Published Ahead of Print on February 13, 2012 as /JCO J Clin Oncol by American Society of Clinical Oncology

Transcription

1 Published Ahead of Print on February 13, 212 as 1.12/JCO The latest version is at JOURNAL OF CLINICAL ONCOLOGY O R I G I N A L R E P O R T Minimal Residual Disease Quantification Is an Independent Predictor of Progression-Free and Overall Survival in Chronic Lymphocytic Leukemia: A Multivariate Analysis From the Randomized GCLLSG CLL8 Trial Sebastian Böttcher, Matthias Ritgen, Kirsten Fischer, Stephan Stilgenbauer, Raymonde M. Busch, Günter Fingerle-Rowson, Anna Maria Fink, Andreas Bühler, Thorsten Zenz, Michael Karl Wenger, Myriam Mendila, Clemens-Martin Wendtner, Barbara F. Eichhorst, Hartmut Döhner, Michael J. Hallek, and Michael Kneba Sebastian Böttcher, Matthias Ritgen, and Michael Kneba, University Hospital of Schleswig-Holstein, Campus Kiel, Kiel; Kirsten Fischer, Günter Fingerle- Rowson, Anna Maria Fink, Clemens- Martin Wendtner, Barbara F. Eichhorst, and Michael J. Hallek, University of Cologne, Cologne; Stephan Stilgenbauer, Andreas Bühler, Thorsten Zenz, and Hartmut Döhner, University of Ulm, Ulm; Raymonde M. Busch, Technical University, Munich, Germany; and Michael Karl Wenger and Myriam Mendila, Hoffmann-La Roche, Basel, Switzerland. Submitted May 7, 211; accepted December 5, 211; published online ahead of print at on February 13, 212. Supported by F. Hoffmann-La Roche, Basel, Switzerland. Presented at the 5th Annual Meeting of the American Society of Hematology, San Francisco, CA, December 6-9, 28. Authors disclosures of potential conflicts of interest and author contributions are found at the end of this article. Corresponding author: Sebastian Böttcher, MD, Second Department of Medicine, University Hospital of Schleswig-Holstein, Campus Kiel, Chemnitzstrasse 33, Kiel, Germany 24116; s.boettcher@med2.uni-kiel.de. 212 by American Society of Clinical Oncology X/12/399-1/$2. DOI: 1.12/JCO A B S T R A C T Purpose To determine the clinical significance of flow cytometric minimal residual disease (MRD) quantification in chronic lymphocytic leukemia (CLL) in addition to pretherapeutic risk factors and to compare the prognostic impact of MRD between the arms of the German CLL Study Group CLL8 trial. Patients and Methods s were prospectively quantified in 1,775 blood and bone marrow samples from 493 patients randomly assigned to receive fludarabine and cyclophosphamide (FC) or FC plus rituximab (FCR). Patients were categorized by MRD into low- ( 1 4 ), intermediate- ( 1 4 to 1 2 ), and high-level ( 1 2 ) groups. Results Low s during and after therapy were associated with longer progression-free survival (PFS) and overall survival (OS; P.1). Median PFS is estimated at 68.7, 4.5, and 15.4 months for low, intermediate, and high s, respectively, when assessed 2 months after therapy. Compared with patients with low MRD, greater risks of disease progression were associated with intermediate and high s (hazard ratios, 2.49 and 14.7, respectively; both P.1). Median OS was 48.4 months in patients with high MRD and was not reached for lower MRD levels. MRD remained predictive for OS and PFS in multivariate analyses that included the most important pretherapeutic risk markers in CLL. PFS and OS did not differ between treatment arms within each MRD category. However, FCR induced low s more frequently than FC. Conclusion s independently predict OS and PFS in CLL. Therefore, MRD quantification might serve as a surrogate marker to assess treatment efficacy in randomized trials before clinical end points can be evaluated. J Clin Oncol by American Society of Clinical Oncology INTRODUCTION With the advent of chemoimmunotherapy, median progression-free survival (PFS) in chronic lymphocytic leukemia (CLL) now ranges from 3.5 to 6.7 years after first-line therapy. 1-3 However, further prolongation of disease control would be highly desirable especially for younger patients. Allogeneic stem-cell transplantation 4 and drugs such as lenalidomide, 5 flavopiridol, 6-8 bendamustine, 9 as well as novel antibodies 1-12 have recently shown promising activity. Unfortunately, some of the most effective therapies are associated with significant toxicities. Therefore, the prediction of the individual remission duration gains importance in avoiding overtreatment in low-risk patients. Deletionsofchromosomes11qand17p, mutated TP53, unmutated IGHV status, ZAP-7 expression, increased serum levels of 2 -microglobulin, and thymidine kinase, as well as advanced clinical stage are associated with poor prognosis. 2,3,13-26 Nevertheless, these established risk features still fail to predict the outcome in substantial numbers of patients. For example, although mutated TP53 and deletion 17p are strongly associated with resistance to nucleoside analogs, the combination of both factors 212 by American Society of Clinical Oncology 1 Copyright 212 by American Society of Clinical Oncology

