Fig. 2
The rank of each study based on manual TLG measurements against the rank of each study based on the CNN-based measurements. This serves to show, that in terms of ranking patients with respect to tumour burden the automated method is similar to a manual analysis