Table 2 Performance of the deep-learning system on the internal test set of 1086 ultrasound images in 632 patients.

From: A deep-learning pipeline to diagnose pediatric intussusception and assess severity during ultrasound scanning: a multicenter retrospective-prospective study

Images

Accuracy (95%CI)

Sensitivity (95%CI)

Specificity (95%CI)

AUC (95%CI)

FK (95%CI)

Average-AUC (95%CI)

FPS Median (range)

NSD

0.971 (0.971–0.972)

0.966 (0.953–0.978)

0.989 (0.976–1.000)

0.977 (0.968–0.986)

0.925 (0.899–0.951)

0.972 (0.954–0.994)

102 (93–109)

NSS

0.971 (0.971–0.972)

0.959 (0.927–0.991)

0.973 (0.963–0.984)

0.966 (0.949–0.983)

0.884 (0.844–0.924)

SSI

0.985 (0.985–0.985)

0.958 (0.923–0.994)

0.989 (0.982–0.995)

0.973 (0.955–0.992)

0.927 (0.891–0.962)

  1. AUC The area under the receiver operating characteristic curve, Average-AUC the arithmetic mean of the AUC for each class, FK Fleiss’ Kappa, FPS Frames per second.
Back to article page