2 Böttcher et al identified only 29% of all fludarabine-resistant patients in a recent investigation. 13 The sensitive quantification of minimal residual disease (MRD) after treatment has been suggested as an alternative means to predict response duration 19,27-32 and overall survival. 19,27,28 However, since the prognostic significance of MRD has never been assessed together with pretherapeutic risk factors such as cytogenetic aberrations in multivariate analyses, the added value of this test remained a matter of controversy. 33,34 Moreover, it was unknown whether or not the prognostic impact of MRD would be independent from the therapeutic regimen used. The German CLL Study Group (GCLLSG) recently published the results of a randomized trial that demonstrated the efficacy of adding rituximab to fludarabine and cyclophosphamide (FC) chemotherapy in treatment-naive patients with CLL. 2 By using prospective flow cytometric MRD quantification in 493 patients from this trial, we found that identical s predict for similar PFS and overall survival (OS) irrespective of treatment arm. We firmly establish MRD as an independent prognostic factor in CLL. PATIENTS AND METHODS Patients MRD was scheduled to be prospectively assessed in German and Austrian patients participating in the GCLLSG CLL8 trial (ClinicalTrials.gov number NCT281918). The protocol was approved by institutional review boards and ethics committees of each participating institution. The patients provided written informed consent to participate in the trial and to undergo MRD testing. 2 According to protocol, staging and parallel MRD testing in peripheral blood (PB) were scheduled before starting the therapy (initial staging), after three cycles of therapy (interim staging), 1 month after the last treatment cycle (initial response assessment), 2 months thereafter (final restaging), and subsequently every 3 months. Patients with complete remission (CR) by clinical examination at initial response assessment were subjected to a bone marrow (BM) examination at final restaging. Details on eligibility, treatment, staging procedures, clinical efficacy, and laboratory and statistical analyses have been reported elsewhere. 2 MRD MRD was quantified by four-color flow cytometry with a sensitivity of at least 1 4 according to the technique previously described and validated against allele-specific oligonucleotide primer real-time quantitative IGH polymerase chain reaction. 35 The method uses an international standardized approach 36 with minor modifications on patient samples received within 48 hours after collection. In particular, tubes that contained more than 2 CLL cells were regarded as positive. A positive MRD result for a sample required positive results from two separate tubes. Negative MRD test results from samples that did not allow the acquisition of at least 2, leukocytes in two tubes were excluded from the analysis. s are reported as fraction of CLL cells of all nucleated cells. MRD analyses were centrally performed in the GCLLSG MRD laboratory in Kiel. Statistical Analysis Quantitative MRD results were categorized into low- ( 1 4 ), intermediate- ( 1 4 to 1 2 ), and high-level ( 1 2 ) groups. Low-level MRD equals MRD negativity according to the current International Workshop on Chronic Lymphocytic Leukemia definition, 37 whereas intermediate and high levels correspond to MRD positive patients. Independent analyses of MRD data were performed at several predefined time points and also separately for PB and BM. Samples from each patient could contribute to up to six MRD analyses: in PB at initial staging, interim staging, initial response assessment, final restaging, and during follow-up (first MRD measurement between 36 and 54 days after final restaging) and also in BM at final restaging. Per protocol, patients were subjected to a BM examination only to confirm a CR; complete responders were therefore over-represented in the cohort of patients who contributed to MRD analyses in BM. Because of increased relapse frequency associated with poor-risk features, patients with a deletion 17p and those without CR were under-represented in the patient cohort that contributed to MRD monitoring during follow-up. B symptoms and a mutated IGHV were more frequent in the MRD group tested at final restaging in PB. Patients assessed for MRD were not significantly different from the remaining patients in the CLL8 trial with respect to age, 2 -microglobulin, thymidine kinase, Binet stage, Eastern Cooperative Oncology Group (ECOG) performance status, treatment regimen, or sex at any of the analyzed time points (data not shown). Fisher s exact and Mann-Whitney U tests were used for analyses of categorical and continuous variables, respectively. OS and PFS were evaluated with a median follow-up of 52.4 months by using Kaplan-Meier estimates and were compared between MRD groups by log-rank test for trend. Cox proportional hazard models with stepwise backward selection were applied to OS and PFS. The models were calculated separately for the individual analyses of sampling time points and sample material. Two-sided significance levels were set at.5. Analyses were done by using SPSS (SPSS, Chicago, IL) and GraphPad Prism (GraphPad Software, La Jolla, CA) software programs. RESULTS Patient Characteristics s were assessed in 1,775 PB and BM samples from 493 patients (6.3% of the total of 817 patients in the CLL8 trial; 9.5% of the 545 Austrian and German patients eligible for MRD monitoring). Patients who contributed to any of the six separate MRD analyses (at different time points and in PB or BM) were more likely to receive the planned six cycles of therapy and comprised more patients with a chromosomal deletion 11q compared with the remaining patients in the CLL8 trial. Patients who underwent MRD monitoring did not differ significantly in any other variable from patients in the CLL8 trial without MRD measurements (Appendix Table A1, online only). The sample acquisition is detailed in Appendix Table A2 (online only). MRD Kinetics Treatment with both FC and FC plus rituximab (FCR) significantly reduced s, but more profound reductions of MRD were observed in patients who received FCR (Fig 1). Whereas tumor load was well balanced before therapy (FC v FCR: v ), the initial three applications of therapy induced a significant, roughly 1-fold difference in median MRD levels