Intraprocedural MRI-based dosimetry during transarterial radioembolization of liver tumours with holmium-166 microspheres (EMERITUS-1): a phase I trial towards adaptive, image-controlled treatment delivery

Abstract

Purpose: Transarterial radioembolization (TARE) is a treatment for liver tumours based on injection of radioactive microspheres in the hepatic arterial system. It is crucial to achieve a maximum tumour dose for an optimal treatment response, while minimizing healthy liver dose to prevent toxicity. There is, however, no intraprocedural feedback on the dose distribution, as nuclear imaging can only be performed after treatment. As holmium-166 (166Ho) microspheres can be quantified with MRI, we investigate the feasibility and safety of performing 166Ho TARE within an MRI scanner and explore the potential of intraprocedural MRI-based dosimetry.

Methods: Six patients were treated with 166Ho TARE in a hybrid operating room. Per injection position, a microcatheter was placed under angiography guidance, after which patients were transported to an adjacent 3-T MRI system. After MRI confirmation of unchanged catheter location, 166Ho microspheres were injected in four fractions, consisting of 10%, 30%, 30% and 30% of the planned activity, alternated with holmium-sensitive MRI acquisition to assess the microsphere distribution. After the procedures, MRI-based dose maps were calculated from each intraprocedural image series using a dedicated dosimetry software package for 166Ho TARE.

Results: Administration of 166Ho microspheres within the MRI scanner was feasible in 9/11 (82%) injection positions. Intraprocedural holmium-sensitive MRI allowed for tumour dosimetry in 18/19 (95%) of treated tumours. Two CTCAE grade 3-4 toxicities were observed, and no adverse events were attributed to treatment in the MRI. Towards the last fraction, 4/18 tumours exhibited signs of saturation, while in 14/18 tumours, the microsphere uptake patterns did not deviate from the linear trend.

Conclusion: This study demonstrated feasibility and preliminary safety of a first in-human application of TARE within a clinical MRI system. Intraprocedural MRI-based dosimetry enabled dynamic insight in the microsphere distribution during TARE. This proof of concept yields unique possibilities to better understand microsphere distribution in vivo and to potentially optimize treatment efficacy through treatment personalization.

Registration: Clinicaltrials.gov, identifier NCT04269499, registered on February 13, 2020 (retrospectively registered).

Keywords: Dosimetry; Holmium; Image-guided; Personalization; SIRT; TARE.

Conflict of interest statement

J. F. W. Nijsen is a co-founder of Quirem Medical which has been acquired by Terumo Europe NV in July 2020. Nijsen has a scientific advisory role and is entitled to certain milestone payments from Terumo which are related to Quirem’s financial, operational and regulatory performance in the future. Furthermore, Nijsen is an inventor on the patents related to radioactive microspheres that are assigned to University Medical Center Utrecht Holding BV, Quirem Medical or BASF Corp. The activities of J. F. W. Nijsen within Quirem Medical are approved and supported by the Board of Directors of the Radboudumc. The other authors declare that they have no conflict of interest.

© 2022. The Author(s).

Figures

Fig. 1

Workflow during the treatment procedure. Top row represents treatment steps at the MRI scanner, and bottom row represents treatment steps at the cone-beam CT. Microspheres were injected in 4 fractions (10% and three times 30% of the microspheres), with imaging for MRI-based dosimetry in between

Fig. 2

Relationship between injected activity and mean tumour and liver doses of two patients, in whom different uptake patterns were observed (A and B). Lines are linear trendlines based on the first 3 data points, and the dashed line indicates where the fourth data point is expected based on the linear trend. A A patient with hepatocellular carcinoma (patient 4) with a linear increase in mean dose in all volumes of interest. B A patient with colorectal cancer metastases (patient 5) with an initial linear increase in mean dose in the tumours (up to 70% injected activity), and then with a decrease in the relative additive dose on the fourth fraction compared to the first three fractions

Fig. 3

99mTc-SPECT-based treatment simulation and MRI-based dose distributions in patient 5 with colorectal carcinoma liver metastases, in two different slices (rows A and B). Column 1: treatment simulation after injection of 99mTc-MAA. The high activity deposition in segment IVa and the gall bladder (arrowheads) and low uptake in the dorsal part of the second tumour (row B) are probably a result of vasospasm during injection. Column 2: T1-weighted MR images with tumours delineated with the dashed line. Columns 3–5: fusion with MRI-based dose maps generated after 40%, 70% and 100% of the activity had been injected. As more activity is administered, the mean tumour dose increases, but the tumour coverage hardly improves

Fig. 4

Digital subtraction angiography of the superior mesenteric artery (SMA) of a patient with breast cancer liver metastases (patient 1). A The chosen injection position proximal in the SMA. The asterisk indicates a branch of the SMA that is further explored in B. B More distal branch of the SMA, with contrast enhancement of tumour vasculature (T) that could have been opted for during image-guided TARE

Fig. 5

99mTc-SPECT-based treatment simulation and MRI-based dose distributions in a tumour that received a low mean dose (breast cancer metastasis, patient 1) via the superior mesenteric artery (SMA), after different fractions of microspheres had been injected. The tumour is delineated with the dashed line. The asterisk in the leftmost MR image indicates a large, confluent tumour, which was vascularized through both the SMA and the right hepatic artery. Dmean indicates the mean dose in the delineated tumour

Fig. 6

A tumour (hepatocellular carcinoma) of patient 4 treated via the right hepatic artery, in which a small part that was vascularized through an aberrant vessel originating from the phrenic artery remained untreated. A MRI-based dose distributions after each of the four fractions of radioactive microspheres had been administered. The tumour is delineated with the dashed line, and the arrow indicates the untreated area. B99mTc-SPECT/CT acquired for treatment simulation. C and D Diffusion-weighted MRI prior to and 3 months after treatment, indicating persistent diffusion restriction in the untreated area (arrow). E Digital subtraction angiography of the aberrant vessel; the asterisk indicates the persistent tumour vasculature. F 166Ho-SPECT/CT after re-treatment of the patient via the aberrant vessel, with high uptake in the untreated area (arrow)