Skip to main content
. 2021 May 12;11(5):870. doi: 10.3390/diagnostics11050870

Figure 2.

Figure 2

The procedure adopted to build model C starting from a “mixed” population (a) of patients with grade 1 (G1) and 2 (G2) pancreatic neuroendocrine neoplasia assessed through the analysis of histological (HS) or biopsy (BS) samples is described. In particular, repeated cross validation is applied 100 times to extract as many radiomic models trained and tested on randomly selected and balanced groups of patients (b) and exploited to build the final radiomic model (c).