Table 4.
Assessment of Top Performing Tumor Features from WES CRCs in Panel-Sequenced CRC, EC, and SST Test Sets
| Tumor feature | CRC, % |
EC, % |
SST, % |
||||||
|---|---|---|---|---|---|---|---|---|---|
| Mean accuracy | 95% CI | Error rate | Mean accuracy | 95% CI | Error rate | Mean accuracy | 95% CI | Error rate | |
| MSMuTect | 27.6 | 12.7–47.2 | 72.4 | 18.2 | 5.2–40.3 | 81.8 | 35.0 | 15.4–59.2 | 65.0 |
| MSIseq | 82.8 | 64.2–94.2 | 17.2 | 68.2 | 45.1–86.1 | 31.8 | 65.0 | 40.8–84.6 | 35.0 |
| MANTIS | 100.0 | 88.1–100.0 | 0.0 | 86.4 | 65.1–97.1 | 13.6 | 85.0 | 62.1–96.8 | 15.0 |
| INDEL count | 27.6 | 12.7–47.2 | 72.4 | 18.2 | 5.2–40.3 | 81.8 | 35.0 | 15.4–59.2 | 65.0 |
| MSISensor | 96.6 | 82.2–99.9 | 3.4 | 77.3 | 54.6–92.2 | 22.7 | 75.0 | 50.9–91.3 | 25.0 |
| TMS ID2 + ID7 | 82.8 | 64.2–94.2 | 17.2 | 63.6 | 40.7–82.8 | 36.4 | 85.0 | 62.1–96.8 | 15.0 |
| TMS SBS20 | 69.0 | 49.2–84.7 | 31.0 | 50.0 | 28.2–71.8 | 50.0 | 40.0 | 19.1–63.9 | 60.0 |
| TMS SBS54 | 51.7 | 32.5–70.6 | 48.3 | 36.4 | 17.2–59.3 | 63.6 | 40.0 | 19.1–63.9 | 60.0 |
| TMB | 44.8 | 26.4–64.3 | 55.2 | 31.8 | 13.9–54.9 | 68.2 | 35.0 | 15.4–59.2 | 65.0 |
| TMS SBS15 | 44.8 | 26.4–64.3 | 55.2 | 27.3 | 10.7–50.2 | 72.7 | 60.0 | 36.1–80.9 | 40.0 |
Table presents the prediction accuracies, error rates, and corresponding 95% CIs for panel-sequenced CRCs, ECs, and SSTs for the top 10 predicting tumor features, MSMuTect, MSIseq, MANTIS, INDEL count, MSISensor, TMS ID2 + ID7, TMS SBS20, TMS SBS54, TMB (mutations/megabase), and TMS SBS15, from WES CRC analysis applied on panel-sequenced CRCs, ECs, and SSTs.
CRC, colorectal cancer; EC, endometrial cancer; ID, small insertion/deletion; INDEL, insertion/deletion; SBS, single-base substitution; SST, sebaceous skin tumor; TMB, tumor mutational burden; TMS, tumor mutational signature; WES, whole-exome sequencing.