A Scoring System for Prediction of FLT3-ITD Positivity in Patients with Newly Diagnosed Acute Myeloid Leukemia.

FLT3-internal tandem duplication (ITD) is found in about 30% of patients with acute myeloid leukemia (AML) at diagnosis and confers a high risk of relapse. Thus allogeneic hematopoietic transplant (HCT) is recommended for these patients in first complete remission (CR) and after HCT they become candidates for trials of FLT3-ITD inhibitors (such as quizartinib) to prevent relapse. However at referral to tertiary centers after reaching CR, FLT3-ITD status at diagnosis is often unknown, complicating decisions about HCT. FLT3-ITDs are known to be associated with a normal karyotype (NK), translocation 6;9 and a high white blood cell (WBC) count, and we hypothesized that assessment of likely FLT3-ITD status at diagnosis in patients presenting in CR not tested at diagnosis would be improved by examining these covariates simultaneously.

Our initial analysis included 434 adult patients with newly diagnosed AML (excluding APL) treated on three SWOG trials (S9031, S9333, and S0106) in whom FLT3-ITD status (positive/negative) was established at diagnosis. Univariate and then multivariate analyses were used to identify covariates independently associated with FLT3-ITD positivity. The relative abilities of these to predict FLT3-ITD positivity were quantified using the area under the receiver operator characteristic curve (AUC); an AUC of 1.0 denotes perfect prediction, whereas an AUC of 0.5 is analogous to a coin flip. The log odds ratios (ORs) from the multivariate models were used to assign a score to each covariate and scores were summed; such that the higher the score, the greater is the likelihood of the FLT3-ITD positivity at diagnosis. We tested the performance of the scoring system in 2 newly-diagnosed populations that had not contributed to the system's development and in whom FLT3-ITD status at diagnosis was known: (a) 210 patients treated at FHCRC (Fred Hutchinson Cancer Research Center) and (b) 1,401 patients treated at MDACC (M.D. Anderson Cancer Center). Covariates examined were: age, sex, performance status (PS), WBC count, platelet count, bone marrow blast percentage, secondary AML, and cytogenetic risk (using SWOG/Eastern Cooperative Oncology Group criteria).

FLT3- ITD was present in 101 of the 434 SWOG patients (23%) in the scoring system development population. The log OR were rounded to the nearest half point to create the scoring system. Only WBC > 20,000 (reference, WBC < 20,000) and cytogenetics (reference, normal) had non-zero scores, which are summarized below:

Scores less than −0.5 were called low, ≥−0.5 and <0.5 intermediate, ≥ 0.5 high.

The AUC was 0.75 and contrasted with 0.66 and 0.69 when only WBC or cytogenetics were considered. However when this system was tested in the FHCRC population (16% FLT3-ITD positive) its AUC was only 0.58, not better than when each covariate was examined separately (AUC 0.54 and 0.6 for WBC and cytogenetics, respectively). Similarly at MDACC (17% FLT3-ITD positive) the system's AUC was 0.68 vs. 0.59 and 0.68 for WBC and cytogenetics, respectively.

Although this scoring system seemed useful tool within the population it was developed (SWOG), such was not the case in two independent cohorts of AML patients with known FLT3-ITD status (FHCRC and MDACC). This indicates that there is no obvious substitute for actual data on FLT3-ITD status.

No relevant conflicts of interest to declare.