Identification of therapy options for rare and resistant gastrointestinal stromal tumors (GIST)

Background: GISTs are predominantly defined by KIT/PDGFRA mutations which are targetable with a range of kinase inhibitors, however the majority become TKI-resistant (TKI-R). Double (KIT/PDGFRA) wildtype (D-WT) GISTs represent a rare subset of GIST patients in need of treatment options. We investigated a commercial database of theranostic biomarkers for the identification of novel therapy options for GIST. Methods: 217 GIST cases were evaluated for D-WT and TKI-R. A multiplatform approach of biomarker testing was used and included a combination of sequencing (NGS, Sanger), protein expression (IHC) and gene amplification (ISH). Results: D-WT (n = 15) and TKI-R (n = 23) (including 7 with resistance mutations in the absence of a primary, activating KIT mutation and 4 PDGFRA D842V) were studied for additional targetable alterations. IHC and ISH tests revealed no overexpression or amplification in cMET, EGFR, or HER2. PTEN was intact (positive expression) in the majority of GISTs (92.9% (13/14) D-WT; 100% (19/19) TKI-R). Mutational screening revealed variants in 6/47 genes (excluding cKIT and PDGFRA), most of which are potentially targetable with therapies currently available, or in clinical trials: PIK3CA, ABL, cMET, JAK3, RB1, and VHL. ABL and JAK3 mutations were exclusively found in the TKI-R subgroup. PD-1 positive tumor infiltrating lymphocytes were found in 33% (1/3 D-WT) and 60% (3/5 TKI-R), while PD-L1 tumor expression was found in 67% (2/3 D-WT) and 40% (2/5 TKI-R). Although chemotherapy has historically elicited poor responses in GIST (non-selected patient trials), we observed a high frequency of low expression of predictive markers for gemcitabine (RRM1) and paclitaxel (TUBB3) (77%, 90%; 57%, 73% for D-WT and TKI-R, respectively) and high frequency of TOPO1 overexpression for irinotecan (57%, 32% in D-WT and TKI-R, respectively) which were recently shown to be cytotoxic in TKI-R GIST cell lines (Boichuk, 2014). Conclusions: A multiplatform approach of theranostic biomarkers identified non-cKIT/PDGFRA therapy options for rare and resistant GIST. Opportunities for investigating new targetable agents and potentially re-visiting cytotoxics with biomarker guidance in these subpopulations are warranted.