between treatment arms (FC v FCR at interim staging: v ; P.1). Both FC and FCR treatment after interim staging further reduced MRD when assessed at initial response assessment or final restaging. Differences between the arms persisted at final restaging, resulting in median s below 1 4 after FCR versus following FC (P.1). PB s below 1 4 at final restaging were measured in 35% of patients treated with FC and in 63% of patients who received FCR (P.1; Fig 1). The proportion of patients presenting with low MRD in PB increased from 17% at interim staging to 49% at final restaging (Table 1). Treatment after interim staging redistributed 5% of all patients into lower MRD categories, whereas only one patient showed increasing tumor load (data not shown). Patients treated with FCR attained lower median s in BM at final restaging ( ) than those who received FC only ( ; P.1; Fig 1). MRD below 1 4 was more frequently by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY

3 MRD as Independent Predictor of PFS and OS in CLL MRD Level < 1-4 N FC FCR Initial *** *** *** *** ** % 26% 33% 6% 35% 63% 22% 52% 28% 44% FC FCR Interim FC FCR IRA FC FCR FR Peripheral Blood FC FCR FU FC FCR FR BM Fig 1. Minimal residual disease (MRD) levels at different time points in patients treated with fludarabine and cyclophosphamide (FC; triangles) and FC plus rituximab (FCR; diamonds). Individual triangles and diamonds symbolize patients; blue lines depict the medians. Top line numbers: total number of patients assessed; bottom numbers: number and percentage of patients with MRD levels 1 4. Comparison between treatment arms: *** P.1; **P.7. Comparisons of consecutive time points within a treatment arm in peripheral blood: initial versus interim and interim versus initial response assessment (IRA): P.1 each in both FC and FCR arms. IRA versus final restaging (FR): not significant in both treatment arms. FR versus follow-up (FU): P.1 (FC) and P.27 (FCR). BM, bone marrow. observed in BM after FCR treatment (44%) than after FC treatment (28%). The only pretherapeutic feature significantly associated with attaining a low at final restaging after FCR treatment was the absence of a deletion 17p, whereas low s more often occurred after FC treatment in patients without chromosomal aberrations according to the Döhner hierarchical model 2 and in those with mutated IGHV genes (Appendix Tables A3 and A4, online only). A significant increase in s appeared in both arms during follow-up, but MRD remained different between the two treatment groups (medians: FC, v FCR, 1 4 ; P.1). Prediction of PFS Low-level MRD results from interim staging onward were significantly associated with longer PFS regardless of sample material and sampling time point (P.1 by log-rank test each; Fig 2). Pair-wise comparisons between the MRD groups demonstrated that each increase in MRD significantly increased the risk for disease progression (Table 1 and Appendix Table A5, online only). Depending on sampling time point and material, median PFS estimates differed from 18.9 to 29.6 months between consecutive groups. The prognosis of patients who belonged to the high MRD group at final restaging was even inferior to the prognosis of patients with identical s at interim staging (median PFS, 15.4 v 23.3 months). However, this difference of 7.9 months is considerably smaller than the difference between median PFS associated with different MRD categories at any given time point (Table 1 and Appendix Table A5). Patients who reached a clinical CR were more likely to present with low-level MRD at final restaging in PB (P.1) and in BM (P.1) compared with partial responders (Appendix Table A6, online only). Nevertheless, even after grouping patients by clinical response, MRD categories assessed in PB and BM still significantly predicted for PFS (log-rank test for trend, Appendix Table A6 and Appendix Fig A1, online only). Both low-level MRD and CR independently predicted for longer PFS. Therefore, median PFS was longer when the same was attained in complete relative to partial responders. FCR more often led to low s compared with FC chemotherapy (Fig 1). However, patients from the two treatment arms who presented with the same s at interim staging or at final restaging had no significantly different risks for disease progression (Fig 3). In contrast, the likelihood of progression significantly differed between all MRD cohorts after additional grouping by treatment arm. Table 1. PFS and OS by MRD Group Assessed at Interim Staging and at Final Restaging in PB MRD Cohort Patients No. % Median (months SE) PFS 75% (months SE) HR 95% CI P Median (months SE) OS 75% (months SE) HR 95% CI P Interim staging N/R N/R 1 4 to N/R v 1 4 to to to v to to to 1 2 v to to Final restaging in PB N/R N/R 1 4 to N/R v 1 4 to to to v to to to 1 2 v to to NOTE. P.1 for OS and PFS at interim staging and final restaging. Abbreviations: HR, hazard ratio; MRD, minimal residual disease; N/R, not reached; OS, overall survival; PB, peripheral blood; PFS, progression-free survival by American Society of Clinical Oncology 3

