The magnetic susceptibility of a vascular access port-a-cath and of surgical ... Diego, USA) and the Ferritometer® at Children's Hospital Oakland (CHO: model.
Interaction of Artificial Metallic Objects with Biosusceptometric Measurements Engelhardt, R.1, Fung, E.B.2, Kelly, P.3, Biehl, T.R.4, Pakbaz, Z.2, Nielsen, P.1, Harmatz, P.2, Fischer, R.1,2. 1
University Hospital Hamburg-Eppendorf, Germany; 2Children’s Hospital & Research Center at Oakland; 3 Children’s Hospital, Cincinatti; 4Virginia Mason Hospital, Seattle, USA.
ABSTRACT In human subjects, metallic objects cause distortions of the magnetic fields used by magnetic resonance imaging (0.5 – 3.0 T) or by SQUID biomagnetic liver susceptometry (0.1 – 30 mT) and may lead to artifacts in the measurement of the relaxation rate or the magnetic susceptibility. In biosusceptometry, the measured signal will depend not only from the magnetic susceptibility of the object, but also from its distance to the sensor assembly, and in case of ferromagnetic objects, from the direction of its remanent field. The magnetic susceptibility of a vascular access port-a-cath and of surgical clips have been measured by a SQUID biosusceptometer. Additionally, the impact from port-a-caths and dental braces on liver iron concentration (LIC) measurements was measured in vivo with respect to their radial distance from the gradiometer center axis. For the port-a-cath, a mean magnetic volume susceptibility of (83.5 ± 0.3)·10-6 SI-units was found, which may be compared with the magnetic susceptibility of titanium at room temperature of (180 ± 2)·10-6 SI and of ferritin/hemosiderin iron as found in the human liver of (1550 ± 50)·10-6 SI. At a radial distance of 5 cm from the gradiometer center axis, the voltage amplitude is similar to the signal generated by a normal liver. Modern surgical clips have nearly no impact on LIC measurements. However, dental braces although further away from the center axis, often superimpose the signal even from an iron overloaded liver. Depending on the Ni-content, these objects reveal ferromagnetic properties, making measurements unstable and analysis difficult. KEY WORDS Implant, port-a-cath, dental brace, staple, surgical clip, magnetic susceptibility, biosusceptometry. INTRODUCTION In human subjects, the interaction of metallic implants such as surgical clips (staples), orthopedic implants, stents, pace makers, dental braces, and vascular access ports with magnetic fields as applied in magnetic resonance imagers (MRI) or in biomagnetic measurements with SQUID biosusceptometers always create technical difficulty. Of greatest concern are ferromagnetic objects because of their excessive magnetic interactions (magnetic volume susceptibility 106 SI-units) and their remanent fields. In the strong magnetic fields of MRI systems (0.5 – 3.0 T) these implants may cause either a hazardous situation due to the generation of torque and deflection forces on these objects [Shellock, 1988] or may at least cause magnetic field distortions leading to image artifacts around these objects. Often the magnetic properties of these implants are not known, especially, older implants and antimagnetic stainless steel implants may contain tiny ferromagnetic contaminants leading to a contraindication of using MRI [Nowak, 1997]. Although, in current practice, most implants are manufactured from titanium and plastic materials, which do not show any deflection nor torque in a magnetic field of 3.0 T (Shellock, 2002), these implants may cause field distortions in the weak magnetic field of a SQUID biosusceptometer (0.1 – 30 mT) leading to artifacts in the measurement of the magnetic susceptibility within the human body. In patients with iron overload from blood transfusions (e.g., thalassemia) often a vascular access port (port-a-cath) is applied for the continuous infusion of the iron chelator deferoxamine or surgical clips (staples) have been used for splenectomy or cholecystotomy. The position of these implants may vary
relative to the field of view (FOV) of a biosuceptometer for the assessment liver iron. Therefore, the impact on the measured magnetic signal will depend not only from the magnetic susceptibility of the implant, but also from its relative distance to the FOV and in the case of ferromagnetic dental braces, also from the orientation of the remanent magnetic field.
