Impact of silicone and metal port-a-cath implants on superficial hyperthermia treatment quality

Abstract

Purpose: A port-a-cath is a device implanted under the skin for continuous drug administration. It is composed of a catheter and a silicone or metal reservoir. A simulation study was done to assess the impact of a port-a-cath implant on the quality of superficial hyperthermia treatments applied using the Lucite cone applicator (LCA). Methods: Specific absorption rate (SAR) and temperature distributions were predicted using SEMCAD-X (version 14.8). We simulated 72 arrangements: two LCA-implant set-ups (central port-a-cath or at an edge below the LCA footprint), six translations of the LCA per set-up, two LCA orientations (Parallel or perpendicular electric field direction) per set-up, two implant materials (silicon or metal) and a control without port-a-cath. Treatment quality was quantified by the average 1 g SAR coverage (CV_25%), i.e. volume within the 25% iso-SAR surface, and the volume within the 40 °C iso-temperature surface (CV_40 °C). Results: CV_25% reduced with a silicon port-a-cath located below the LCA footprint. In the worst scenario, only 64% of the CV_25% of the control set-up was achieved. For a metal port-a-cath below the LCA aperture, dramatic reductions of CV_25% were predicted: worst scenario down to 12.1% of the control CV_25%. For the CV_40 °C the worst case values were 74.5% and 6.5%, for silicon and metal implants, respectively. Conclusions: A silicone port-a-cath below the LCA had a smaller effect on treatment quality than a metal implant. Based on this study we recommend verifying heating quality by 3D patient-specific treatment planning when a port-a-cath is located below the footprint of the applicator.

• Implants
• port-a-cath
• superficial hyperthermia
• thermal modelling
• treatment planning

Introduction

In hyperthermia, some implants, e.g. a pacemaker, metallic implants larger than surgical clips are considered as contraindications for treatment due to the risk of high and very localized hot spots [1]. In general, three types of implants can be differentiated: (1) metallic implants, (2) active medical devices and (3) passive non-metallic implants. Metallic implants can interact strongly with the electromagnetic (EM) field generated by the applicators, whereby the degree of interaction strongly depends on the direction of the electric (E)-field, as well as of the electrical dimensions of the device [2]. Lee et al. [3] concluded that metallic implants with dimensions much shorter than the wavelength can cause significant changes in the power deposition. Active medical devices pose a dual risk: malfunction due to its electromagnetic compatibility (pacemakers), and hot-spots by interaction with the EM-field [4,5]. For passive non-metallic implants such as silicone breast implants and port-a-cath systems, less is known about how they affect the specific absorption rate (SAR) distribution and their associated risk for inducing localised low or high local SAR values. To our knowledge, there is no publication addressing the influence of port-a-cath systems on quality of hyperthermia treatment.

In this study, we investigated whether a port-a-cath can be considered a general exclusion factor for superficial hyperthermia treatment with the Lucite cone applicator (LCA). The study was initiated by the referral of a patient with a radiotherapy target volume deeper than 4 cm, i.e. our standard inclusion parameter, and a metal/silicone port-a-cath implant in close proximity to the hyperthermia treatment field. Superficial hyperthermia treatment planning initially showed a good target coverage at 25% of the maximum SAR level, i.e. CV_25% = 75%. In addition, no high SAR was predicted near the port-a-cath, which was confirmed by the absence of pain complaints near the implant during the four treatments. Since port-a-caths are generally treated as an exclusion criterion we investigated with a simulation study the impact of the port-a-cath implant on hyperthermia treatment quality for a number of applicator implant scenarios.

Materials and methods

A port-a-cath is a small medical device that is implanted beneath the skin; its typical dimensions are 30.5 mm (base) × 14.7 mm (height) with a weight of 10 g and circular shape. It has two principal functions: (1) for drug administration for chemotherapy, and (2) to extract blood samples. These devices are composed of a catheter and a reservoir made of silicone or metal (stainless steel or titanium). The drug is injected through the reservoir to the patient. Figure 1 shows an example of a silicone and metal port-a-cath.

Figure 1. Examples of a metallic and silicone port-a-cath which can be implanted in patients treated with hyperthermia [6].

Patient model

The 3D patient model was reconstructed with computed tomography (CT) data set composed of 158 slices with a slice thickness of 2.5 mm. In the CT, the port-a-cath is visible as an implant at the left side of the patient’s thorax. Muscle, fat, lung, bone, air and the port-a-cath were taken into account to generate the 3D patient model. The control model was generated using the same CT data set but the port-a-cath was segmented as fatty tissue. Figure 2 shows a CT slice in which the port-a-cath is visible as well as the segmentation of port-a-cath, bone, lung, muscle and fatty tissues.

