Molecular and Cytogenetic Characteristics of Acute Myeloid Leukemia (AML) and Myelodysplastic Syndrome (MDS) in Patients Exposed to Chemotherapy (CT) and/or Radiation (XRT)

Background: Therapy-related myeloid neoplasm (t-MNs) are caused by exposure to CT and/or XRT and present with high rates of adverse cytogenetics (CG). However there are pts who present with non-adverse CG making the definition of T-MN controversial. Correlation between exposure to specific CT, molecular profile, and impact on overall survival (OS) has not been established.

Methods: 171 CT/XRT exposed pts with AML/MDS from Weill Cornell and Northwestern University were evaluated for disease characteristics, CT/XRT treatment, and survival. Study is ongoing and we plan report on appx 500 CT/XRT exposed pts with AML/MDS. CT exposure was categorized as: Alkylators, Topoisomerase inhibitors (TI), Anthracyclines, Taxanes, Anti-metabolites and Platinum compounds. Mutations were classified as: Signaling/kinase pathway (FLT3, KRAS, NRAS, PTPN11, NF1), Epigenetic modifiers (DNMT3A, IDH1, IDH2, TET2, ASXL1, EZH2), NPM1, Transcription factors (CEBPA, RUNX1, GATA2), Tumor suppressors (TP53), Spliceosome (SRSF2, U2AF1, SF3B1, ZRSR2), Cohesin (RAD21, STAG1, STAG2, SMC1A, SMC3) and secondary type mutations (ASXL1, BCOR, EZH2, SF3B1, SRSF2, STAG2, U2AF1, ZRSR2).

Analyses: Chi-square and Fisher’s Exact test was used for comparisons of categorical and continuous factors. Cox Regression model was used to evaluate the effect of gene mutations and CG risk factors on OS. All p-values are two-sided with significance at 0.05 alpha level (SAS Version 9.4, SAS Institute, Inc., Cary, NC).

Results: Among 171 with CT/XRT exposure, median age at diagnosis was 70 yrs (AML: 76%; MDS 24%). 56%, 20% and 24% had exposure to CT only, XRT only, and CT+XRT, respectively. In t-AML, 52% and 48% had adverse and non-adverse CG, respectively, with no difference in t-MDS (50% in both).

Pts (AML/MDS) with previous CT/CT+XRT were more likely to have adverse CG compared to XRT-treated pts (56% vs. 33%, HR 4.16, p <0.0001). With regards to specific CT exposure, presence of adverse CG was higher in pts exposed to alkylators (OR 3.11, p=0.0002), anti-metabolites (OR 2.55, p=0.02), TI (OR 1.9, p=0.05) and platinums (OR 2.40, p=0.01); this was not observed with anthracyclines and taxanes. Characteristics of pts presenting with adverse CG vs. not are shown in Table 1 along with OS with mutation classes. Age-adjusted 5-year OS of adverse CG vs. non-adverse-risk pts exposed to CT/XRT was less than 5% vs. 30%. Median OS of pts exposed to CT/XRT vs XRT alone was 14 months (95% C.I. 11-18) vs 16 months (95% C.I 10-44). Only 25.88 % of CT/XRT exposed AML pts with non-adverse CG were classified as ELN adverse (2022).

Mutation frequencies by CT exposure are detailed in Table 2. TP53 mutations were more common in pts exposed to anti-metabolites (OR 2.71, C.I. 1.28-5.72, p=0.01) and platinums (OR 2.12, C.I. 1.09-4.13, p=0.03) (comparison with other CTs) while epigenetic modifiers were less likely with anthracyclines, alkylators, platinums, TI, and taxanes. Secondary type mutations were not enriched in pts exposed to any particular CT class (alkylators 25%, TI 30%, antimetabolites 28%, cisplatin 24%, and taxanes 26%). When stratified by latency from CT/XRT exposure (<5 vs > 5 yrs; Table 1), we found that adverse CG were seen in 47.22 % and 54.55 % of pts when disease was diagnosed <5 and >5 yrs, respectively. Frequency of TP53 mutations were higher with latency <5 yrs, while epigenetic mutations were more likely for >5 yrs latency.

Finally we looked at treatment in a subset of pts (n=98). 57.5% of pts were treated with HMA+venetoclax, 20.6% with 7+3, 8.8% with HIDAC regimen, 14.7% with CPX-351, 17.5% with HMA alone. CR/CRi was attained by 46.4% (53% in HMA+V vs. 39% others) and 28.8% were primary refractory. Among pts <75 yrs (n=118), 38.98% were consolidated with HSCT. Analyses comparing to pts w/o CT/XRT exposure will be presented.

Conclusions: In our pts with AML/MDS exposed to previous chemotherapy/XRT, we found (i) True adverse CG, TP53 and secondary type mutations were present in a subset of AML/MDS, thus suggesting that not all exposure has a t-MN like signature and therapy relatedness needs to be redefined by biological subgroups rather than exposure alone. (ii) OS of non-adverse CG (AML) was comparable to historical data in denovo AML again suggesting that not all AML/MDS after chemo/XRT exposure behaves like t-MN (iii) We found an association of shorter latency and antimetabolite and platinum exposure to TP53 mutations and needs further validation in a larger dataset.