Informing materials: drugs as tools for exploring cancer mechanisms and pathways

Groundwork for AI: Enforcing a benchmark for neoantigen prediction in personalized cancer immunotherapy

This article expands on recent studies of machine learning or artificial intelligence (AI) algorithms that crucially depend on benchmark datasets, often called 'ground truths.' These ground-truth datasets gather input-data and output-targets, thereby establishing what can be retrieved computationally and evaluated statistically. I explore the case of the Tumor nEoantigen SeLection Alliance (TESLA), a consortium-based ground-truthing project in personalized cancer immunotherapy, where the 'truth' of the targets-immunogenic neoantigens-to be retrieved by the would-be AI algorithms depended on a broad technoscientific network whose setting up implied important organizational and material infrastructures. The study shows that instead of grounding an undisputable 'truth', the TESLA endeavor ended up establishing a contestable reference, the biology of neoantigens and how to measure their immunogenicity having slightly evolved alongside this four-year project. However, even if this controversy played down the scope of the TESLA ground truth, it did not discredit the whole undertaking. The magnitude of the technoscientific efforts that the TESLA project set into motion and the needs it ultimately succeeded in filling for the scientific and industrial community counterbalanced its metrological uncertainties, effectively instituting its contestable representation of 'true' neoantigens within the field of personalized cancer immunotherapy (at least temporarily). More generally, this case study indicates that the enforcement of ground truths, and what it leaves out, is a necessary condition to enable AI technologies in personalized medicine.

Organizing precision medicine: A case study of Memorial Sloan Kettering Cancer Center's engagement in/with genomics

Recent decades have seen a dramatic rise of in the number of initiatives designed to promote precision oncology, a domain that has played a pioneering role in the implementation of post-genomic approaches and technologies such as innovative clinical trial designs and molecular profiling. In this paper, based on fieldwork carried out at the Memorial Sloan-Kettering Cancer Center from 2019 onwards, we analyze how a world-leading cancer center has adapted, responded, and contributed to the challenge of "doing" precision oncology by developing new programs and services, and building an infrastructure that has created the conditions for genomic practices. We do so by attending to the "organizing" side of precision oncology and to the nexus between these activities and epistemic issues. We situate the work that goes into making results actionable and accessing targeted drugs within the larger process of creating a precision medicine ecosystem that includes purpose-built institutional settings, thus simultaneously experimenting with bioclinical matters and, reflexively, with organizing practices. The constitution and articulation of innovative sociotechnical arrangements at MSK provides a unique case study of the production of a large and complex clinical research ecosystem designed to implement rapidly evolving therapeutic strategies embedded in a renewed and dynamic understanding of cancer biology.

Healthcare policy by other means: Cancer clinical research as "oncopolicy"

Social studies of biomedicine often focus on how exogenous policies shape the medical domain. While policy agendas no doubt affect complex biomedical projects, in the present paper we analyze a different dynamic, namely how oncologists enact policy as part of several flagship precision oncology endeavors. Empirically, the article focuses on the U.S. TAPUR trial, the Dutch DRUP trial, and the Canadian CAPTUR trial, which have recently been joined by similar Scandinavian studies. Taken together, these trials represent innovative forms of clinical research that, beyond their varying experimental nature, have been designed to transform the evidential processes to provide access to biomarker-driven treatments. Along with gathering evidence on effectiveness of off-label targeted therapies, their explicit goals include the recentering of a major professional organization around research, and the reframing of healthcare as a learning system seamlessly connecting epistemic, organizational, and economic issues. Accordingly, we analyze the design and implementation of these trials as a form of (onco)policy by other means.

Actionable data for precision oncology: Framing trustworthy evidence for exploratory research and clinical diagnostics

Huge amounts of genomic data produced by researchers around the world undermine data-centred discovery and therapeutic development. This paper considers how researchers make decisions about the actionability of specific datasets and the conditions that allow such data to be trusted. We discuss the case of COSMIC, a leading cancer genomics database which aggregates a large amount of sources. We research what the actionability of cancer data means in different situations of use, contrasting exploratory and diagnostics research. They highlight different questions and concerns upon genomic data use in medical research. At the same time, strategies and justifications pursued to evaluate and re-use can also share important similarities. To explain differences and similarities, we argue for an understanding of actionability and trust in data that depends on the goals and resources within the situation of inquiry, and the social epistemology of standards.

Genomic expertise in action: molecular tumour boards and decision-making in precision oncology

The recent development of cancer precision medicine is associated with the emergence of 'molecular tumour boards' (MTBs). Attended by a heterogenous set of practitioners, MTBs link genomic platforms to clinical practices by establishing 'actionable' connections between drugs and molecular alterations. Their activities rely on a number of evidential resources - for example databases, clinical trial results, basic knowledge about mutations and pathways - that need to be associated with the clinical trajectory of individual patients. Experts from various domains are required to master and align diverse kinds of information. However, rather than examining MTBs as an institution interfacing different kinds of expertise embedded in individual experts, we argue that expertise is the emergent outcome of MTBs, which can be conceptualised as networks or 'agencements' of humans and devices. Based on the ethnographic analysis of the activities of four clinical trial MTBs (three in France and an international one) and of two French routine-care MTBs, the paper analyses how MTBs produce therapeutic decisions, centring on the new kind of expertise they engender. The development and activities of MTBs signal a profound transformation of the evidentiary basis and processes upon which biomedical expertise and decision-making in oncology are predicated and, in particular, the emergence of a clinic of variants.

Multi-polar scripts: Techno-regulatory environments and the rise of precision oncology diagnostic tests

The paper examines the development and marketing of five multi-gene tests, a.k.a. as tumor signatures, designed to aid clinicians and cancer patients in therapeutic decision-making, and, in particular, to avoid overtreatment. We build on a 2011 paper that investigated the emergence of this new domain by opening the 'black box' of two pioneering tests and analyzing the hybrid, scientific-regulatory 'scripts' that were built into them. In subsequent years, second-generation tests, produced by a diverse blend of academic and commercial initiatives, have become available, and they all built into their scripts the lessons learned from their predecessors. The present paper confirms the heuristic value of the initial script-analysis but expands it to consider the multi-polar nature of the space within which multigene tests mutually position themselves. We examine how the tests were first problematized - i.e. how they described and prescribed the kind of world in which they would operate - and how their initial problematization was re-specified following the emergence of a comparative arena and their resulting informational enrichment. In parallel, we explore valuation processes, i.e. the evolving definition of the set of referents against which the assays are mutually compared, and the debates about the appropriate criteria for doing so. We note that the cancer diagnostic industry is involved in the reconfiguration of the multi-polar environment defined by socio-technical, techno-scientific, and regulatory matters of concern that seamlessly blend commercial and scientific considerations.