- Email: [email protected]
Understanding the Relationship Between Real-Time Thermal Imaging and Thermal Damage Estimate During Magnetic Resonance–Guided Laser Interstitial Thermal Therapy
Journal Pre-proof Understanding the Relationship Between the Real-Time Thermal Imaging and the Thermal Damage Estimate during MR-guided Laser Interstitial Thermal Therapy Sean M. Munier, MD, Elizabeth E. Ginalis, BS, Akshay N. Desai, MS, Shabbar F. Danish, MD PII:
S1878-8750(19)32950-X
DOI:
https://doi.org/10.1016/j.wneu.2019.11.110
Reference:
WNEU 13779
To appear in:
World Neurosurgery
Received Date: 24 August 2019 Revised Date:
18 November 2019
Accepted Date: 19 November 2019
Please cite this article as: Munier SM, Ginalis EE, Desai AN, Danish SF, Understanding the Relationship Between the Real-Time Thermal Imaging and the Thermal Damage Estimate during MRguided Laser Interstitial Thermal Therapy, World Neurosurgery (2019), doi: https://doi.org/10.1016/ j.wneu.2019.11.110. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier Inc.
Understanding the Relationship Between the Real-Time Thermal Imaging and the Thermal Damage Estimate during MR-guided Laser Interstitial Thermal Therapy Sean M Munier, MD; Elizabeth E Ginalis, BS; Akshay N Desai, MS; Shabbar F Danish, MD Department of Neurosurgery, Rutgers University, Robert Wood Johnson Medical School, New Brunswick, New Jersey, United States Corresponding Author: Sean M. Munier, MD 10 Plum St. 5th Floor New Brunswick, NJ 08901 Phone: 201-446-2156 Email: [email protected] Keywords: Laser interstitial thermal therapy; LITT; Magnetic resonance thermometry; MRTI; Thermal ablation parameters
Conflicts of Interest: Dr. Danish is a consultant for Medtronic and has received educational honoraria.
Disclosure of Funding: None.
Running Title: MRT Imaging Boundary Quantification
1
ABSTRACT
2
Objectives: Magnetic Resonance-guided Laser Interstitial Thermal Therapy (MRgLITT) is a minimally
3
invasive procedure that utilizes intraoperative magnetic resonance thermometry (MRT) to generate a
4
thermal damage estimate (TDE) of the ablative area. This study compared areas produced by the MRT
5
heat map to the system-generated TDE produced by the Visualase software.
6 7
Methods: All ablations were performed using the Visualase MRI-Guided Laser Ablation System. MRT
8
heat-map and TDE were quantified using MATLAB version R2014a. The TDE was compared to the
9
summed area of the green, yellow, and red areas (HM63.9) and the summed area of the light blue, green,
10
yellow, and red areas (HM50.4) produced by the MRT heat map.
11 12
Results: 56 patients undergoing MRgLITT were examined. The mean TDE produced was 236 mm2
13
(SEM = 9.5). The mean HM63.9 was 231 mm2 (SEM = 8.7), and the mean HM50.4 was 370 mm2 (SEM =
14
12.8). There was no significant difference between the TDE and HM63.9 (p = 0.51) . There was a
15
significant difference between TDE and HM50.4 (p < 0.001) and between HM63.9 and HM50.4 (p < 0.001).
16 17
Conclusions: The system-generated TDE consistently remains contained within the boundaries of the
18
MRT heat map. At standard factory settings, the TDE and the area produced within the periphery of the
19
HM63.9 are similar in magnitude. The light blue portion of the MRT heat map may serve as an additional
20
means of predicting when critical structures may be at risk during laser ablation if exposed to further
21
thermal stress.
22 23 24 25 26 27 28 29 30 31 32 33 34
2 1
INTRODUCTION
2
Magnetic resonance-guided laser interstital thermal therapy (MRgLITT) is a minimally invasive
3
neurosurgical technique utilized in the treatment of many intracranial pathologies. One of the key features
4
of LITT is the ability to monitor ablation in real-time through magnetic resonance thermometry (MRT)
5
and the system-generated thermal damage estimate (TDE) (Figure 1). This estimate is produced using the
6
Arrhenius equation, which utilizes the temperature data recorded by the MRI machine to estimate the
7
tissue boundaries where cell death has occurred.1 Temperature data is produced under MRT, which is
8
recorded by the MRI machine through measurement of proton resonance frequency (PRF).2 Measurement
9
of PRF is based on the property of protons to resonate at a baseline frequency given a constant set of
10
parameters, including temperature. When temperature increases, as it does in LITT, hydrogen bonds
11
between water molecules are stretched, resulting in a change in PRF.2 The ability of MRI machines to
12
quantify these changes in PRF is what facilitates intraoperative MRT, and thus allows for a predictive
13
estimate of where tissue necrosis is occurring during LITT in real time.
14 15
To date, MRT produced by the MRI machine has primarily been used as a means of setting temperature
16
limits for the procedure that trigger the laser to automatically deactivate should the center or periphery of
17
the ablation exceed certain operator-determined temperatures. The manufacturer of the technology
18
advises a central temperature limit of 90°C to prevent production of steam and a peripheral temperature
19
limit of 45°C to prevent damage to healthy tissue.3 However, since no studies have analyzed
20
intraoperative MRT data relative to intraoperative TDE, MRT data currently offers little practical value to
21
the operator during the procedure. In this study, we aimed to understand the relationship of the real-time
22
thermal imaging to the thermal damage estimate in an effort to provide a practical and useful
23
interpretation of intraoperative MRT thermal data that may better inform intraoperative decision making.
24 25
METHODS
26
Patient Selection
27
We retrospectively reviewed and included 56 patients who underwent MRgLITT for various intracranial
28
procedures at our institution. Inclusion criteria for the study required single-laser catheter use, available
29
TDE and MRT data, and MRT temperature map settings with the following parameters: dark blue, >
30
37°C; light blue, > 50.4°C; green, > 63.9°C; yellow, > 77.3°C; red, > 90.8°C (with said temperatures
31
referring to tissue temperature within that color region). For the purpose of this study, only ablations with
32
a homogenous thermal distribution and without signal artifact were considered appropriate for analysis.
