Table 3.
Clinical classification performance of machine learning model.
| Group | ML model results (n, %) | Pathological results (n, %) | Recognition rate (%) | P value | ||||
|---|---|---|---|---|---|---|---|---|
| F0-1 | F2 | F3 | F4 | |||||
| Training cohort (N = 618) | F0-1 | 302(48.87) | 346(56) | 87.28 | 0.581 | |||
| F2 | 116(18.77) | 162(26.2) | 71.60 | |||||
| F3 | 45(7.28) | 54(8.7) | 83.33 | |||||
| F4 | 49(7.93) | 56(9.1) | 87.5 | |||||
| Validation cohort (N = 571) | F0-1 | 251(43.96) | 319(55.8) | 78.68 | 0.819 | |||
| F2 | 126(22.07) | 151(26.4) | 83.44 | |||||
| F3 | 46(8.06) | 49(8.6) | 93.87 | |||||
| F4 | 47(8.23) | 52(9.1) | 90.38 | |||||
| Study cohort (N = 1189) | F0-1 | 553(46.51) | 665(55.9) | 83.15 | 0.776 | |||
| F2 | 242(20.35) | 313(26.3) | 77.32 | |||||
| F3 | 91(7.65) | 103(8.7) | 88.34 | |||||
| F4 | 96(8.07) | 108(9.1) | 88.89 | |||||
ML machine learning.