Table 2. Prediction performance of IS-mod (percentile), IS-like (percentile), 3-category, and 2-category IS-mod in two cohorts.

Variables	Discovery cohort				Validation cohort
	C-index (95% CI)	AUC at 5-year (95% CI)	iAUC		C-index (95% CI)	AUC at 5-year (95% CI)	iAUC
IS, Immunoscore; cat, category; IS-mod, modified Immunoscore; IS-like, Immunoscore-like; IS-mod25%,70%, 3-category IS-mod using 25% and 70% as thresholds; IS-mod25%, 2-category IS using 25% as threshold; 95% CI, 95% confidence interval; AUC, area under the receiver operating characteristic curve; iAUC, integrated AUC.
IS-mod (percentile)	0.676 (0.608−0.743)	0.706 (0.623−0.789)	0.657		0.671 (0.557−0.786)	0.705 (0.593−0.818)	0.704
IS-like (percentile)	0.661 (0.594−0.729)	0.682 (0.601−0.763)	0.654		0.657 (0.553−0.760)	0.710 (0.599−0.821)	0.664
3-cat IS-mod	0.643 (0.581−0.706)	0.676 (0.600−0.752)	0.614		0.623 (0.518−0.727)	0.666 (0.565−0.767)	0.513
2-cat IS-mod	0.653 (0.597−0.710)	0.683 (0.616−0.751)	0.638		0.666 (0.580−0.752)	0.690 (0.607−0.773)	0.643
3-cat IS-mod25%,70%	0.602 (0.539−0.665)	0.611 (0.535−0.687)	0.591		0.600 (0.500−0.700)	0.609 (0.500−0.719)	0.526
2-cat IS-mod25%	0.573 (0.521−0.625)	0.579 (0.516−0.642)	0.553		0.568 (0.461−0.674)	0.574 (0.458−0.691)	0.487