Table 1.
Base-case and probabilistic sensitivity analysis on outcome measures.
| Outcomes | DGS | IHC | MSI | MSIGS | ||||
| Base Case | (95% CI) | Base Case | (95% CI) | Base Case | (95% CI) | Base Case | (95% CI) | |
| Effectiveness | ||||||||
| Sensitivity of protocol | 99.90% | (99.57–99.93%) | 80.56% | (73.81–81.97%) | 82.50% | (76.15–84.03%) | 85.04% | (79.40–86.82%) |
| Specificity of protocol | 99.50% | (97.28–99.85%) | 99.98% | (99.89–99.99%) | 99.99% | (99.92–100.00%) | 99.95% | (99.68–99.98%) |
| Number of true LS cases expected to be identified * | 30 | (23–37) | 24 | (17–30) | 25 | (18–30) | 26 | (19–31) |
| Number of missed LS cases | 0 | (0–0) | 6 | (5–9) | 5 | (4–8) | 4 | (4–7) |
| Number of unexplained dMMR cases | NA | NA | 44 | (36–46) | 27 | (19–33) | 102 | (98–128) |
| Costs | ||||||||
| Cost per protocol for a 1000hypothetical cohort ($ millions) | $0.94 | ($0.51–$1.69) | $0.49 | ($0.38–$0.60) | $0.57 | ($0.44–$0.71) | $0.61 | ($0.47–$0.79) |
| Cost per CRC case screened | $940 | ($514–$1687) | $485 | ($375–$598) | $569 | ($437–$710) | $605 | ($471–$788) |
| Efficiency | ||||||||
| Cost per true LS case identified | $31,365 | ($16,618–$60,814) | $20,082 | ($14,604–$29,676) | $22,988 | ($17,006–$33,986) | $23,726 | ($17,336–$35,748) |
| Outcomes | IHCDS | MSIDS | MSIGSDS | TSGS | ||||
| Base Case | (95% CI) | Base Case | (95% CI) | Base Case | (95% CI) | Base Case | (95% CI) | |
| Effectiveness | ||||||||
| Sensitivity of protocol | 80.56% | (73.96–81.93%) | 82.50% | (76.11–84.12%) | 85.04% | (79.30–86.95%) | 99.42% | (96.55–99.63%) |
| Specificity of protocol | 99.98% | (99.88–99.99%) | 99.99% | (99.93–100.00%) | 99.95% | (99.68–99.98%) | 96.58% | (96.46–96.60%) |
| Number of true LS cases expected to be identified * | 24 | (17–29) | 25 | (18–30) | 26 | (19–31) | 30 | (22–37) |
| Number of missed LS cases | 6 | (5–9) | 5 | (4–8) | 4 | (3–7) | 0 | (0–1) |
| Number of unexplained dMMR cases | 13 | (11–17) | 13 | (8–15) | 51 | (39–59) | 43 | (31–57) |
| Costs | ||||||||
| Cost per protocol for a 1000 hypothetical cohort ($ millions) | $0.52 | ($0.40–$0.64) | $0.59 | ($0.46–$0.73) | $0.68 | ($0.55–$0.86) | $3.11 | ($2.26–$3.94) |
| Cost per CRC case screened | $517 | ($405–$636) | $588 | ($460–$728) | $679 | ($547–$857) | $3114 | ($2263–$3940) |
| Efficiency | ||||||||
| Cost per true LS case identified | $21,396 | ($16,244–$30,736) | $23,771 | ($17,751–$33,880) | $26,624 | ($19,803–$39,286) | $104,394 | ($76,520–$150,355) |
Note: * The number of true LS cases expected to be identified is 30 (20–40) in the hypothetical cohort of 1000 CRC patients based on LS prevalence of 3% (2–4%). LS, Lynch syndrome; CRC, colorectal cancer; DGS, direct germline sequencing; IHC, immunohistochemistry; MSI, microsatellite instability; MSIGS, MSI to germline sequencing; IHCDS, IHC with double somatic; MSIDS, MSI with double somatic; MSIGSDS, MSI to germline sequencing with double somatic; TSGS, tumor sequencing to germline sequencing.