"Just Leave the Patient as a Number": Algorithmic Renderings of Cancer

• In Event: Emergent Technologies in Biomedicine and Healthcare: Boundaries, Tensions, and Possibilities - II

Abstract

This paper explores how whole genome sequencing (WGS) technologies are changing cancer and oncology practices, which in turn refiguring cancer, health and the body. Human genomics is essentially an emergent form of medical (big) data. My fieldwork sites involve a cancer genomic clinical trial, entitled Personalized OncoGenomics (POG), in Vancouver, British Columbia, which pioneers in WGS of metastatic cancer patients. Through my fieldwork I examine how genomic sequencing analysis, a form of machine learning, is changing oncology practice. More particularly the paper highlights how these new techniques of machine learning are engendering a new algorithmic approach of practicing oncology as well as shifting levels of priority and authority in clinical decision-making. I conducted a set of 35 interviews with a group of multidisciplinary medical experts including doctors, pathologists, bioinformaticians, genome scientists, and genetic counsellors. I also conducted a two-month participation observation at POG tumour board and ethics meetings, which represent a space for knowledge co-construction of cancer genomics. The ethnographic data shows that machine learning techniques of cancer render the cancer body numerical and quantifiable. These algorithmic approaches leave out social elements of the patient, such as race and social class, and in turn regroup and reclassify patients and individuals into baskets of treatments based on the logic of probability and inference. The paper hopes to shed light on three important changes in oncology engendered by genomic technologies: structural, epistemological and ontological, a trilogy that underlies the majority of research in the field of science and technology studies.

Author

• Tiên-Dung Hà, Stanford University