Why Precision Imaging Spreads Unevenly: Management Science Offers an Answer

Precision medicine is transforming medical imaging through artificial intelligence, quantitative biomarkers and hybrid modalities such as PET-CT, yet these tools reach patients unevenly across regions and institutions. A new conceptual article in Frontiers in Medicine argues that the usual explanations—interoperability problems, validation gaps and trust deficits—miss the point. What determines whether an innovation reaches patients equitably, the authors contend, is not the technology itself but the way it travels through the health system, and imaging innovations have been travelling in the least equitable way possible.

The article, written by Renaud Guignard, Bénédicte Geffroy and Valérie Fernandez and published on 20 October 2026, draws on management science and the sociology of professions to distinguish three modes of spread. Diffusion is the spontaneous propagation of an innovation through informal professional networks—what the authors call letting it happen. Dissemination is a deliberate, directed effort to spread an innovation toward a defined population—helping it happen. Implementation is the organizational work of turning an adopted innovation into everyday routine—making it happen. Imaging innovations, they argue, have travelled mainly by diffusion, which follows existing specialist density, funding rules and like-with-like professional networks, and therefore reproduces the inequalities of access it inherits instead of correcting them.

The theoretical contribution is grounded in a bibliometric synthesis of sixty peer-reviewed documents spanning 2001 to 2025, assembled from the Scopus database using an AI-assisted application called Artirev. Co-citation analysis identified two founding poles in the literature: one treating adoption as an individual clinician’s decision, anchored in classic diffusion theory, and another treating adoption as a system-level institutional process, anchored in the work of Greenhalgh, Berwick and Rogers. Bibliographic coupling then sorted recent work into three active research streams covering the spread and translation of innovation, its embedding in organizations, and professional networks and peer effects. The authors are careful to note that the mapping step produced groupings, not mechanisms; the interpretive and theoretical steps were their own.

A central analytical move in the article is the decomposition of adoption. In a diagnostic service that does not treat its own patients, adoption actually covers three decisions taken by different actors: acquisition, where the organization commits to the technology; provision, where the imaging specialist determines what the examination actually contains; and referral, where the referring clinician decides whether the examination is requested at all. Equipment registries record acquisitions and activity statistics record referrals, but provision—the one decision that determines what the examination actually delivers—leaves no trace in routine administrative data. The authors stress this is a statement about routine data, not about research, since provision can be and has been studied qualitatively.

The argument is illustrated throughout by a three-study empirical programme on PET-CT in France. The first study mapped a coherent nationwide expansion of PET-CT supply alongside persistent, unexplained regional disparities in actual use. The second examined how nuclear medicine physicians appropriate a technology that straddles two professional jurisdictions and found the decisive tension to lie within the specialty itself, between a closure strategy confining expertise to the metabolic domain and an opening strategy incorporating radiological interpretation. The third turned to the referring oncologists, radiation oncologists and surgeons and found adoption to be contested, relationally mediated and constrained by the report itself. Read together, the authors say, these studies locate the obstacle neither in the technology nor in any single professional group, but in the articulation between them.

PET-CT is a technically revealing case. The scanner acquires two complementary images in a single session: the PET component maps metabolic activity, with a radiopharmaceutical concentrating in tissues of elevated turnover so tumours appear as focal areas of uptake, while the CT component maps anatomy, locating those areas and characterizing their size and relations to adjacent structures. The CT component may be acquired either as a low-dose localizer or as a full diagnostic study with intravenous contrast. France is unusual in that the two components fall within the competence of two distinct specialties, nuclear medicine and radiology. The device operates under a nuclear medicine authorization, and neither the authorization regime nor the payment nomenclature recognizes a diagnostic CT reading as part of the act being performed.

The medical report emerges in the article as the boundary object where professional jurisdictions meet. A report that separates metabolic from anatomical findings does not merely fail to transfer knowledge across the boundary; the authors argue it inscribes a distribution of competences and configures the referring clinician as someone who will require a further examination. They propose that a boundary object must be plastic, robust and transformative, and add a fourth property they call decidability: the artefact must contain enough to close the clinical question that prompted it, so that no residual uncertainty justifies ordering the examination it was meant to replace. An object may satisfy the first three properties and still fail on this last one—which, they argue, is precisely why the legacy examination survives alongside the new one.

The article also introduces a cross-level proposition about professional boundaries. The same demarcation may be simultaneously regulatory, jurisdictional and relational—one stratum at each level of organization. The authors propose that the resistance of a boundary to innovation increases with the number of strata that sustain it: where a single demarcation is upheld at all three levels at once, no single-level intervention can displace it. This, they suggest, explains why purely regulatory reforms and purely relational devices have both proved disappointing in hybrid imaging. The proposition carries a scope condition: it concerns innovations whose value depends on competences held by more than one professional jurisdiction. An innovation internal to a single jurisdiction may spread through diffusion alone.

At the macro level, the authors examine how hospitals frequently adopt technologies through institutional isomorphism—emulating prestigious institutions or complying with mandates—rather than to meet specific clinical needs, creating a risk of cosmetic adoption in which expensive equipment is purchased to signal modernity but not integrated into care. Activity statistics count PET-CT examinations without recording whether the CT component was acquired diagnostically and interpreted, so an examination in which the CT serves only for attenuation correction is registered identically to a fully integrated one. Measuring diffusion by volume therefore conceals a gap between the hybrid act and the hybrid act in name only. In France, early regional advantages became entrenched as adoption and multidisciplinary-team discussion fed back into further adoption, widening the gap with slower regions.

The article’s final imperative is exnovation: the active decommissioning of obsolete or low-value legacy practices, rather than layering new technology on old. In France, referring clinicians frequently request both the hybrid examination and a conventional CT, and the fragmented report itself generates additional examinations to answer the questions the PET raised. The authors conclude that equity and sustainability in precision imaging are governance choices, not by-products of technological progress, and that a health system can count installed scanners and performed examinations and still have no record of what those examinations actually contained. They extend the claim beyond imaging to molecular diagnostics, pathology in precision oncology and AI-assisted decision support, and propose that radioligand therapy—which displaces the demarcation between diagnosis and treatment—offers the most direct prospective test of their framework.

Subject of Research: How diffusion, dissemination and implementation shape equitable adoption of precision imaging innovations such as PET-CT

Article Title: From disrupted routines to sustainable practice: integrating management sciences to navigate the innovation journey of precision imaging

Article References: Guignard, R., Geffroy, B., & Fernandez, V. (2026). From disrupted routines to sustainable practice: integrating management sciences to navigate the innovation journey of precision imaging. Frontiers in Medicine, 13, Article 1965552. https://doi.org/10.3389/fmed.2026.1965552

Image Credits: AI Generated

DOI: 10.3389/fmed.2026.1965552

Keywords: precision imaging, PET-CT, diffusion, dissemination, implementation, exnovation, professional boundaries, medical report, health equity, management science, nuclear medicine, bibliometrics