MRI Device Compliance Martin Vogel, PhD Kimberley Poling Application Engineering Team Eastern USA

Transcription

1 MRI Device Compliance Martin Vogel, PhD Kimberley Poling Application Engineering Team Eastern USA 1 ANSYS, Inc. September 26,

2 Overview High simulation efficiency for MRI Method that enables a non-ee to analyze the thermal effects of an MRI scan on an implant 2 ANSYS, Inc. September 26,

3 Main tools HFSS 3D Full-wave Electromagnetic simulator Designer Circuit simulator that interacts with HFSS ANSYS Professional NLT Thermal simulator that uses electromagnetic losses from HFSS as heat loads 3 ANSYS, Inc. September 26,

4 1. Simulation efficiency Minimize RAM and simulation time 4 ANSYS, Inc. September 26,

5 Generic birdcage coil Ring with 24 ports, phases deg Ring with 24 ports, phases deg Around the coil is a cylindrical shield with open ends. 5 ANSYS, Inc. September 26,

6 Saving computer resources A birdcage coil can be created with many, e.g. 48, sources that impose the required rotating field. A realistic birdcage coil has few sources and many capacitors. Such a coil can be simulated with less resources. With a couple of ANSYS tools, we have created such realistic coils, one for 1.5 T and one for 3 T. 6 ANSYS, Inc. September 26,

7 Realistic coil feed network Feed points simplified and optimized with ports and capacitors Bot_1:1 Top_1:1 C4 Bot_2:1 Top_2:1 C44 PNUM=1 RZ=50ohm IZ=0ohm 0 C6 Cap C8 Cap C20 Cap C5 C7 Cap Cap Bot_3:1 Bot_4:1 Bot_5:1 Bot_6:1 Bot_7:1 Bot_8:1 Top_3:1 Top_4:1 Top_5:1 Top_6:1 Top_7:1 Top_8:1 C43 Cap C41 Cap Cap C42 Cap C40 Cap C57 0 PNUM=3 RZ=50ohm IZ=0ohm PNUM=2 RZ=50ohm IZ=0ohm C18 Cap C16 Cap C26 Cap C24 Cap Cap C38 C36 Cap C34 Cap C19 Bot_9:1 Bot_10:1 C17 Cap Bot_11:1 Bot_12:1 CapC237 Bot_13:1 Bot_14:1 C25 Cap Bot_15:1 Bot_16:1 C23 Cap Bot_17:1 C240 Bot_18:1 C39 Cap Bot_19:1 Cap Bot_20:1 C37 Cap Bot_21:1 Bot_22:1 C35 Cap Bot_23:1 Bot_24:1 Cap Top_9:1 C56 Top_10:1 Top_11:1 Cap C54 Top_12:1 Top_13:1 Cap C64 Top_14:1 Cap Top_15:1 C62 Top_16:1 Top_17:1 Cap C60 Top_18:1 Top_19:1 Cap C70 Top_20:1 Top_21:1 Cap C68 Top_22:1 Top_23:1 Cap C66 Top_24:1 Cap Cap C55 Cap C53 Cap C63 Cap C61 Cap C59 Cap C69 Cap C67 Cap C PNUM=4 RZ=50ohm IZ=0ohm Cap Cap Feed points are driven with phases 0, 90, 180 and 270 degrees. 7 ANSYS, Inc. September 26,

8 S 11 coils with capacitors MHz 63.8 MHz 8 ANSYS, Inc. September 26,

9 Fine-tune capacitors for MRI image quality B+/B- ratio is at least T pf ANSYS, Inc. September 26, 3 T 26.5 pf

10 Rotating H with capacitors Feed points Capacitors have been finetuned in HFSS for best rotating H after optimization in Designer. 10 ANSYS, Inc. September 26,

11 RAM 12.5 GB Apply this design technique to birdcage coil with human-body model Elapsed time 58 minutes 11 ANSYS, Inc. September 26,

12 Rotating B for body in birdcage with 48 sources Compare with results of other simulations. 12 ANSYS, Inc. September 26,

13 RAM 5.5 GB Rotating B for homogeneous body in birdcage with 4 sources and 44 capacitors is quite similar. Elapsed time 22 minutes 13 ANSYS, Inc. September 26,

