Fig. 7: Performance on other cancer entities. | Nature Communications

Fig. 7: Performance on other cancer entities.

From: Teacher-student collaborated multiple instance learning for pan-cancer PDL1 expression prediction from histopathology slides

Fig. 7

a Histogram of model performance on the other 11 cancer entities indicated by AUC, accuracy, sensitivity and specificity. The average performance of the group with PDL1 diagnostics is displayed in the leftmost column. w/PDL1 means the tumors with PDL1 as an established biomarker. b Scatter plot showing the correlation between the overall PDL1 expression level and AUC performance. Box plots displayed alongside the Y and X axes, represent the distribution of tumour-specific AUCs and the PDL1 FPKM, respectively. The central line in each box represents the median value. The box spans the interquartile range (IQR), with its lower and upper boundaries marking the 25th and 75th percentiles, respectively. Whiskers represent the maximum and minimum values. In the AUC box plots, the groups are divided into those with established PDL1 diagnostics (n = 9) and those without (n = 11). For the PDL1 FPKM box plots, the sample sizes are n = 3830 and n = 3386 for the two groups, respectively. Groups with and without PDL1 diagnostics are distinguished by dark blue and light blue markers in the data. w/, with; w/o, without. The error band represents a 95% confidence interval, calculated using bootstrap methods. Source data are provided as a Source Data file.

Back to article page