TY - GEN
T1 - Persistent homology based characterization of the breast cancer immune microenvironment
T2 - 36th International Symposium on Computational Geometry, SoCG 2020
AU - Aukerman, Andrew
AU - Carrière, Mathieu
AU - Chen, Chao
AU - Gardner, Kevin
AU - Rabadán, Raúl
AU - Vanguri, Rami
N1 - Publisher Copyright:
© Andrew Aukerman, Mathieu Carrière, Chao Chen, Kevin Gardner, Raúl Rabadán, and Rami Vanguri; licensed under Creative Commons License CC-BY
PY - 2020/6/1
Y1 - 2020/6/1
N2 - Persistent homology is a common tool of topological data analysis, whose main descriptor, the persistence diagram, aims at computing and encoding the geometry and topology of given datasets. In this article, we present a novel application of persistent homology to characterize the spatial arrangement of immune and epithelial (tumor) cells within the breast cancer immune microenvironment. More specifically, quantitative and robust characterizations are built by computing persistence diagrams out of a staining technique (quantitative multiplex immunofluorescence) which allows us to obtain spatial coordinates and stain intensities on individual cells. The resulting persistence diagrams are evaluated as characteristic biomarkers of cancer subtype and prognostic biomarker of overall survival. For a cohort of approximately 700 breast cancer patients with median 8.5-year clinical follow-up, we show that these persistence diagrams outperform and complement the usual descriptors which capture spatial relationships with nearest neighbor analysis. This provides new insights and possibilities on the general problem of building (topology-based) biomarkers that are characteristic and predictive of cancer subtype, overall survival and response to therapy.
AB - Persistent homology is a common tool of topological data analysis, whose main descriptor, the persistence diagram, aims at computing and encoding the geometry and topology of given datasets. In this article, we present a novel application of persistent homology to characterize the spatial arrangement of immune and epithelial (tumor) cells within the breast cancer immune microenvironment. More specifically, quantitative and robust characterizations are built by computing persistence diagrams out of a staining technique (quantitative multiplex immunofluorescence) which allows us to obtain spatial coordinates and stain intensities on individual cells. The resulting persistence diagrams are evaluated as characteristic biomarkers of cancer subtype and prognostic biomarker of overall survival. For a cohort of approximately 700 breast cancer patients with median 8.5-year clinical follow-up, we show that these persistence diagrams outperform and complement the usual descriptors which capture spatial relationships with nearest neighbor analysis. This provides new insights and possibilities on the general problem of building (topology-based) biomarkers that are characteristic and predictive of cancer subtype, overall survival and response to therapy.
KW - Persistence diagrams
KW - Topological data analysis
UR - https://www.scopus.com/pages/publications/85086505233
U2 - 10.4230/LIPIcs.SoCG.2020.11
DO - 10.4230/LIPIcs.SoCG.2020.11
M3 - Conference contribution
AN - SCOPUS:85086505233
T3 - Leibniz International Proceedings in Informatics, LIPIcs
BT - 36th International Symposium on Computational Geometry, SoCG 2020
A2 - Cabello, Sergio
A2 - Chen, Danny Z.
PB - Schloss Dagstuhl- Leibniz-Zentrum fur Informatik GmbH, Dagstuhl Publishing
Y2 - 23 June 2020 through 26 June 2020
ER -