SYNOPSIS Detection of Minimal Residual Disease in Acute

Detection of minimal residual disease in acute lymphoblastic leukemia: the St Jude experience

After many years of laboratory and preclinical studies, assess- or T cell receptor (TCR) gene rearrangements in each patient’s ment of minimal residual disease (MRD) during treatment is malignant cells at the time of diagnosis. In many cancer cen- now widely accepted as a method of determining the effec- ters this procedure has been simplified by automation. At St tiveness of therapy for ALL. The use of MRD assays to guide Jude, for example, antigen-receptor gene configuration is therapy promises to further increase cure rates by preventing identified in ALL samples at the time of diagnosis by using a overtreatment of patients with highly responsive disease and relatively small panel of consensus primers and then per- undertreatment of those with more aggressive leukemia. forming direct automated sequencing of the PCR products (in Although they are clinically informative, current MRD assays the institution’s DNA chemistry core facility). After the leuke- have not been perfected, and their optimal clinical use has mia-specific sequences are determined, the corresponding not yet been established. probes are synthesized, also in the core facility. This process The most reliable methods of MRD detection in ALL include requires approximately 5 days. During the subsequent 2 to 10 flow cytometric profiling of aberrant immunophenotypes, days, the optimal PCR conditions for specific amplification of polymerase chain reaction (PCR) amplification of antigen- the leukemic rearrangement are identified. When the receptor genes, and PCR amplification of fusion transcripts. At remission sample arrives in the laboratory, assessment of MRD St Jude, all three methods are used to study MRD. Conven- requires only 5 to 6 h. tional karyotyping and fluorescence in situ hybridization, with There has been some debate about the optimal method of a sensitivity of 1% to 5%, are occasionally useful for clarifying quantifying MRD with PCR studies of antigen-receptor genes. the nature of morphologically suspect blast cells, but cannot We prefer limiting-dilution PCR based on detection of a single reliably detect submicroscopic leukemia. leukemic gene rearrangement by using two rounds of PCR Detection of MRD presents some unique technical chal- amplification. At the sample dilution at which both positive lenges that go beyond those of the individual method used. and negative results occur, the percentage of leukemic cells Therefore, investigators who have technical expertise, but can be estimated by using Poisson statistics. Limiting-dilution little experience in MRD detection, should exercise caution PCR is relatively rapid (5 to 6 h), has a uniform sensitivity, by following well-established methodological protocols or by and provides quantitative assessment of MRD. The main extensively testing newly developed approaches. If results drawback to this method is the need for a relatively large obtained in different centers are to be comparable, the uni- number of dilution replicates for sample analysis. formity of MRD measurements must be ensured. In this regard, Real-time PCR technology has great potential for MRD an international collaboration between St Jude and other cen- assessment. It ensures quantification of samples during the ters will use modern Internet-based communication methods exponential phase of PCR and eliminates processing of to exchange and review flow cytometry files used in MRD samples after amplification. There are, however, some disad- studies.1 The efforts of the Berlin–Frankfurt–Mu¨nster (BFM) vantages associated with this technology. The cost of synthes- Consortium to standardize PCR assays among participating izing a junction-specific fluorogenic probe is approximately centers are also commendable. 20 times that of synthesizing an unmodified oligonucleotide. Current strategies for flow cytometric detection of MRD rely This cost could be reduced through alternative PCR strategies, on combinations of leukocyte markers that do not normally such as those using a limited number of consensus probes, occur in cells of the peripheral blood and bone marrow. We positioned at the 3Ј end of V regions or J regions. In addition, identify such leukemia-associated phenotypes by quadruple- at present there are rather strict limitations on the nucleotide color staining techniques and include in our panels only composition of the probes and amplification primers which marker combinations that allow the detection of one leukemic may restrict the sensitivity of the procedure. cell among 104 or more normal cells.2 We can currently study Limited applicability and difficult quantitation of residual approximately 90% of ALL cases with this level of sensitivity. leukemia are the main limitations of MRD assays based on We have pioneered the use of DNA microarrays to identify RT-PCR amplification of fusion transcripts. Only approxi- new leukemia markers,3 with the aim of increasing the num- mately 40% of ALL cases have specific chromosomal aber- ber of patients that can be studied and simplify the complex rations with well-defined breakpoint-fusion regions. Quanti- antibody panels currently used. In this regard, investigators fication of PCR products has been significantly improved by should beware that simplification without extensive testing of the development of real-time PCR technology. Nevertheless, remission samples is likely to lead to an unacceptable number the relationship between the number of transcripts and the of false-negative results. number of residual tumor cells may be difficult to establish, The greatest obstacle to the use of PCR amplification as a because it is not yet clear whether the levels of transcripts routine MRD assay is that it requires the identification of Ig fluctuate during therapy. Moreover, the susceptibility of mRNA to degradation remains a potential problem. At our institution, MRD studies are performed primarily by flow cytometry and by PCR amplification of IgH genes. To Correspondence: D Campana, Department of Hematology-Oncology, St. Jude Children’s