Mutation burden and other molecular markers of prognosis in the QUASAR2 clinical trial of colorectal cancer treated with curative intent


Year:

Session type:

Theme:

Enric Domingo1,Carme Camps1,Pamela Kaisaki1,Marie Parsons2,Dmitri Mouradov2,Melissa Pentony1,Seiko Makino1,Michelle Palmieri3,Robyn Ward4,Nicholas Hawkins4,Peter Gibbs5,Hanne Askautrud6,Dahmane Oukrif7,Haitao Wang1,Joe Wood8,Elaine Johnstone1,Claire Palles1,David Church1,Marco Novelli7,Havard Danielsen6,Jon Sherlock8,David Kerr1,Rachel Kerr1,Oliver Sieber2,Jenny Taylor1,Ian Tomlinson1
1University of Oxford,2The Walter and Eliza Hall Institute of Medial Research,3The Walter and Eliza Hall Institute of Medical Research,4University of Queensland,5Royal Melbourne Hospital,6Oslo University Hospital,7University College London,8Thermo Fisher Scientific

Abstract

Background

The decision to give adjuvant chemotherapy after resection of stage II/III colorectal cancer (CRC) is based mainly on pathological factors such as tumour and nodal stage, with microsatellite instability (MSI) the only molecular marker routinely used. However, patient outcomes remain variable and stratification needs to be improved. While several relatively large studies have assessed molecular indicators of colorectal cancer (CRC) prognosis, most analyses have been restricted to a handful of markers.

Method

In 511 stage II/III CRCs from the QUASAR2 clinical trial, we assessed a panel of 82 cancer genes for mutations and additionally derived a measure of overall mutation burden.

Results

We initially refined the molecular pathways of CRC carcinogenesis. Pathogenic mutations in the genes NF1 and NRAS from the mitogen-activated protein kinase pathway co-occurred, as did SMAD4 and BRAF mutations, whilst there was a negative association between mutations in DNA damage response genes TP53 and ATM. Using relapse-free survival as an endpoint, TP53, KRAS, BRAF and GNAS mutations were independently associated with poor prognosis. In contrast, mutation burden was associated with good prognosis, even after excluding MSI+ and POLE-mutant ultramutator tumours. We successfully validated these associations in an additional Australian set of 296 stage II/III CRCs. In an extended analysis of 1,732 QUASAR2 and Australian CRCs for which KRAS, BRAF and MSI status was available, KRAS and BRAF mutations were specifically associated with poor prognosis in MSI- cancers, but not in MSI+.

Conclusion

In summary, a multi-gene panel has identified new prognostic associations in CRC with both TP53 mutation and mutation burden, and confirmed associations with KRAS and BRAF. We conclude that even a modest-sized gene panel can provide important information for use in clinical practice, and may allow existing and novel therapies to be targeted to poor prognostic sub-groups while sparing patients with good outcomes unnecessary and toxic treatment.