Identification of patients with ductal carcinoma in situ at high risk of postoperative upstaging: A comprehensive review and an external (un)validation of predictive models developed

Background: Between 8% and 56% of pure ductal carcinoma in situ are upstaged to microinvasive or invasive carcinoma on definitive pathological examination. The first objective of this study was to perform a comprehensive review of the literature on factors associated with increased risk of pre-operative underestimation. The second objective was to perform an external validation of the predictive models developed to enable their use in daily practice if relevant. Materials and methods: A literature search using Medline was undertaken. For each model selected, external validation within the study cohort was undertaken. The study cohort consisted of patients with histologically proven ductal carcinoma in situ who underwent surgical treatment at a French referral centre for cancer treatment between January 2007 and November 2018. Results: Two hundred and thirteen articles were identified; of these, 34 articles focused on factors associated with pre-operative underestimation of invasive carcinoma, 11 studies were identified as predictive models, and three studies were selected for external validation within the study cohort. Four hundred and eighty-eight eligible patients were identified in the study cohort, with an underestimation rate of 9.2%. The Jakub nomogram concordance index was 0.45 [95% confidence interval (CI) 0.39–0.51], the Park nomogram concordance index was 0.57 (95% CI 0.48–0.55), and the Coufal nomogram concordance index was 0.52 (95% CI 0.48–0.55). Conclusion: While the literature is rich on this topic, this review clearly highlights the lack of consensus regarding parameters associated with underestimation. It was not possible to validate previously published models for use in daily practice.