Table 1 Prevalence of PD-L1 expression by artificial intelligence-powered and manual scoring.
Evaluable samples for AI-powered assessment—no. | Prevalence—no. (%) | P value | ||||
|---|---|---|---|---|---|---|
AI-powereda | Manuala | |||||
1% cutoff | ||||||
<1% | ≥1% | <1% | ≥1% | |||
CheckMate 057 (NSCLC) | 194 | 19 (9.8) | 175 (90.2) | 95 (49.0) | 99 (51.0) | <0.01 |
CheckMate 275 (UC) | 241 | 75 (30.1) | 166 (68.9) | 128 (53.1) | 113 (46.8) | <0.01 |
CheckMate 067 (MEL; NIVO) | 265 | 91 (34.3) | 174 (65.7) | 104 (39.3) | 161 (60.8) | 0.28 |
CheckMate 067 (MEL; NIVO + IPI) | 262 | 101 (38.6) | 161 (61.5) | 117 (44.7) | 145 (55.3) | 0.18 |
CheckMate 238 (MEL) | 377 | 70 (18.6) | 307 (81.4) | 118 (31.3) | 259 (68.7) | <0.01 |
CheckMate 141 (SCCHN) | 153 | 88 (57.5) | 65 (42.5) | 69 (45.1) | 84 (54.9) | 0.04 |
5% cutoff | ||||||
<5% | ≥5% | <5% | ≥5% | |||
CheckMate 026 (NSCLC) | 254 | 21 (8.3) | 233 (91.7) | 58 (22.8) | 196 (77.2) | <0.01 |
CheckMate 057 (NSCLC) | 194 | 84 (43.3) | 110 (56.7) | 117 (60.3) | 77 (39.7) | <0.01 |
CheckMate 275 (UC) | 241 | 151 (62.7) | 90 (37.3) | 167 (69.3) | 74 (30.7) | 0.15 |
CheckMate 067 (MEL; NIVO) | 265 | 162 (61.1) | 103 (38.9) | 189 (71.3) | 76 (28.7) | 0.02 |
CheckMate 067 (MEL; NIVO + IPI) | 262 | 170 (64.9) | 92 (35.1) | 197 (75.2) | 65 (24.8) | 0.01 |
CheckMate 238 (MEL) | 377 | 143 (37.9) | 234 (62.1) | 238 (63.1) | 139 (36.9) | <0.01 |
CheckMate 141 (SCCHN) | 153 | 109 (71.2) | 44 (28.8) | 101 (66.0) | 52 (34.0) | 0.39 |
