared almost as poorly as those with >5% blast cells in the bone marrow (ie induction failures). Sequential monitoring of MRD can improve A prerequisite for applying MRD measurements in clinical the clinical utility of risk assessment still further. Additional trials is that data should be available for all patients. This studies are needed to determine the critical levels of MRD at obstacle could be overcome by increasing the numbers of leu- various times of treatment and whether therapeutic inter- kemic cell markers and target genes amenable to flow cyto- vention based on MRD findings can improve clinical outcome. metric and PCR analyses, respectively. An alternative strategy Leukemia (2000) 14, 783–785. would be to apply these methods in tandem. Indeed, we have Keywords: minimal residual leukemia; complete remission; flow cytometry; polymerase chain reaction found good concordance between results obtained with the two techniques.11 Tandem application of flow cytometry and PCR testing would not only allow one to study marrow samples from all patients, but may also reduce the likelihood Introduction of false-negative results due to changes in the predominant immunophenotypes or gene rearrangements during the The initial response to remission induction therapy is one of disease course. the most important prognostic factors in acute lymphoblastic leukemia (ALL). Patients who respond slowly have a high risk of relapse,1,2 while those who fail to attain a complete MRD levels of clinical importance remission within 4 to 6 weeks of treatment have a particularly dismal prognosis.3 Complete remission is currently defined as The number of residual leukemic cells varies widely among restoration of normal hematopoiesis with a blast cell fraction patients, raising the issue of which levels are in fact predictive of less than 5% by light microscopic examination of the bone of clinical outcome. In one study, an MRD level of 0.5 to marrow. This relatively liberal definition reflects the difficulty − 5 × 10 5 cells during remission did not correlate with sub- of identifying leukemic lymphoblasts in regenerating marrow sequent relapse,12 suggesting that the residual cells were either by morphologic criteria alone. Thus, patients with nearly 5% preleukemic or in a nonproliferative state, or perhaps were leukemic blast cells in their marrow specimens can harbor as eventually eradicated by the patient’s immune system. many as 1010 leukemic cells, and their treatment may not dif- Although not yet confirmed by others, this finding suggests fer from that of patients with much larger reductions in the −5 4 that MRD levels of 10 or lower are not clinically relevant. By leukemic cell burden. Conversely, normal hematopoietic − contrast, levels of 10 4 or greater correlated with an increased progenitors (‘hematogones’), which may represent 5% or more relapse hazard in three recent studies.7–9 Patients without of the cells in regenerating marrow, can be erroneously − detectable MRD or with a level below 10 4 had a 3-year interpreted as a sign of residual leukemia. cumulative risk of relapse of less than 10%, compared with With use of flow cytometry (to detect aberrant − 56% to 78% for those with 10 2 or more residual blasts immunophenotypes) or polymerase chain reaction (PCR) (Table 1). Only two of our nine patients with high MRD levels analysis (to detect rearrangements of antigen receptor genes), at the end of induction are alive and in remission at 2.5 and it is now possible to search for minimal residual disease 4.5 years post-diagnosis (the second child underwent hemato- (MRD) during clinical remission in approximately 90% of poietic stem cell transplantation in first remission). Patients in this category appear to fare as poorly as induction failures Correspondence: C-H Pui, St Jude Children’s Research Hospital, 332 identified by traditional criteria and clearly should be North Lauderdale, Memphis, TN 38105, USA; Fax: 1 901 521 9005 managed as very high-risk cases. Received 21 January 2000; accepted 26 January 2000 The clinical importance of MRD levels between 10−4 and New definition of remission in childhood ALL C-H Pui and D Campana 784

Figure 1 Patterns of early cellular responses to antileukemic therapy. In 49% of the patients, 2 weeks of remission induction therapy reduced the MRD level to below 10−4 by ﬂow cytometry (estimated total-body leukemic cell burden, ,108). At 6 weeks of therapy, 26% of the patients had an MRD level of less than 10−4. Thus, the immunologic remission rate was 75%. Another 12% of the patients required from 7 to 22 weeks of therapy to attain such a remission, with all remaining patients requiring more than 22 weeks.

