An International Multi-Center Study To Define the Application of Microarray-Based Gene Expression Profiling in the Diagnosis and Sub-Classification of Leukemia (MILE Study): Analysis of Completed Stage I with 2030 Patients Achieved a 95.4% Accuracy.

In 2005, the MILE (Microarray Innovations in LEukemia) study research program was initiated with 11 centers: 7 from the European Leukemia Network (ELN, WP13), 3 from the US, and 1 in Singapore. The first stage was designed to assess the ability of GEP to classify leukemias into 17 recognized categories as compared to conventional diagnostic assays (“Gold Standards”). These include: mature B-ALL with t(8;14), pro-B-ALL with t(11q23)/MLL, c-ALL/pre-B-ALL with t(9;22), T-ALL, ALL with t(12;21), ALL with t(1;19), ALL with hyperdiploid karyotype, c-ALL/pre-B-ALL without specific genetic abnormalities, i.e. t(9;22), AML with t(8;21), AML with t(15;17), AML with inv(16)/t(16;16), AML with t(11q23)/MLL, AML with normal karyotype, AML with complex karyotype, CML, CLL, and MDS, as well as non-leukemic and healthy bone marrow samples as controls. We now report the data from the completed MILE Stage I, which included 2030 adult and pediatric samples utilizing HG U133 Plus 2.0 (Affymetrix). 102 (5%) were removed from the study for protocol violations, most often for an incomplete “Gold Standard”. All microarrays were also assessed for their technical performance (e.g. micorarray scaling factor, signal to noise ratio, etc.). In total only 36 samples (1.9%) failed these technical quality criteria. Cross validation accuracy (average of three 30-fold cross validations) of the remaining 1892 MILE Stage I samples is 95.4% concordant with the “Gold Standard” diagnosis. In ten classes the concordance was ≥97% i.e. for pro-B-ALL with t(11q23)/MLL, T-ALL, AML with t(8;21), AML with inv(16)/t(16;16), AML with normal karyotype, CLL, CML and MDS, as well as for the healthy or non-leukemia samples. The following classes showed lower sensitivities (range 74%-91%): mature B-ALL with t(8;14), c-ALL/pre-B-ALL with t(9;22), ALL with t(1;19), ALL with hyperdiploid karyotype, c-ALL/pre-B-ALL without specific abnormalities, AML with t(11q23)/MLL, and AML with complex karyotype. This can largely be explained by biological heterogeneity and weak “gold standard” definitions. It is notable that all these classes showed specificities above 98.4%. The second stage now prospectively process additional 1000 samples on a custom designed microarray using 1,449 probe sets only. So far 755 samples have been analyzed and will represent an independent and blinded test set for the algorithms developed in stage I. In conclusion, the MILE research study confirms that standardized microarray-based GEP may accurately classify leukemia samples into known diagnostic and prognostic sub-categories with a low technical failure rate and a very high accuracy.