4 Böttcher et al A < to < B < to < C < to < D < to < Fig 2. Progression-free survival (PFS) in patients with chronic lymphocytic leukemia grouped by minimal residual disease (MRD) levels assessed (A) in peripheral blood (PB) at interim staging, (B) in PB at final restaging, (C) in bone marrow at final restaging, and (D) in PB during follow-up. The inserted pie charts represent the frequency distributions. P.1 for all analyses by log-rank test. Only the differences between low- and intermediate-level cohorts at interim staging currently lack statistical significance when assessed separately by treatment arm, likely reflecting a low event number. Similarly, the prognostic significance of MRD was largely independent from treatment when determined at initial response assessment, at final restaging in BM, or during follow-up (data not shown). The significance of MRD for PFS was tested in Cox regression analyses separately for four time points after initiation of treatment and separately for PB and BM (Table 2 and Appendix Table A7, online only). Parameters initially included in the models because of significance for PFS in univariate analyses were clinical response, deletion 17p, IGHV mutational status, application of full number of treatment cycles, FCR treatment, and thymidine kinase levels. In addition, we included 2 -microglobulin and pretherapeutic WBC because those variables were predictive in a multivariate model for the CLL8 cohort as a whole. 2 MRD remained significant for PFS after backward selection in all separate analyses, being the parameter most closely associated with PFS in most of them. Additional independent predictors for PFS at least at certain time points were clinical response, deletion 17p, the applied treatment cycle number, and thymidine kinase levels. Prediction of OS Higher s at all time points were associated with significantly shorter OS (log-rank P.1 each; Fig 4 and data not shown). Patients with high s are more likely to die compared with patients from both intermediate- and low-level groups (Table 1 and Appendix Table A5). The differences between low- and intermediatelevel MRD cohorts are not statistically significant with current followup. Median OS was longer in patients in whom s of more than 1 2 were observed at interim staging compared with patients in whom, even at final restaging, such high levels were still detectable (69.2 months v 48.4 months). Paralleling our results for PFS, the OS by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY

5 MRD as Independent Predictor of PFS and OS in CLL A.2 and treatment FC 1-2 FC 1-4 to < 1-2 FC < 1-4 FCR 1-2 FCR 1-4 to <1-2 FCR < 1-4 B.2 and treatment FC 1-2 FC 1-4 to < 1-2 FC < 1-4 FCR 1-2 FCR 1-4 to <1-2 FCR < C D.2 and treatment FC 1-2 FC 1-4 to < 1-2 FC < 1-4 FCR 1-2 FCR 1-4 to <1-2 FCR < and treatment FC 1-2 FC 1-4 to < 1-2 FC < 1-4 FCR 1-2 FCR 1-4 to <1-2 FCR < FC, 45% FCR, 22% FC, 48% FCR, 53% FC, 8% FCR, 26% FC, 18% FCR, 13% FC, 47% FCR, 24% FC, 35% FCR, 63% OS (months) OS (months) E F to < 1-2 < 1-4 MRD Category to < 1-2 < 1-4 MRD Category Fig 3. (A, B) Progression-free survival (PFS) and (C, D) overall survival (OS) grouped by treatment regimen and by minimum residual disease (MRD) category assessed at interim staging (A, C, E) and at first restaging (B, D, F) in peripheral blood. Histograms E and F depict the frequency distributions of MRD groups at both analysis A, time points. Percentages in the histograms denote frequency within a treatment arm. FC, fludarabine and cyclophosphamide; FCR, FC plus rituximab by American Society of Clinical Oncology 5

6 Böttcher et al Table 2. Multivariate Analysis of the Effects of Prognostic Factors on PFS and OS in Combination With PB MRD Levels Assessed at Interim Staging and at Final Restaging Variable Interim Staging (n 286) PFS Final Restaging (n 254) Interim Staging (n 274) OS Final Restaging (n 243) HR 95% CI P HR 95% CI P HR 95% CI P HR 95% CI P Overall to 1 2 v to to to to v to to to to Clinical response Overall PR v CR to to to NR v CR to to to 1.71 ECOG PS to to Del(17p) present to to to to IGHV unmutated to to Thymidine kinase 1 U/L to No. of treatment cycles to NOTE. Blank spaces denote the lack of a significant association in multivariate analysis. Abbreviations: CR, complete response; ECOG, Eastern Cooperative Oncology Group; HR, hazard ratio; MRD, minimal residual disease; NR, nonresponder; OS, overall survival; PB, peripheral blood; PFS, progression-free survival; PR, partial response; PS, performance score. was comparable once patients from both treatment arms were categorized according to the during and after therapy. MRD status itself was predictive for survival (Fig 4). Besides MRD, all variables that showed associations with survival in univariate analyses (age, 2 -microglobulin, thymidine kinase, IGHV status, deletion 17p, clinical response, ECOG performance status, application of all treatment cycles) and treatment regimen (significant in the CLL8 cohort as a whole 2 ) were included in Cox regression models for OS. After backward selection, MRD levels remained significant predictors for survival at all time points and regardless of sample material (Table 2 and Appendix Table A7). The only other variable always predictive was the deletion 17p, whereas ECOG performance status, clinical response, IGHV status, thymidine kinase, 2 -microglobulin, and age appeared as independent survival predictors at certain time points only. DISCUSSION This large multicentric study assesses for the first time (to the best of our knowledge) the prognostic significance of MRD in CLL in a randomized trial. Measurements at consecutive time points during and after therapy gave insights into MRD kinetics and allowed us to compare the prognostic significance of MRD at different stages of therapy. Achieving low-level MRD was associated with a comparable clinical benefit for patients in both treatment arms. Patients who attained low-level MRD by FC chemotherapy had PFS similar to that of patients who achieved the low CLL cell levels with FCR. The superiority of the more active FCR regimen over FC was reflected by a greater chance to achieve low-level disease. Thus, the profound reduction of tumor load and not the treatment regimen by which