Extended Data Fig. 8: Weights for the DyAM Rad+IHC-A + Gen+PD-L1 model. | Nature Cancer

Extended Data Fig. 8: Weights for the DyAM Rad+IHC-A + Gen+PD-L1 model.

From: Multimodal integration of radiology, pathology and genomics for prediction of response to PD-(L)1 blockade in patients with non-small cell lung cancer

Extended Data Fig. 8

Parameter weight results for combined radiomic, intensity-based PD-L1 texture, genomic and TPS features for the DyAM Rad+IHC-G + Gen model. The box-and-whisker plots show the mean and the interquartile range (25–75%) as boxes, the minimum and maximum as whiskers up to 1.5 times the IQR and datapoints outside the IQR range as diamonds, for N = 10 fits. The adjoined bar plot shows the number of times, for a given feature, the following criteria were satisfied: (1) if applicable, the feature was selected as part of the radiomics filtering process, (2) the magnitude of the feature’s weight was at least 1 standard deviation from zero and (3) the magnitude of the feature’s weight was greater than 0.01. Only coefficients selected more than once are shown, up to 100.

Source data

Back to article page