33
All subjects were from a single institution and operated on by a single surgeon, the senior author. All
3 1
patients underwent the procedure as part of the routine clinical care algorithm at our institution and were
2
enrolled in an institutional review board-approved protocol with their consent at our institution.
3 4
Laser Ablation Procedure
5
The ablation procedure has previously been described.4 Briefly, each patient undergoes preoperative CT
6
and MRI. The images are loaded onto a stereotactic planning software and fused to allow for planning of
7
the laser catheter trajectory.4 Bone implanted fiducials are fixed to the calvarium to allow for stereotactic
8
registration with the system software. After determination of appropriate laser trajectory, a small stab
9
incision is made and a burr hole is drilled over the entry point using a 3.2 mm twist drill. A bone anchor is
10
introduced into the entry site to secure the laser catheter. The laser catheter is introduced through the bone
11
anchor to a predetermined length into the ablation target. A 980-nm diffusing tip diode laser catheter is
12
utilized. The patient is transported to the MRI suite where the procedure is performed.
13 14
The procedure is performed under the guidance of MRT, which allows for intraoperative monitoring of
15
the ablation boundaries in real time. Using the Visualase software, the operator identifies safety margins
16
at the periphery and center of the intended target such that the peripheral healthy tissue temperature does
17
not surpass an operator-set temperature (typically 45°C) and the lesion center does not exceed 90°C to
18
avoid steam formation (Figure 2). Should either of these temperature limits be reached, the laser is
19
automatically deactivated. The system workstation provides the operator with information via two
20
dynamic images: one image displays the MRT data produced in the form of a heat map, where different
21
colors correspond to different tissue temperature thresholds, which are displayed in the upper right corner
22
of the MRT image (Figure 3). The right image in figure 1 displays the real-time software-generated
23
thermal damage estimate (TDE), which provides a calculated approximation of where tissue necrosis has
24
occurred. In all cases considered, these temperature thresholds on the MRT were set at factory-standard
25
settings seen in Figure 3. This estimated area is produced by the software every 5-7 seconds and appears
26
on the workstation as an orange area superimposed over the MRI images, allowing the operator to
27
determine when the target in question has been ablated and to terminate the ablation at their discretion.
28
The ablation area expands outward from the laser catheter in an ellipsoidal fashion, with expansion in any
29
single direction depending on tissue composition and target morphology.5 Following the procedure, the
30
patient undergoes postoperative MRI to verify that complete target destruction has occurred.
31 32
Imaging Analysis
33
Intraoperative MRT and TDE images were acquired from the Visualase console. Cases included were
34
those performed under the most recently updated factory-standard settings for the MRT heat map
4 1
temperature thresholds. The final image produced by the software prior to laser deactivation was analyzed
2
for each case included. Imaging areas were quantified using MATLAB version R2014a (Mathworks, Inc,
3
Natick, Massachusetts). The program individually quantifies the number of pixels of each color during the
4
course of the ablation. Two specific areas from the MRT heat map were analyzed. The first area included
5
all colors within the heat map (HM50.4) and is produced by the summed pixel count of the light blue,
6
green, yellow, and red areas. The second area included the summed pixel counts of the green, yellow, and
7
red areas (HM63.9). Each pixel is equivalent to 0.9 mm2. For TDE, the number of orange pixels were
8
individually quantified. Figure 4 delineates the specific areas being analyzed.
9 10
Data Analysis
11
All analyses were performed with GraphPad Prism version 7.0 (GraphPad Software, San Diego,
12
California). A student’s t-test analysis was used to detect differences between TDE, HM63.9, and HM50.4.
13 14
RESULTS
15
We evaluated a total of 56 cases. Operative indications included cavernoma, ependymoma, epilepsy,
16
glioblastoma multiforme, primary glioma, recurrent glioma, recurrent meningioma, and recurrent
17
metastasis. Operative indications are summarized in Table I.
18 19
The average power used across all cases was 11.35 W (SEM = 0.25). The average duration across all
20
ablations was 134 seconds (SEM = 11.70). The mean TDE produced was 236 mm2 (SEM = 9.5). The
21
mean HM63.9 was 231 mm2 (SEM = 8.7), and the mean HM50.4 was 370 mm2 (SEM = 12.8) (Figure 5).
22
There was no significant difference between the TDE and HM63.9 (p = 0.51). There was a significant
23
difference between the TDE and HM50.4 (p < 0.001) and between HM63.9 and HM50.4 (p < 0.001). The
24
TDE filled an average of 62.8% of HM50.4, while HM63.9 filled an average of 64.4% HM50.4.
25 26
DISCUSSION
27
Over the last decade, numerous studies have examined laser interstitial thermal therapy (LITT) as
28
treatment for a broad range of intracranial pathologies, finding it to be an efficacious and appropriate
29
technique for certain patient populations.4, 6-9 Recently, greater attention has turned towards examining the
30
technical aspects of the procedure, such as intraoperative ablation dynamics and accuracy of the TDE
31
compared with postoperative MRI sequences. For example, one previous study found that ablation of
32
previously heat-damaged tissue results in a reduction of ablation rate and ablation area, and that each
33
successive thermal dose in a series of thermal doses results in a decreased ablation rate relative to that of
34
the previous ablation.10 Other studies have compared the software-generated TDE to postoperative MRI,
5 1
finding that intraoperative TDE strongly correlates with area of ablation seen on postoperative imaging.11
2
Still, no study to date has examined the MRT heat map in comparison to the TDE, which could
3
potentially provide a simple and practical application of intraoperative MRT data for operators that may
4
aid in operative decision making. In this study, we compared the TDE area with different color-coded
5
areas on the MRT heat map to evaluate the relationship between the two. We found that there is no
6
difference between the TDE area and the area contained within the peripheral border of the green area on
7
the MRT heat map (HM63.9), suggesting that these two areas are essentially equal in magnitude.
8
Furthermore, the TDE was found to be contained within peripheral boundary of the light blue area in
9
100% of cases included, and on average, filled 62.8% of the HM50.4.