14 Is it possible to simulate just the blue volume and IMPOSE fields from a separate coil simulation? YES, you can do that! 14 ANSYS, Inc. September 26,

15 With imposed fields, the rotating B again looks very similar! RAM 1.6 GB Elapsed time 7.3 minutes 15 ANSYS, Inc. September 26,

16 Now that we have such an efficient method, use it with a complicated heterogeneous human-body model and a pacemaker. RAM 4.3 GB Elapsed time 20 minutes 17 ANSYS, Inc. September 26,

17 Local SAR in cut plane (3T coil) 18 ANSYS, Inc. September 26,

18 R13 Schematic 19 ANSYS, Inc. September 26,

19 Thermal result Maximum occurs on neck, not at implant 20 ANSYS, Inc. September 26,

20 Contours of C and higher Tissue around pacemaker 21 ANSYS, Inc. September 26,

21 Pacemaker, lead and heart 22 ANSYS, Inc. September 26,

22 2. Thermal effects of MRI scan on implant Minimize the implant designer s exposure to electromagnetics 23 ANSYS, Inc. September 26,

23 Generic birdcage coil Around the coil is a cylindrical shield with open ends. 24 ANSYS, Inc. September 26,

24 Where should implants be placed? Implants should have worst-case location and orientation. Next few slides illustrate how this is achieved. 25 ANSYS, Inc. September 26,

25 Phantom placement B is almost uniform within coil (left). However, E is not (right). Make sure implants be located in a region with relative high E-field. B E 26 ANSYS, Inc. September 26,

26 Worst-case SAR: implant near phantom wall; does not stick out of RF coil 27 ANSYS, Inc. September 26,

27 Worst case: implant parallel with electric-field vectors 28 ANSYS, Inc. September 26,

28 Workflow for implant testing Have reusable coil+phantom model in HFSS Have phantom+parameterized implant(s) in separate geometry system This separate geometry will be exposed to the fields from the coils. Thermal simulator will determine temperatures of phantom+implant(s) due to RF losses. Next slides show this in more detail. 29 ANSYS, Inc. September 26,

29 Workflow in R14 Schematic No need to open HFSS. CAD 30 ANSYS, Inc. September 26,

30 Create in other CAD tool import with parameters into prepared HFSS design 31 ANSYS, Inc. September 26,

31 Transfer geometry to thermal simulator No more need for File/Export and File/Import Note the parameters from The CAD tool are available! 32 ANSYS, Inc. September 26,

32 Library of thermal material properties is available 33 ANSYS, Inc. September 26,

33 Have prepared a subset with the same material names as in HFSS. 34 ANSYS, Inc. September 26,

34 Objects come in with correct material assignment! 35 ANSYS, Inc. September 26,

35 Drag and Drop HFSS Solution to Thermal Setup Drop and Drag Resulting Schematic with extra Link 36 ANSYS, Inc. September 26,

36 Assign Losses from HFSS to Objects in Thermal 37 ANSYS, Inc. September 26,

37 Effect implants on local SAR 38 ANSYS, Inc. September 26,

38 Temperatures on implants after 900 s 39 ANSYS, Inc. September 26,

39 Add final maximum temperature as an output parameter 40 ANSYS, Inc. September 26,

40 Input parameter: Implant length Output parameter: Max. temperature 41 ANSYS, Inc. September 26,

41 Attach DX to investigate design variations efficiently 42 ANSYS, Inc. September 26, Vary length and inspect final temperatures

42 Maximum temperature as a function of implant length (at reduced power, not to be compared with previous result) Length implant (mm) 43 ANSYS, Inc. September 26,

43 Observations Temperatures tend to decrease when implant is buried deeper. This makes sense, as electromagnetic fields decay with depth. For small implants, temperatures increase with implant length. This makes sense, because a longer implant can function as a more effective receiving antenna (as long as implant λ/2 at this frequency and in this environment). For large implants, field inhomogeneity makes predictions harder to make. Results are sensitive to the gel s electrical and thermal material properties. 44 ANSYS, Inc. September 26,

44 Summary We have demonstrated (1) A way to achieve very efficient simulations; (2) A simulation flow that enables non-em experts to determine MRI-induced heating for different implant designs. 45 ANSYS, Inc. September 26,