Table 3 Performance on the external test based on input modality.

From: A self-supervised multimodal deep learning approach to differentiate post-radiotherapy progression from pseudoprogression in glioblastoma

Modality

AUC

Accuracy

Sensitivity

Specificity

MRI

0.659 ± 0.065

0.620 ± 0.119

0.454 ± 0.342

0.822 ± 0.160

(0.717)

(0.700)

(0.545)

(0.857)

Clinical Data

0.698 ± 0.026

0.630 ± 0.072

0.618 ± 0.121

0.644 ± 0.083

(0.727)

(0.700)

(0.636)

(0.778)

Multimodal

0.748 ± 0.038

0.670 ± 0.090

0.618 ± 0.118

0.733 ± 0.079

(0.753)

(0.750)

(0.727)

(0.800)

  1. Results after ensembling with soft voting are provided in parentheses.