Research Hospital, 332 North Lauderdale, determine how well measurements obtained by these tech- Memphis TN 38105–2794, USA; Fax: 901 495 3749 niques were correlated, we studied serial dilutions of normal Received 27 March 2000; accepted 11 October 2000 and leukemic cells by both flow cytometry and PCR amplifi- Synopses 279 cation of IgH genes.4 We found the two methods to be highly is less extensive. Therefore, oncologists may be reluctant to concordant (r2 = 0.962). We then examined 62 bone marrow abandon the established prognostic factors. At present, a samples collected from children with ALL during clinical prudent course would be to combine MRD with clinical and remission.4 In 12 samples, both techniques detected MRD lev- biological parameters for comprehensive risk assignment of els у1in104. The percentages of leukemic cells measured children with ALL. by the two methods correlated well (r2=0.978). Of the remain- ing 50 samples, 48 had MRD levels Ͻ1in104. Results were discordant in only two samples: PCR detected two in 104 leu- Acknowledgements kemic cells in one sample and five in 104 in the other, whereas the flow cytometric assay was negative. Both patients This work was supported by grants CA60419, CA52259, remain in remission by clinical, flow cytometric, and molecu- CA21765 and CA20180 from the National Cancer Institute, lar criteria, one 18 months and one 28 months after testing. and by the American Lebanese Syrian Associated Charities In childhood ALL, results of studies based on flow cytometry (ALSAC). or PCR amplification of antigen receptor genes are remarkably concordant,5–7 suggesting that MRD studies with these D Campana Departments of Hematology-Oncology methods should be incorporated into treatment protocols. A GAM Neale and Pathology, St Jude Children’s single measurement at an informative time-point during ther- E Coustan-Smith Research Hospital, and University of apy may be sufficient for most patients. For example, patients C-H Pui Tennessee, Memphis, Tennessee, USA with B-lineage ALL who show no MRD at the end of remission induction have only a 5.8% 5-year cumulative incidence of relapse.8 The outlook for patients who do have MRD at the References end of induction therapy depends on the quantity and the per- sistence of MRD.8 The presence of 1% or more leukemic cells 1 Lorenzana R, Coustan-Smith E, Antillon F, Ribeiro RC, Campana D. Simple methods for the rapid exchange of flow cytometric data is strongly associated with subsequent relapse and indicates between remote centers. Leukemia 1999; 14: 336–337. high-risk ALL. Measurements at subsequent time-points may 2 Campana D, Coustan-Smith E. Detection of minimal residual dis- also help to evaluate risk. In our series, patients whose MRD ease in acute leukemia by flow cytometry. Cytometry 1999; 38: disappeared by week 14 of continuation therapy had a cumu- 139–152. lative incidence of relapse similar to that of patients who had 3 Chen J-S, Coustan-Smith E, Suzuki T, Neale GA, Pui CH, Campana D. Identification of novel leukemia markers by comparative analy- no MRD at any time after induction therapy. By contrast, if sis of gene expression with DNA microarrays. Blood 1999; 94: MRD persisted during continuation chemotherapy, the risk of 657a. subsequent relapse continued to increase as time elapsed. 4 Neale GA, Coustan-Smith E, Pan Q, Chen X, Gruhn B, Stow P, Thus, patients who have positive MRD findings at the end of Behm FG, Pui CH, Campana D. Tandem application of flow cyto- remission induction therapy may benefit from more frequent metry and polymerase chain reaction for comprehensive detection MRD testing during clinical remission. of minimal residual disease in childhood acute lymphoblastic leukemia. Leukemia 1999; 13: 1221–1226. Clearly, the potential benefits of MRD monitoring should 5 Coustan-Smith E, Behm FG, Sanchez J, Boyett JM, Hancock ML, be extended to all patients. Because we have found no single Raimondi SC, Rubnitz JE, Rivera GK, Sandlund JT, Pui CH, Cam- MRD detection technique that can be used for all patients, pana D. Immunological detection of minimal residual disease in multiple methods must be used for universal monitoring of children with acute lymphoblastic leukaemia. Lancet 1998; 351: MRD. At St Jude, we use flow cytometry and PCR amplifi- 550–554. cation of antigen-receptor genes simultaneously. Since we 6 Cave H, van der Werff ten Bosch, Suciu S, Guidal C, Waterkeyn C, Otten J, Bakkus M, Thielemans K, Grandchamp B, Vilmer E. started using this strategy, we have been able to conduct MRD Clinical significance of minimal residual disease in childhood studies in 80 consecutive ALL cases. This approach should acute lymphoblastic leukemia. European Organization for also prevent false-negative results caused by changes in Research and Treatment of Cancer – Childhood Leukemia Cooper- immunophenotype or predominant antigen-receptor gene ative Group. N Engl J Med 1998; 339: 591–598. clones during the course of the disease. 7 van Dongen JJ, Seriu T, Panzer-Grumayer ER, Biondi A, Pongers- Willemse MJ, Corral L, Stolz F, Schrappe M, Masera G, Kamps It remains to be decided how MRD assays should be used WA, Gadner H, Van Wering ER, Ludwig W-D, Basso G, de Bruijn to guide treatment. Should the results of these assays outweigh MAC, Cazzaniga G, Hettinger K, van der Does-van den Berg A, traditional prognostic features? Because prognostic features Hop WCJ, Riehm H, Bartram CR. Prognostic value of minimal predict response to therapy, whereas MRD assays measure it, residual disease in acute lymphoblastic leukaemia in childhood. it would be logical to give more weight to the latter. However, Lancet 1998; 352: 1731–1738. the association between response to treatment and clinical 8 Coustan-Smith E, Sancho J, Hancock ML, Boyett JM, Behm FG, Raimondi SC, Sandlund JT, Rivera GK, Rubnitz JE, Ribeiro RC, Pui and biologic parameters such as age, leukocyte count, and CH, Campana D. Clinical importance of minimal residual disease genetic features is well established, even with different treat- in childhood acute lymphoblastic leukemia. Blood 2000; 96: ment protocols, whereas the available information about MRD 2691–2696.

Leukemia