METHODS An apheresis port-a-cath and surgical clips (staples) have been measured by a SQUID biosusceptometer (Ferritometer®). The port-a-cath (Triumph I, Horizon Medical Products Inc., USA) manufactured from titanium and polysulfone has a volume of 2.9 cm3, which can be approximated by a hemisphere with a radius of 1.4 cm. Surgical clips were made of titanium with a length of 10 mm and a volume of 10 mm3 (TIM-20, Ethicon Endo-Surgery Inc., USA). The two LTC-SQUID biosusceptometers, Hamburg Biosusceptometer (UKE: Biomagnetic Technologies Inc., San Diego, USA) and the Ferritometer® at Children’s Hospital Oakland (CHO: model 5700, Tristan Technologies Inc., San Diego, USA) have been described elsewhere [Paulson, 1991] [Starr, 2000]. The sensor assembly, situated in a vacuum space inside the dewar tail of the biosusceptometers, comprise two sets of superconducting magnetic field coils (1st order gradiometer) and 2 (UKE) or 3 (CHO) different symmetrically coaxial detector coils (2nd order gradiometers A, C and planar B). The underlying principles of liver iron measurements with SQUID biosusceptometers have been published over the last 20 years [Farrell, 1983] [Fischer, 1998]. Usually, a measurement is performed by moving an object or patient away from the sensor assembly either versus air or water as reference medium. However, also horizontal scanning versus air can be applied. In a first setup, the magnetic susceptibility of the port-a-cath and the surgical clip was measured on an empty wine bottle, which produced nearly no magnetic signature when measured versus air, below the sensor assembly of the Ferritometer® at CHO. Measurements were performed by vertical movement and horizontal scanning versus air reference. In a second setup, biomagnetic liver susceptometry (BLS) with and without a port-a-cath, with and without dental braces was performed. At CHO, a port-a-cath was attached to the skin of an iron overloaded patient at a standard medio-clavicular position on the right upper thorax, which is 10 – 20 cm away from the position of liver iron measurements. At UKE, a patient (β-thalassemia major) was measured before and after dismounting the dental braces. RESULTS AND DISCUSSION In a first test, the port-a-cath was scanned in the horizontal plane (x-direction) at two distances relative to the lower pick-up coils of the sensor assembly (z = 2.4 and 3.4 cm). Thereby, the setup was moved manually by 20 cm within 10 seconds, which corresponds with the inner diameter of the bellows of the water coupling membrane. The SQUID voltage patterns for the most critical detector with the largest gradiometer pick-up coil diameter are shown in Figure 1. Close to the dewar tail, the sensitivity volume of the sensor assembly is more restricted by the radius of the field coil of 1.7 cm, while further away, the radius of the gradiometer A of 2.5 cm determines the measurable signal from the port-a-cath in a range of x = ± 5 cm. This coincides with the outer diameter of the dewar tail. The signal patterns of the other two detector channels were even more restricted by the smaller size of their gradiometer coil radii. The magnetic susceptibility was measured directly by lowering the port-a-cath in the central position (x = 0). The transformation of the vertical distance z into magnetic flux integrals for the 3 SQUID detector channels was performed by a hemisphere with a volume of the port. A mean magnetic volume susceptibility of (83.5 ± 0.3) ·10-6 SI-units was achieved, which may be compared with the susceptibility of titanium at room temperature of (180 ± 2)·10-6 SI. The size of the susceptibility clearly shows the
SQUID A Voltage
1. 5
1. 0
0.5
0.0 - 10
-8
-6
-4
-2
0
2
4
6
8
10
x-scan [cm]
Figure 1. Horizontal scans of a vascular access porta-cath at sensor distance z = 2.4 cm (thin solid line), at z = 3.1 cm (thick solid line) and of 3 surgical clips at z = 2.6 cm (dashed line).