this reduction is induced is the key factor for durable remissions. Although MRD has already been used as an indicator for treatment efficacy, 12,14,38-43 we herein prove for the first time that the method is able to identify a clinically superior treatment arm of a randomized trial. The availability of MRD data shortly after treatment is important, because with more effective treatment regimens, PFS will be evaluable only after long observation periods. Our data predict that in future randomized trials, the treatment arm with longer PFS will be identified by lower MRD, provided that the treatment subsequent to the analysis time point is comparable between the arms. PFS depends on tumor load after therapy and regrowth kinetics. We therefore expect equal MRD levels in future trials to forecast similar remission duration in patients with CLL with comparable biologic risk profile and regrowth kinetics. In addition to the efficacy of the treatment regimen, the number of treatment cycles likely also influences the MRD response at the end of treatment. We observed MRD kinetics proving the added efficacy of treatment cycles applied after interim staging. The data therefore corroborate the application of six cycles as treatment standard for patients with CLL with risk features comparable to CLL8. However, this finding does not rule out that those patients who achieve low-level MRD after the first three treatment cycles are overtreated if they receive the full six treatment cycles. In keeping with this hypothesis, we observed similar PFS in patients who had already achieved low s after three treatment cycles compared with those who required the full treatment to attain this status. It is therefore possible that the additional therapy after interim staging did not contribute to treatment efficacy in patients with an excellent MRD response early in the course of therapy. A randomized trial comparing MRD-guided with fixed cycle therapy would be required to eventually prove that hypothesis. The study corroborates the concept that low s are a desirable goal of CLL therapy, as suggested by previous publications. 19,33,34 Patients were comprehensively characterized for clinical and biologic risk markers allowing investigation of the added prognostic impact of MRD for the first time. The analyses demonstrate an by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY

7 MRD as Independent Predictor of PFS and OS in CLL A B < to < < to < C OS (months) D OS (months) < to < < to < OS (months) OS (months) Fig 4. Overall survival (OS) in patients with chronic lymphocytic leukemia grouped by minimal residual disease (MRD) levels assessed (A) in peripheral blood (PB) at interim staging, (B) in PB at final restaging, (C) in bone marrow (BM) at final restaging, and (D) in PB during follow-up. The inserted pie charts represent the frequency distributions. P.1 for all analyses by log-rank test. independent and clinically relevant prognostic significance of MRD on OS and PFS in multivariate models. Consequently, MRD is not merely a surrogate marker for biologically defined CLL subgroups, but it can be viewed as the integrator of an array of treatment, host, and disease-specific features in individual patients. This probably explains why MRD belonged to the factors that were most closely associated with the risks of progression and of dying. Although the data presented herein are already mature for PFS and for OS in poor-risk patients, additional correlations between s and OS are expected with longer follow-up in good-risk patients. Nonadherence to the protocol in this multicenter trial is the principal reason why samples were not available at all time points from all eligible patients. This fact notwithstanding, there was no major bias in risk features between patients who could be monitoredformrdandthosewhowerenot, sothatdataarevalidforthecll8 trial as a whole. Low-level MRD does not equal complete disease eradication, but it is an important prognostic factor in a noncurative treatment setting. We and others have reported direct measurements of s below 1 4 before. 28,35 Moreover, we observed increasing s during follow-up, even in patients with low-level MRD after therapy, thus corroborating earlier reports. 19,27 Different quantitative levels within the range of MRD positivity are at least as important for prognosis as the distinction between MRD negativity (below 1 4 by current definition 37 ) and MRD positivity, thus supporting earlier results from a small study. 28 MRD was predictive in partial and complete responders in our trial, but similar s predicted for better outcome in patients with a CR compared with patients who have a partial response. In summary, our data emphasize the prognostic significance of the quality of remission after therapy. Patients with low tumor load, as quantified by a combination of MRD and clinical staging, enjoy long PFS. The results are in concordance with earlier reports showing an advantageous clinical course in good responders, when assessed by clinical staging, 1,44 and sensitive 19,27-32,45-48 or relatively insensitive 1,44,49,5 MRD determinations by American Society of Clinical Oncology 7