10 11
The findings of this study are meaningful for several reasons. The finding of concordance between the
12
TDE and HM63.9 implies that under the Arrhenius model, any area of tissue reaching a temperature
13
threshold of 63.9°C for any period of time will be included in the TDE and is therefore nonviable. This
14
result is not surprising, as previous studies have found that the critical temperature at which tissue
15
necrosis occurs within seconds is somewhere between 55-60°C.12 At these temperatures, cell death occurs
16
via thermal coagulation and protein denaturation, both of which are irreversible and result in certain cell
17
death.13, 14 Achieving cell death with temperatures as low as 43 to 45°C is also possible, though the
18
duration of time required to induce irreversible damage ranges from 25 minutes to several hours
19
depending on tissue type and therapeutic circumstances.14 This concept is critical to consider, and in this
20
context, the finding that the TDE always falls within the peripheral border of the light blue portion of the
21
heat map becomes relevant. The light blue portion of the heat map represents a gradient of temperatures
22
ranging from 50.3°C at the periphery to 63.9°C more centrally, at which point, the green portion of the
23
heat map begins. Because the TDE always falls within the peripheral border of the light blue portion of
24
the heat map, this implies that at a minimum, the most peripheral portion of the heat map is able to
25
tolerate these elevated temperatures without undergoing irreversible cell damage for at least some period
26
of time. As such, operators can be confident that any area of parenchyma not contained within the
27
boundaries of the light blue portion of the heat map will not be covered by the system-generated TDE.
28
Thus, the MRTI heat map itself can then be used as a secondary means of surveilling healthy brain tissue
29
outside of the TDE and can help operators anticipate when those healthy areas will be at risk of necrosis.
30 31
The results of this study are not meant to be interpreted as a test of Arrhenius equation. Rather, these
32
findings should serve as a practical interpretation the MRT heat map, which previously offered little
33
tangible value intraoperatively. Based on the results of this study, operators may now use the MRT data to
34
supplement the information provided by the TDE. Specifically, because the TDE and HM63.9 are
6 1
essentially equal, the light portion of the heat map serves as a predictor of when critical structures are at
2
risk of being damaged if exposed to further thermal stress. In other words, the light portion of the heat
3
map informs operators of where in the parenchyma the TDE will expand to in the subsequent frame if the
4
laser remains activated. As such, operators can use this information to deactivate the laser before damage
5
occurs to critical structures rather than guess as to how far the TDE will expand in the subsequent frame
6
provided by the software 5-7 seconds later.
7 8
LIMITATIONS
9
The primary limitation within this study concerns the fact that the measurements of the TDE and MRT
10
heat map were taken at a single time point at the very end of the ablation. Previous studies have evaluated
11
the temporal profile of tissue damage by assessing the relationship between thermal damage estimate
12
(TDE) and time, finding that expansion of the ablation area can either follow logistic or negative
13
exponential growth as it approaches a finite maximum.15 Considering this, the measurements within this
14
study were taken while the ablation had reached or was approaching its finite maximum. As such, the
15
findings of this study primarily apply to the later stages of TDE growth and may not be applicable to the
16
early course of the ablation where TDE is increasing exponentially. Nonetheless, the correct time point at
17
which the ablation volume should be documented is still a matter of discussion.
18 19
A second limitation to consider is that these findings rely strongly on the accuracy of the TDE produced
20
by the system-generated software. McNichols et al. first investigated the accuracy of the TDE in canine
21
cerebral tissue by evaluating histopathological specimens, finding a strong correlation between the LITT
22
predicted ablation extent and postoperative tissue analysis.16 This conclusion was further supported in a
23
study that evaluated whether the TDE correlates with ablative area seen on post-operative MRI in
24
perfused intracranial pathologies, once again finding that the TDE is in fact an accurate and reliable
25
measure of ablative area.17 Without a histologic analysis in human subjects, it remains difficult to
26
optimally verify the accuracy of the TDE produced by the ablation software. However, these studies
27
provide a reasonable foundation to support that the software utilized in LITT is in fact an accurate and
28
reliable measure of ablation extent.
29 30
A final limitation to consider is the lack of stratification by pathology. This study included patients
31
undergoing ablation for cavernoma, ependymoma, epilepsy, GBM, primary glioma, recurrent glioma,
32
recurrent meningioma, and recurrent metastasis. Due to sample size, there were a limited number of
33
patients within each pathology. As such, a sub-analysis examining the data within each pathology would
34
be underpowered, though it is possible that different trends would be observed within different
7 1
pathologies. However, previous studies have shown that TDE expansion is consistent across all
2
pathologies, demonstrating that the TDE consistently follows negative exponential growth as it
3
approaches a finite maximum.15 This is reassuring that further stratification would not demonstrate
4
different findings across pathologies. Still, a study reexamining the data with further consideration to
5
pathology would be useful in the future. However, at present, it is reasonable to assume that these
6
findings would remain consistent regardless of disease process.
7 8
CONCLUSIONS
9
During LITT therapy, the MRT heat map produced alongside the system-generated TDE can be a useful
10
tool for anticipating what the TDE will look like with respect to the real-time thermal imaging. At
11
standard factory settings, the TDE and the area produced within the periphery of the green area are similar
12
in magnitude. As such, the light blue portion of the MRT heat map may serve as an additional means of
13
predicting when critical structures may be at risk during laser ablation if exposed to further thermal stress.
