Use of predictive models (PMs) in colorectal cancer (CRC) for diagnosis of Lynch syndrome (LS): A process measure for quality of care.
116 Background: LS diagnosis (Dx) in CRC patients (probands or Pds) and their first degree relatives (FDRs) impacts the management, outcomes and quality of care. Lack of a uniform approach to screening in academic and community centers is an impediment. Recent studies advocate universal immunohistochemistry (IHC) testing for Pds; however, these analyses exclude clinical criteria e.g., Amsterdam (Ams), Bethesda (Beth), and PMs, e.g., MMRpro (Mpro), MMRpredict (Mpre), and PREMM (PRE) due to concerns for reliability. This comprehensive comparison of all LS screening strategies (STs) aims to identify a cost effective process measure that addresses this need. Methods: We performed a cost effectiveness analysis with a societal viewpoint using TreeAge software. 21 STs for Pd and general population (GP) screening in a population of 100,000 were examined assuming a 3% LS prevalence (Prev) in Pd and 0.23% in GP, 5 FDRs per LS (Dx), 50% LS Prev in FDRs, and 90% germline testing (GT) compliance in Pds and GP and 52% in FDRs. Sensitivity, life years gained (LYG), and incremental cost effectiveness ratios (ICERs) were calculated. Results: See table. Conclusions: This study suggests that Mpro is a cost-effective first step in screening for LS in Pds, and its routine use may be considered as a possible process measure for quality of care in CRC patients. Up-front IHC +/- BRAF, GT could be reserved for Pds where history is incomplete. [Table: see text]