Cellular and Cell-Free Genomic Characterization of Patients Receiving Avatrombopag and Immunosuppressive Therapy for Treatment-Naïve Severe Aplastic Anemia

Background

The efficacy and safety of avatrombopag, a second generation thrombopoietin-receptor agonist, in the treatment of newly diagnosed severe aplastic anemia (sAA) is being evaluated in the DIAAMOND-Ava-FIRST trial. Herein, we report somatic and germline molecular findings of this study in which treatment-naïve patients received avatrombopag in combination with immunosuppressive therapy (IST; equine ATG and ciclosporin).

Methods

Enrolled patients (pts) received IST plus oral avatrombopag at a maximum dose of 60mg for an initial six months (mths; standard), with those achieving a partial response eligible to receive a further six mths of avatrombopag (extended). At baseline, occult germline disease was screened for with use of a 37-gene inherited bone marrow failure (IBMF) NGS hybridisation-based capture panel. Cellular (bone marrow aspirate) and cell free (peripheral blood) DNA samples were obtained at baseline, 6, 12, 18 and 24 mths and were analysed using a targeted unique molecular index-corrected hematological malignancy NGS panel. Whole genome NGS copy number variation (CNV) analysis was performed by comparing read counts from on and off target reads to a pooled reference comprising normal samples to correct for enrichment and sequencing biases. Avatrombopag therapy has now been completed for enrolled pts, with a median follow up of 20.4 mths (IQR (8.9, 23.4)).

Results

56 eligible pts (male:female 1.33:1) with a median age 58 years (yrs) (range 18-78 yrs) were enrolled. IBMF testing did not reveal occult germline disease in any pt (two pts had non-suspicious variants of uncertain significance (in GATA2 and RTEL1)). Somatic profiling at baseline (excluding PIGA), demonstrated that 15 pts (27%) had somatic mutations; 9 (16%) had 1 mutation, 5 (9%) had 2 mutations, and 1 pt (2%) had more than 2 mutations. Patients with detectable somatic mutations were typically older compared to those with no detectable mutations (median age 65 v 50 yrs). DNMT3A was the most frequently mutated gene at baseline, followed by PIGA, ASXL1, and TET2. 47 pts had both baseline and 6 mth somatic data available, with mutations present in 21 (45%) of these pts at 6 mths. Newly emergent mutations at 6 mths were most frequent in ASXL1, DNMT3A, PIGA and U2AF1. The presence of mutations at baseline did not correlate with complete response (p=0.49) or overall response (p=0.31) at 6 mths.

13 pts received extended avatrombopag therapy for a total of 12 mths, with this extended therapy group demonstrating enrichment of somatic mutations at baseline (≥1 mutation 58% vs 20% standard, ≥2 mutations 42% vs 0% standard) and at 12 mths (≥1 mutation 83% v 48% standard, ≥2 mutations 67% v 20% standard) and a trend to higher median variant allele frequencies compared with the standard therapy arm (14.2% v 8% at 12 mths and 24.2% v 12.1% at 24 mths, p=0.08). 2 pts developed myelodysplastic syndrome (MDS); at 12 mths and 18 mths, with both pts harbouring ASXL1 mutations at the time of MDS development. One of the pts who developed MDS had received extended avatrombopag therapy to 12 mths.

22/53 pts (42%) had a PNH clone by flow cytometry reported at baseline; of these 8 (36%) had a PIGA mutation detected. There were no PIGA mutations detected in the absence of a PNH clone. Baseline NGS CNV analysis was performed in 52 pts and detected trisomy 8 in 1 pt (concordant with conventional cytogenetics (CG) result), acquired loss of heterozygosity in 1p, 6p and 13q in 1 ptBCOR duplication in 1 pt and a duplication of Xq21.33 involving DIAPH2 in 1 pt. CG at baseline failed in 23/56 pts (41%), with no detectable chromosomal aberrations from NGS CNV analysis in this group.

Finally, we analysed cell-free DNA from the cohort which showed broad concordance between mutations detected in the cellular and cell free compartments, however the cell free DNA compartment samples did reveal further clonal complexity at baseline in 5/46 (11%) pt samples with mutations detected that were not detected in the cellular compartment (in TP53, BRAF, MAP2K1 and DNMT3A).

Conclusion

Molecular profiling of pts with sAA treated with avatrombopag demonstrated that clonal hematopoiesis is common at baseline, with an increase in pts with clonal hematopoiesis observed at 6 and 12 months after commencing therapy, with enrichment of somatic mutations in those who received extended duration of avatrombopag. Moreover, further genomic complexity may be revealed in this subgroup from cell-free DNA analysis.