Table 5.

Model Performance

Goodness of Fit Test	Data Set	HAP (SE)	mHAP‐III (SE)	Pre‐TACE‐Predict Model (SE)	Post‐TACE‐Predict Model (SE)
Harrell’s C index	Training	0.616 (0.010)	0.651 (0.011)	0.682 (0.010)	0.723 (0.013)
	Internal validation	0.624 (0.009)	0.649 (0.010)	0.659 (0.010)	0.693 (0.016)
	External validation (Eastern)	0.640 (0.012)	0.687 (0.012)	0.707 (0.012)	0.730 (0.011)
	External validation (Western)	0.597 (0.015)	0.618 (0.016)	0.613 (0.017)	0.631 (0.017)
Gönen & Heller’s K	Training	0.592 (0.010)	0.633 (0.010)	0.651 (0.010)	0.680 (0.012)
	Internal validation	0.598 (0.010)	0.617 (0.010)	0.623 (0.010)	0.654 (0.013)
	External validation (Eastern)	0.605 (0.013)	0.655 (0.011)	0.667 (0.012)	0.681 (0.012)
	External validation (Western)	0.581 (0.014)	0.545 (0.023)	0.587 (0.016)	0.596 (0.016)
Royston‐Sauerbrei’s $R_D^2$	Training	0.078 (0.015)	0.132 (0.021)	0.181 (0.020)	0.262 (0.034)
	Internal validation	0.087 (0.016)	0.111 (0.020)	0.120 (0.020)	0.185 (0.030)
	External validation (Eastern)	0.096 (0.023)	0.184 (0.024)	0.209 (0.028)	0.243 (0.034)
	External validation (Western)	0.059 (0.023)	0.050 (0.019)	0.058 (0.022)	0.073 (0.026)

SEs were estimated from 200 bootstrap samples.

Abbreviation: SE, standard error.