Radiomics has produced tens of thousands of publications yet almost no handcrafted radiomic signatures in routine clinical use, and the reasons are increasingly understood to be problems of reproducibility and clinical translation rather than of algorithms. This critical narrative review argues that the field systematically generates paper-grade evidence—findings sufficient to publish—far faster than decision-grade evidence—findings sufficient to change clinical practice. Drawing on meta-scientific research, we describe seven fragility mechanisms (publication bias, analytical flexibility, underpowering, HARKing [hypothesizing after the results are known], citation distortion, cognitive bias, and misaligned incentives) and show why radiomics is structurally exposed to all of them simultaneously: high-dimensional feature spaces, acquisition-dependent measurement instability, segmentation variability, retrospective single-centre data, small samples, and leakage-prone validation. We then summarise empirical evidence on the radiomics literature, which remains pervaded by suboptimal methodological quality, near-absent negative results, limited external validation, sparse calibration and clinical-utility assessment, low data and code sharing, and a measurable retraction signal. We interpret these patterns as the output of a self-reinforcing system rather than isolated errors, and argue that better algorithms alone cannot resolve them. Finally, we argue that closing this gap requires not better models but evidentiary discipline: the consistent, enforceable application of standards the field already has, and the calibration of published claims to the strength of the underlying evidence.

Radiomics has produced tens of thousands of publications yet almost no tools in routine clinical use, and the reasons are increasingly understood to be problems of reproducibility and clinical translation rather than of algorithms. This critical narrative review argues that the field systematically generates paper-grade evidence—findings sufficient to publish—far faster than decision-grade evidence—findings sufficient to change clinical practice. Drawing on meta-scientific research, we describe seven fragility mechanisms (publication bias, analytical flexibility, underpowering, HARKing [hypothesizing after the results are known], citation distortion, cognitive bias, and misaligned incentives) and show why radiomics is structurally exposed to all of them simultaneously: high-dimensional feature spaces, acquisition-dependent measurement instability, segmentation variability, retrospective single-centre data, small samples, and leakage-prone validation. We then summarise empirical evidence on the radiomics literature, which remains pervaded by suboptimal methodological quality, near-absent negative results, limited external validation, sparse calibration and clinical-utility assessment, low data and code sharing, and a measurable retraction signal. We interpret these patterns as the output of a self-reinforcing system rather than isolated errors, and argue that better algorithms alone cannot resolve them. Finally, we argue that closing this gap requires not better models but evidentiary discipline: the consistent, enforceable application of standards the field already has, and the calibration of published claims to the strength of the underlying evidence.

Diagnostic imaging is central to contemporary healthcare, but increasing demand has intensified concerns regarding inappropriate and low-value examinations. Imaging appropriateness should be understood as a multidimensional quality framework rather than a simple reduction in examination volume. This narrative review synthesizes evidence on clinical outcomes, radiation protection, healthcare costs, organizational efficiency, equity, incidental findings, and environmental sustainability. A structured PubMed search was complemented by targeted database and institutional searches and selective citation tracking. Low-value imaging may disrupt diagnostic pathways through diagnostic uncertainty, unnecessary follow-up, avoidable radiation exposure, additional costs, waiting-list pressure, and environmental burden. Mitigation strategies include referral guidelines, clinical decision support, radiologist-led vetting, standardized incidental findings management, and audit and feedback. Their effects vary by setting and outcome: improved referral scores do not necessarily reduce utilization or improve patient outcomes. Environmental estimates depend on imaging modalities, electricity supply, and assessment boundaries, while proposed artificial intelligence applications require prospective validation. We translate this evidence into an operational pathway linking examination justification, modality and protocol selection, reporting, follow-up, and service evaluation. Imaging appropriateness therefore requires reducing low-value examinations while preserving diagnostic safety and timely, equitable access to clinically beneficial imaging.