8 Böttcher et al On the basis of the data presented here, MRD quantification in CLL qualifies as a parameter to compare treatment efficacy between the arms of randomized trials before the availability of clinical end point data. Even more importantly, MRD might guide maintenance and consolidation strategies, thus making a step forward toward tailored treatment strategies in this disease. AUTHORS DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a U are those for which no compensation was received; those relationships marked with a C were compensated. For a detailed description of the disclosure categories, or for more information about ASCO s conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors. Employment or Leadership Position: Günter Fingerle-Rowson, F. Hoffmann-La Roche (C); Michael Karl Wenger, F. Hoffmann-La Roche (C); Myriam Mendila, F. Hoffmann-La Roche (C) Consultant or Advisory Role: Stephan Stilgenbauer, F. Hoffmann-La Roche (C); Clemens-Martin Wendtner, F. Hoffmann-La Roche (C); Michael J. Hallek, F. Hoffmann-La Roche (C); Michael Kneba, F. Hoffmann-La Roche (C) Stock Ownership: Michael Karl Wenger, F. Hoffmann-La Roche; Myriam Mendila, F. Hoffmann-La Roche Honoraria: Sebastian Böttcher, F. Hoffmann-La Roche; Stephan Stilgenbauer, F. Hoffmann-La Roche; Thorsten Zenz, F. Hoffmann-La Roche; Clemens-Martin Wendtner, F. Hoffmann-La Roche; Barbara F. Eichhorst, F. Hoffmann-La Roche; Michael J. Hallek, F. Hoffmann-La Roche; Michael Kneba, F. Hoffmann-La Roche Research Funding: Sebastian Böttcher, F. Hoffmann-La Roche; Matthias Ritgen, F. Hoffmann-La Roche; Stephan Stilgenbauer, F. Hoffmann-La Roche; Clemens-Martin Wendtner, F. Hoffmann-La Roche; Barbara F. Eichhorst, F. Hoffmann-La Roche, Mundipharma; Michael J. Hallek, F. Hoffmann-La Roche; Michael Kneba, F. Hoffmann-La Roche Expert Testimony: None Other Remuneration: Kirsten Fischer, F. Hoffmann-La Roche; Anna Maria Fink, Travel grants: F. Hoffmann-La Roche AUTHOR CONTRIBUTIONS Conception and design: Sebastian Böttcher, Matthias Ritgen, Stephan Stilgenbauer, Raymonde M. Busch, Michael Karl Wenger, Myriam Mendila, Barbara F. Eichhorst, Michael J. Hallek, Michael Kneba Financial support: Michael Karl Wenger Provision of study materials or patients: Sebastian Böttcher, Stephan Stilgenbauer, Günter Fingerle-Rowson, Clemens-Martin Wendtner, Hartmut Döhner, Michael J. Hallek, Michael Kneba Collection and assembly of data: Sebastian Böttcher, Matthias Ritgen, Kirsten Fischer, Stephan Stilgenbauer, Günter Fingerle-Rowson, Anna Maria Fink, Andreas Bühler, Thorsten Zenz, Clemens-Martin Wendtner, Hartmut Döhner Data analysis and interpretation: Sebastian Böttcher, Matthias Ritgen, Stephan Stilgenbauer, Raymonde M. Busch, Michael Karl Wenger, Myriam Mendila, Barbara F. Eichhorst, Michael J. Hallek, Michael Kneba Manuscript writing: All authors Final approval of manuscript: All authors REFERENCES 1. Tam CS, O Brien S, Wierda W, et al: Longterm results of the fludarabine, cyclophosphamide, and rituximab regimen as initial therapy of chronic lymphocytic leukemia. Blood 112:975-98, Hallek M, Fischer K, Fingerle-Rowson G, et al: Addition of rituximab to fludarabine and cyclophosphamide in patients with chronic lymphocytic leukaemia: A randomised, open-label, phase 3 trial. Lancet 376: , Woyach JA, Ruppert AS, Heerema NA, et al: Chemoimmunotherapy with fludarabine and rituximab produces extended overall survival and progression-free survival in chronic lymphocytic leukemia: Long-term follow-up of CALGB study J Clin Oncol, 29: , Dreger P, Döhner H, Ritgen M, et al: Allogeneic stem cell transplantation provides durable disease control in poor-risk chronic lymphocytic leukemia: Long-term clinical and MRD results of the German CLL Study Group CLL3X trial. Blood 116: , Chen CI, Bergsagel PL, Paul H, et al: Singleagent lenalidomide in the treatment of previously untreated chronic lymphocytic leukemia. J Clin Oncol 29: , Byrd JC, Lin TS, Dalton JT, et al: Flavopiridol administered using a pharmacologically derived schedule is associated with marked clinical efficacy in refractory, genetically high-risk chronic lymphocytic leukemia. Blood 19:399-44, Lin TS, Ruppert AS, Johnson AJ, et al: Phase II study of flavopiridol in relapsed chronic lymphocytic leukemia demonstrating high response rates in genetically high-risk disease. J Clin Oncol 27: , Phelps MA, Lin TS, Johnson AJ, et al: Clinical response and pharmacokinetics from a phase 1 study of an active dosing schedule of flavopiridol in relapsed chronic lymphocytic leukemia. Blood 113: , Knauf WU, Lissichkov T, Aldaoud A, et al: Phase III randomized study of bendamustine compared with chlorambucil in previously untreated patients with chronic lymphocytic leukemia. J Clin Oncol 27: , Wierda WG, Kipps TJ, Mayer J, et al: Ofatumumab as single-agent CD2 immunotherapy in fludarabine-refractory chronic lymphocytic leukemia. J Clin Oncol 28: , Stilgenbauer S, Zenz T, Winkler D, et al: Subcutaneous alemtuzumab in fludarabinerefractory chronic lymphocytic leukemia: Clinical results and prognostic marker analyses from the CLL2H study of the German Chronic Lymphocytic Leukemia Study Group. J Clin Oncol 27: , Hillmen P, Skotnicki AB, Robak T, et al: Alemtuzumab compared with chlorambucil as first-line therapy for chronic lymphocytic leukemia. J Clin Oncol 25: , Zenz T, Eichhorst B, Busch R, et al: TP53 mutation and survival in chronic lymphocytic leukemia. J Clin Oncol 28: , Bosch F, Abrisqueta P, Villamor N, et al: Rituximab, fludarabine, cyclophosphamide, and mitoxantrone: A new, highly active chemoimmunotherapy regimen for chronic lymphocytic leukemia. J Clin Oncol 27: , Byrd JC, Gribben JG, Peterson BL, et al: Select high-risk genetic features predict earlier progression following chemoimmunotherapy with fludarabine and rituximab in chronic lymphocytic leukemia: Justification for risk-adapted therapy. J Clin Oncol 24: , Oscier D, Wade R, Davis Z, et al: Prognostic factors identified three risk groups in the LRF CLL4 trial, independent of treatment allocation. Haematologica 95: , Oscier DG, Gardiner AC, Mould SJ, et al: Multivariate analysis of prognostic factors in CLL: Clinical