Skip to main content

Table 1 Performance analysis of the SVM-based multi-view and single-source pipelines for adjuvant treatment response

From: Radiotranscriptomics of non-small cell lung carcinoma for assessing high-level clinical outcomes using a machine learning-derived multi-modal signature

 

AUC

SN

SPC

Multi-view

0.72 ± 0.08

0.58 ± 0.2

0.61 ± 0.2

Multi-view score

0.74 ± 0.06

0.65 ± 0.08

0.62 ± 0.1

Deep features

0.69 ± 0.1

0.29 ± 0.2

0.84 ± 0.15

Deep feature score

0.69 ± 0.09

0.71 ± 0.19

0.57 ± 0.2

Radiomics

0.68 ± 0.1

0.72 ± 0.2

0.51 ± 0.15

Radiomic score

0.71 ± 0.08

0.70 ± 0.15

0.56 ± 0.18

Transcriptomics

0.64 ± 0.11

0.69 ± 0.2

0.38 ± 0.2

Transcriptomics score

0.66 ± 0.1

0.63 ± 0.15

0.55 ± 0.13

  1. The following metrics represent the mean ± standard deviation of 100 iterations for each pipeline. SVM support vector machine, AUC area under curve, SN sensitivity, SPC specificity
Back to article page

BioMedical Engineering OnLine

ISSN: 1475-925X

Contact us