14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
8 1
REFERENCES
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43
1. Pearce J. Mathematical models of laser-induced tissue thermal damage. Int J Hyperthermia. 2011;27(8): 741-750. https://doi.org/10.3109/02656736.2011.580822. 2. Rieke V, Butts Pauly K. MR thermometry. Journal of Magnetic Resonance Imaging: An Official Journal of the International Society for Magnetic Resonance in Medicine. 2008;27(2): 376-390. 3. Riordan M, Tovar-Spinoza Z. Laser induced thermal therapy (LITT) for pediatric brain tumors: case-based review. Transl Pediatr. 2014;3(3): 229-235. https://doi.org/10.3978/j.issn.2224-4336.2014.07.07. 4. Jethwa PR, Barrese JC, Gowda A, Shetty A, Danish SF. Magnetic resonance thermometryguided laser-induced thermal therapy for intracranial neoplasms: initial experience. Neurosurgery. 2012;71(1 Suppl Operative): 133-144; 144-135. https://doi.org/10.1227/NEU.0b013e31826101d4. 5. McGahan JP, Griffey SM, Budenz RW, Brock JM. Percutaneous ultrasound-guided radiofrequency electrocautery ablation of prostate tissue in dogs. Academic radiology. 1995;2(1): 61-65. 6. Curry DJ, Gowda A, McNichols RJ, Wilfong AA. MR-guided stereotactic laser ablation of epileptogenic foci in children. Epilepsy Behav. 2012;24(4): 408-414. https://doi.org/10.1016/j.yebeh.2012.04.135. 7. Esquenazi Y, Kalamangalam GP, Slater JD, et al. Stereotactic laser ablation of epileptogenic periventricular nodular heterotopia. Epilepsy Res. 2014;108(3): 547-554. https://doi.org/10.1016/j.eplepsyres.2014.01.009. 8. Willie JT, Laxpati NG, Drane DL, et al. Real-time magnetic resonance-guided stereotactic laser amygdalohippocampotomy for mesial temporal lobe epilepsy. Neurosurgery. 2014;74(6): 569-584; discussion 584-565. https://doi.org/10.1227/NEU.0000000000000343. 9. Gonzalez-Martinez J, Vadera S, Mullin J, et al. Robot-assisted stereotactic laser ablation in medically intractable epilepsy: operative technique. Neurosurgery. 2014;10 Suppl 2: 167-172; discussion 172-163. https://doi.org/10.1227/NEU.0000000000000286. 10. Munier SM, Hargreaves EL, Patel NV, Danish SF. Ablation dynamics of subsequent thermal doses delivered to previously heat-damaged tissue during magnetic resonance–guided laser-induced thermal therapy. Journal of neurosurgery. 2018;1(aop): 1-8. 11. Patel NV, Jethwa PR, Shetty A, Danish SF. Does the real-time thermal damage estimate allow for estimation of tumor control after MRI-guided laser-induced thermal therapy? Initial experience with recurrent intracranial ependymomas. J Neurosurg Pediatr. 2015;15(4): 363371. https://doi.org/10.3171/2014.10.PEDS13698. 12. McDannold N. Quantitative MRI-based temperature mapping based on the proton resonant frequency shift: review of validation studies. International journal of hyperthermia. 2005;21(6): 533-546. 13. Dewhirst MW, Viglianti B, Lora-Michiels M, Hanson M, Hoopes P. Basic principles of thermal dosimetry and thermal thresholds for tissue damage from hyperthermia. International journal of hyperthermia. 2003;19(3): 267-294. 14. Thomsen S. Pathologic analysis of photothermal and photomechanical effects of laser– tissue interactions. Photochemistry and photobiology. 1991;53(6): 825-835.
9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46
15. Sun XR, Patel NV, Danish SF. Tissue Ablation Dynamics During Magnetic ResonanceGuided, Laser-Induced Thermal Therapy. Neurosurgery. 2015;77(1): 51-58; discussion 58. https://doi.org/10.1227/NEU.0000000000000732. 16. McNichols RJ, Gowda A, Kangasniemi M, Bankson JA, Price RE, Hazle JD. MR thermometry-based feedback control of laser interstitial thermal therapy at 980 nm. Lasers in Surgery and Medicine: The Official Journal of the American Society for Laser Medicine and Surgery. 2004;34(1): 48-55. 17. Patel NV, Frenchu K, Danish SF. Does the Thermal Damage Estimate Correlate With the Magnetic Resonance Imaging Predicted Ablation Size After Laser Interstitial Thermal Therapy? Operative Neurosurgery. 2017;15(2): 179-183.
10 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Figures Figure 1 – Example of an ablation with real-time magnetic resonance thermal imaging (left) and thermal damage estimate (right). These images are produced by the system software every 7 seconds during the procedure, allowing for real-time monitoring of ablation boundaries. Figure 2 – Example of MRT imaging prior to initiation of ablation. The numbers overlying the image represent the operator-set temperature thresholds. The “2” is set at the center of the target and set to a temperature threshold of 90°C, where the laser will automatically deactivate should it reach that limit at any point during the ablation. The “5” and “6” represent peripheral temperature thresholds set at 50°C to protect surrounding healthy tissue from excess heat exposure. Figure 3 – Intraoperative MRT imaging of a metastatic tumor (left). The colors correspond to the different temperature ranges (upper right corner) at which the parenchyma has been heated to. The image on the right shows a magnified image of the heat map to further depict the different temperature zones reached during the ablation. Figure 4 – MRT imagine with outlines denoting the different areas analyzed. HM50.4 is all area included within the red outline. HM63.9 is outlined by the blue line and includes all area interior to that line. Figure 5 – Measured areas (mm2) of TDE, HM63.9, and HM50.4.
Table I: Patient Information and Operative Indications Characteristic No. (%) of patients; n = 56 Age group (years) <18 0 (0.0) 18-44 8 (14.3) 45-64 27 (48.2) 65+ 21 (37.5) Gender Male 23 (41) Female 33 (59) Indication for Ablation Cavernoma 1 (1.8) Ependymoma 5 (8.9) Epilepsy 5 (8.9) GBM 6 (10.8) Primary Glioma 5 (8.9) Recurrent Glioma 11 (19.6) Recurrent Meningioma 2 (3.6) Recurrent Metastasis 21 (37.5) Table I: Patient information and Operative Indications
Author Contributions -
-
Sean Munier o Development of project idea, data acquisition, manuscript preparation, editing, statistical analysis Elizabeth Ginalis o Data acquisition, manuscript editing Akshay Desai o Data acquisition, statistical analysis Shabbar Danish o Project development, senior surgeon in all cases, manuscript editing
Abbreviations List: CT: computed tomography HM50.4: heat map 50.4 HM63.9: heat map 63.9 LITT: laser interstitial thermal therapy
MRgLITT: magnetic resonance guided laser interstitial thermal therapy MRI: magnetic resonance imaging MRT: magnetic resonance thermometry MRTI: magnetic resonance thermal imaging PRF: proton resonance frequency TDE: thermal damage estimate SEM: standard error mean
Declaration of interests: Dr. Danish is a consultant for Medtronic and has received educational honoraria.
