Worldwide, cancer is a major cause of health impairment and premature death.Footnote 1 With the rise in life expectancy across the globe, cancer incidence and mortality rates are estimated to increase substantially in the decades to come (Global Burden of Disease Cancer Collaboration et al. 2018; GBD 2017 Causes of Death Collaborators 2018). Efforts aimed at improving clinical management of cancer are extensive (Eckhouse et al. 2008; van de Loo et al. 2012). At the heart of this exertion translational cancer research (Cambrosio et al. 2006) and the imaginaries of precision medicine and precision oncology are taking shape and gaining traction (Hamburg and Collins 2010; National Research Council (US) 2011; Mirnezami et al. 2012; Collins and Varmus 2015; Celis and Heitor 2019).

Precision oncology adheres to the prevailing conceptual understanding of what cancer is: a clonal disease, caused by acquisition and accumulation of genetic alterations in cells, ultimately resulting in disruption of normal cell function (Nowell 1976; Vogelstein and Kinzler 1993; Hanahan and Weinberg 2000; Garraway et al. 2013). On the premise that these molecular events are patient specific and at the core of the causality of cancer precision oncology proposes a change of cancer management along two dimensions: (i) from groups to individuals and (ii) from morphological to molecular classification. Through identification of causally contributing molecular mechanisms the goal is to enable precise disease categorisation, prediction, prevention, early detection and targeted treatment; providing the right treatment to the right patient at the right time (Mirnezami et al. 2012; Tsimberidou et al. 2014; Collins and Varmus 2015; Ashley 2016) (Fig. 1).

Fig. 1
The diagram represents the precision medicine initiative launched by Barak Obama, in 2015. It enables categorisation, prediction, prevention, detection, and treatment.

Precision medicine is a relatively new term overlapping and gradually replacing the preceding term “personalised medicine”. According to the Precision Medicine Initiative launched by Barak Obama in 2015 precision medicine is “an emerging approach for disease treatment and prevention that takes into account individual variability in genes, environment, and lifestyle for each person”. The approach is founded on the assumption that interindividual heterogeneity result in suboptimal utility of medical interventions. Through disease stratification and tailoring of medical care outcomes can improve. The illustrations are adapted from various online sources based on the google image search “precision medicine”

Full size image

The integration of precision oncology related approaches in standard patient care is an ongoing process, resulting in shifts in diagnostic thresholds, formation of novel disease subcategories and adaptation of new treatment strategies (Jameson and Longo 2015). Currently, however, only a limited fraction of cancer patients is estimated to benefit from this line of approaches (Marquart et al. 2018). Based on the limited progress so far some investigators and clinicians have even challenged the validity, utility and sustainability of precision oncology all together (Prasad 2016; Prasad et al. 2016; Marquart et al. 2018). The tension between the current status of precision oncology and the optimism related to future benefits of this strategy is an important motivation for this work. In what follows is a brief outline of the emergence of precision medicine and precision oncology.

Precision Medicine – Tradition, Evidence, Reason and Ambition

The advancement towards increased precision in medicine and oncology can be seen as a continuation of the direction modern medicine has had since its conception (Le Fanu 2000). A recent analysis of ancient Hippocratic texts identified that inter-individual heterogeneity was recognised already 2500 years ago. This suggests that individually tailored treatment and medical care always has been a fundamental feature of applied medicine (Konstantinidou et al. 2017). It is, however, only throughout the last two centuries molecular mechanisms underlying this inter-individual heterogeneity have begun to be revealed. Technological progress has allowed a gradual increase in resolution in the exploration of both human physiology as well as pathology. Disease classification systems as well as clinical practices have evolved in close relationship with methodological advancements. This development is characterised by gradual shifts in dimensionality from the clinical and macro-anatomical organisation and understanding of human maladies to tissue centred approaches, followed by increasing attention on cells and subcellular components as the origin of pathology (Keating and Cambrosio 2001).

