Skip to main content

Table 5 Model performance metrics for the prediction of LNM in PTC

From: Analysis of features of papillary thyroid carcinoma on color Doppler ultrasound images: implications for lymph node metastasis

Model

Cutoff

Accuracy

(95%CI)

Sensitive

(95%CI)

Specificity

(95%CI)

Positive Predictive Value

(95%CI)

Negative Predictive Value

(95%CI)

Tumor diameter(mm)

0.568

0.634(0.633–0.636)

0.414(0.334–0.494)

0.859(0.802–0.916)

0.750(0.655–0.845)

0.589(0.522–0.656)

Age(years)

0.557

0.676(0.674–0.677)

0.593(0.513–0.673)

0.761(0.690–0.831)

0.717(0.636–0.797)

0.647(0.574–0.719)

Garland pattern

0.430

0.544(0.542–0.545)

0.924(0.881–0.967)

0.155(0.095–0.214)

0.528(0.466–0.589)

0.667(0.506–0.828)

Number of lesions

0.526

0.571(0.570–0.573)

0.441(0.361–0.522)

0.704(0.629–0.779)

0.604(0.511–0.697)

0.552(0.480–0.625)

Model2

0.422

0.700(0.699–0.702)

0.593(0.727–0.859)

0.606(0.525–0.686)

0.673(0.602–0.743)

0.741(0.662–0.821)

Model1

0.573

0.704(0.702–0.705)

0.693(0.513–0.773)

0.817(0.753–0.881)

0.768(0.690–0.846)

0.663(0.593–0.733)

  1. Model 2 includes variables Age, Tumor diameter(mm) and Number of lesions
  2. Model 1 adds Garland pattern on the basis of Model, Bold values indicate the best-performing model under the same evaluation criteria