Translating prognostic quantification of c-MYC and BCL2 from tissue microarrays to whole slide images in diffuse large B-cell lymphoma using deep learning

Table 1 Performance of AB-MIL c-MYC and BCL2 scoring as well as double-expressor performance on TMAs

Marker	Method	Pearson correlation	Sensitivity	Specificity
c-MYC	Baseline	0.842 [0.781,0.892]	0.596 [0.400,0.778]	0.993 [0.978,1.00]
c-MYC	AB-MIL	0.862 [0.797,0.907]	0.702 [0.519,0.865]	0.966 [0.933,0.993]
BCL2	Baseline	0.928 [0.902,0.950]	0.862 [0.784,0.928]	0.950 [0.896,0.987]
BCL2	AB-MIL	0.905 [0.860,0.940]	0.885 [0.819,0.946]	0.949 [0.892,0.989]
Double-expressor	Baseline	-	0.44 [0.200,0.684]	1.00 [1.00,1.00]
Double-expressor	AB-MIL	-	0.711 [0.500,0.887]	0.974 [0.945,0.994]

Pearson correlation, sensitivity, and specificity are reported along with 95% confidence intervals in brackets. Average pooling is reported as a baseline comparison method

ISSN: 1746-1596