Figure 2. Patient model generation. (A) CT slice in which the port-a-cath is indicated by the arrow, (B) segmented CT slice in which the port-a-cath (grey), lung (blue), bone (yellow), muscle (brown) and fatty tissue (orange) are segmented. For the control model, the port-a-cath is segmented as fatty tissue.

LCA and port-a-cath locations

Figure 3 shows the schemes to describe the locations of the LCA with respect to the port-a-cath for each case analysed. Two LCA-implant set-ups were analysed: (A) a port-a-cath at the edge below the LCA footprint or (B) a port-a-cath at the centre below the LCA footprint. The port-a-cath was of course fixed at the same place in all scenarios, while the LCA was moved six times per LCA-implant set-up (locations 1–6: 0, 45, 70, 80, 90, and 100 mm from the LCA central plane) to cover a total distance of 100 mm. In locations 1 and 2 the port-a-cath was located inside the treatment field of the LCA (see Figure 3). For locations 3 -5, the port-a-cath was located outside the treatment field of the LCA but within the region covered by the water bolus (WB). Finally, in location 6, the port-a-cath was located out of the region covered by the WB (see Figure 3). These last models were implemented in order to see how far the LCA should be from the port-a-cath in order to have no impact on the SAR and the temperature distributions. Two LCA orientations (Parallel or perpendicular E-field direction) per LCA implant set-up, two implant materials (silicon or metal) and the control (no port-a-cath) were included in the analysis.

Figure 3. Schematic description of the LCA locations with respect to the port-a-cath for each case analysed. Blue line represents the water bolus, black lines represent the footprint of the LCA, red dashed lines represent the Lucite in the applicator, and the small circle represents the port-a-cath. All the parameters included in the analysis are described: (1) The LCA- port-a-cath clinical set-ups (A) at the edge, and (B) centrally below the LCA footprint, (2) the LCA translations (only four different locations of the LCA are depicted as examples), (3) the LCA orientations (parallel and perpendicular) and (4) the different port-a-cath configurations: no port-a-cath, silicone, and metal.

Electromagnetic models

SAR distributions were computed using the finite difference time domain (FDTD) solver of SEMCAD-X (version 14.8.5, Schmid & Partner Engineering, Zurich, Switzerland). The EM-field was generated by a voltage edge source placed between the capacitive feeding pin and the inner wall of the LCA waveguide. Details of the FDTD model including the LCA validation are described extensively in the literature [7–10]. Each simulation set-up includes the LCA, the water bolus (180 × 180 × 20 mm³), and the patient model with and without the port-a-cath. The port-a-cath was assumed to be made of silicone or titanium, which was simulated as a lossy metal using the surface impedance boundary condition (SIBC) in SEMCAD-X. Table 1 shows the dielectric and thermal properties of tissues used in EM and thermal simulations [11,12]. Where εr is the relative permittivity; σ (S/m) the electrical conductivity, ρ (kg/m³) the tissue density, c (J/kg K) the specific heat capacity, k (W/m K) the thermal conductivity, ω (mL/min kg) is the blood perfusion and Q (W/kg) is the heat generation rate.

Table 1. Dielectric and thermal tissues properties used for the models at 433 MHz [11,12].

Tissue	εr	σ (S/m)	ρ (kg/m^3)	c (J/kg K)	k (W/m K)	Q Heat generation rate (W/kg)	Temperature-dependent ω (37–48 °C) (mL/min kg)
Fat	11.6	0.08	911	2348	0.21	0.50	ωf_ITIS = 32.7–65.4
Muscle	56.9	0.80	1090	3421	0.49	0.96	ωm_ITIS = 39.0–347.6
Bone	13.1	0.09	1595	1313	0.32	0.15	–
Lung	23.6	0.38	750	3886	0.38	6.20	–
Silicone port-a-cath	12.0	0.05	1000	850	0.20	–	–
Metal port-a-cath (titanium: SIBC)	–	2.38e6	4507	520	0.5	–	–

SIBC, surface impedance boundary condition; ω_{f_ITIS,} fat perfusion from ITIS database; ω_{m_ITIS}, muscle perfusion from ITIS database.

