Skip to main content
. 2025 Jun 24;1(1):3. doi: 10.1186/s44398-025-00003-8

Table 2.

AUROC across 2-year endpoints and cohort for the obtained radiomics models alongside p-value of one-sided permutation tests with alternative hypothesis of superiority of MIL models over radiomics models

Endpoint Cohort AUROC [95% CI] 2D MIL AUROC [95% CI] radiomics p-value
2-year OS RADCURE test 0.75 [0.69–0.80] 0.74 [0.68–0.80] 0.50
2-year OS HN1 0.77 [0.68–0.85] 0.74 [0.64–0.85] 0.50
2-year OS HN-PET-CT 0.84 [0.77–0.90] 0.68 [0.56–0.79] 0.012
2-year LRC RADCURE test 0.75 [0.68–0.81] 0.69 [0.62–0.76] 0.15
2-year LRC HN1 0.66 [0.52–0.78] 0.74 [0.61–0.84] 0.84
2-year LRC HN-PET-CT 0.72 [0.61–0.80] 0.64 [0.53–0.74] 0.15
2-year FFDM RADCURE test 0.78 [0.71–0.84] 0.71 [0.64–0.77] 0.21
2-year FFDM HN1 0.75 [0.55–0.92] 0.68 [0.52–0.83] 0.48
2-year FFDM HN-PET-CT 0.71 [0.59–0.80] 0.67 [0.66–0.79] 0.53