Characterizing Surveillance Recommendations From National Comprehensive Cancer Network Guidelines
A partir d'une revue systématique des directives du "National Comprehensive Cancer Network" sur les tumeurs solides chez l'adulte (43 directives), cette étude examine les caractéristiques des directives existantes en matière de surveillance du cancer (nombre, niveau de preuve...) afin d'identifier des pistes d'amélioration
Introduction : Cancer surveillance strategies aim to detect recurrence in patients treated with curative intent. Despite a strong theoretical rationale for the benefit of surveillance, empirical evidence does not demonstrate a survival benefit from routine follow-up testing.1 Instead, known harms can occur without improvements in mortality or quality of life.2-4 Despite growing scrutiny of this unfavorable risk-benefit ratio, the scope of surveillance recommendations in the US remains poorly defined. We characterized existing cancer surveillance guidelines to highlight areas of ambiguity, identify potential cases of oversurveillance, and suggest opportunities to improve current recommendations.
Methods : Between June and July 2025, we systematically reviewed the National Comprehensive Cancer Network (NCCN) Clinical Practice Guidelines in Oncology (NCCN Guidelines) for solid-organ cancers in adults. Two reviewers (A.S.B. and A.K. or M.B.) independently extracted surveillance recommendations following definitive treatment and recorded their NCCN category of evidence. Recommendations were classified by stratification type (eg, risk, stage) and modality (eg, imaging, laboratory testing), and assessed for frequency, duration, and individualization to patient-specific factors. Recommendations addressing only metastatic, recurrent, or progressive disease were excluded. This study followed the Standards for Quality Improvement Reporting Excellence version 2.0 (SQUIRE) reporting guideline for quality improvement studies and was exempt from institutional review board review based on the University of California, San Francisco self-certification protocol.
Results : Forty-three guidelines provided 483 recommendations across 99 cancer types (Table 1). Four hundred fifty recommendations (93%) of recommendations were assigned category 2A evidence (ie, lower-level evidence with uniform consensus). Although 327 recommendations (68%) were stratified by stage, treatment, or a combination of factors, only 116 (24%) called for individualizing surveillance testing to patient-specific factors (eg, measuring a tumor marker only when initially elevated).
JAMA Network Open , article en libre accès, 2025