Heterogeneity of recurrences in rectal cancer: application of population models facilitates personalized medicine

Abstract

AbstractDue to heterogeneity in presentation and outcome, patients with metastatic disease cannot be considered a single group. The timing, location and combinations of recurrences determine the feasibility of treatment of the individual patient in an era in which the options for local and systemic treatment have expanded. Studies investigating this complexity are hampered by the lack of both large cohorts and adequate methods.In a well-defined cohort of rectal cancer patients from a randomized clinical trial, with long standardized follow-up, we applied spatial projection models derived from population ecology to overcome the complexity problem. We describe the recurrence patterns in detail and performed stochastic simulation experiments resulting in 1.5 million evaluable patients. The risk of subsequent recurrences was dependent on the presentation of the first recurrent event and decreased with increasing recurrence-free interval. The risk of local recurrence for the median patient (65.8 years, pT3 adenocarcinoma) was threefold increased after the development of rare metastases. The risk of development of rare metastases was increased after the development of other extrahepatic metastases.Our cross-disciplinary approach delivers insights allowing for the development of personalized strategies for (local) treatment of recurrent disease, as well as for surveillance strategies that may potentially impact large patient cohorts. In this proofof-principle study we demonstrate the feasibility of spatial projection models for cancer research.