Abstract IA15: Promise and challenges of pediatric cancer PDX models

Abstract

Abstract The Pediatric Preclinical Testing Program (PPTP) has evaluated almost 97 single agents and several combinations over the last 10 years. A total of 89 xenograft models that encompass solid tumors, brain tumors, and acute leukemias have been used as a screen to identify novel agents that could be accelerated for clinical trial. With few exceptions, all models were directly established in mice from patient biopsies, and approximately half of the models were established from patients who were refractory or had relapsed. Expression analysis showed that tumor models clustered with their appropriate clinical histotype, and exome sequencing showed mutation frequency and identity to be similar to published genomic data for patient cancers. The models accurately identified known active agents that are used as standard of care therapy for solid tumors and leukemias, thus validating the models. Criteria for assessing response similar to clinical criteria were developed. Of the 97 agents tested, twenty-four signaling inhibitors (ligand or receptor binding antibodies, kinase inhibitors etc.) were evaluated. With the exception of inhibitors of Aurora kinase A, and Polo-like kinase 1, the overall objective response rate (ORR) was <2%. The low response rate is probably an accurate reflection of likely ORR in clinical trials, as the models identified selumetinib in an astrocytoma with BRAF (V600E) mutation, dasatinib in a Ph+ ALL, sorafenib in a Flt3 mutant ALL, and crizotinib in an ALK mutant neuroblastoma. The models identified several cytotoxic agents as having broad-spectrum antitumor activity (eribulin, alisertib, PR-104), with small molecule and antibody conjugated antimitotic agents showing high-level activity in many models (eribulin in sarcoma and ALL, glembatumab in osteosarcoma). The major issue is that the models overpredict drug activity due to tolerance of the host (mouse) compared to patients. Thus, it is imperative to adjust dose/schedule to accurately reproduce clinical exposures. Another issue to be addressed is whether the models adequately encompass genomic and epigenomic heterogeneity. With increasing evidence of molecular subtypes within histotypes previously considered as relatively homogeneous, it is clear that additional models need to be established. We have analyzed 2134 tumor/drug studies undertaken as part of the PPTP to determine whether alternative experimental designs can be used that would allow an increased genomic diversity to be interrogated using current resources. Citation Format: Peter J. Houghton, Raushan T. Kurmasheva. Promise and challenges of pediatric cancer PDX models. [abstract]. In: Proceedings of the AACR Special Conference: Patient-Derived Cancer Models: Present and Future Applications from Basic Science to the Clinic; Feb 11-14, 2016; New Orleans, LA. Philadelphia (PA): AACR; Clin Cancer Res 2016;22(16_Suppl):Abstract nr IA15.