S1878-8750(19)32950-X
DOI:
https://doi.org/10.1016/j.wneu.2019.11.110
Reference:
WNEU 13779
To appear in:
World Neurosurgery
Received Date: 24 August 2019 Revised Date:
18 November 2019
Accepted Date: 19 November 2019
Please cite this article as: Munier SM, Ginalis EE, Desai AN, Danish SF, Understanding the Relationship Between the Real-Time Thermal Imaging and the Thermal Damage Estimate during MRguided Laser Interstitial Thermal Therapy, World Neurosurgery (2019), doi: https://doi.org/10.1016/ j.wneu.2019.11.110. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier Inc.
Understanding the Relationship Between the Real-Time Thermal Imaging and the Thermal Damage Estimate during MR-guided Laser Interstitial Thermal Therapy Sean M Munier, MD; Elizabeth E Ginalis, BS; Akshay N Desai, MS; Shabbar F Danish, MD Department of Neurosurgery, Rutgers University, Robert Wood Johnson Medical School, New Brunswick, New Jersey, United States Corresponding Author: Sean M. Munier, MD 10 Plum St. 5th Floor New Brunswick, NJ 08901 Phone: 201-446-2156 Email: [email protected] Keywords: Laser interstitial thermal therapy; LITT; Magnetic resonance thermometry; MRTI; Thermal ablation parameters
Conflicts of Interest: Dr. Danish is a consultant for Medtronic and has received educational honoraria.
Disclosure of Funding: None.
Running Title: MRT Imaging Boundary Quantification
1
ABSTRACT
2
Objectives: Magnetic Resonance-guided Laser Interstitial Thermal Therapy (MRgLITT) is a minimally
3
invasive procedure that utilizes intraoperative magnetic resonance thermometry (MRT) to generate a
4
thermal damage estimate (TDE) of the ablative area. This study compared areas produced by the MRT
5
heat map to the system-generated TDE produced by the Visualase software.
6 7
Methods: All ablations were performed using the Visualase MRI-Guided Laser Ablation System. MRT
8
heat-map and TDE were quantified using MATLAB version R2014a. The TDE was compared to the
9
summed area of the green, yellow, and red areas (HM63.9) and the summed area of the light blue, green,
10
yellow, and red areas (HM50.4) produced by the MRT heat map.
11 12
Results: 56 patients undergoing MRgLITT were examined. The mean TDE produced was 236 mm2
13
(SEM = 9.5). The mean HM63.9 was 231 mm2 (SEM = 8.7), and the mean HM50.4 was 370 mm2 (SEM =
14
12.8). There was no significant difference between the TDE and HM63.9 (p = 0.51) . There was a
15
significant difference between TDE and HM50.4 (p < 0.001) and between HM63.9 and HM50.4 (p < 0.001).
16 17
Conclusions: The system-generated TDE consistently remains contained within the boundaries of the
18
MRT heat map. At standard factory settings, the TDE and the area produced within the periphery of the
19
HM63.9 are similar in magnitude. The light blue portion of the MRT heat map may serve as an additional
20
means of predicting when critical structures may be at risk during laser ablation if exposed to further
21
thermal stress.
22 23 24 25 26 27 28 29 30 31 32 33 34
2 1
INTRODUCTION
2
Magnetic resonance-guided laser interstital thermal therapy (MRgLITT) is a minimally invasive
3
neurosurgical technique utilized in the treatment of many intracranial pathologies. One of the key features
4
of LITT is the ability to monitor ablation in real-time through magnetic resonance thermometry (MRT)
5
and the system-generated thermal damage estimate (TDE) (Figure 1). This estimate is produced using the
6
Arrhenius equation, which utilizes the temperature data recorded by the MRI machine to estimate the
7
tissue boundaries where cell death has occurred.1 Temperature data is produced under MRT, which is
8
recorded by the MRI machine through measurement of proton resonance frequency (PRF).2 Measurement
9
of PRF is based on the property of protons to resonate at a baseline frequency given a constant set of
10
parameters, including temperature. When temperature increases, as it does in LITT, hydrogen bonds
11
between water molecules are stretched, resulting in a change in PRF.2 The ability of MRI machines to
12
quantify these changes in PRF is what facilitates intraoperative MRT, and thus allows for a predictive
13
estimate of where tissue necrosis is occurring during LITT in real time.
14 15
To date, MRT produced by the MRI machine has primarily been used as a means of setting temperature
16
limits for the procedure that trigger the laser to automatically deactivate should the center or periphery of
17
the ablation exceed certain operator-determined temperatures. The manufacturer of the technology
18
advises a central temperature limit of 90°C to prevent production of steam and a peripheral temperature
19
limit of 45°C to prevent damage to healthy tissue.3 However, since no studies have analyzed
20
intraoperative MRT data relative to intraoperative TDE, MRT data currently offers little practical value to
21
the operator during the procedure. In this study, we aimed to understand the relationship of the real-time
22
thermal imaging to the thermal damage estimate in an effort to provide a practical and useful
23
interpretation of intraoperative MRT thermal data that may better inform intraoperative decision making.
24 25
METHODS
26
Patient Selection
27
We retrospectively reviewed and included 56 patients who underwent MRgLITT for various intracranial
28
procedures at our institution. Inclusion criteria for the study required single-laser catheter use, available
29
TDE and MRT data, and MRT temperature map settings with the following parameters: dark blue, >
30
37°C; light blue, > 50.4°C; green, > 63.9°C; yellow, > 77.3°C; red, > 90.8°C (with said temperatures
31
referring to tissue temperature within that color region). For the purpose of this study, only ablations with
32
a homogenous thermal distribution and without signal artifact were considered appropriate for analysis.
33
All subjects were from a single institution and operated on by a single surgeon, the senior author. All
3 1
patients underwent the procedure as part of the routine clinical care algorithm at our institution and were
2
enrolled in an institutional review board-approved protocol with their consent at our institution.
