Tumor infiltrating lymphocyte clusters are associated with response to immune checkpoint inhibition in BRAF V600E/K mutated malignant melanomas

Abstract

Patients with metastasized malignant melanomas (MM) are regularly treated with immune checkpoint inhibitors (CPI). Within our study, we evaluated the predictive value of tumor infiltrating lymphocyte (TIL) clusters in primary MM and its association to molecular subtypes to predict response to CPI treatment. A cohort of 90 MM patients who received CPI treatment were collected from a single center, as well as a validation cohort of 351 patients from the TCGA database (SKCM) who received standard of care. A deep-convolutional-neural network (U-Net) was trained to detect viable tumor areas on H&E whole-slide-images, following a quantitative detection of TILs with help of a separate additional neural network. The number of TIL clusters was associated with response to CPI in 90 MM patients (AUC = 0.6), even more pronounced within the sub-cohort of BRAF V600^E/K-mutated MM patients (AUC = 0.7, n = 32). Interestingly, the TIL clusters in NRAS-mutated as well as wildtype MM (BRAF-wt, NRAS-wt) tumors, did not demonstrate a predictive value of CPI response (AUC = 0.5, n = 25). Moreover, PD-L1 expression had a limited predictive value within our cohort. In parallel, within an independent cohort of MM patients (TCGA, n = 351), the number of TIL clusters was associated with improved survival in BRAF V600^E/K mutated MM (p < 0.0001, n = 164) but neither in NRAS-mutated (55.7 months vs. 63.0 months, respectively, p = 0.590, n = 85) nor BRAF/NRAS-wildtype MM patients (52.4 months vs. 47.4 months, respectively, p = 0.581, n = 104). While TILs in MM have been associated with improved survival, we show-for the first time-that TIL clusters are associated with response to immunotherapy in BRAF V600^E/K mutated MM.

Figure 1

Detection of viable tumor areas and TIL clusters using regular H&E whole-slide-images of malignant melanomas. Illustration of tumor segmentation using regular H&E stains of cases of malignant melanoma, with visualizations of TIL detection and TIL clustering using single images and combined overlays. (A) H&E stains of one representative case of malignant melanoma with segmentation of viable tumor areas with help of a deep convolutional neural network (U-Net). The yellow line represents the segmentation of tumor area, while necrotic areas are left out. (B) Subsequent TIL detection within viable tumor areas and illustration of TILs using a density heatmap where red indicates high density of TILs and blue indicates low density, according to the legend. (C) Visualization of TIL cluster detection. Black circles indicate individual TILs and colored areas highlight distinct TIL cluster that had been assigned using HDBSCAN. (D) Overlay of the images from panels A-C with transparent layers combining H&E image, TIL density heatmap and TIL clusters. (E) Panel of BRAF-mut (V600^E/K) NRAS-mut and BRAF/NRAS-wildtype samples, where the overlay is shown in the upper panel and the given TIL density heatmap, TIL clustering and H&E images are shown below.

Figure 2

Predictive value of TIL clusters in malignant melanoma subsets. Overview of predictive value of TIL clusters according to molecular subtypes of malignant melanoma using individual ROC curves and a nomogram combining several attributes and their predictive value to predict response to CPI treatment. (A) Area under the curve for TIL clusters for different mutation types of MM tumors (BRAF-mut; BRAF V600^E/K). (B) Area under the curve for PD-L1 scoring for different MM mutation subtypes. (C) Nomogram for the whole cohort of MM patients (n = 90). (D) Nomogram for the BRAF-mutated (BRAF-mut; BRAF V600^E/K) sub cohort (n = 32) of MM patients. The positive values (points) indicate the prediction towards responses, while negative values reflect the predictive value for therapy resistance.

Figure 3

Prognostic value of TIL clusters within molecular subtypes of malignant melanomas. Kaplan–Meier curve of TCGA MM patients (n = 351) stratified for mutational subtypes (BRAF-mutated, NRAS-mutated, and BRAF/NRAS wildtype). Inflamed tumors are indicated according to their molecular subtype, while inflamed tumors are defined as tumors with TIL clusters above the median value as a cutoff. The median survival time for each molecular subtype and inflammatory status is provided using a vertical line and indicated in letters according to their color code. The risk table is shown below for the given subgroups.