ESMO 2017-MOLECULAR ANALYSIS OF TUMORS

Large scale comprehensive genomic profiling of more than one hundred thousand advanced cancers identified potentially actionable oncogenic rearrangements including kinase domain duplications (KDD) and kinase fusions that may have therapeutic implications, researchers reported during the ESMO 2017, the Annual Congress of the European Society for Medical Oncology in Madrid, Spain.

KDD have not been characterised as extensively as fusions or other kinase-related genomic alterations seen in cancer. Kinase fusions are an important class of targetable oncogenes that are associated with both haematopoietic malignancies and solid tumours. Recently, oncogenic KDD in BRAF and EGFR were reported in the context of responses to tyrosine kinase inhibitors (TKI)^1,2.

Laurie M. Gay, Senior Scientist in Pathology, Foundation Medicine, Cambridge, USA and a team of researchers assessed the frequency of kinase-related rearrangements across hundreds of advanced cancer types to determine the landscape of oncogenic fusions, KDD, and non-canonical rearrangements.

The investigators performed comprehensive genomic profiling (CGP) on DNA and/or RNA from 114,200 solid tumours or haematological malignancy samples, evaluating up to 406 cancer-related genes and selected introns from up to 31 genes that are commonly rearranged in cancer. RNA sequencing could also be done for 265 genes for some cases.  

KDD and fusions were detected in genes that were specific to the tumour type and/or location

Fusions, kinase domain duplications, and non-canonical rearrangements were seen in all kinases, but the relative distribution of rearrangement type varied widely by gene. For example, kinase domain duplications were common for EGFR, but rare for ALK or ROS1. Those two genes were far more likely to be observed as canonical fusions, like those in non-small cell lung cancer.

© Laurie Gay. 

Across thelarge series of samples,KDD were detected in 598 (0.62%) samples. Among the genes that were found to harbour KDD were many targetable kinases, including BRAF, EGFR, FGFR1/2/3/4, RET, ERBB2, MET, ALK, ROS1, NTRK1/2, and the PDGFRA/B genes.

The proportion of samples harbouring KDD differed among cancer types, with 2.7% of 6317 brain tumour samples found to have KDD affecting genes such as EGFR, BRAF, PDGFRA, and FGFR3.

Extracranial tumours were also found to harbour KDD. RET alterations were observed in 13 to 16% of breast, lung, and thyroid KDD-positive cases. MET was implicated in 15 to 20% of uterine and brain KDD-positive cases, and ALK was the gene most often affected in 54% of lung KDD-positive cases.

KDD that were possibly related to tyrosine kinase inhibitor (TKI) resistance were found in BRAF V600E-positive melanoma and ALK-related non-small cell lung cancer (NSCLC).

KFN were most often detected in ALK, FGFR2/3, RET, and ROS1, with fusions in each gene found in 48 to 57 different tumour types. As observed previously, fusion partners varied widely by tumour site. For example, ROS1 fusions with GOPC predominated in gliomas and colorectal cancer, whereas TFG-ROS1 fusions were most common in sarcomas, and CD74- and EZR-ROS1 fusions in NSCLC.

Conclusions

The authors summarised that KDD are enriched in brain tumours and found in other diverse tumour types. KDD may also underlie acquired resistance.

Recurrent fusions are found widely, with gene partners that vary by cancer subtype or tumour site.

These findings have clinical implication since index cases have demonstrated clinical responses to matched TKIs suggest that KDD, fusions and non-canonical rearrangements can be targeted therapeutically with currently available TKIs in many histological subtypes of cancer.

Ultan McDermott of the Wellcome Trust Sanger Institute who discussed the study findings said although there is a clear trajectory towards eventual whole genome sequencing in cancer patients, a number of technical, computational and financial issues need to be resolved. Until then, hybrid capture technologies can resolve important gaps in our knowledge around prevalence of druggable rare genomic events across all common cancer types, driver landscape of every rare tumour, and evolution of drug resistance mutations over time in cancer patients.