3 4
Laser Ablation Procedure
5
The ablation procedure has previously been described.4 Briefly, each patient undergoes preoperative CT
6
and MRI. The images are loaded onto a stereotactic planning software and fused to allow for planning of
7
the laser catheter trajectory.4 Bone implanted fiducials are fixed to the calvarium to allow for stereotactic
8
registration with the system software. After determination of appropriate laser trajectory, a small stab
9
incision is made and a burr hole is drilled over the entry point using a 3.2 mm twist drill. A bone anchor is
10
introduced into the entry site to secure the laser catheter. The laser catheter is introduced through the bone
11
anchor to a predetermined length into the ablation target. A 980-nm diffusing tip diode laser catheter is
12
utilized. The patient is transported to the MRI suite where the procedure is performed.
13 14
The procedure is performed under the guidance of MRT, which allows for intraoperative monitoring of
15
the ablation boundaries in real time. Using the Visualase software, the operator identifies safety margins
16
at the periphery and center of the intended target such that the peripheral healthy tissue temperature does
17
not surpass an operator-set temperature (typically 45°C) and the lesion center does not exceed 90°C to
18
avoid steam formation (Figure 2). Should either of these temperature limits be reached, the laser is
19
automatically deactivated. The system workstation provides the operator with information via two
20
dynamic images: one image displays the MRT data produced in the form of a heat map, where different
21
colors correspond to different tissue temperature thresholds, which are displayed in the upper right corner
22
of the MRT image (Figure 3). The right image in figure 1 displays the real-time software-generated
23
thermal damage estimate (TDE), which provides a calculated approximation of where tissue necrosis has
24
occurred. In all cases considered, these temperature thresholds on the MRT were set at factory-standard
25
settings seen in Figure 3. This estimated area is produced by the software every 5-7 seconds and appears
26
on the workstation as an orange area superimposed over the MRI images, allowing the operator to
27
determine when the target in question has been ablated and to terminate the ablation at their discretion.
28
The ablation area expands outward from the laser catheter in an ellipsoidal fashion, with expansion in any
29
single direction depending on tissue composition and target morphology.5 Following the procedure, the
30
patient undergoes postoperative MRI to verify that complete target destruction has occurred.
31 32
Imaging Analysis
33
Intraoperative MRT and TDE images were acquired from the Visualase console. Cases included were
34
those performed under the most recently updated factory-standard settings for the MRT heat map
4 1
temperature thresholds. The final image produced by the software prior to laser deactivation was analyzed
2
for each case included. Imaging areas were quantified using MATLAB version R2014a (Mathworks, Inc,
3
Natick, Massachusetts). The program individually quantifies the number of pixels of each color during the
4
course of the ablation. Two specific areas from the MRT heat map were analyzed. The first area included
5
all colors within the heat map (HM50.4) and is produced by the summed pixel count of the light blue,
6
green, yellow, and red areas. The second area included the summed pixel counts of the green, yellow, and
7
red areas (HM63.9). Each pixel is equivalent to 0.9 mm2. For TDE, the number of orange pixels were
8
individually quantified. Figure 4 delineates the specific areas being analyzed.
9 10
Data Analysis
11
All analyses were performed with GraphPad Prism version 7.0 (GraphPad Software, San Diego,
12
California). A student’s t-test analysis was used to detect differences between TDE, HM63.9, and HM50.4.
13 14
RESULTS
15
We evaluated a total of 56 cases. Operative indications included cavernoma, ependymoma, epilepsy,
16
glioblastoma multiforme, primary glioma, recurrent glioma, recurrent meningioma, and recurrent
17
metastasis. Operative indications are summarized in Table I.
18 19
The average power used across all cases was 11.35 W (SEM = 0.25). The average duration across all
20
ablations was 134 seconds (SEM = 11.70). The mean TDE produced was 236 mm2 (SEM = 9.5). The
21
mean HM63.9 was 231 mm2 (SEM = 8.7), and the mean HM50.4 was 370 mm2 (SEM = 12.8) (Figure 5).
22
There was no significant difference between the TDE and HM63.9 (p = 0.51). There was a significant
23
difference between the TDE and HM50.4 (p < 0.001) and between HM63.9 and HM50.4 (p < 0.001). The
24
TDE filled an average of 62.8% of HM50.4, while HM63.9 filled an average of 64.4% HM50.4.
25 26
DISCUSSION
27
Over the last decade, numerous studies have examined laser interstitial thermal therapy (LITT) as
28
treatment for a broad range of intracranial pathologies, finding it to be an efficacious and appropriate
29
technique for certain patient populations.4, 6-9 Recently, greater attention has turned towards examining the
30
technical aspects of the procedure, such as intraoperative ablation dynamics and accuracy of the TDE
31
compared with postoperative MRI sequences. For example, one previous study found that ablation of
32
previously heat-damaged tissue results in a reduction of ablation rate and ablation area, and that each
33
successive thermal dose in a series of thermal doses results in a decreased ablation rate relative to that of
34
the previous ablation.10 Other studies have compared the software-generated TDE to postoperative MRI,
5 1
finding that intraoperative TDE strongly correlates with area of ablation seen on postoperative imaging.11
2
Still, no study to date has examined the MRT heat map in comparison to the TDE, which could
3
potentially provide a simple and practical application of intraoperative MRT data for operators that may
4
aid in operative decision making. In this study, we compared the TDE area with different color-coded
5
areas on the MRT heat map to evaluate the relationship between the two. We found that there is no
6
difference between the TDE area and the area contained within the peripheral border of the green area on
7
the MRT heat map (HM63.9), suggesting that these two areas are essentially equal in magnitude.
8
Furthermore, the TDE was found to be contained within peripheral boundary of the light blue area in
9
100% of cases included, and on average, filled 62.8% of the HM50.4.