2.5
SQUID-Voltage [V]
2.0
Patient with dental braces
1.5 Fluxintegral calculation
1.0 Patient without dental braces
0.5 Fluxintegral calculation
0.0 0.1
0.2
0.3
0.4
Distance -1 [cm -1]
Figure 2. Biomagnetic liver susceptometry (SQUID channel A) in a patient with (circles) and without dental braces (triangles). With braces, the fluxintegrals do not fit the measured data, while, without braces the fluxintegral fit is perfect.
paramagnetic property of the investigated porta-cath model and may be compared with the susceptibility of the paramagnetic ferritin/hemosiderin iron in the human liver of (1550 ± 50) ·10-6 SI. The in vivo experiment with the port-a-cath mounted at the standard medio-clavicular position of the right upper thorax on a patient (β-thalassemia major, 31 y, mean LIC = 1100 ± 200 µg/gliver) revealed no significant magnetic contribution. With and without port-a-cath, a LIC of 1063 ± 166 and 1171 ± 128 µg/gliver was determined, which is in the range of variation for this patient. In a recent measurement, a patient with an implanted port-a-cath at an even closer distance of 6-7 cm was measured without any significant additional magnetic contribution. This is in agreement with the findings from the horizontal scanning. The influence of dental braces on a SQUIDmeasurement in a patient is shown in Figure 2. The measurement with braces show a strong distortion on the measured SQUID-voltage (circles). The measured data disagree with the calculated fluxintegrals (line through the data) indicated by a huge chi-square (~1200). The data without braces 5 weeks later (triangles) show no magnetic distortion and the chi-square (~5) indicates a good agreement between the measured data and the fluxintegral calculation. In the final analysis of BLS-measurements, weak magnetic distortions can be considered by a one parameter error-function. In the shown measurement with braces, the huge error term would cause the final LIC to range between 1450 and 2700 µg/gliver. In the subsequent BLS-measurement without braces, the final analysis of the LIC results in 1120 ± 120 µg/gliver. Further magnetic fieldmeasurements by a Hall probe revealed a remanent magnetic field on the braces clearly above 2 mT.
However, patients with dental braces of smaller ferromagnetic signature (micro Tesla) can be measured, although in practice, this may become difficult to judge in advance.
ACKNOWLEDGEMENTS We would like to thank Hannes Nowak for helpful discussion. This work was supported in part by Cooley’s Anemia Foundation. REFERENCES Farrell DE, Tripp JH, Brittenham GM, Danish EH, Harris JW, Tracht AE. A clinical system for accurate assessment of tissue iron concentration. In: Romani GL, Williamson SJ, editors. Proceedings of the 4th International Workshop on Biomagnetism. Il Nuovo Cimento 1983; 2D:582-93. Fischer R. Liver iron susceptometry. In: Andrä W, Nowak H, editors. Magnetism in Medicine - A Handbook. Berlin: Wiley-VCH; 1998. p. 286-301. Nowak H, Michaelsen S, Giessler F, Huonker R, Haueisen J, Kaiser WA. Patient screening for MRI with SQUID biomagnetometers. Biomed Tech 1997;42 Suppl:233-4. Paulson DN, Fagaly RL, Toussaint RM, Fischer R. Biomagnetic susceptometer with SQUID instrumentation. IEEE Trans Magnetics 1991;MAG-27: 3249-52. Shellock FG. MR imaging of metallic implants and materials: a compilation of the literature. Am J Roentgenol 1988;151:811-4. Shellock FG. Magnetic resonance safety update 2002: implants and devices. J Magn Reson Imaging 2002;16:485-96. Starr TN, Fischer R, Ewing T, Longo F, Engelhardt R, Trevisiol E, Fagaly RL, Paulson DN, Piga A. A new generation SQUID biosusceptometer. In: Nenonen J, Ilmoniemi RJ, Katila T, editors. Biomag 2000 – Proceedings of the 12th International Conference on Biomagnetism. Helsinki: University of Technology; 2001. p. 986-9.