stage, IGVH gene mutational status, and loss or mutation of the p53 gene are independent prognostic factors. Blood 1: , Keating MJ, Flinn I, Jain V, et al: Therapeutic role of alemtuzumab (Campath-1H) in patients who have failed fludarabine: Results of a large international study. Blood 99: , Moreton P, Kennedy B, Lucas G, et al: Eradication of minimal residual disease in B-cell chronic lymphocytic leukemia after alemtuzumab therapy is associated with prolonged survival. J Clin Oncol 23: , Döhner H, Stilgenbauer S, Benner A, et al: Genomic aberrations and survival in chronic lymphocytic leukemia. N Engl J Med 343: , Orchard JA, Ibbotson RE, Davis Z, et al: ZAP-7 expression and prognosis in chronic lymphocytic leukaemia. Lancet 363:15-111, Rassenti LZ, Jain S, Keating MJ, et al: Relative value of ZAP-7, CD38, and immunoglobulin mutation status in predicting aggressive disease in chronic lymphocytic leukemia. Blood 112: , Rassenti LZ, Huynh L, Toy TL, et al: ZAP-7 compared with immunoglobulin heavy-chain gene mutation status as a predictor of disease progression in by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY

9 MRD as Independent Predictor of PFS and OS in CLL chronic lymphocytic leukemia. N Engl J Med 351:893-91, Del Principe MI, Del Poeta G, Buccisano F, et al: Clinical significance of ZAP-7 protein expression in B-cell chronic lymphocytic leukemia. Blood 18: , Gonzalez D, Martinez P, Wade R, et al: Mutational status of the TP53 gene as a predictor of response and survival in patients with chronic lymphocytic leukemia: Results from the LRF CLL4 trial. J Clin Oncol 29: , Crespo M, Bosch F, Villamor N, et al: ZAP-7 expression as a surrogate for immunoglobulinvariable-region mutations in chronic lymphocytic leukemia. N Engl J Med 348: , Rawstron AC, Kennedy B, Evans PA, et al: Quantitation of minimal residual disease levels in chronic lymphocytic leukemia using a sensitive flow cytometric assay improves the prediction of outcome and can be used to optimize therapy. Blood 98:29-35, Moreno C, Villamor N, Colomer D, et al: Clinical significance of minimal residual disease, as assessed by different techniques, after stem cell transplantation for chronic lymphocytic leukemia. Blood 17: , Lamanna N, Jurcic JG, Noy A, et al: Sequential therapy with fludarabine, high-dose cyclophosphamide, and rituximab in previously untreated patients with chronic lymphocytic leukemia produces highquality responses: Molecular remissions predict for durable complete responses. J Clin Oncol 27: , Ysebaert L, Gross E, Kühlein E, et al: Immune recovery after fludarabine-cyclophosphamide-rituximab treatment in B-chronic lymphocytic leukemia: Implication for maintenance immunotherapy. Leukemia 24: , Castro JE, James DF, Sandoval-Sus JD, et al: Rituximab in combination with high-dose methylprednisolone for the treatment of chronic lymphocytic leukemia. Leukemia 23: , Maloum K, Settegrana C, Chapiro E, et al: IGHV gene mutational status and LPL/ADAM29 gene expression as clinical outcome predictors in CLL patients in remission following treatment with oral fludarabine plus cyclophosphamide. Ann Hematol 88: , Goodman M: Alemtuzumab until minimal residual disease for chronic lymphocytic leukemia: Is it a new standard? J Clin Oncol 23: , 25; author reply Moreno C, Ritgen M, Rawstron A: Is MRD eradication a desirable goal in CLL? Best Pract Res Clin Haematol 23:97-17, Böttcher S, Stilgenbauer S, Busch R, et al: Standardized MRD flow and ASO IGH RQ-PCR for MRD quantification in CLL patients after rituximabcontaining immunochemotherapy: A comparative analysis. Leukemia 23:27-217, Rawstron AC, Villamor N, Ritgen M, et al: International standardized approach for flow cytometric residual disease monitoring in chronic lymphocytic leukaemia. Leukemia 21: , Hallek M, Cheson BD, Catovsky D, et al: Guidelines for the diagnosis and treatment of chronic lymphocytic leukemia: A report from the International Workshop on Chronic Lymphocytic Leukemia updating the National Cancer Institute- Working Group 1996 guidelines. Blood 111: , Bosch F, Ferrer A, Villamor N, et al: Fludarabine, cyclophosphamide, and mitoxantrone as initial therapy of chronic lymphocytic leukemia: High response rate and disease eradication. Clin Cancer Res 14: , Hillmen P, Cohen DR, Cocks K, et al: A randomized phase II trial of fludarabine, cyclophosphamide and mitoxantrone (FCM) with or without rituximab in previously treated chronic lymphocytic leukaemia. Br J Haematol 152:57-578, Cazin B, Divine M, Leprêtr S, et al: High efficacy with five days schedule of oral fludarabine phosphate and cyclophosphamide in patients with previously untreated chronic lymphocytic leukaemia. Br J Haematol 143:54-59, Del Poeta G, Del Principe MI, Buccisano F, et al: Consolidation and maintenance immunotherapy with rituximab improve clinical outcome in patients with B-cell chronic lymphocytic leukemia. Cancer 112: , Montillo M, Tedeschi A, Miqueleiz S, et al: Alemtuzumab as consolidation after a response to fludarabine is effective in purging residual disease in patients with chronic lymphocytic leukemia. J Clin Oncol 24: , Faderl S, Wierda W, O Brien S, et al: Fludarabine, cyclophosphamide, mitoxantrone plus rituximab (FCM-R) in frontline CLL 7 Years. Leuk Res 34: , Kay NE, Geyer SM, Call TG, et al: Combination chemoimmunotherapy with pentostatin, cyclophosphamide, and rituximab shows significant clinical activity with low accompanying toxicity in previously untreated B chronic lymphocytic leukemia. Blood 19:45-411, Bosch F, Ferrer A, López-Guillermo A, et al: Fludarabine, cyclophosphamide and mitoxantrone in the treatment of resistant or relapsed chronic lymphocytic