10 11
The findings of this study are meaningful for several reasons. The finding of concordance between the
12
TDE and HM63.9 implies that under the Arrhenius model, any area of tissue reaching a temperature
13
threshold of 63.9°C for any period of time will be included in the TDE and is therefore nonviable. This
14
result is not surprising, as previous studies have found that the critical temperature at which tissue
15
necrosis occurs within seconds is somewhere between 55-60°C.12 At these temperatures, cell death occurs
16
via thermal coagulation and protein denaturation, both of which are irreversible and result in certain cell
17
death.13, 14 Achieving cell death with temperatures as low as 43 to 45°C is also possible, though the
18
duration of time required to induce irreversible damage ranges from 25 minutes to several hours
19
depending on tissue type and therapeutic circumstances.14 This concept is critical to consider, and in this
20
context, the finding that the TDE always falls within the peripheral border of the light blue portion of the
21
heat map becomes relevant. The light blue portion of the heat map represents a gradient of temperatures
22
ranging from 50.3°C at the periphery to 63.9°C more centrally, at which point, the green portion of the
23
heat map begins. Because the TDE always falls within the peripheral border of the light blue portion of
24
the heat map, this implies that at a minimum, the most peripheral portion of the heat map is able to
25
tolerate these elevated temperatures without undergoing irreversible cell damage for at least some period
26
of time. As such, operators can be confident that any area of parenchyma not contained within the
27
boundaries of the light blue portion of the heat map will not be covered by the system-generated TDE.
28
Thus, the MRTI heat map itself can then be used as a secondary means of surveilling healthy brain tissue
29
outside of the TDE and can help operators anticipate when those healthy areas will be at risk of necrosis.
30 31
The results of this study are not meant to be interpreted as a test of Arrhenius equation. Rather, these
32
findings should serve as a practical interpretation the MRT heat map, which previously offered little
33
tangible value intraoperatively. Based on the results of this study, operators may now use the MRT data to
34
supplement the information provided by the TDE. Specifically, because the TDE and HM63.9 are
6 1
essentially equal, the light portion of the heat map serves as a predictor of when critical structures are at
2
risk of being damaged if exposed to further thermal stress. In other words, the light portion of the heat
3
map informs operators of where in the parenchyma the TDE will expand to in the subsequent frame if the
4
laser remains activated. As such, operators can use this information to deactivate the laser before damage
5
occurs to critical structures rather than guess as to how far the TDE will expand in the subsequent frame
6
provided by the software 5-7 seconds later.
7 8
LIMITATIONS
9
The primary limitation within this study concerns the fact that the measurements of the TDE and MRT
10
heat map were taken at a single time point at the very end of the ablation. Previous studies have evaluated
11
the temporal profile of tissue damage by assessing the relationship between thermal damage estimate
12
(TDE) and time, finding that expansion of the ablation area can either follow logistic or negative
13
exponential growth as it approaches a finite maximum.15 Considering this, the measurements within this
14
study were taken while the ablation had reached or was approaching its finite maximum. As such, the
15
findings of this study primarily apply to the later stages of TDE growth and may not be applicable to the
16
early course of the ablation where TDE is increasing exponentially. Nonetheless, the correct time point at
17
which the ablation volume should be documented is still a matter of discussion.
18 19
A second limitation to consider is that these findings rely strongly on the accuracy of the TDE produced
20
by the system-generated software. McNichols et al. first investigated the accuracy of the TDE in canine
21
cerebral tissue by evaluating histopathological specimens, finding a strong correlation between the LITT
22
predicted ablation extent and postoperative tissue analysis.16 This conclusion was further supported in a
23
study that evaluated whether the TDE correlates with ablative area seen on post-operative MRI in
24
perfused intracranial pathologies, once again finding that the TDE is in fact an accurate and reliable
25
measure of ablative area.17 Without a histologic analysis in human subjects, it remains difficult to
26
optimally verify the accuracy of the TDE produced by the ablation software. However, these studies
27
provide a reasonable foundation to support that the software utilized in LITT is in fact an accurate and
28
reliable measure of ablation extent.
29 30
A final limitation to consider is the lack of stratification by pathology. This study included patients
31
undergoing ablation for cavernoma, ependymoma, epilepsy, GBM, primary glioma, recurrent glioma,
32
recurrent meningioma, and recurrent metastasis. Due to sample size, there were a limited number of
33
patients within each pathology. As such, a sub-analysis examining the data within each pathology would
34
be underpowered, though it is possible that different trends would be observed within different
7 1
pathologies. However, previous studies have shown that TDE expansion is consistent across all
2
pathologies, demonstrating that the TDE consistently follows negative exponential growth as it
3
approaches a finite maximum.15 This is reassuring that further stratification would not demonstrate
4
different findings across pathologies. Still, a study reexamining the data with further consideration to
5
pathology would be useful in the future. However, at present, it is reasonable to assume that these
6
findings would remain consistent regardless of disease process.
7 8
CONCLUSIONS
9
During LITT therapy, the MRT heat map produced alongside the system-generated TDE can be a useful
10
tool for anticipating what the TDE will look like with respect to the real-time thermal imaging. At
11
standard factory settings, the TDE and the area produced within the periphery of the green area are similar
12
in magnitude. As such, the light blue portion of the MRT heat map may serve as an additional means of
13
predicting when critical structures may be at risk during laser ablation if exposed to further thermal stress.