Table 1 Cumulative Risk of Relapse at 3 Years Postdiagnosis, than those with an MRD level <10−4 at week 7 post- According to MRD Level on Completion of Remission Induction a remission, who in turn had a better outcome than patients Therapy with persistently higher levels of MRD. In an ongoing prospective study, we have noted several dis- Ref. Low (<10−4) Intermediate High (>10−2) . −4 crete patterns of leukemic cell clearance that may have clini- ( 10 but 14 ,10−2) cal relevance (Figure 1). Almost half of our patients lacked measurable disease by ﬂow cytometric analysis after only 2 Coustan-Smith et al7b 5% (123) 24% (33) 56% (9) weeks of remission induction therapy. MRD was detected in Van Dongen et al8 <9% (104) 32% (38) 67% (27) 25% of the patients upon completion of the induction phase Cave´ et al9 <9% (88) <17% (30) <78% (15) (week 6), in 13% during week 14 of continuation treatment, and 4% during week 32. Virtually all of the patients with MRD aNumbers of patients at risk are given in parentheses. at week 32 have relapsed, while those with MRD at week 14 bUpdated analysis. but not week 32 have had a relatively low relapse hazard (<14% with 3 years of follow-up). It is still uncertain whether

−2 −3 the rapid disappearance of MRD during the first 2 weeks of 10 is less certain. Possibly, a level of 10 could be used to remission induction confers a particularly favorable prognosis, discriminate between patients with lower and higher risks of allowing treatment reduction. relapse. We and others observed that patients with MRD at or above the 10−3 level following consolidation therapy (1 to 3 months post-remission) have an especially high rate of relapse,7–9 almost comparable to that of patients classified as Recommendations induction failures. In any case, accurate quantification of MRD levels appears to be a critical component of sound clini- Now that cure rates in childhood ALL are approaching 80%,10 cal management. Flow cytometry is well suited for this pur- we should begin to apply immunologic or molecular defi- pose because it allows direct enumeration of viable leukemic nitions of complete remission in risk assessment. The findings cells and normal cells.6 The task is more difficult with PCR summarized here indicate that an MRD level of less than 10−4 analysis, which relies on the measurement of a PCR product leukemic cells upon completion of induction therapy predicts rather than cell counting;4 however, ‘real-time’ PCR tech- a very favorable prognosis. Sequential monitoring of MRD nology may substantially improve the utility of this method.13 may not be necessary for these patients. In our experience, only two of 123 such patients had detectable MRD before overt relapse.14 The value of MRD at the end of therapy is Sequential monitoring of MRD questionable because a negative result would not preclude subsequent relapse.7,8 Hence, the disappearance of MRD Until recently, very little was known about the kinetics of leu- should not be used as a reason for elective cessation of kemic cell clearance and its clinical significance. A European therapy. multicenter study found that the outcome of therapy correlates Patients with 10−2 or more leukemic cells during any phase well with the rapidity of leukemia reduction.8 Patients without of remission should be regarded as having a very high risk of detectable MRD at the end of the induction phase fared better relapse. Further studies are needed to define the significance

