DCE-MRI following intraperitoneal gadodiamide injection for multimodal imaging of the tumor microenvironment

Objectives In vivo multimodality DCE-MRI/[¹⁸F]-FMISO-PET imaging of the tumor microenvironment provides many advantages as data on vasculature/blood flow can be combined with information on hypoxia. However, to date DCE-MRI relies on successful intravenous (iv) cannulation to administer the contrast agent which can be problematic in mice and injection errors are often not recognised. Hence, intraperitoneal (ip) injections might present a pragmatic and easy alternative.

Methods Having established equivalent uptake of gadodiamide (Gd, 10 mg) in DCE-MRI (T1-weighted) for the iv and ip administration routes, static [¹⁸F]-FMISO-PET imaging at 90 min p.i. was performed, followed immediately by DCE-MRI using an ip placed catheter. Parametric maps of initial slope, maximal intensity and area-under-curve (AUC) were created by non-compartmental analysis and AUC maps were co-registered to [¹⁸F]-FMISO-PET images. Tumors were harvested and EF5 staining and autoradiography performed on representative tumor slices. The murine adenocarcinoma NT subcutaneous xenograft model was used.

Results ip cannulation was quick and technically simple. Tumor uptake of Gd at steady state was equivalent for both ip and iv dosing routes but was slower for the ip route. The latter resulted in an increased measurable linear range for initial slope and in more data points for analysis of well perfused/vascularised tumor regions. [¹⁸F]-FMISO autoradiography correlated well with EF5 staining for hypoxia (r=0.952). Hypoxic tumors showed heterogeneous Gd influx following ip injection which is reflected in the ip AUC parametric maps. Co-registering these maps to [¹⁸F]-FMISO-PET images allowed classification of tumor areas into (a) poorly perfused hypoxic; (b) poorly perfused non-hypoxic; and (c) well vascularised/perfused non-hypoxic areas. In contrast, the non-hypoxic tumors were homogenous for Gd-influx and showed negligible [¹⁸F]-FMISO uptake.

Conclusions Multimodality imaging using ip DCE-MRI and [¹⁸F]-FMISO-PET imaging provides supplementary data and allows characterisation of the tumor microenvironment.

Research Support CRUK award C5255/A12678 and OCIC award C5255/A10339