14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
8 1
REFERENCES
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43
1. Pearce J. Mathematical models of laser-induced tissue thermal damage. Int J Hyperthermia. 2011;27(8): 741-750. https://doi.org/10.3109/02656736.2011.580822. 2. Rieke V, Butts Pauly K. MR thermometry. Journal of Magnetic Resonance Imaging: An Official Journal of the International Society for Magnetic Resonance in Medicine. 2008;27(2): 376-390. 3. Riordan M, Tovar-Spinoza Z. Laser induced thermal therapy (LITT) for pediatric brain tumors: case-based review. Transl Pediatr. 2014;3(3): 229-235. https://doi.org/10.3978/j.issn.2224-4336.2014.07.07. 4. Jethwa PR, Barrese JC, Gowda A, Shetty A, Danish SF. Magnetic resonance thermometryguided laser-induced thermal therapy for intracranial neoplasms: initial experience. Neurosurgery. 2012;71(1 Suppl Operative): 133-144; 144-135. https://doi.org/10.1227/NEU.0b013e31826101d4. 5. McGahan JP, Griffey SM, Budenz RW, Brock JM. Percutaneous ultrasound-guided radiofrequency electrocautery ablation of prostate tissue in dogs. Academic radiology. 1995;2(1): 61-65. 6. Curry DJ, Gowda A, McNichols RJ, Wilfong AA. MR-guided stereotactic laser ablation of epileptogenic foci in children. Epilepsy Behav. 2012;24(4): 408-414. https://doi.org/10.1016/j.yebeh.2012.04.135. 7. Esquenazi Y, Kalamangalam GP, Slater JD, et al. Stereotactic laser ablation of epileptogenic periventricular nodular heterotopia. Epilepsy Res. 2014;108(3): 547-554. https://doi.org/10.1016/j.eplepsyres.2014.01.009. 8. Willie JT, Laxpati NG, Drane DL, et al. Real-time magnetic resonance-guided stereotactic laser amygdalohippocampotomy for mesial temporal lobe epilepsy. Neurosurgery. 2014;74(6): 569-584; discussion 584-565. https://doi.org/10.1227/NEU.0000000000000343. 9. Gonzalez-Martinez J, Vadera S, Mullin J, et al. Robot-assisted stereotactic laser ablation in medically intractable epilepsy: operative technique. Neurosurgery. 2014;10 Suppl 2: 167-172; discussion 172-163. https://doi.org/10.1227/NEU.0000000000000286. 10. Munier SM, Hargreaves EL, Patel NV, Danish SF. Ablation dynamics of subsequent thermal doses delivered to previously heat-damaged tissue during magnetic resonance–guided laser-induced thermal therapy. Journal of neurosurgery. 2018;1(aop): 1-8. 11. Patel NV, Jethwa PR, Shetty A, Danish SF. Does the real-time thermal damage estimate allow for estimation of tumor control after MRI-guided laser-induced thermal therapy? Initial experience with recurrent intracranial ependymomas. J Neurosurg Pediatr. 2015;15(4): 363371. https://doi.org/10.3171/2014.10.PEDS13698. 12. McDannold N. Quantitative MRI-based temperature mapping based on the proton resonant frequency shift: review of validation studies. International journal of hyperthermia. 2005;21(6): 533-546. 13. Dewhirst MW, Viglianti B, Lora-Michiels M, Hanson M, Hoopes P. Basic principles of thermal dosimetry and thermal thresholds for tissue damage from hyperthermia. International journal of hyperthermia. 2003;19(3): 267-294. 14. Thomsen S. Pathologic analysis of photothermal and photomechanical effects of laser– tissue interactions. Photochemistry and photobiology. 1991;53(6): 825-835.
9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46
15. Sun XR, Patel NV, Danish SF. Tissue Ablation Dynamics During Magnetic ResonanceGuided, Laser-Induced Thermal Therapy. Neurosurgery. 2015;77(1): 51-58; discussion 58. https://doi.org/10.1227/NEU.0000000000000732. 16. McNichols RJ, Gowda A, Kangasniemi M, Bankson JA, Price RE, Hazle JD. MR thermometry-based feedback control of laser interstitial thermal therapy at 980 nm. Lasers in Surgery and Medicine: The Official Journal of the American Society for Laser Medicine and Surgery. 2004;34(1): 48-55. 17. Patel NV, Frenchu K, Danish SF. Does the Thermal Damage Estimate Correlate With the Magnetic Resonance Imaging Predicted Ablation Size After Laser Interstitial Thermal Therapy? Operative Neurosurgery. 2017;15(2): 179-183.
10 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Figures Figure 1 – Example of an ablation with real-time magnetic resonance thermal imaging (left) and thermal damage estimate (right). These images are produced by the system software every 7 seconds during the procedure, allowing for real-time monitoring of ablation boundaries. Figure 2 – Example of MRT imaging prior to initiation of ablation. The numbers overlying the image represent the operator-set temperature thresholds. The “2” is set at the center of the target and set to a temperature threshold of 90°C, where the laser will automatically deactivate should it reach that limit at any point during the ablation. The “5” and “6” represent peripheral temperature thresholds set at 50°C to protect surrounding healthy tissue from excess heat exposure. Figure 3 – Intraoperative MRT imaging of a metastatic tumor (left). The colors correspond to the different temperature ranges (upper right corner) at which the parenchyma has been heated to. The image on the right shows a magnified image of the heat map to further depict the different temperature zones reached during the ablation. Figure 4 – MRT imagine with outlines denoting the different areas analyzed. HM50.4 is all area included within the red outline. HM63.9 is outlined by the blue line and includes all area interior to that line. Figure 5 – Measured areas (mm2) of TDE, HM63.9, and HM50.4.
Table I: Patient Information and Operative Indications Characteristic No. (%) of patients; n = 56 Age group (years) <18 0 (0.0) 18-44 8 (14.3) 45-64 27 (48.2) 65+ 21 (37.5) Gender Male 23 (41) Female 33 (59) Indication for Ablation Cavernoma 1 (1.8) Ependymoma 5 (8.9) Epilepsy 5 (8.9) GBM 6 (10.8) Primary Glioma 5 (8.9) Recurrent Glioma 11 (19.6) Recurrent Meningioma 2 (3.6) Recurrent Metastasis 21 (37.5) Table I: Patient information and Operative Indications
Author Contributions -
-
Sean Munier o Development of project idea, data acquisition, manuscript preparation, editing, statistical analysis Elizabeth Ginalis o Data acquisition, manuscript editing Akshay Desai o Data acquisition, statistical analysis Shabbar Danish o Project development, senior surgeon in all cases, manuscript editing
Abbreviations List: CT: computed tomography HM50.4: heat map 50.4 HM63.9: heat map 63.9 LITT: laser interstitial thermal therapy
MRgLITT: magnetic resonance guided laser interstitial thermal therapy MRI: magnetic resonance imaging MRT: magnetic resonance thermometry MRTI: magnetic resonance thermal imaging PRF: proton resonance frequency TDE: thermal damage estimate SEM: standard error mean
Declaration of interests: Dr. Danish is a consultant for Medtronic and has received educational honoraria.