The concept of “molecular” disease was first put forward in 1949 by Pauling and colleagues in the Science paper “Sickle Cell Anemia, a Molecular Disease”. The authors hypothesised the genetic basis for the condition, and experimentally explored the aberrant protein product responsible for erythrocyte “sickling” (Pauling et al. 1949). In the decades that followed and up until the present genotype-phenotype relationships have been confirmed to account for a myriad of human traits and disease phenotypes (Buniello et al. 2019). The idea of precision medicine gradually emerged from this body of knowledge. It was, however, in relation to the planning and execution of the “Human Genome Project”, formally commenced in 1990, that the vision of precision medicine was truly articulated (Collins 1999). The Human Genome Project was a milestone in the development of the implicit idea of “precision medicine” into a recognizable sociotechnical imaginary: a shared vision, ambition and commitment, co-created and co- maintained by experts and policy makers (Jasanoff and Kim 2015; Tarkkala et al. 2019). The goal of the “Human Genome Project”, providing a complete sequence of the human genome, was ambitious and required considerable financial and intellectual investment. The legitimacy of this publicly funded venture was rationalised through postulations of significant scientific, medical, and societal advancements (National Research Council (US) 1988). Francis CollinsFootnote 2 put it like this: “Scientists wanted to map the human genetic terrain, knowing it would lead them to previously unimaginable insights, and from there to the common good. That good would include a new understanding of genetic contributions to human disease and the development of rational strategies for minimising or preventing disease phenotypes altogether” (Collins 1999).Footnote 3

Since 1999 the imaginary of precision medicine has matured and expanded beyond its initial scope to propose fundamental changes not only in how diseases are to be managed but also to be categorised and understood. In 2011 the National Research Council (US) released the report “Toward Precision Medicine: Building a Knowledge Network for Biomedical Research and a New Taxonomy of Disease”. Here the authors commend the development of a new taxonomy of human diseases, predominantly based on intrinsic biology and causal molecular disease mechanisms rather than signs and symptoms (National Research Council (US) 2011). The term “precision medicine” has since then rapidly been integrated in the biomedical and biotechnological scientific literature (Fig. 2) as well as the political, regulatory and public discourse (Blasimme and Vayena 2016). The uptake has been substantially intensified by the launch of the “Precision Medicine Initiative” by Barack Obama in 2015, aimed at accelerating the translation of biomedical science to improved clinical outcomes (Collins and Varmus 2015).

Fig. 2
A histogram illustrates the distribution of papers from the year 1980 to 2020. In 2010, the limits start at 50. The maximum limit is 4000. In 2020, the range goes to 2500.

Histogram of the temporal distribution of papers associated with the term “Precision medicine” in the search-engine Pubmed (10.07.2019)

Full size image

Precision Oncology – Expectations and Realisations

Considered a genetic and molecular disease cancer served as an example of the hypothesised future significance of the Human genome project as well as the transition towards a molecular based disease taxonomy (National Research Council (US) 1988, 2011). Precision medicine in relation to cancer management has been characterised by a strong emphasis on inter-individual variability of genes, and is often referred to as genomics-driven cancer medicine (Garraway et al. 2013). Medical strategies related to precision oncology are profoundly tied to postulations of genetic causality in cancer development. The idea of a monoclonal origin of cancer suggest that the cellular mass of an individual tumour share molecular characteristics involved in pathogenesis. Observations of cellular dependency of mutated or aberrantly expressed gene-products for both initiation and maintenance of malignant phenotypes support this idea and led to the postulation and experimental verification of “oncogene addiction” (Weinstein 2002). This provided a strong rational for the possibility of classifying various cancers with respect to their molecular origin, as well as molecular targeted treatment strategies. The feasibility of this approach was confirmed in the early 2000s based on several unprecedented clinical success stories, including molecular targeted therapy in chronic myeloid leukaemia (Deininger et al. 2005), gastrointestinal stromal tumours (DeMatteo 2002) and a molecular defined sub-group of breast cancer (Slamon et al. 2001).

Identification of shared molecular “drivers” in cancer cells originating from discrete cell-types and diverse tissues led to the hypothesis that this approach may be scalable, perhaps even to all cancers and all cancer patients (Tsimberidou et al. 2014). Instead of managing cancers in accordance with their macro-and microanatomical origin treatment could be guided by genomic profiling (Garraway et al. 2013). Recently, therapeutic compounds based on molecular defined indications, rather than tissue or histology, such as pembrolizumab, were subject to regulatory approvalFootnote 4 (Lemery et al. 2017; Scott 2019). This development can be seen as a sizeable stride towards making such an approach become standard of care.

