Fig. 6

Arterial and venous sampling. Whole blood SUVs (mean ± SD) obtained from venous and arterial samples, in patients off (a) and on (b) erlotinib therapy. Bland-Altman plots demonstrating the arterial-venous parent fraction difference (%) per mean parent fraction value in patients off (c) and on (d) erlotinib therapy. In patients(E−) and patients(E+), a bias of −4.8 ± 4.6 and −0.7 ± 3.5 % was seen, respectively. Dotted lines indicate the mean bias and the 95 % limits of agreement. Next, Bland-Altman plots demonstrating the arterial-venous TBR_40–50 difference (%) per mean V _T value, in ten patients(E−) (e) and 11 patients(E+) (f) that had evaluable arterial and venous TBR values. The horizontal line indicates zero difference between arterial and venous TBR_40–50 measures, values above the zero difference line indicate lower venous TBR_40–50 values (i.e., an underestimation). In patients(E−) and patients(E+), a bias (i.e., average of the differences) of 26 ± 12 and 9 ± 9 % was seen, respectively. Dotted lines indicate the mean bias and the 95 % limits of agreement. Abbreviations: TBR tumor-to-blood ratio, SUV standardized uptake value, E− without erlotinib therapy, E+ with erlotinib therapy