Genomic and immune heterogeneity are associated with differential responses to therapy in melanoma

Abstract

Appreciation for genomic and immune heterogeneity in cancer has grown though the relationship of these factors to treatment response has not been thoroughly elucidated. To better understand this, we studied a large cohort of melanoma patients treated with targeted therapy or immune checkpoint blockade (n = 60). Heterogeneity in therapeutic responses via radiologic assessment was observed in the majority of patients. Synchronous melanoma metastases were analyzed via deep genomic and immune profiling, and revealed substantial genomic and immune heterogeneity in all patients studied, with considerable diversity in T cell frequency, and few shared T cell clones (<8% on average) across the cohort. Variables related to treatment response were identified via these approaches and through novel radiomic assessment. These data yield insight into differential therapeutic responses to targeted therapy and immune checkpoint blockade in melanoma, and have key translational implications in the age of precision medicine.

Fig. 1

Differential intrapatient responses to targeted therapy and immune checkpoint blockade are widespread in patients with synchronous melanoma metastases. a Change in tumor size from baseline in a cohort of 30 patients with synchronous melanoma metastases treated with BRAF/MEK-inhibitor combination first-line therapy. b Change in tumor size from baseline in a cohort of 30 patients with synchronous melanoma metastases treated with PD-1 checkpoint blockade first-line therapy. Representative CT scans showing differential intrapatient responses to therapy in two patients treated with c BRAF-inhibitor therapy and d PD-1 blockade

Fig. 2

Molecular heterogeneity in synchronous melanoma metastases. a Overall mutational analysis and overlap in targeted therapy, immune checkpoint blockade, and treatment-naïve patients and representative patients from each treatment background presented as percentage of shared (purple) and unique (blue and red) NSEM between synchronous metastases. b Aggregate genomic data showing number of somatic NSEM, percent unique and shared in targeted therapy, immune checkpoint blockade, and treatment-naïve patients. c Predicted neoantigens in a representative targeted therapy, immune checkpoint blockade, and treatment-naïve patients based on their respective IC50 values. Shown are neoantigens shared (gray) and unique (blue or red) in each metastasis within patients presenting an IC50 < 500 nM

Fig. 3

Immune heterogeneity in synchronous melanoma metastases. a Flow cytometry demonstrating the relative contribution of each immune cell subset as a percentage of total CD45⁺ cells within synchronous metastases in a representative targeted therapy, immune checkpoint blockade, and treatment-naïve patient. b Aggregate flow cytometric profiling data for all targeted therapy, immune checkpoint blockade, and treatment-naïve patients. c Immune score as calculated from gene expression profiling data. d Aggregate data showing TCR clonality in each metastasis. e Aggregate TCR sequencing data showing the percent of shared T cells detected in synchronous metastases within each patient as a % of total T cell clones. Center value represents mean and error bar represents SD. f Aggregate TCR sequencing data showing unique T cell clones within the top 5, 2.5, 1, 0.5% and 100 most prevalent T cell clones per patient

Fig. 4

Genomic and immune heterogeneity are associated with differential tumor growth, and response to targeted therapy and immune checkpoint blockade. Genomic and immune data were studied, and a somatic NSEM and CD8%, b CD8% and TCR clonality and c somatic NSEM and TCR clonality were plotted to show correlation. Agglomerative Ward’s hierarchical clustering of individual samples based on genomic (d) and immune (e) parameters. f Average clustering branch length of genomic and immune parameters of different regions of the same metastasis (red) and different metastases in the same patient (blue). g Average clustering branch length of genomic and immune parameters based on treatment background. h Correlation between radiographic change in tumor size from baseline and radiomic texture analysis features such as entropy (blue) and homogeneity (red). i Percentage of patients in whom the best and worst responding synchronous lesions within each patient show the highest entropy, energy, dissimilarity, homogeneity, and contrast by texture analysis