Leukemia New definition of remission in childhood ALL C-H Pui and D Campana 785 of intermediate MRD levels (.10−4 but ,10−2) during early SE. Induction failure in acute lymphoblastic leukemia of child- remission. hood. Cancer 1999; 85: 1395–1404. In theory, more rapid and complete clearance of leukemic 4 Campana D, Pui C-H. Detection of minimal residual disease in acute leukemia: methodologic advances and clinical significance. blast cells with treatment could forestall the emergence of Blood 1995; 85: 1416–1434. drug-resistant subclones, resulting in a more favorable clinical 5 Foroni L, Harrison CJ, Hoffbrand AV, Potter MN. Investigation of outcome. Intensification of early postremission therapy minimal residual disease in childhood and adult acute lymphobl- improves the clinical outcome in patients with slow responses astic leukaemia by molecular analysis. Br J Haematol 1999; 105: to remission induction therapy (more than 25% bone marrow 7–24. blasts on day 7),15 but randomized trials are needed to ascer- 6 Campana D, Coustan-Smith E. Detection of minimal residual disease in acute leukemia by flow cytometry. Cytometry 1999; 38: tain whether similar therapy would benefit patients with high- 139–152. risk ALL defined by MRD assays. Investigators should continue 7 Coustan-Smith E, Behm FG, Sanchez J, Boyett JM, Hancock ML, to search for novel leukemic cell markers with high sensitivity Raimondi SC, Rubnitz JE, Rivera GK, Sandlund JT, Pui C-H, Cam- and specificity. To this end, we have identified several useful pana D. Immunological detection of minimal residual disease in new markers using cDNA microarrays, enabling us to apply children with acute lymphoblastic leukaemia. Lancet 1998; 351: flow cytometric MRD detection methods in over 90% of ALL 550–554. 16 8 van Dongen JJM, Seriu T, Panzer-Gru¨mayer ER, Biondi A, Pongers- cases. Finally, an immunologic or molecular definition of Willemse MJ, Corral L, Stolz F, Schrappe M, Masera G, Kamps complete remission could speed the development of optimal WA, Gadner H, van Wering ER, Ludwig W-D, Basso G, de Bruijn risk-directed therapy for children with ALL, boosting cure rates MAC, Gazzaniga G, Hettinger K, van der Does-van den Berg A, closer to 100%. Hop WCJ, Riehm H, Bartram CR. Prognostic value of minimal residual disease in acute lymphoblastic leukaemia in childhood. Lancet 1998; 352: 1731–1738. 9 Cave´ H, van der Werff ten Bosch J, Suciu S, Guidal C, Waterkeyn Acknowledgements C, Otten J, Bakkus M, Thielemans K, Grandchamp B, Vilmer E. Clinical significance of minimal residual disease in childhood acute lymphoblastic leukemia. New Engl J Med 1998; 339: This work was supported by grants CA 60419, CA 21765, and 591–598. CA 20180 from the National Cancer Institute, by a Center of 10 Pui C-H, Evans WE. Acute lymphoblastic leukemia. New Engl J Excellence grant from the State of Tennessee, and by the Med 1998; 339: 605–615. American Lebanese Syrian Associated Charities (ALSAC). 11 Neale GAM, Coustan-Smith E, Pan Q, Chen X, Gruhn B, Stow P, Behm FG, Pui C-H, Campana D. Tandem application of flow cytometry and polymerase chain reaction for comprehensive detection of minimal residual disease in childhood acute lym- Note added in proof phoblastic leukemia. Leukemia 1999; 13: 1221–1226. 12 Roberts WM, Estrov Z, Ouspenskaia MV, Johnston DA, McClain KL, Zipf TF. Measurement of residual leukemia during remission In a recent study, all of the 14 patients with a MRD level of in childhood acute lymphoblastic leukemia. New Engl J Med 10−4 or less at day 15 of induction therapy remained in con- 1997; 336: 317–323. tinuous complete remission with a median follow-up duration 13 Pongers-Willemse MJ, Verhagen OJHM, Tibbe GJM, Wijkhuijs AJ, of 5 years 6 months (Panzer-Gru¨mayer ER et al. Blood 2000; de Haas V, Roovers E, van der Schoot CE, van Dongen JJ. Real- time quantitative PCR for the detection of minimal residual disease 95: 790–794). in acute lymphoblastic leukemia using junctional region specific TaqMan probes. Leukemia 1998; 12: 2006–2014. 14 Coustan-Smith E, Behm FG, Sancho J, Boyett JM, Hancock ML, Rivera GK, Rubnitz JE, Sandlund JT, Pui C-H, Campana D. References Sequential monitoring of minimal residual disease in childhood acute lymphoblastic leukemia. Blood 1999; 94 (Suppl. 1): 626a. 1 Gaynon PS, Desai AA, Bostrom BC, Hutchinson RJ, Lange BJ, 15 Nachman JB, Sather HN, Sensel MG, Trigg ME, Cherlow JM, Nachman JB, Reaman GH, Sather HN, Steinherz PG, Trigg ME, Lukens JN, Wolff L, Uckun FM, Gaynon PS. Augmented post- Tubergen DG, Uckun FM. Early response to therapy and outcome induction therapy for children with high-risk acute lymphoblastic in childhood acute lymphoblastic leukemia. A review. Cancer leukemia and a slow response to initial therapy. New Engl J Med 1997; 80: 1717–1726. 1998; 338: 1663–1671. 2 Lilleyman JS. Clincial importance of speed of response to therapy 16 Chen J-S, Coustan-Smith E, Suzuki T, Neale GA, Pui C-H, Cam- in childhood lymphoblastic leukaemia. Leuk Lymphoma 1998; 31: pana D. Identification of novel leukemia markers by comparative 501–506. analysis of gene expression with DNA microarrays. Blood 1999; 3 Silverman LB, Gelber RD, Young ML, Dalton VK, Barr RD, Sallan 94 (Suppl. 1): 657a.

Leukemia