Interleaved mapping of temperature and longitudinal relaxation rate to monitor drug delivery during magnetic resonance-guided high-intensity focused ultrasound-induced hyperthermia

Objectives Magnetic resonance-guided high-intensity focused ultrasound (MR-HIFU) is a method to heat lesions noninvasively to a stable, elevated temperature and a well-suited method to induce local hyperthermia (41°C-43°C) in deep-seated tissues. Magnetic Resonance (MR) imaging provides therapy planning on anatomical images and offers temperature feedback based on near-real-time MR thermometry. Although constant acquisition of MR thermometry data is crucial to ensure prolonged hyperthermia, it limits the freedom to perform measurements of other MR parameters, which are of interest during hyperthermia treatments. In image-guided drug delivery applications, co-encapsulation of paramagnetic MR contrast agents with a drug inside temperature-sensitive liposomes (TSLs) allows to visualize hyperthermia-triggered drug delivery through changes of the longitudinal relaxation rate R1. While the drug accumulates in the heated tumor tissue, R1 changes can be used for an estimate of the tumor drug concentration. The main objective of this study was to demonstrate that interleaved MR sequences are able to monitor temperature with an adequate temporal resolution and could give a reasonable estimate of the achieved tumor drug concentration through R1 changes. To this aim, in vitro validation tests and an in vivo proof-of-concept study were performed. Materials and Methods All experiments were performed on a clinical 3-T MR-HIFU system adapted with a preclinical setup. The validity of the R1 values and the temperature maps stability were evaluated in phantom experiments and in ex vivo porcine muscle tissue. In vivo experiments were performed on rats bearing a 9L glioma tumor on their hind limb. All animals (n = 4 HIFU-treated, n = 4 no HIFU) were injected intravenously with TSLs co-encapsulating doxorubicin and gadoteridol as contrast agent. The TSL injection was followed by either 2 times 15 minutes of MR-HIFU-induced hyperthermia or a sham treatment. R1 maps were acquired before, during, and after sonication, using a single slice Inversion Recovery Look-Locker (IR-LL) sequence (field of view [FOV], 50 × 69 mm2; in-plane resolution, 0.52 × 0.71 mm2; slice thickness, 3 mm; 23 phases of 130 milliseconds; 1 full R1 map every 2 minutes). The R1 maps acquired during treatment were interleaved with 2 perpendicular proton resonance frequency shift (PRFS) MR thermometry slices (dynamic repetition time, 8.6 seconds; FOV, 250 × 250 mm2; 1.4 × 1.4 mm2 in-plane resolution; 4 mm slice thickness). Tumor doxorubicin concentrations were determined fluorometrically. Results In vitro results showed a slight but consistent overestimation of the measured R1 values compared with calibrated R1 values, regardless whether the R1 was acquired with noninterleaved IR-LL or interleaved. The average treatment cell temperature had a slightly higher temporal standard deviation for the interleaved PRFS sequence compared with the noninterleaved PRFS sequence (0.186°C vs 0.101°C, respectively). The prolonged time in between temperature maps due to the interleaved IR-LL sequence did not degrade the temperature stability during MR-HIFU treatment (Taverage = 40.9°C ± 0.3°C). Upon heat treatment, some tumors showed an R1 increase in a large part of the tumor while other tumors hardly showed any ΔR1. The tumor doxorubicin concentration showed a linear correlation with the average ΔR1 during both sonications (n = 8, R2adj = 0.933), which was higher than for the ΔR1 measured after tumor cooldown (averaged for both sonications, n = 8, R2adj = 0.877). Conclusions The new approach of interleaving different MR sequences was applied to simultaneously acquire R1 maps and PRFS thermometry scans during a feedback-controlled MR-HIFU-induced hyperthermia treatment. Interleaved acquisition did not compromise speed or accuracy of each scan. The ΔR1 acquired during treatment was used to visualize and quantify hyperthermia-triggered release of gadoteridol from TSLs and better reflected the intratumoral doxorubicin concentrations than the ΔR1 measured after cooldown of the tumor, exemplifying the benefit of interleaving R1 maps with temperature maps during drug delivery. Our study serves as an example for interleaved MR acquisition schemes, which introduce a higher flexibility in speed, sequence optimization, and timing.