Skip to main content

Table 3 The performance of qualitative and quantitative variables for predicting esophageal fistula in the primary and validation cohorts

From: Quantitative CT analysis to predict esophageal fistula in patients with advanced esophageal cancer treated by chemotherapy or chemoradiotherapy

Cohorts

Variables

AUC

Cutoff

SEN

SPE

PPV

NPV

ACU

Primary cohort

Y1

0.946 (0.902–0.990)

2.1a

0.885

0.933

0.828

0.959

0.922

Y2

0.953 (0.909–0.997)

2.58a

0.923

0.947

0.862

0.973

0.941

Trachea invasion

0.709 (0.579–0.838)

Yes

0.444

0.973

0.857

0.839

0.833

Tumor range

0.756 (0.660–0.853)

3/4–1

0.926

0.587

0.49

0.973

0.676

T stage

0.726 (0.600–0.852)

T4b

0.519

0.933

0.737

0.843

0.824

Location

0.696 (0.582–0.809)

Mid-thorax

0.778

0.613

0.885

0.42

0.657

Deep ulcer

0.874 (0.782–0.966)

Yes

0.815

0.933

0.815

0.933

0.902

Validation cohort

Y1

0.841 (0.758–0.924)

2.1a

0.63

0.84

0.586

0.863

0.784

Y2

0.917 (0.864–0.969)

2.58a

0.63

0.827

0.567

0.861

0.775

Trachea invasion

0.524 (0.392–0.655)

Yes

0.154

0.893

0.333

0.753

0.696

Tumor range

0.648 (0.531–0.745)

3/4–1

0.741

0.569

0.37

0.854

0.598

T stage

0.562 (0.429–0.696)

T4b

0.778

0.893

0.429

0.761

0.716

Location

0.584 (0.470–0.699)

Mid-thorax

0.63

0.587

0,354

0.815

0.598

Deep ulcer

0.767 (0.649–0.885)

Yes

0.667

0.88

0.148

0.867

0.676

  1. Abbreviations: Y1 Joint predictive efficiency of quantitative CT analysis, Y2 Joint predictive efficiency of quantitative CT analysis added with T stage, AUC Area under the curve, SEN Sensitivity, SPE Specificity, PPV Positive predict value, NPV Negative predict value, ACU Accuracy
  2. aFor quantitative parameters, the value >cutoff value indicated oesophageal fistula