Table 3.

Diagnostic performance for sarcopenia prediction in the evaluation set (%, μ ± σ).

Models	ACC	SEN	SPE	PPV	NPV	AUC	P-value*	P-value**
Clinical	81.54 ± 4.83	67.38 ± 11.7	85.59 ± 8.96	61.84 ± 13.9	90.44 ± 2.37	79.47 ± 3.27
USG	69.73 ± 5.47	67.64 ± 4.56	70.32 ± 7.69	40.17 ± 4.84	88.40 ± 1.59	66.83 ± 3.76
FLF	83.55 ± 4.32	60.47 ± 10.6	90.14 ± 7.65	67.96 ± 13.0	89.04 ± 2.22	77.71 ± 3.88	< 0.001+	< 0.001+
SLF-SR	73.58 ± 4.96	68.39 ± 9.57	75.07 ± 8.42	45.94 ± 9.30	89.44 ± 2.00	74.70 ± 3.68	< 0.001+	< 0.001+
SLF-LSE	82.53 ± 6.34	60.92 ± 12.9	88.70 ± 11.3	68.83 ± 17.5	89.13 ± 2.44	78.35 ± 3.93	< 0.001+	< 0.001+
SLF-AAC	79.39 ± 7.62	61.78 ± 12.9	84.43 ± 13.0	62.39 ± 20.0	88.85 ± 2.30	76.74 ± 3.83	< 0.001+	< 0.001+
SLF-RF	77.53 ± 7.09	62.37 ± 12.4	81.86 ± 11.3	54.49 ± 15.1	88.64 ± 2.68	71.16 ± 5.63	< 0.001+	< 0.001+
SLF-SVM	80.20 ± 6.23	56.83 ± 14.1	86.87 ± 10.7	61.32 ± 15.4	87.89 ± 2.98	69.92 ± 7.84	< 0.001+	< 0.001+
SLF-AB	83.17 ± 5.51	59.97 ± 9.83	89.80 ± 8.80	67.47 ± 14.2	88.57 ± 5.14	74.78 ± 4.39	< 0.001+	< 0.001+

*μ = mean, σ = standard deviation, ACC = accuracy, SEN = sensitivity, SPE = specificity, PPV = positive predictive value, NPV = negative predictive value, and AUC = area under the receiver operating characteristics (ROC) curve, FLF = feature-level fusion, SLF-SR = score-level fusion by the sum-rule, SLF-LSE = score-level fusion by the least squares estimation, SLF-AAC = score-level fusion by area above the ROC curve optimization, SLF-RF = score-level fusion by random forest, SLF-SVM = score-level fusion by support vector machine, SLF-AB = score-level fusion by AdaBoost.

* Comparison of accuracy between the clinical features and after fusion (paired t-test).

** Comparison of accuracy between the USG features and after fusion (paired t-test).

⁺ Statistically significant.