New Definition of Treatment Response in Adult Acute Lymphoblastic Leukemia (ALL): Use of Molecular Markers for Minimal Residual Disease (MRD).

Abstract 90^FN2

Response to treatment measured by cytologic bone marrow evaluation has been so far the most important prognostic factor in adult ALL, and cytological complete remission (cyCR) and relapse (cyREL) are accepted endpoints for clinical trials. However, measurement of MRD is a more sensitive approach for assessment of response. In the past decade it has been confirmed that MRD is an independent, strong prognostic factor in ALL. The most widely applied method is based on quantative PCR evaluation of individual rearrangements of TCR and Ig genes. Methodology, interpretation and definitions are highly standardised and confirmed in different labs (e.g. van der Velden et al, Leukemia, 2007). Therefore it is essential to define and validate MRD based response criteria – namely complete molecular/complete MRD response (molCR), molecular failure/MRD persistence (molFAIL) and molecular relapse/MRD relapse (molREL) - as endpoints for clinical trials.

Since 1999 the German Multicenter Study Group for Adult ALL (GMALL) has conducted two consecutive studies (GMALL 06/99 and 07/2003) with a risk and MRD adapted treatment strategy and prospective MRD evaluation. 1489 evaluable pts (15–55 y) with Ph-negative standard (SR) and high risk (HR) ALL have been included so far. MRD evaluation as described above has taken place in central labs. For analysis of molCR MRD evaluation was restricted to pts with cyCR (prerequisite: at least one marker with sensitivity and quantitative range of minimum 10⁻⁴ for negative results). MolCR was defined as negative MRD status after induction therapy. MolREL was defined as reappearance of MRD above 10⁻⁴ after achievement of molCR.

MolCR after induction: The overall cyCR rate was 89% with differences between SR and HR (92% vs 84%;p<.0001) and age < vs >35 yrs (90% vs 86%;p=.01) but not B vs T-lineage (89% vs 88%). As expected the molCR was lower (69%) in 479 evaluable pts. Significant differences were seen between SR vs HR (77% vs 46%;p<.0001) and B vs T-lineage (63% vs 79%;p=.0005). The most rapid decrease of MRD was observed after induction phase I (34% molCR) and phase II (69%), whereas the effect of first consolidation was limited (72%) indicating a progressive selection of chemotherapy resistant leukemic clones. MolCR evaluation was also used to assess treatment modifications such as asparaginase intensification or addition of rituximab to induction therapy.

Prognostic impact of molCR:SR pts were scheduled for chemotherapy. Those with molCR had a superior overall survival (OS) when compared to molFAIL (67% vs 38%;p<.0001) which was mainly due to a lower relapse rate (RR) (26% vs 59%;p<.0001) leading to better remission duration (RD) (57% vs 29%;p<.0001). MolFAIL pts had a better OS with SCT in CR1 than without (60% vs 27%;p=.09).

All HR pts were scheduled for SCT in CR1. OS was superior for MolCR vs MolFAIL (66% vs 42%; p=.003). The RR was higher for molFAIL vs molCR on chemotherapy (66% vs 18%;p=.04) and after SCT (34% vs 11%;p=.04). Even after SCT pts with molCR had a better OS (75% vs 58%; p=.03).

MolREL: The incidence and outcome of molREL was evaluated in 432 pts with MRD tests during and after treatment. 36 molREL without additional extramedullary relapse were observed. The OS was 36% and the probability of cyREL was 82% with median of 75 d from molREL to cyREL. All pts not treated at molREL relapsed (N=12). Only 1/11 (9%) pts who were transferred to SCT in cyCR1 after additional treatment, including experimental drugs, relapsed compared to 21/25 (84%) pts without SCT in cyCR1 (p<.0001).

Overall these data show in the largest cohort of adult ALL pts with MRD analysis reported so far, the proof of principle for molecular response evaluation. In a study with high cyCR rate it allows to detect differences between subgroups and treatment approaches and thereby refines the evaluation of treatment efficacy. It is a clinically relevant endpoint since molFAIL is associated with significantly poorer RD and OS even after subsequent SCT. Detection of molFAIL and molREL identifies pts with resistance to conventional drugs and the need for targeted, experimental drugs and SCT in cyCR1 before occurrence of cyREL. Molecular response evaluation with standardised methods provides new refined endpoints for future trials, including pivotal trials with new drugs with the major advantage of early decision points and strong correlation to clinical outcome.

Supported by Deutsche Krebshilfe 702657Ho2 and BMBF 01GI9971/8