Research Article
Alginate moulding: an empirical method for magnetic resonance imaging/positron emission tomography co-registration in a tumor rat model

https://doi.org/10.1016/j.nucmedbio.2008.04.002Get rights and content

Abstract

Background and Purpose

In the experimental field of animal models, co-registration between positron emission tomography (PET) and magnetic resonance imaging (MRI) data still relies on non-automated post-processing using sophisticated algorithms and software developments. We assessed the value of an empirical method using alginate moulding for PET-MR co-registration in a tumor rat model.

Methods

Male WAG/RijHsd rats bearing grafted syngenic rhabdomyosarcoma were examined under general anesthesia by MRI using a clinical whole-body 3-T system equipped with a sensitivity-encoding four-channel wrist coil and by a small animal PET system using labelled [18F]-fluorocholine as tracer. An alginate mould including a system of external fiducials was manufactured for each animal, allowing strict immobilization and similar positioning for both modalities. Fourteen rats (27 tumors) had only one MR/PET imaging session. Five rats (9 tumors) had a similar MR/PET session before and 3 days after external radiation therapy (13 Gy in one fraction) using the same mould. Co-registration was performed using the Pmod release 2.75 software (PMOD Technologies, Ltd., Adliswil, Switzerland) with mutual information algorithm.

Results

The manufacture of the alginate moulds was easy and innocuous. Imaging sessions were well tolerated. PET-MR co-registration based on mutual information was perfect at visual examination, which was confirmed by the superimposition of external fiducials on fused images. Reuse of the same mould for the post-therapeutic session was feasible 3 days after the pre-therapeutic one in spite of tumor growth.

Conclusion

The empirical method using alginate moulding with external fiducials for PET-MR co-registration in a rodent tumor model was feasible and accurate.

Introduction

Magnetic resonance imaging (MRI) and positron emission tomography (PET) have nowadays become synergistic tools in clinical oncology by giving complementary anatomical and metabolic information which can be superimposed in a so-called co-registration process [1], [2], [3]. The detection and quantification of abnormal metabolic activity within tumoral subareas and/or normal-appearing tissues allow better delineation of the true tumoral volume [4]. Intense research has led to major improvements in both technologies. The accuracy of the co-registration process between anatomical and metabolic images is crucial and requires an easy-to-perform and robust method. The goal has now been achieved for clinical activity by hybrid PET–CT systems enabling fully automated fusion of CT and PET images [5], [6], [7], [8]. However, combined MR–PET systems are more difficult to build because of the low compatibility between positron-detecting crystals and the magnetic environment; prototypic clinical head PET–MR systems are nowadays running [9], [10]. In the experimental field, significant advancements have been made regarding the dedication of MR and PET micro-systems to small animal experimentation. But co-registration of the two modalities still relies on non-automated post-processing using sophisticated algorithms and software developments [11], [12], [13]. We hereby describe an empirical easy-to-use and cheap technique for PET-MR co-registration in a cancer animal model.

Section snippets

Ethical issues

The experiment was carried out in compliance with our institutional ethics committee for laboratory animal experimentation (CE Accred. No. UCL/MD/2007/021).

Animal model

Male WAG/RijHsd rats (Harlan Nederland, Horst, The Netherlands) weighting 250 to 300 g were housed three per cage and received food and water ad libitum. Fragments from a syngenic rhabdomyosarcoma of approximately 1 mm3 were grafted subcutaneously in both thighs under general anesthesia using intraperitoneal injection of ketamine and

Monophasic cohort

Visual assessment was performed to determine whether the contours of the tumors perfectly superimposed on the PET-MR fused images and whether exact superimposition of the external fiducials had been obtained by running “freely” the algorithm without any manual correction.

Biphasic cohort

Similarly to the previous group, visual assessment was performed to verify the superimposition of the tumor contours and of the external fiducials for post-treatment MR and PET image sets. In addition, pre- and post-therapeutic

Feasibility and safety of anesthesia and alginate mould preparation

The technique of general anesthesia was safe and well tolerated by the animals: none of them died from anesthesia or during the period of 3 days in-between the pre- and the post-therapeutic imaging sessions. The alginate mould did not cause respiratory disturbances because the thorax of the animals was located outside it. The procedure to immobilize the rats and to manufacture their own alginate mould was judged easy-to-perform, fast, safe and reproducible by the two operators who manufactured

Discussion

The need for co-registration of PET (metabolic) and CT/MR (anatomic) data has become axiomatic in clinical oncology. Many technological improvements have been done to allow easy-to-use and robust co-registration in the clinical day-to-day practice. But in the experimental field using animal models, PET-MR co-registration still remains a manual process and relies on crafty handwork and sophisticated algorithms and software options aimed at superimposing images acquired in different positions [11]

Acknowledgments

This work was supported by a grant from the Fond National de la Recherche Scientifique (FNRS-FWO) of Belgium. We are indebted to Willy Landuyt, PhD (Experimental Radiobiology/LEO, Katholiek Universiteit Leuven, Belgium), who provided us with the tumor model and brought us his invaluable experience in tumor model handling. We thank Philippe Nuytten, MD, Fellow in Stomatology at our institution, who taught us the use of the alginate.

References (16)

There are more references available in the full text version of this article.

Cited by (15)

  • Enhanced choline metabolism in a rodent rhabdomyosarcoma model: Correlation between RT-PCR and translational 3 T H-MRS

    2012, Magnetic Resonance Imaging
    Citation Excerpt :

    Two data sets were obtained in this way from similarly located tumour sub-areas at the macroscopic level, thereby enabling correlations between methods. The methods for manufacturing the alginate mould and performing H-MRS on full-body anesthetized animals have been described in preliminary works [23]. Although the tumours were carefully excised and sliced after euthanasia, 50 out of the 89 voxels obtained in H-MRS (56%) were excluded from analysis because of necrosis precluding biochemical analyses.

  • Subject specific hand and forearm musculoskeletal 3D geometries using high-resolution MR images

    2019, Computer Methods in Biomechanics and Biomedical Engineering: Imaging and Visualization
  • Linking imaging to omics utilizing image-guided tissue extraction

    2018, Proceedings of the National Academy of Sciences of the United States of America
View all citing articles on Scopus
View full text