(P 121) USING IN VITRO MODELS TO PREDICT IMATINIB RESPONSES IN PDGFRA-MUTANT GASTROINTESTINAL STROMAL TUMOR

Objective: Gastrointestinal stromal tumor (GIST) is the most common sarcoma with 4,000-6,000 cases diagnosed in the United States annually. The two most common GIST subtypes harbor oncogenic mutations in KIT or platelet derived growth factor receptor alpha (PDGFRA) receptor tyrosine kinases, respectively. The use of tyrosine kinase inhibitors (TKIs) has revolutionized treatment for patients with advanced GIST. While the use of type II TKIs such as imatinib has been successful for the treatment of KIT-mutant GIST, they are thought to be ineffective for many PDGFRA-mutant GIST. Notably, the most common PDGFRA driver mutation, exon 18 D842V, is notoriously resistant to imatinib and other type II TKIs, leading to historically poor patient outcomes. Recently, the FDA approved avapritinib, a type I TKI, for treatment of patients with advanced PDGFRA exon 18-mutant GIST. However, in addition to D842V, there are numerous other PDGFRA exon 18 mutations observed in GIST. From a curated data set of a large cohort of PDGFRA-mutant GISTs, over 80% of these mutations occurred in exon 18 (n=302/363), and include many unique, non-D842V exon 18 mutations (n=102/302) with point mutations and/or complex indels. Aside from D842V, it is largely unknown how these mutations respond to imatinib; currently most exon 18 mutations are assumed to be imatinib resistant. In previous work, we biochemically profiled a few of the most frequently observed mutations in our patient cohort. Contrary to previous assumptions, we observed that some of these exon 18 mutations were imatinib sensitive. We also noticed that imatinib sensitivity seemed to be dependent on the properties of the residue in the 842-codon position, which encodes a key auto-inhibitory residue.

Methods: As it is not feasible to model each reported mutation, we hypothesized that the imatinib sensitivity of PDGFRA exon 18 mutations could be predicted based on the 842 residue. We tested the predictive potential of this residue by creating mutagenic libraries that included every amino acid substitution at the 842 position, including both point-mutation as well as deletion mutation constructs that recapitulate patient derived mutations. We expressed each mutation in Ba/F3 cells, and determined the oncogenic potential of each mutation and assessed imatinib sensitivity via immunoblotting for phosphorylated-PDGFRA.

Results: To date, 45% of the 21 tested mutations were sensitive to imatinib (IC50 < 150nM) in our assay, further debunking the belief that all PDGFRA exon 18 mutations are imatinib resistant. Of the mutations tested so far, we found that hydrophobic residues at the 842 codon position conferred imatinib resistance (IC50 > 150nM), and amino acids from other classes (polar, uncharged, positively charged, special case) were imatinib sensitive (IC50 < 150nM).

Conclusion: Based on our data, over half of the unique and complex mutations observed in our PDGFRA-mutant GIST patient cohort would be predicted to be imatinib sensitive. Our predictive model has the potential to identify which patients with PDGFRA exon 18 mutant-GIST would benefit from front-line imatinib therapy as an alternative to avapritinib. Conversely, we could also identify those patients who would require treatment with front-line avapritinib. While avapritinib was recently approved for first-line treatment for all PDGFRA exon 18 mutant GIST, imatinib, which has been used clinically for over 20 years, has a superior safety and tolerability profile. In addition, if a patient progresses on imatinib, they will have access to additional FDA-approved treatment options (e.g. sunitinib, regorafenib, ripretinib), which is not currently the case for those patients who experience tumor progression during treatment with front-line avapritinib. Treating eligible patients with imatinib first is a paradigm shifting approach has the potential to transform patient care and clinical outcomes by optimizing the sequencing of TKI treatments for PDGFRA-mutant GIST.