Table 3. Performance of individual parameters in differentiating between enhanced and unenhanced MS lesions assessed by non-parametric receiver operating characteristic (ROC) curves.

The significant p-values observed show that individual texture parameters are able to differentiate between the two types of MS lesions. Eight texture parameters displayed a level of individual performance that was at least ‘good’. None of these eight parameters was found to be significantly better performing than the other.

	AUC	Se (%)	Sp (%)	Cut-off	p-value 1
Energy *	0.805	65.9	91.9	22.5	<0.0001
Entropy *	0.800	59.1	97.3	211	<0.0001
Contrast	0.798	70.5	81.1	5.28	<0.0001
Homogeneity *	0.809	61.4	97.3	119	<0.0001
Correlation	0.789	61.4	91.9	44.3	<0.0001
Inv. Diff. Moment *	0.806	59.1	97.3	122	<0.0001
Sum average	0.763	61.4	86.5	143	<0.0001
Sum variance	0.638	63.6	63.2	67.6	0.0264
Difference variance	0.736	54.5	91.9	98.2	<0.0001
ADC	0.583	51.2	78.4	937	0.2061
SRE	0.770	61.4	86.5	0.0038	<0.0001
LRE	0.761	68.2	78.4	259	<0.0001
GLN	0.754	63.6	81.1	40.9	<0.0001
RLN *	0.835	75.0	83.8	14.2	<0.0001
RP *	0.800	63.6	94.6	0.723	<0.0001
LGRE	0.764	56.8	94.6	0.80	<0.0001
HGRE *	0.805	65.9	91.9	2.63	<0.0001
SRLGE	0.738	79.5	64.9	0.004	<0.0001
SRHGE *	0.800	56.8	97.3	0.008	<0.0001
LRLGE	0.778	68.2	78.4	189	<0.0001
LRHGE	0.526	22.7	94.6	492	0.6845

¹ Parameters performing significantly better than a random classifier (p ^{(AUC > 0.5)} < 0.0167).

* Parameters with AUC ≥ 0.8 considered for a pair-wise comparison of performance.