109P The added clinical value of AI-assisted histopathology analyses in ICI response prediction in patients with advanced melanoma

Abstract

108P Total N463 Age (N, %) 70 266 (57%) >70 197 (43%) Sex (N, %) Male 297 (64%) Female 166 (36%) Smoking status (N, %) Never 25 (6%) Current 145 (31%) Former 254 (55%) Unknown 39 (8%) PS ECOG (N, %) 0 187 (40%) 1 217 (47%) 2 59 (13%) Histology (N, %) Squamous 75 (16%) Non-squamous 388 (84%) PD-L1 (N, %) <1% 65 (14%) 1-49% 136 (30%) 50% 233 (50%) Unknown 29 (6%) KRAS (N, %) Wild type 185 (40%) Mutated 176 (38%) Unknown 102 (22%) M stage (N, %) M0 5 (1%) M1a 98 (21%) M1b 87 (19%) M1c 273 (59%) Median follow-up (months, IQR) 11.99 (4.43;27.23)Median PFS (months, 95% CI) 5.85 (5.03;6.77)Median OS (months, 95% CI) 12.98 (10.58;15.14) Conclusions:Our workflow for TTE ML-m building is solid and well-performing with real-world data.Future developments will expand the dataset and include other data sources (i.e.imaging/radiomics).

Ähnliche Arbeiten