Digital Pathology AI: The Accuracy–Clinical Utility Gap

Research Highlight PN The accuracy–utility gap Why AI accuracy does not guarantee clinical benefit in digital pathology

Full Article

The following abstract is drawn from a recently published paper in Journal of Pathology Informatics | ScienceDirect.com by Elsevier. We invite you to read the full paper and join the conversation, become a member of the Pathology News community to share your thoughts, ask questions, and engage with others around this work.

Authors: Romina Battiato^a, Felipe Montoya Silva^a, Mylene Cabrera Morales^a, Deyana Salazar Romero^a, Martín H. Benites^b,c

Abstract

Artificial intelligence (AI) in digital pathology has repeatedly achieved expert-level diagnostic accuracy under controlled validation conditions. However, its clinical translation—reflected in regulatory approval, routine adoption, changes in diagnostic decisions, or demonstrated patient benefit—remains inconsistent. This viewpoint argues that this discrepancy arises because diagnostic accuracy is a property of a model under specified conditions, whereas clinical utility emerges from the entire sociotechnical pathway in which the model is deployed. Building on recent work on standardization, system selection, institutional validation, and trial governance, we integrate these complementary layers into a continuous tissue-to-outcome pathway comprising six stages: tissue, image, model, workflow, decision, and outcome, with each stage explicitly linked to the claim it can support. Variability introduced at any stage may remain undetected and propagate downstream, potentially affecting clinical decisions and patient care. Accordingly, evidence generated at one stage cannot substitute for that required to support claims at subsequent stages. We therefore propose a stage-linked evidence map that identifies the principal risk and evidence required at each stage, while making explicit which stronger downstream claims remain unsupported. This framework enables reviewers and institutions considering adoption to identify where a study has addressed the accuracy–utility gap and where uncertainty remains. Evaluating AI across the complete pathway reframes the evidentiary requirements for clinical translation: high diagnostic accuracy is necessary, but insufficient, to establish meaningful clinical utility.

Read the full article: The accuracy–utility gap: Why AI accuracy does not guarantee clinical benefit in digital pathology – ScienceDirect

^aServicio de Anatomía Patológica, Hospital Regional Dr. Franco Ravera Zunino, Av. Lib-ertador Bernardo O’Higgins 3065, Rancagua, Región del Libertador General Bernardo O’Higgins, Chile

^bCentro de Pacientes Críticos, Clínica Las Condes, Santiago, Chile

^cEscuela de Medicina, Facultad de Medicina, Universidad Finis Terrae, Santiago, Chile

Summary

The article highlights the persistent gap between the diagnostic accuracy of artificial intelligence (AI) in digital pathology and its inconsistent clinical translation, despite AI models often achieving expert-level performance in controlled settings. The authors argue that while diagnostic accuracy is a model-specific property, clinical utility emerges from the broader sociotechnical pathway in which the AI system operates. They propose a six-stage "tissue-to-outcome" framework—comprising tissue, image, model, workflow, decision, and outcome—to systematically evaluate AI’s real-world impact. Each stage introduces potential variability that, if unaddressed, may propagate downstream, affecting diagnostic decisions and patient outcomes. The authors emphasize that evidence at one stage (e.g., model accuracy) cannot substitute for the evidence required to support claims at subsequent stages (e.g., workflow integration or patient benefit). To address this, they introduce a stage-linked evidence map, outlining key risks and evidentiary needs at each step while clarifying which downstream claims remain unsupported. This framework helps institutions and reviewers assess whether a study has bridged the accuracy-utility gap or left critical uncertainties unresolved. The takeaway is that high diagnostic accuracy alone is insufficient to ensure meaningful clinical benefit; instead, AI adoption demands comprehensive validation across the entire pathway to demonstrate real-world utility and minimize unintended consequences in patient care.