Exploring mechanisms of response and resistance to immuno-oncology approaches in melanoma
![Thumbnail](/bitstream/handle/internal/6248/Ben%20Shum%20MDres%20thesis.pdf.jpg?sequence=3&isAllowed=n)
View/ Open
Date
2024-05-21ICR Author
Author
Turajlic S
Shum, B
Turajlic, S
Type
Thesis or Dissertation
Metadata
Show full item recordAbstract
Background
Immuno-oncology strategies have transformed the prognosis of advanced melanoma; in particular, immune checkpoint inhibitors (ICIs) provide durable response in a subset of patients, while tebentafusp (tebe) is the first treatment to improve survival in metastatic uveal melanoma (mUM). However, many patients still do not benefit and no biomarkers are ready for use in the clinic. To date, proposed biomarkers are constrained by cohort heterogeneity or small sample size, with limited reproducibility and scalability. Through whole genome sequencing
(WGS) paired with real-world clinical data, and post-mortem sampling, I aim to explore the mechanisms of response and resistance to both treatments.
Method
In the Genomics England cohort, I co-ordinated clinical data collation from 13 NHS trusts for 247 patients, to maximise insights from correlation with WGS, including prognostic markers for survival from primary diagnoses. I identified clinical and genomic predictors for outcome following first line (1L) ICIs, including a multivariate model for response. In the PEACE post-mortem cohort, I conducted a multi-omic investigation into mechanisms of tebentafusp resistance using comprehensive multi-regional sampling in 12 patients with mUM.
Conclusions
Real-world clinical data can unleash the potential for insights from WGS, revealing a rich resource of prognostic and predictive markers. We identify numerous independent predictors of outcome following 1L ICI, in particular neoantigens and copy number alterations. We show that different subtypes of copy number loss can associate with opposing biological sequelae, and uncover a potential interaction with genomic imprinting that can influence ICI outcomes. Following logistic regression, I present a predictive model for 1L ICI response with AUC 0.91. Strikingly, neoantigen burden, escape from nonsense mediated decay mutations and tumour purity are independent predictors that withstand multivariate analyses across 3 clinical endpoints. In PEACE, tebe treated metastases sampled at post-mortem had a higher immune infiltrate and lower gp100 expression than tebe naïve metastases, generating hypotheses for potential mechanisms of action and resistance. In a pilot analysis, multi-regional sampling revealed lower expression of antigen presentation machinery in liver metastases, which may facilitate immune escape and contribute to mUM’s distinct hepatic organotropism. Together, this work highlights the central role of melanoma immunogenicity and host immune response as determinants of therapeutic outcome in advanced melanoma.
Collections
Research team
Melanoma & Kidney Cancer
Language
eng
License start date
2024-05-21
Citation
2024
Publisher
Institute of Cancer Research (University Of London)