Machine learning analyses of highly-multiplexed immunofluorescence identifies distinct tumor and stromal cell populations in primary pancreatic tumors

Abstract

Background: Pancreatic ductal adenocarcinoma (PDAC) is a formidable challenge for patients and clinicians.

Objective: To analyze the distribution of 31 different markers in tumor and stromal portions of the tumor microenvironment (TME) and identify immune cell populations to better understand how neoplastic, non-malignant structural, and immune cells, diversify the TME and influence PDAC progression.

Methods: Whole slide imaging (WSI) and cyclic multiplexed-immunofluorescence (MxIF) was used to collect 31 different markers over the course of nine distinctive imaging series of human PDAC samples. Image registration and machine learning algorithms were developed to largely automate an imaging analysis pipeline identifying distinct cell types in the TME.

Results: A random forest algorithm accurately predicted tumor and stromal-rich areas with 87% accuracy using 31 markers and 77% accuracy using only five markers. Top tumor-predictive markers guided downstream analyses to identify immune populations effectively invading into the tumor, including dendritic cells, CD4+ T cells, and multiple immunoregulatory subtypes.

Conclusions: Immunoprofiling of PDAC to identify differential distribution of immune cells in the TME is critical for understanding disease progression, response and/or resistance to treatment, and the development of new treatment strategies.

Keywords: Pancreatic ductal adenocarcinoma (PDAC); Whole Slide Imaging (WSI); machine-learning; multiplexed-immunofluorescence (MxIF); tumor microenvironment (TME).

Fig. 1.

Representative images from cyclic multiplexed-immunofluorescence (MxIF) labeling of a primary PDAC tumor with image registration. A) Images showing distinct marker distributions in the same area across staining round 3, quenching and re-scanning at the same intensity, and re-staining round 4. Scale bar = 20 µm. B) Unaligned image of DAPI staining of the same area from round 1 (green) and 9 (red) . C) Alignment (overlaid yellow) of round 1 and 9 DAPI using image registration software. Scale bar = 100 µm. D) Aligned composite image showing a subset of markers from different rounds of the MxIF panel. Scale bar = 20 µm.

Fig. 2.

Overview of the cyclic MxIF, WSI, and image analysis pipeline. After MxIF processing and pathological validation, images are aligned and classified for tumor-stromal content. Tumor and stromal areas were quantified separately and trained via the RFA. A limited branching decision tree is shown.

Fig. 3.

Random forest outputs showing contributions of prediction driving markers in all, and those unique to representative long- and short-term survivor PDACs. A) Top 15 tumor-stroma predictors in all PDACs, excluding known cancer markers. B) Top 15 tumor-stroma predictors for long-term survivor (RAP 116), excluding cancer markers. C) Top 15 tumor-stroma predictors in short-term survivor (RAP 80), excluding cancer markers. D) Average marker rankings following jackknife resampling for top 15 markers.

Fig. 4.

Distribution of tumor-associated immune cells identified in PDAC TMEs using top ten tumor-predicting markers and unsupervised clustering. A) t-SNE plot of clustered cell groups (numbered) using RFA-identified antigens. B) Heatmap of the average cellular antigen intensity in groupings (see 4A) identified via Louvain clustering. Initial groupings shown in parentheses. C) Stacked bar plot of cell types identified with Louvain clustering of tumor and stroma tissues. Cell group colors matched in A and C.

Fig. 5.

Neighborhood analysis of all samples. A) t-SNE plot of spatial microenvironments (neighborhoods) and tissue distribution of neighborhoods between samples. Neighborhoods were classified by the distribution of cell types identified in Fig. 4. Red outline shows neighborhoods consistently recovered through all samples. B) Stacked bar plot of the cellular makeup of all neighborhoods. Approximate tissue distribution (bottom) from Supplemental Fig. 2, red indicates common neighborhoods. * indicates neighborhoods of future downstream focus. C) Correlation matrix of relationship between all cells in neighborhoods. Size and color indicate the Pearson’s correlation coefficient, positive (green), and negative (red) correlations are shown. Significant correlation (p < 0.05).

Fig. 6.

Sub-Gating of CD4+ and CD8+ T Cells, DCs, and NK cells. A) CD8+ T cell sub-gating using Louvain clustering (t-SNE plot) and GZMB distribution in CD8+ T cells (feature plot). B) Heatmaps of average marker expression in CD4+ T cells, CD8+ T cells, DC 2, and NK 2 cells. C) Stacked bar plot of CD4+ T cell, CD8+ T cell, DC 2, and NK 2 cell sub-gated cell groups in tumor and stroma regions in all patients. * = p < 0.05.

Fig. 7.

Differential expression of tumor-associated immune cells in short-term (ST) and long-term (LT) PDAC survivors. Tumor- and stroma-specific cells samples sub-gated for DC 2 (A), CD4+ T cells (B), CD8+ T cells (C), and NK 2 cells (D).