To analyse the effect of the port-a-cath on the SAR distributions, three different configurations were simulated: (1) patient without port-a-cath (control model), (2) patient with a silicone port-a-cath and (3) patient with a metal port-a-cath. Usually at the Erasmus MC Cancer Institute a hyperthermia treatment is applied four times; each treatment lasts 1 h [1]. For a more uniform overall SAR coverage, the E-field direction of every LCA is rotated 90 degrees for each consequent treatment [1,13]. In this study two different E-field directions of the LCA were analysed: (1) E-field direction parallel to the body axis, and (2) E-field direction perpendicular to the body axis. In addition, six different locations of the LCA relative to the port-a-cath position in the patient were included in our study. Hence, a total of 2 × 6 × 2 × 3 = 72 scenarios (LCA-port-a-cath controls set up × LCA translations × LCA orientations × Port-a-cath configurations) have been simulated.

Thermal models

Temperature distributions for each scenario were simulated using the Pennes [14,15] bioheat transfer equation-based transient-state solver until steady state was reached. All thermal models were made using the thermal solver of SEMCAD-X. The bioheat transfer equation is described by Equation 1(1) , where c (J/kg K) represent specific heat capacity, ρ (kg/m³) density, k (W/m K) thermal conductivity, c_b (J/kg K) specific heat capacity of blood, ω (mL/min kg) blood perfusion rate, T_b (K) temperature of the blood, Q (W/m³) metabolic heat generation rate and SAR (W) the specific absorption rate. (1)

The radiofrequency power was increased, per scenario, to obtain a maximum temperature of 43 °C in normal tissue, i.e. the maximum temperature allowed during the treatment [16], in all scenarios (no port-a-cath, silicone port-a-cath, and metal port-a-cath). The SAR pattern generated by the LCA in the EM simulations was used as the input source for thermal simulations. A mixed boundary condition was used, i.e. the heat flux depended on the local temperature and its equilibrium to the temperature of the environment based on a convection coefficient (h). This boundary condition was appropriate to model effects like convection. The convection coefficient of the array-specific water bolus was from Van der Gaag et al. [17]. For the single LCA this resulted in modelling the water bolus as a boundary condition with a convection coefficient of h = 152 W (m²K)⁻¹ and an outer temperature of 37 °C. Tissue perfusion was implemented as a temperature-dependent variable, with values ranging between 39.0 to 347.6 mL/min kg for muscle and 32.7 to 65.4 mL/min kg for fat. In cases in which the port-a-cath was present, its thermal properties were taken into account (see Table 1).

Evaluation

SAR and thermal distributions

For SAR distributions, the total volume enclosed by the 25% iso-SAR surface (i.e. the volume having a SAR of ≥25% of the maximum SAR in the patient body expressed as: coverage volume 25% (CV_25%)) was calculated and temperature distributions were quantified by the total volume of tissue with a temperature over 40 °C (coverage volume over 40 °C, CV_40 °C).

Results

Figure 4, shows an example of the predicted SAR distributions for the LCA set-up A, location 2. Relative SAR distributions are presented as a colour wash overlay on the patient’s CT for all six configurations, i.e. none (control), a silicone port-a-cath, or a metal port-a-cath at the edge of the applicator footprint. The graphs visualise that the SAR distributions for the control and silicone port-a-cath configuration are quite similar for both E-field directions. A small difference occurred due to the exact location of the maximum SAR (different CT-slice). Only for the case with the metal port-a-cath and an E-field direction perpendicular to the body axis was a clearly different SAR distribution predicted, with maximum SAR hot spots located in the direct vicinity of the metal port-a-cath. Note also, the extension of the CV_25% in the deep muscle tissue and even in lung tissue at depths greater than 4 cm.

Figure 4. Predicted SAR distributions for set-up A, location 2 (schematically shown on the left) and E-field directions perpendicular and parallel to the body axis. Colour wash SAR distributions are for no port-a-cath (i.e. the control distribution with the port-a-cath replaced by fat tissue), the silicone port-a-cath, and the metal port-a-cath.

For a more quantitative evaluation, the SAR and temperature distributions are aggregated into the earlier defined CV_25% and CV_40 °C and reported in Figures 5 and 6 as absolute numbers and in Table 2 as relative numbers to clearly show the dependence of the set-up configuration on the effective treatment volume.

Figure 5. Volume covered by the 25% of the maximum SAR normalised to 1 W (CV_25%); both E-field directions are depicted for each scenario: (A) set-up A, (B) set-up B. The dashed black line at 400 mL represents the volume of 4 cm tissue under the LCA aperture (10 × 10 cm).

