In non-small cell lung cancer (NSCLC) patients off erlotinib therapy, positron emission tomography (PET) using [¹¹C]erlotinib distinguished epidermal growth factor receptor (EGFR) mutations from wild-type EGFR. However, tumor uptake of [¹¹C]erlotinib during erlotinib therapy is unknown. Therefore, the aims of this study were to evaluate tumor [¹¹C]erlotinib uptake in NSCLC patients both on and off erlotinib therapy, to evaluate the effect of erlotinib therapy on tumor perfusion and its correlation to tumor [¹¹C]erlotinib uptake, and also, to investigate simplified uptake parameters using arterial and venous blood samples.

Ten patients were to be scanned twice with a 1–2-week interval, i.e., on (E+) and off (E−) erlotinib therapy. Each procedure consisted of a low-dose CT scan, a 10-min dynamic [¹⁵O]H₂O PET scan, and a 60-min dynamic [¹¹C]erlotinib PET scan with arterial and venous sampling at six time points. In patients(E+), the optimal compartment model was analyzed using Akaike information criterion. In patients(E−), the uptake parameter was the volume of distribution (V _T), estimated by using metabolite-corrected plasma input curves based on image-derived input functions and discrete arterial and venous blood samples. Tumor blood flow (TBF) was determined by rate constant of influx (K1) of [¹⁵O]H₂O using the 1T2k model and correlated with V _T and K1 values of [¹¹C]erlotinib. The investigated simplified parameters were standardized uptake value (SUV) and tumor-to-blood ratio (TBR) at 40–60 min pi interval.

Of the 13 patients included, ten were scanned twice. In patients(E+), [¹¹C]erlotinib best fitted the 2T4k model with V _T. In all patients, tumor V _T(E+) was lower than V _T(E−) (median V _T(E−) = 1.61, range 0.77–3.01; median V _T(E+) = 1.17, range 0.53–1.74; P = 0.004). Using [¹⁵O]H₂O, five patients were scanned twice. TBF did not change with erlotinib therapy, TBF showed a positive trend towards correlation with [¹¹C]erlotinib K1, but not with V _T. TBR_40–50 and TBR_50–60, using both arterial and venous sampling, correlated with V _T(E−) (all r _s >0.9, P < 0.001), while SUV did not. In patients off and on therapy, venous TBR underestimated arterial TBR by 26 ± 12 and 9 ± 9 %, respectively.

In patients on erlotinib in therapeutic dose, tumor V _T decreases with high variability, independent of tumor perfusion. For simplification of [¹¹C]erlotinib PET scanning protocols, both arterial and venous TBR 40–60 min post injection can be used; however, arterial and venous TBR values should not be interchanged as venous values underestimate arterial values.

Registered at the Netherlands Trial Registry: NTR3670 .