While tissue agnostic indications strongly enforce the implementation of molecular profiling of all cancer patients it has been challenging to demonstrate that broad genetic testing followed by rationally selected therapeutic compounds generally lead to superior outcomes compared to current evidence-based practices (Le Tourneau et al. 2015; Stockley et al. 2016; Massard et al. 2017; Rodon et al. 2019; Rothwell et al. 2019). Experience from multiple trials as well as general estimates suggest that currently only a small percentage of cancer patients with advanced stage disease are eligible and will benefit from genome-informed therapy. Furthermore, the magnitude of clinical benefit that can be attributed to biomarker matched interventions is sobering. So far, it is a matter of additional months of life (Marquart et al. 2018; Sicklick et al. 2019), rather than years or decades, as has been achieved in chronic myeloid leukaemia, gastrointestinal stromal tumours and some patients with breast cancer (Slamon et al. 2001; DeMatteo 2002; Deininger et al. 2005).

The limited benefit of precision oncology may in part be accounted for by lack of knowledge as well as restrictions in technology, availability of therapeutic compounds and investigation in suboptimal study populations. Discovery of novel targets, development of better technological solutions, increased availability of therapeutic compounds, improved clinical infrastructure, and therapeutic repositioning to earlier disease stages may all contribute to further progress of this approach. However, more than 20 years have passed since precision oncology related approaches were first projected to result in substantial benefit (Collins 1999). It seems timely to re-explore the theoretical foundations as well as the issues at stakes and matters of concern of precision oncology, as this volume endeavours to do.

Precision Medicine and the Complexity of Biological Systems

The central dogma in molecular biology is the unidirectional flow of information from genes to proteins; the “genotype-phenotype relationship” (Crick 1958). This bottom-up model has dominated the experimental work of biomedicine as well as the interpretation of observational data. While nobody in their right mind might reject the value and validity of the fundamentals of molecular biology, it is becoming clear that a view on cancer as a simple “genotype-phenotype relationship” in linear unidirectional terms is far too simplistic (Bertolaso 2016). I have argued elsewhere that metazoan cell identity, cell state and cell fate are determined by numerous intrinsic and extrinsic factors (Engen 2020). For any given cell the selection of potential cell identities and cell states is intrinsically defined by the cell’s genetic material, the DNA. Through quantitative or qualitative alterations of complex gene-interactions a somatic mutation can reshape the trajectories of cell fate. Through emergence of new molecular features mutations in the regulatory part of the DNA or mutated gene products can open up unconventional transcriptional states resulting in novel cellular properties. Cell identity and cell state is further strongly influenced by the line of descent of the cell, defining its epigenetic configuration and confining its potential differentiation paths. Fundamentally, metazoan cells are, however, neither self-sufficient nor self-governing. Metazoan cells are collective in nature, and every new cell develops into being profoundly embedded in context. Networks of cells co-produce and co-maintain tissue and organ integrity, and collectively perform plastic transformations in response to perturbations. Through interactions like physical contact, autocrine, paracrine and endocrine signals, the collective of cells co-operate through continuously modifying their individual epigenomes and transcriptomes, in response to their surroundings (Bertolaso and Dieli 2017). Under these premises, cancer, although frequently described as a “genetic” disease, more fundamentally is a manifestation of aberrant cell behaviour. Although mutations can change the boundary conditions for a cell’s repertoire of potential phenotypic expression the effect of a mutation on a cell is profoundly relational. As a cell or a line of descending cells phenotypically diversify by expressing non-canonical transcriptional states it is in part the conditions of the environment that defines if the change is beneficial or deleterious. The emergence of novel cellular properties can accordingly never be fully understood or accounted for at the cellular or sub-cellular level. The gene-environment provides a dynamic and relational substrate where the meaning of the gene variant is defined. As neoplastic properties emerge by force of gene-gene-environment alterations the most relevant question may not be how mutations arise and translate to change but how the gene gene-environment relationship restrict the potential translational effect of novel gene-variants. Indeed, the nature of the “phenotype – emergent genotype relation” appears as a highly promising field to explore in cancer. Genotype emergence can under these premises be predictive of disease trajectories through association rather than through causation. The question is what concept of precision medicine one may retain if it is increasingly understood that the disease is a result of stochastic and emergent properties rather than deterministic linear causation.

