Concordance between Bone Marrow and Peripheral Blood Samples for Assessment of FLT3 Internal Tandem Duplication (ITD) Mutations: Data from Patients Screened for Participation in Quantum-R, a Global, Randomized, Open-Label, Phase 3 Study Examining the Effect of Quizartinib Monotherapy Vs Salvage Chemotherapy on Overall Survival in Patients with FLT3 - ITD - Mutated AML Who Are Refractory to or Have Relapsed after First-Line Therapy

Introduction

Acute myeloid leukemia (AML) is a heterogeneous disease with multiple factors influencing long-term outcome. FLT3 (FMS-like tyrosine kinase 3) mutations (predominantly internal tandem duplication [ITD]) are reported in ~25% of patients with AML and have been associated with poorer outcomes compared with FLT3 non-mutated AML (Kottaridis et al. Blood. 2001). Targeting these mutations with tyrosine kinase inhibitors has become an active area of research. However, patients must undergo molecular testing to identify the presence of this mutation. The specimen of choice for such testing is bone marrow (BM; aspirate or biopsy), as the site of origin of the disease. However, BM testing is invasive and sometimes cannot be performed repeatedly in a timely manner. A simple technique using easily obtainable biological specimens from a source such as peripheral blood (PB) may be desirable if the results accurately reflect status in the active disease site (i.e., BM). We investigated whether a PB specimen could be utilized for FLT3 -ITD determination and if the characteristics of the ITD mutations identified in PB were consistent with those in BM samples from the same patient.

Methods

QuANTUM-R is a global, randomized, phase 3 study to assess the efficacy of quizartinib, a selective FLT3 inhibitor, in patients with FLT3 -ITD-mutated relapsed or refractory (R/R) AML. Paired samples of BM and PB were collected from individual patients as part of the screening process for QuANTUM-R. One to 3 mL each of BM and PB were collected on the same day (window Day -14 to Day 0) from each patient prior to receipt of any study treatment. A statistically powered number of patients (minimum 95) with paired BM and PB samples were selected during a defined period of screening in the phase 3 study. Samples were subjected to DNA isolation followed by PCR and fragment size analysis. Each sample was assessed for the presence of ITD mutation; moreover, the length and allelic frequency of the ITD mutation were measured for each sample. The assay was validated down to a limit of detection (LoD) of 1%. Correlation between the data obtained from BM and PB was assessed using Deming regression analysis and Bland-Altman plot constructed to assess concordance between the 2 sample types.

Results

Paired samples from 107 patients (median age 58y; 32% were ≥65y; 57% were female) were included. Sixteen samples had no ITD mutations in either BM or PB; 3 had ITD only in the BM sample with no corresponding ITD in the paired PB sample (despite having circulating blasts), and 88 had ITD in both BM and PB. In all 88 cases, the length of the ITD was identical in both specimen types. The sizes of ITD observed in this patient population were between 9 and 198 bases; 21/88 samples with ITD demonstrated the presence of multiple ITDs, with samples from 2 patients exhibiting 3 distinct ITDs, each of whose sizes were the same in BM and PB. Across all samples with ITD, the measured ITD allelic ratio ranged from 1% to 96%, with a similar or slightly lower ratio measured in the PB compared to the BM. In 23 (26%) cases, the allelic ratio measured in the PB was lower than that of the BM by ≥10%. Blast counts for these samples were also measured and there was no correlation between blast count and allelic ratio in either specimen type. High correlation between PB and BM samples for the presence of FLT3 -ITD was observed when regressing PB Ratio on BM Ratio via Deming regression analysis, and the following values were obtained: R² = 0.866; 95% CI for the slope, 0.98 to 1.12; and Bland-Altman difference plot for allelic ratio against difference between PB and BM showed good random scatter around average bias of -5.2%, with 95% CI including 0.0% (95% CI, -27.6 to 17.2).

Conclusions

FLT3 -ITD mutation testing is performed in all AML patients at diagnosis for prognostic purposes and to guide therapeutic decisions. As FLT3 inhibitors are developed for clinical use, regular monitoring of patients' residual disease burden/response to therapy through assessment of FLT3 status may become an important element of monitoring benefit from effective therapy. Mutation assessment is currently performed on BM aspirate or biopsy based on the common belief that PB is not adequate for this assessment. Our results of FLT3 -ITD mutation analysis in paired BM and PB samples collected from >100 patients with R/R AML show that PB specimens have a high degree of concordance with BM specimens for assessment of FLT3 mutation while being less invasive.