Molecular profiling and genomic substratification to personalise therapeutic decision making in cancer
Davey, Matthew
Davey, Matthew
Loading...
Identifiers
http://hdl.handle.net/10379/17273
https://doi.org/10.13025/17895
https://doi.org/10.13025/17895
Repository DOI
Publication Date
2022-08-16
Type
Thesis
Downloads
Citation
Abstract
Breast cancer management recognises that the unique genetic expression profiles of each cancer enables substratification into clinically and therapeutically distinct molecular subgroups. There are pragmatic ways of subclassifying cancer subtypes; multigene expression assays, such as the 21-gene Recurrence Score© (RS), personalise therapeutic strategy through guiding chemoendocrine prescription based on relapse risk. Other biomarkers, such as mi(cro)RNA and pathological complete response (pCR), differentiate subgroups, guide treatment algorithms, and predict outcomes. Herein, these are extensively investigated. This study validated the local use of RS within our tertiary referral centre, by highlighting the individualisation of chemoendocrine prescription based on each patient’s relapse risk. Additionally, the expansion of indications for RS into novel settings was assessed, such as guiding neoadjuvant endocrine therapy use, estimating locoregional recurrence risk, and counselling BRCA carriers and male patients in relation to their relapse risk. A novel user-friendly nomogram was also developed for use as a surrogate prediction model of RS, with strong diagnostic test accuracy (area under curve (AUC): 0.740). This study supports using miRNA as diagnostic, prognostic, and therapeutic biomarkers in breast and colorectal (CRC) cancers. A novel five-miRNA signature was developed to aid CRC diagnosis and validated in an independent cohort (AUC: 0.830). Thereafter, the value of miR-195 and miR-135b to predict disease recurrence in CRC was successfully explored (P=0.001 & P=0.017). Validation of the ICORG10/11 translational research trial (NCT01722851) was performed and identified Let-7a and miR-145 as predictive of pCR to neoadjuvant therapies (NAT) in HER2+ and luminal A breast cancers (P=0.037 & P=0.027). Regression tree analyses elicited a relative clinical cut-off of ≤0.222 for miR-145 in predicting recurrence in patients from ICORG10/11 at 9 years follow-up (P=0.012). While most breast cancer patients experience favourable outcomes, a proportion will suffer recurrence, with mortality typically ensuing. This study established the prognostic value of achieving a pCR to NAT in HER2+ disease in our own series (hazard ratio (HR): 0.470, 95% confidence interval (CI): 0.222–0.994) and through systematic review (HR: 0.67, 95% CI: 0.60–0.74), supporting the use of pCR as the primary analytical endpoint in this phase of neoadjuvant randomized clinical trials. Finally, a panel of miRNA predictive of pCR to NAT in HER2+ disease was identified with intention for future use in our research facility. This study successfully highlighted the value of interrogating biomarkers (i.e.: RS, miRNAs, and pCR) to predict and enhance oncological outcomes. We intend to maintain this momentum – foundations for the next Irish translational research trial (the IRELAND trial) have been laid from this work. The hypothesis supporting novel biomarker discovery as a strategy of individualising therapies within current oncological practice continues to look promising.
Funder
Publisher
NUI Galway