Cost-Effectiveness and Diagnostic Effectiveness Analyses of Multiple Algorithms for the Diagnosis of Lynch Syndrome

Abstract

Background & Aims The optimal algorithm to identify Lynch syndrome (LS) among patients with colorectal cancer (CRC) is unclear. The definitive test for LS, germline testing, is too expensive to be applied in all cases. Initial screening with the Revised Bethesda Guidelines (RBG) cannot be applied in a considerable number of cases due to missing information. Methods We developed a model to evaluate the cost-effectiveness of 10 strategies for diagnosing LS. Three main issues are addressed: modeling estimates (20–40%) of RBG applicability; comparing sequential or parallel use of MSI and IHC; and a threshold analysis of the charge value below which universal germline testing becomes the most cost effective strategy. Results LS detection rates in RBG-based strategies decreased to 64.1%-70.6% with 20% inapplicable RBG. The strategy that uses MSI alone had lower yield but also lower cost than strategies that use MSI sequentially or in parallel with IHC. The use of MSI and IHC in parallel was less affected by variations in the sensitivity and specificity of these tests. Universal germline testing had the highest yield and the highest cost of all strategies. The model estimated that if charges for germline testing drop to $633–1518, universal testing of all newly diagnosed CRC cases becomes the most cost effective strategy. Conclusions The low applicability of RBG makes strategies employing initial laboratory based testing more cost effective. Of these strategies, parallel testing with MSI and IHC offers the most robust yield. With a considerable drop in cost, universal germline testing may become the most cost effective strategy for the diagnosis of LS.

Universal Panel Testing of Pancreatic Cancer Cases for Cancer Predisposition

“…http://dx.doi.org/10.1101/195537 doi: bioRxiv preprint first posted online Sep. 28, 2017; concluded that universal testing of CRC patients for LS based on sequencing alone will become more cost-effective than the two-step test when the cost of MMR gene sequencing drops to or below $609 USD, echoing a similar finding by Gould-Suarez et al [59,60] Based on our results, universal testing of pancreatic cancer patients using a panel test would identify pathogenic variants or MMR HiP-VUS in 1.4% of pancreatic cancer; this is quite comparable to currently accepted two-step universal testing of CRC patients, which identifies pathogenic MMR variants in about 1.2% of patients. [59] In a recent study, 11.8% of unselected patients with metastatic prostate cancer were found to carry pathogenic variants in DNA-repair genes.…”

Evolving approach and clinical significance of detecting DNA mismatch repair deficiency in colorectal carcinoma

“…A recent cost-effectiveness analysis estimated that if charges for germline testing dropped to $633–1518, universal germline testing of all newly diagnosed colorectal cancer cases would be more cost-effective than the various commonly used screening strategies. 64 …”

Identifying Persons with Lynch Syndrome: Why and How?

“…In this issue of Digestive Diseases and Sciences, GouldSuarez et al [5] use decision analytic modeling to compare the cost-effectiveness of several strategies to diagnose LS among patients presenting with colorectal cancer. The strategies considered rely on first applying the revised Bethesda guidelines (RBG), or first testing colorectal cancers for features suggestive of LS [abnormal immunohistochemistry (IHC) or microsatellite instability (MSI)], or proceeding directly to germline testing for mutations in the DNA MMR genes.…”

“…It is more difficult to judge how much we should we be willing to pay per incremental case of LS detected without knowing the downstream outcomes and costs. Figure 1 displays graphically the clinical (LS cases detected) and economic (total cost) outcomes for each strategy presented in Table 1 of the study by Gould-Suarez et al [5]. The slope connecting two points reflects the value of the ICER as the incremental cost per additional LS case detected by the more effective versus the less effective strategy.…”

“…Also, germline testing is assumed to detect almost all cases of LS, with no consideration of ''Lynch-like'' cases without a germline mutation as Fig. 1 Graphical representation of the effectiveness and cost for each strategy depicted in Table 1, based on the study by Gould-Suarez et al [5]. Each point represents an individual strategy, numbered as in Table 1.…”