Scalable Health Labs

Towards Bio-Behavioral Medicine

Cancer Imaging Analytics

Spatial computation of intratumoral T cells correlates with survival of patients with pancreatic cancer

Julienne L. Carstens, Pedro Correa de Sampaio, Dalu Yang, Souptik Barua, Huamin Wang, Arvind Rao, James P. Allison, Valerie S. LeBleu and Raghu Kalluri, “Spatial computation of intratumoral T cells correlates with survival of patients with pancreatic cancer,” Nature Communications, April 2017.


Souptik Barua, Graduate Student, ECE, Rice University

Dr. Arvind Rao, Professor, Bioinformatics and Computational Biology, M.D Anderson Cancer Center; ECE, Rice University

Dr. Raghu Kalluri, Professsor, Cancer Biology, M.D Anderson Cancer Center; Bioengineering, Rice University

THE PROBLEM: The exact nature and dynamics of pancreatic ductal adenocarcinoma (PDAC) immune composition remains largely unknown. The goal of the study was to understand and characterize the spatial tumor-immune interactions occurring in PDAC, to determine optimal immunotherapy treatments. In addition to simple statistics such as counts, we present a novel application of spatial analysis developed in ecological statistics for the PDAC problem.

OUR SOLUTION: We used a bivariate Ripley’s L-function to characterize the spatial distribution of intratumoral T cells with respect to cancer cells. A high infiltration of an immune cell type results in a high Area Under the Curve (AUC) of the L-function curve, and vice-versa. We further investigated the potential of using the L-function AUC as an independent predictor of patient outcome.

Figure 1: The two different infiltration scenarios in b) result in two different kinds of L-function curves. A high infiltration results in a high area under the curve (AUC), while a low infiltration results in a smaller AUC, as shown in a).

MAJOR FINDINGS: Cancer cell-adjacent cytotoxic T cells correlate with survival Spatial distribution of cytotoxic Tcells in proximity to cancer cells (within ) correlates with increased overall patient survival. Further exploration of our spatial framework may improve our understanding of how specific stromal composition could impact T-cell activity, with potential impact on the optimization of immune-modulatory therapies.

Figure 2: Patients with higher cytotoxic T-cell infiltration have a significantly higher overall survival time (p = 0.038). N indicates the number of patients in the two AUC groups.