Sarcomas are a highly diverse group (over 70 types) of rare (1% of all cancers) bone and soft tissue cancers that are overrepresented in children (incidence 20%) and adolescents/young adults (incidence 10%). Their diversity and rarity make it very difficult to undertake robust clinical trials of new therapies, necessitating the development of alternative models that base treatment/trial selection on the molecular profile of each patient’s specific sarcoma. We undertook a pilot study of 12 sarcoma patients (covering 9 subtypes) with metastatic disease that had failed first line treatment, by developing patient derived xenografts (PDX) and using whole exome sequencing (WES) of each patient’s germline, tumour and their PDX (N. Calvert, 2018, Clin. Sarcoma Res. 8:e4). Tumour-specific actionable driver variants and chemotherapy sensitivity/resistance signatures were identified by a bioinformatic pipeline. Selected/available treatment regimens indicated from the genomic analysis were tested on PDX cells in vitro and in pre-clinical PDX mouse trials.
The WES pipeline identified actionable therapeutics (both cancer trial therapeutics with matched indicator gene variants and clinically approved drugs with tumour-enriched target variants) for all 12 patients. We were able to establish PDXs for 7 of the 12 patients. Driver variants in the tumours were maintained in the PDXs, as was their clonal diversity, over the three passages utilized for treatment testing (in vitro and in vivo). Additional variants accumulated in PDXs over time, with those derived from tumours with signatures for defective DNA repair showing the highest rate of de novo variants. Selected predicted patient specific treatments underwent in vitro sensitivity testing (PDX cells) and in vivo efficacy analysis (PDX-bearing mice), which showed concordance between variant identified treatments and their efficacy in vitro and in vivo.