The Imaginary of Precision Medicine and Unintended Consequences

There is no sharp demarcation between precision and non-precision related medical approaches. Rather, precision oncology is in many regards a continuation of the reductionist biomedical traditions that emerged in the nineteenth century and came to dominate in the twentieth century. As such, precision medicine is not best understood as evolution of practices, but rather as an expansion from medical practice to a techno-scientific imaginary. Although precision medicine has been presented as a societal endeavour, it is not the destination but rather a technical solution to a political oriented objective: namely to improve public health (Fig. 3). Precision oncology is as such a means to this end and not a goal in and of itself. The question is therefore not if precision oncology is feasible, but if it is “feasible enough”, as to be both desirable, viable and sustainable within certain frames. This is a combined scientific, medical, political, economic and ethical question. As a medical, political, economic, and societal aim the intended and unintended consequences related to precision medicine far exceed the sum of measurable effects of single medical interventions. The hope, vision and objective of precision medicine shape research objectives, policy agendas (Horgan et al. 2015), legal and regulatory frameworks, health care delivery systems and public expectations across the world (Tarkkala et al. 2019).

Fig. 3
The diagram illustrates the techno-scientific solution of medicine. Desirability-sustainability, ethics, and feasibility.

Precision medicine is a techno-scientific solution to a political problem

Full size image

A well cited review paper on acute myeloid leukaemia (AML), a severe cancer disease of the blood, summarised the molecular knowledge base with regards to AML pathophysiology and concluded: “Hopefully, this new biological information will contribute to less empirical approaches to treatment” (Ferrara and Schiffer 2013). This statement embodies what seems to be the prevailing mindset of the field. Precision medicine promotes a substantial change in the foundation of clinical decision making, characterised by increased reliance on bio-plausibility and de-evaluation of evidence. This is reflected in oncological practice. Stakeholders of precision oncology advocate increased pace in the translation and implementation of novel “promising” agents. This has resulted in deregulation and reduced evidence requirements for marked authorisation of novel drugs, including increased reliance on single arm studies as well as poorly validated surrogate endpoints (Chen et al. 2019; Gyawali et al. 2019; Hilal et al. 2019; Zettler et al. 2019). Based on a systematic evaluation of cancerdrugs approved by the European Medicines Agency in the period 2009–2013, Davies et al. demonstrate that the majority of novel oncological agents were approved with no clear evidence of clinical benefit and that evidence of clinically meaningful utility remained unfounded a minimum of 3 years after approval. Quantifiable benefits were marginal, and the estimated median life expansion provided when documented was only 2.7 months (Range:1–5.8 months) (Davis et al. 2017). Early marked-approval is further dis-incentivising for execution of confirmatory well powered randomised studies, resulting in absent quantification of comparative effectiveness. Lack of high-level evidence further affect the possibility and validity of cost-benefit analysis (DeLoughery and Prasad 2018). Low-grade evidence paradoxically increase uncertainty in clinical decision making (Moscow et al. 2018). Despite the failing empirical foundation (Djulbegovic and Ioannidis 2019) uptake of low-grade evidence in clinical care is substantial. There is an increase in use of off-label targeted therapy (Saiyed et al. 2017), despite evidence suggesting inefficiency (Le Tourneau et al. 2015). Some countries and health care delivery systems even have aliquoted funds for such practises, like the National Health Service (NHS, UK) Cancer Drug Funds. A recent analysis of the use of such solutions suggest no meaningful societal or patient value gain (Aggarwal et al. 2017).

These sobering results suggest that there is an increasing distance between the expectations and realisations of precision oncology. Repercussions of this friction are currently materialising across a wide range of medical as well as social domains (Fojo et al. 2014; Bowen and Casadevall 2015; MacLeod et al. 2016). The gradual implementation of low value precision oncology related strategies has contributed to a situation where the total financial burden of cancer treatment and cancer care is rapidly spiralling out of control (Aggarwal et al. 2014). This has resulted in significant financial toxicity for cancer patients (Knight et al. 2018). In settings characterised by resource constraint this has further generated restrictions in the priority setting, which ultimately result in reduced availability of novel therapeutic agents within the frames of both public healthcare systems as well as from insurance providers. With these agents being available in the free market an increasing discrepancy in access to care is materialising. Individual cancer patients no longer only fight their disease, they also battle public institutions or insurers for access to treatment (Aggarwal et al. 2014). Desperation, fear and increasing inequity may negatively influence phenomenological aspects of living and dying from cancer. Such experiences may further contribute in shaping public discourse, conceivably resulting in justified erosion of trust in scientific knowledge, medicine and policy makers. Cancer was always a medical phenomenon as well as an existential and cultural one. It is increasingly becoming political, financial and social, involving a myriad of actors, issues and concerns that literally are matters of life and death.