Figure 6. Volume of tissue at temperatures higher than 40 °C; both E-field directions are shown for each scenario. (A) CV_40°C for scenarios in which the port-a-cath was located at the edge of the LCA. (B) CV_40°C for scenarios in which the port-a-cath was located at the centre line of the LCA. The solid black line represents the volume of 4 cm tissue under the LCA aperture (10 × 10 cm).

Table 2. Relative volume enclosed by the 25% iso-SAR surface (CV_25%) and the relative volume of tissue over 40 °C (CV_40°C) normalised to the control model (no port-a-cath) for a port-a-cath at the edge or the centre of the LCA with six locations and two E-field LCA orientations. A CV_25% or CV_40°C of 75% or larger is considered as a threshold for an acceptable tissue coverage. The bold numbers represent the scenarios in which the port-a-cath is considered a contraindication for hyperthermia treatment (i.e. CV_25% or CV_40°C is <75%).

Location (%)		6		5		4		3		2		1
E-field direction		//	┴	//	┴	//	┴	//	┴	//	┴	//	┴
No port-a-cath
reference A&B	CV25%	100	100	100	100	100	100	100	100	100	100	100	100
	CV40°C	100	100	100	100	100	100	100	100	100	100	100	100
Silicone
A: LCA edge	CV25%	115	80	89	77	78	90	78	98	79	96	86	95
	CV40°C	105	100	106	100	104	98	105	103	109	96	86	98
B: LCA centre	CV25%	81	119	92	145	82	90	84	89	64	74	89	78
	CV40°C	111	97	88	104	116	110	115	95	76	75	98	96
Metal
A: LCA edge	CV25%	115	83	89	81	78	94	77	100	79	44	17	14
	CV40°C	104	100	106	102	102	103	99	80	64	16	8	15
B: LCA centre	CV25%	80	116	93	133	94	88	98	100	39	13	12	16
	CV40°C	111	94	87	95	116	109	118	71	42	7	10	14

LCA, Lucite cone applicator; //, parallel E-field direction; ┴, perpendicular E-field direction.

When evaluating and comparing the CV_25% and CV_40 °C for the various configurations we recall that the anatomy below the applicator is different for each LCA location and for each port-a-cath position (edge or centre), inevitably causing an anatomy-dependent variation in the absolute value of the CV_25% and CV_40 °C readings. This ‘anatomy’ effect is clearly visible in Figures 5 and 6.

Overall, the calculated data in Table 2 shows that the impact of a port-a-cath located just outside the footprint of the applicator on the resulting SAR or temperature distributions is less than 20% for CV_25% and less than 10% for CV_40 °C in 27 of the 32 configurations investigated. In general, CV_40 °C is less sensitive for the precise configuration (E-field direction, silicone or metal port-a-cath), with a maximum decrease of 13% and a maximum increase of 18%.

For a port-a-cath located below the footprint of the applicator the impact is, however, more substantial and more variable. For the silicon port-a-cath the largest reduction of CV_25% (−36%, E_//; −26%, E_⊥) and CV_40 °C (−24%, E_//; −25%, E_⊥) is found at location 2 with the port-a-cath at the centre position. For the other three configurations with the silicone port-a-cath under the applicator footprint, the differences are smaller: in four out of six locations the difference in CV_40 °C is less than −4%, in one +9% and in one −14%. The presence of a metal port-a-cath under the applicator footprint has much more dramatic effects. A metal port-a-cath at location 1 (centrally in the heating field) causes both CV_25% and CV_40 °C to drop to values lower than 20% of those without a port-a-cath. For the metal port-a-cath at location 2 the decrease in CV_25% and CV_40 °C is smaller but still substantial and resulting in clinically ineffective heating volumes.

Further, as can be seen in Figure 5(B), rotating the E-field direction has a substantial impact on the resulting effective SAR volume enclosed by the 25% iso-SAR surface. This effect is, however, seen independently of the presence of any type of port-a-cath. When comparing data shown in Figures 5 and 6 it becomes clear that due to thermal convection and conduction the effective treatment volume based on CV_40 °C is for all investigated configurations larger than that based on CV_25%.

Figure 7 shows the power levels obtained from the optimisation to achieve a maximum temperature of 43 °C in tissue. Power levels for all scenarios are shown. For the control and the silicone models, the power level was always in the same range (∼120 W). However, for worst case scenario metal implant the power level needed to achieve 43 °C was 39 W, indicating the existence of localised SAR hot spots, with an associated risk to overheat tissue locally around the port-a-cath while under-treating the rest of the treatment volume. From the best case scenario with a perpendicular E-field direction it is possible to observe that a power of 64 W (as in real hyperthermia treatments) should be sufficient to achieve 43 °C in healthy tissue.

