Chromosome genomic array testing (CGAT) improves diagnostic yield in patients with AML and MDS and has the potential for better risk stratefication - an NCI myelomatch and SWOG report

Abstract Introduction: Chromosome genomic array testing (CGAT) has been widely used in clinical genomic laboratories for over a decade due to its robust performance in identifying submicroscopic copy number aberrations (CNAs) and copy-neutral loss-of-heterozygosity (cnLOH). Retrospective analysis demonstrated the prognostic significance of clonal aberrations uniquely identified by CGAT, especially 13q cnLOH (homozygous FLT3-ITD) and complex genome, defined as three or more CNAs/cnLOH (Gronseth 2015 Cancer). In addition, 4q cnLOH (homozygous TET2) and KMT2A-ITD are associated with a poor prognosis (Raychaudhuri 2024 ASH). Here, we aimed to assess the prevalence of CGAT abnormalities in newly diagnosed AML and MDS patients enrolled in the NCI myeloMATCH clinical trial who underwent central lab testing per SWOG master screening and reassessment protocol (MSRP), to determine the value added by CGAT in a prospective clinical trial. Methods: All patients in the myeloMatch MSRP from May 2024 to May 2025 were included in this analysis. CGAT was performed using CytoScan HD Array, in conjunction with the 8-probe FISH panel for inv(3)/t(3;3), t(6;9), t(8;21), t(9;22), t(v;11q23), t(15;17), inv(16)/t(16;16), or TP53 deletion. We compared CGAT results with those from karyotype analysis, FISH, and NCI Myeloid Assay (NMA) performed at the myeloMATCH Molecular Diagnostics Network (MDNet) labs to determine the unique value of CGAT from comprehensive correlations. Risk stratification was based on ELN2017 per myeloMATCH-IDE. Results: Of 468 total patients enrolled in the first year of myeloMATCH, 54 (11.5%) with disqualifying diagnoses were excluded while 7 acute promyelocytic leukemia (APL) were retained for the purpose of this study. Median age was 64.1 (range 18.4-95.8) at screening; 43% were female. Valid karyotype and FISH were obtained in 405 diagnostic AML (88.4%) and MDS (11.6%) samples, including 353 bone marrow and 52 peripheral blood samples. CGAT was successful in 100% of samples tested; in 35 samples CGAT was not performed due to insufficient material. Abnormal rate was 67.0% by CGAT, compared to 59.0% and 50.5% by karyotype and FISH, respectively. In 286 (77.3%) of 370 total diagnostic samples with all three testing results available, at least one test showed clonal aberration(s). The most common CGAT aberrations were 5q- (17.0%), 7q- (11.2%), 12p- (8.5%), 18q- (7.6%), and +8, 11q+, 17p- each in 7.1%; similar to karyotype findings. The most common cnLOH involved chromosome arms 13q (3.6%), 11p (2.7%), and 17p (2.7%). Using the 370 patients as the denominator, we further evaluated the result subgroups. Among patients with normal karyotype/FISH results, 58 (15.7%) were abnormal by CGAT while 84 (22.7%) were normal. In patients with normal karyotype but abnormal FISH with translocations, 5 (1.4%) demonstrated CNAs while 3 (0.8%) did not. Among patients with both abnormal karyotype and FISH results, 146 (39.5%) were also abnormal by CGAT with a significant number of additional unique findings by CGAT, and 30 (8.1%) showed normal CGAT result. In patients with abnormal karyotype but normal FISH, CGAT was abnormal in 38 (10.3%) and normal in 6 (1.6%). Of the 58 patients with abnormal CGAT but normal karyotype/FISH, KMT2A-PTD accounted for 31% (n=18), 4 of which also showed 11q cnLOH, implying the homozygous status of KMT2A-PTD. Collectively, cnLOH was seen in 41 (71%) with 31 (53%) as the sole aberrations without CNAs. Chromosome arms with the most prevalent cnLOH were 13q (n=11, 19%, with homozygous FLT3-ITD), 1p or 11p or 11q (each n=5, 8.6%), 19q (n=4, 6.9%, with homozygous CEBPA mutation), and 4q (n=3, with homozygous TET2 mutation). Among patients with intermediate risk cytogenetics, 10 of 53 (19%) patients with abnormal karyotype showed complex genome by CGAT, suggesting an adverse risk. Combining these patients and patients with normal karyotype described above, CGAT findings would have altered risk stratification in 11.4% overall patients. Conclusion: The first year myeloMATCH study demonstrated that CGAT significantly improved diagnostic yield for AML/MDS patients in ~20% patients and may alter ELN risk in 11%. We recommend routine use of CGAT for intermediate risk patients at diagnosis. Confirmation of the prognostic significance of CGAT-identified lesions will improve current risk stratification tools important for precision medicine. Funding: NIH/NCI grants U10CA180888, U10CA180819