Metastatic forms of cancer account for the majority of cancer-related deaths and yet is rarely diagnosed^[1-3]. The disseminated nature of metastatic foci often limits the efficacy of existing diagnostic options. Accordingly, there is an urgent unmet medical need.

Hyperpolarized ¹³C metabolic imaging is a promising new imaging technique that may be able to improve the characterization of both primary and metastatic neoplastic diseases^[4]. Conventionally, hyperpolarized ¹³C substrates are administered via intravenous (i.v.) injection, with the intention of allowing the substrates to gain access to tumour tissue by way of vasculature. Nonetheless, the majority of newly established metastatic tumour foci are non-vascularized, and are consequently not detected within the short lifetime of ¹³C signal enhancement using conventional i.v. delivery of hyperpolarized ¹³C substrate.

In this proof-of-concept study, we investigated the possibility of detecting tumour nodules, independently of their level of vascularization, by employing intraperitoneal (i.p.) delivery of hyperpolarized [1-¹³C]pyruvate using a newly developed murine model of peritoneal carcinomatosis^[5]. Imaging of hyperpolarized [1-¹³C]pyruvate administered i.p. and converted [1-¹³C]lactate in a mouse model of abdominal metastasis was demonstrated with both 2D single time point and 3D dynamic imaging.

Tumour bearing mice were imaged 14 days after i.p. implantation of carcinoembryonic antigen (CEA) expressing MC38.CEA murine colonic carcinoma cells^[5]. Each mouse was placed supine in a dual-tuned ¹H/¹³C mouse coil (GE Healthcare) in a GE MR750 3T scanner. Neat [1-¹³C]pyruvic acid (Isotec) with 15 mM OX63 (Oxford Instruments) was polarized at 0.8 K and 5 T using a GE SpinLab DNP Polarizer. Via 24G catheter inserted through the abdominal wall, 500 µL of 80 mM pre-polarized [1-¹³C]pyruvate was injected over 12 s into the peritoneal cavity. Two imaging regimes are demonstrated: single time point 2D CSI (80 ms TR, 5 kHz bandwidth, 256 points, 6.4 cm FOV, 6 mm slice, 16×16 matrix, 10° tips) and dynamic 3D EPI with spectral-spatial excitation (TR/TE = 56/25.6 ms, FOV 64×8×6 cm, 5 mm isotropic resolution, 80°/9° net excitation of lactate/pyruvate per volume). Imaging commenced 60 s after the start of i.p. injection.Data from 5 animals are shown. Coronal 2D CSI of 2 tumour-bearing mice and 1 normal mouse without tumours (control) are shown in Figure 1 with anatomical references from ¹H T₂w FSE and postmortem dissection showing tumour nodules in the abdomen. Signal from [1-¹³C]lactate was observed in both tumour-bearing mice, but not in the control mouse. Maximum intensity projections of volumetric lactate and pyruvate data from dynamic 3D EPI of a tumour-bearing mouse are shown in Figure 2. Pyruvate and lactate are mostly localized in the anterior half of the mouse. Dynamic data from 3D EPI of a second tumour-bearing mouse were quantified within one axial slice and shown as a function of time in Figure 3.Signal quantification of the dynamic imaging data showed saturation of lactate signal beyond 3 volumes each imaged with net excitations of 80°, which suggested slower uptake and/or conversion of pyruvate to lactate by tumour nodules. In contrast with intravenous (i.v.) delivery of ¹³C substrate, where perfusion is an important parameter^[6], the primary mechanism of substrate uptake by abdominal tumour nodules is presumably diffusion from the bolus in the peritoneal cavity. Higher SNR from the dynamic EPI protocol may be achievable with smaller net excitations on lactate and longer repetition intervals to allow more [1-¹³C]lactate to accumulate between imaging volumes. The in vivo pyruvate signal exhibited a decay resembling a monoexponential function with a time constant of 54 s, which is similar to the in vitro pyruvate T₁. In addition to minimal uptake of substrate by major organs such as the heart and kidneys, there is no paramagnetic relaxation from deoxyhemoglobin in blood. The longer in vivo apparent decay constant may potentially allow an extended time window of metabolic monitoring and is one advantage of i.p. substrate administration over i.v. injection. Further optimization of the dynamic imaging protocol may allow improved characterization of the metabolic kinetics in poorly perfused metastatic tumour nodules in the abdomen, especially at longer time scales typically inaccessible by i.v. substrate delivery of hyperpolarized substrates.A novel route of [1-¹³C]pyruvate administration via intraperitoneal injection has been demonstrated with converted [1-¹³C]lactate observed in a mouse model of abdominal metastasis. With a 54 s in vivo pyruvate time constant approaching the in vitro T₁, metabolic imaging of longer time windows previously inaccessible with intravenous administration may be possible. Intraperitoneal administration of hyperpolarized ¹³C substrates is a promising complementary technique well suited for observing poorly vascularized metastatic nodules in the abdomen.