Figure 7. Power levels needed to achieve a maximum temperature of 43 °C in tissue. The level power needed for both E-field directions is reported as well as both control set-ups. (A) and (B) describe the power level for scenarios in which the port-a-cath was at the edge of the LCA. (C) and (D) describe the power level for scenarios in which the port-a-cath was at the centre of the LCA.

Discussion

To our knowledge this is the first study in which the effect of a metal or silicone port-a-cath in the E-field of a superficial hyperthermia applicator (LCA) has been investigated. Until recently, superficial hyperthermia is mostly applied in combination with radiotherapy. Although only a small number of these patient have a port-a-cath in the re-irradiation field, a good knowledge on how and to what extent a metal or silicone port-a-cath disturbs the SAR distribution of an applicator is important for a safe and effective superficial hyperthermia treatment. Further, we notice there is a growing discussion [18] towards the combination of hyperthermia with chemotherapy [19–23]. In that patient group, inclusion of a port-a-cath in the hyperthermia treatment field might be a more regular occurrence, and hence stresses the need to know how close to the port-a-cath we can heat.

Other studies investigating the effect of implants (e.g. metallic, silicone) are more focused on the risk of overheating, i.e. maximum SAR values and maximum temperature increases. This is a good approach for assessing the risk of toxicity but not when the intention is to investigate the feasibility of treating the tumour despite the presence of a port-a-cath.

This study shows that the difference in SAR and thermal distributions between a model with and without a silicone port-a-cath is small, and it is possible to treat a patient with this kind of implant. It was demonstrated that the E-field parallel to the body axis had less impact over SAR and thermal distributions compared to an E-field perpendicular to the body axis. Heating very near to a metal port-a-cath is strongly discouraged as regions with very high localised SAR hot spots exist at the edges of a metal port-a-cath. However, if the hyperthermia treatment is applied under controlled conditions; for instance by using a slightly larger distance between the applicator and the port-a-cath, it might also be possible to treat patients with a metal port-a-cath without risks. From Figures 5 and 6 it can be seen that metal port-a-caths placed outside the hyperthermia treatment field, i.e. port-a-cath not under the LCA footprint, excludes the risk for local hot-spots. We found that for a port-a-cath localised outside the hyperthermia treatment field (after location 3) the volumes of tissue enclosed by the CV_25% iso-SAR surface and by the CV_40 °C iso-temperature surface were only marginally different between the control model and the models with a silicone or metallic port-a-cath.

With respect to the levels of power used to achieve 43 °C in healthy tissue, we predicted that for metal implants in the worst case the power level needed to achieve 43 °C was 39 W. In this case a power of 120 W (i.e. the power needed in the control and silicone models) will cause severe localised overheating around a metal port-a-cath. Thus these data also show that a metal port-a-cath under the applicator footprint is a contraindication for hyperthermia treatments. For the best case scenario with a perpendicular E-field direction, we predicted that 64 W is enough to achieve 43 °C in healthy tissue, i.e. a power level comparable to the average power of 56.4 W [16] as applied in actual clinical hyperthermia treatments.

Conclusion

This study indicates that the predicted volumes enclosed by the CV_25% or CV_40 °C iso-SAR or temperature surface for a situation without or with a silicone port-a-cath are comparable for port-a-cath locations near or within the hyperthermia treatment field. Only in cases in which the silicone port-a-cath is located at the centre of the LCA, can this implant be considered as a contraindication; in such cases treatment planning should be done to ensure the coverage of the tissue. However, a metal port-a-cath should be considered a strong contraindication when the port-a-cath is located within the hyperthermia treatment field (i.e. below the applicator footprint). Such configuration leads to high local SAR hot spots and a dramatic reduction in the effective treatment volume expressed as the volume enclosed by the CV_25% or CV_40 °C iso-surfaces. For the worst configuration this might result in an effective treatment volume of only 12.1% of the total volume for the control model. If the metal port-a-cath is localised out of the hyperthermia treatment field, the patient can be treated without significant degradation of treatment quality (an effective treatment volume of 77% of the total volume for the control model was achieved in the worst case).

Declaration of interest

The authors of Erasmus MC are supported by the Dutch Cancer Society (grant number EMCR2007-3837) and the Mexican National Council of Science and Technology. The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this paper.