Prognostic Impact of Multiparameter Flow Cytometry in Patients Analyzed for Suspected MDS.

Abstract Abstract 2806 Background: Multiparameter flow cytometry (MFC) is increasingly used to evaluate patients with suspected myelodysplastic syndromes (MDS). The prognostic impact of distinct MFC findings has been controversial yet. Aims: Assess the respective impact of different antigen expression aberrancies on overall survival (OS) in suspected MDS in relation to established prognostic parameters. Methods: We studied 804 patients (pts) who were analyzed for suspected MDS by cytomorphology (CM), cytogenetics and MFC in parallel (f/m 463/341; median age 70 yrs, range 2–89). Pts had been included in a previous study evaluating the diagnostic role of MFC (Kern et al., Cancer 2010). Data on OS was available in all pts (median OS 6.2 yrs, median follow-up 3.2 yrs). CM revealed MDS in 493 (61.3%) pts; 170 (21.1%) pts had evidence of dysplasia which was not sufficient to diagnose MDS by CM; in 141 (17.5%) pts MDS was excluded by CM. Karyotypes were favorable / intermediate / unfavorable according to IPSS in 684 (85.1%) / 89 (11.1%) / 31 (3.9%) pts. MFC was performed following recent ELN Working Group recommendations (Westers et al., Leukemia 2012) in myeloid progenitor cells (MPC), granulocytes, monocytes and erythroid cells. MFC parameters included, each compared to normal bone marrow, increased or decreased expression of antigens, expression of normally not expressed antigens, aberrant expression pattern of pairs of antigens as well as cross-lineage expression of lymphatic antigens. Results: Considering all 804 pts, i.e. regardless of confirmation of MDS by CM, the following MFC parameters were associated with OS: MPC >5% (p<0.001, hazard ratio (HR) 2.5), expression of CD5 (p<0.001, HR 4.1), CD56 (p=0.043, HR 2.0), CD7 (p=0.015, HR 2.2) in MPC; reduced side-scatter signal (p<0.001, HR 1.9), aberrant CD13/CD16 expression pattern (p=0.007, HR 1.4), aberrant CD11b/CD16 expression pattern (p=0.003, HR 1.6), expression of CD56 (p<0.001, HR 2.1), reduced expression of CD33 (p=0.018, HR 1.6) in granulocytes; expression of CD56 (p<0.001, HR 1.6) in monocytes; reduced expression of CD71 (p<0.001, HR 2.1) in erythroid cells. A score was devised calculating for each pt the sum of all HRs for the respective parameters found positive. Pts were separated into 4 groups: group 1 (n=263 pts) had a score of 0; group 2 (n=259) had a score >0 and below the median (2.135); group 3 (n=149) had a score above the median and below the 75th percentile (4.961); group 4 had a score above the 75thpercentile. Significant differences in OS were observed: OS at 4 years in groups 1, 2, 3, and 4 amounted to 82.4%, 67.1%, 54.7%, and 36.2%, respectively (p=0.001 for 1 vs 2, p=0.022 for 2 vs 3, p=0.003 for 3 vs 4, p<0.001 for all other comparisons). Cox analysis of the MFC score revealed a significant association with OS (p<0.001, HR 1.4 per group). Other parameters univariably related to OS were: diagnosis of MDS by CM (p<0.001, HR 2.2), percentage of bone marrow blasts by CM (p<0.001, HR 1.9 per 10% increment), cytogenetic group according to IPSS (p<0.001, HR 5.1 per group), WBC count (p<0.001, HR 1.2 per 10 G/L increment), hemoglobin level (p<0.001, HR 0.8 per g/L), platelet count (p<0.001, HR 1.4 per 100 G/L increment), and age (p<0.001, HR 1.5 per decade). Multivariable Cox analysis including diagnostic markers as covariates revealed an independent impact on OS for all of them: MFC score (p<0.001, HR 1.3), diagnosis of MDS by CM (p=0.003, HR 1.4), bone marrow blasts by CM (p=0.015, HR 1.4), and cytogenetics according to IPSS (p<0.001, HR 2.9). The addition of age or of peripheral blood counts as covariates into the multivariable analysis still indicated an independent impact of the MFC score on OS. Limiting the multivariable analysis to cases with MDS proven by CM still resulted in an independent impact of the MFC score on OS (p=0.001, HR 1.2). Interestingly, if only cases were considered with signs of dysplasia by CM, which are not sufficient to diagnose MDS, multivariable analysis also revealed an independent impact of the MFC score on OS (p=0.016, HR 1.3). Conclusions: The present data indicates that MDS-related findings by MFC provide prognostic information not just in pts with MDS proven by CM but also in a comprehensive cohort of pts being diagnosed for suspected MDS. Furthermore, even in pts with evidence of dysplasia not sufficient to diagnose MDS by CM a prognostic impact of MFC was demonstrated. This data thus suggests to integrate MFC into the diagnostic work-up of pts with suspected MDS. Disclosures: Kern: MLL Munich Leukemia Laboratory: Equity Ownership. Haferlach:MLL Munich Leukemia Laboratory: Equity Ownership. Alpermann:MLL Munich Leukemia Laboratory: Employment. Schnittger:MLL Munich Leukemia Laboratory: Equity Ownership. Haferlach:MLL Munich Leukemia Laboratory: Equity Ownership.