leukaemia. Br J Haematol 119: , Wierda W, O Brien S, Wen S, et al: Chemoimmunotherapy with fludarabine, cyclophosphamide, and rituximab for relapsed and refractory chronic lymphocytic leukemia. J Clin Oncol 23:47-478, Montillo M, Tedeschi A, Petrizzi VB, et al: An open-label, pilot study of fludarabine, cyclophosphamide, and alemtuzumab in relapsed/refractory patients with B-cell chronic lymphocytic leukemia. Blood 118: , Badoux XC, Keating MJ, Wang X, et al: Fludarabine, cyclophosphamide, and rituximab chemoimmunotherapy is highly effective treatment for relapsed patients with CLL. Blood 117: , Keating MJ, O Brien S, Albitar M, et al: Early results of a chemoimmunotherapy regimen of fludarabine, cyclophosphamide, and rituximab as initial therapy for chronic lymphocytic leukemia. J Clin Oncol 23: , O Brien SM, Kantarjian HM, Thomas DA, et al: Alemtuzumab as treatment for residual disease after chemotherapy in patients with chronic lymphocytic leukemia. Cancer 98: , 23 Acknowledgment We thank the patients who participated in the CLL8 trial and the physicians who treated them. We also thank Stephan Zurfluh and Jamie Wingate for their excellent support during the conduct of this trial. We thank Elke Harbst, Jamileh Hanani, Maike Starken, Lada Henseleit, and Linda Falck for excellent technical assistance as well as Anne Westermann and Cora Heiligensetzer for data handling by American Society of Clinical Oncology 9

10 Böttcher et al Appendix Table A1. Demographic Characteristics and Risk Features of Patients in Whom MRD Assessments Were Performed (n 493) Compared With the Remaining CLL8 Patients (n 324) Characteristic/Risk Feature MRD Assessments No MRD Assessments No. % No. % FC treatment Age, years Female sex Binet A or B ECOG PS No B symptoms WBC at diagnosis below /L Mutated IGHV Serum thymidine kinase 1. U/L Serum 2 -microglobulin 3.5 mg/l Complete clinical response Del(17p) Del(11q) Trisomy Del(13q) No cytogenetic abnormalities according to the hierarchical model Planned six cycles received Abbreviations: CLL8, German Chronic Lymphocytic Leukemia Study Group CLL8 trial; ECOG, Eastern Cooperative Oncology Group; FC, fludarabine and cyclophosphamide; MRD, minimal residual disease; PS, performance score. P Table A2. Preanalytical Sample Flow Per Analysis Time Point Variable Initial Staging Interim Staging Initial Response Assessment Final Restaging in PB Final Restaging in BM Total German and Austrian CLL8 patients Without any response assessment Suitable aberrations for MRD detection Progressive disease CR patients npr patients 17 Samples expected Samples received Collection-analysis interval 48 hours Sensitivity of analysis inferior to Samples analyzed Percent of samples expected Follow-Up NOTE. Samples from the German and Austrian patients in the German Chronic Lymphocytic Leukemia Study Group CLL8 trial were eligible for analysis. Patients with progressive disease, without any response assessment (n 45) or without a suitable marker for flow cytometric minimal residual disease (MRD) analysis (n 3) were not included in the analysis. Peripheral blood (PB) samples from all other patients were expected for MRD monitoring at interim staging, initial response assessment, and final restaging. Bone marrow (BM) samples at final restaging were scheduled for MRD monitoring in patients considered to be in complete response (CR) at initial response assessment. Follow-up MRD quantification in PB was scheduled for patients in CR or nodular partial remission (npr). MRD results on samples with delayed arrival ( 48 hours) or insufficient sensitivity due to acquisition of less than 2, leukocytes/per tube (sensitivity inferior to 1 4 ) were not included in this analysis. Note that investigators sent more BM samples than expected (because suspected CR at initial response assessment could not be confirmed at final restaging) as well as follow-up samples from patients with CR, PR, and npr by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY

11 MRD as Independent Predictor of PFS and OS in CLL Table A3. Proportions of Patients Who Achieved an MRD Level Below 1 4 in PB at Final Restaging by Treatment Arm and Pretherapeutic Risk Features FCR Treatment FC Treatment Variable MRD 1 4 MRD 1 4 MRD 1 4 MRD 1 4 No. % No. % P No. % No. % P Age, years Sex Female Male Binet A/B C ECOG PS B symptoms Absent Present WBC at diagnosis /L /L IGHV Mutated Unmutated Thymidine kinase 1. U/L U/L microglobulin 3.5 mg/l mg/l Del(17p) Absent Present Del(11q) Absent Present Del(13q) Absent Present Trisomy 12 Absent Present Aberration (Döhner model) Any aberration No aberration Abbreviations: ECOG, Eastern Cooperative Oncology Group; FC, fludarabine and cyclophosphamide; FCR, FC plus rituximab; MRD, minimal residual disease; PB, peripheral blood; PS, performance score by American Society of Clinical Oncology 11

12 Böttcher et al Table A4. Proportions of Patients Who Achieved an MRD Level Below 1 4 in BM at Final Restaging by Treatment Arm and Pretherapeutic Risk Features FCR Treatment FC Treatment Variable MRD 1 4 MRD 1 4 MRD 1 4 MRD 1 4 No. % No. % P No. % No. % P Age, years Sex Female Male Binet A/B C ECOG PS B symptoms Absent Present WBC at diagnosis /L /L IGHV Mutated Unmutated Thymidine kinase 1. U/L U/L microglobulin 3.5 mg/l mg/l Del(17p) Absent Present Del(11q) Absent Present Del(13q) Absent Present Trisomy 12 Absent Present Aberration (Döhner model) Any aberration No aberration Abbreviations: BM, bone marrow; ECOG, Eastern Cooperative Oncology Group; FC, fludarabine and cyclophosphamide; FCR, FC plus rituximab; MRD, minimal residual disease; PS, performance score by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY