Lasers in Surgery and Medicine 46:558–562 (2014)
MR-Guided Laser-Induced Thermotherapy in Ex Vivo Porcine Kidney: Comparison of Four Different Imaging Sequences Stefan Vallo, MD,1 Katrin Eichler, MD, PD,2 Kathrin Kelly,1 Boris Schulz, MD, PD,2 Georg Bartsch, MD, PD,1 Axel Haferkamp, MD,1 Thomas J Vogl, MD,2 and Stephan Zangos, MD2 1 Department of Urology, University Hospital Frankfurt, Frankfurt am Main, Germany 2 Department of Diagnostic and Interventional Radiology, University Hospital Frankfurt, Frankfurt am Main, Germany
Purpose: To evaluate the clinical value of different magnetic resonance imaging (MRI) sequences for a realtime thermo-monitoring during laser-induced thermotherapy (LITT) in kidneys. Methods: Twenty-eight ex vivo pig kidneys were treated with laser ablation under MR guidance in a high-field MR scanner (Magnetom Espree or Avanto Fit, Siemens, Germany). For the thermal ablation of the kidney, a neodymium yttrium-aluminum-garnet (Nd:YAG) laser was used in combination with a special protective catheter (length 43 cm, 4 French) which is sealed at the distal end. First, ablation was performed for 7, 10, and 13 minutes using FLASH sequences for investigation of time-dependent growth of lesion size. In the second step, we evaluated the optimal imaging sequence during a 7 minutes ablation of the kidney and after cooling using four different MR sequences (Haste, FLASH, radial VIBE, and Caipirinha DIXON). Results: Macroscopic lesion volume increased from 3,784 1,525 mm3 to 7,683 5,756 mm3 after the ablation from 7 to 13 minutes and MR volume ranged from 2,107 1,674 mm3 to 2,934 1,549 mm3 after the ablation from 7 to 13 minutes. During ablation, FLASH (132 34%) and radial VIBE (120 43%) sequences displayed lesion volumes most efficiently with a trend to overestimation. The Caipirinha DIXON (323 24%) sequence overestimated the volumes significantly during real-time monitoring. The volumes measured by MRI with FLASH (61 30%), Haste (67 28%), or radial VIBE (48 14%) sequences after cooling of the kidney after ablation were always underestimated. The Caipirinha DIXON (142 2%) sequence still overestimated the lesion volume after cooling of the kidney. Conclusion: LITT is a feasible ablation modality in kidney tissue. Moreover, macroscopic and MR lesion volume increases time-dependently. For online monitoring, radial VIBE and FLASH sequences seem to be most efficient. Lasers Surg. Med. 46:558–562, 2014. ß 2014 Wiley Periodicals, Inc. Key words: laser ablation; LITT; MRI; real-time monitoring; renal cancer; thermometry ß 2014 Wiley Periodicals, Inc.
INTRODUCTION The number of detected renal cell carcinomas (RCCs) has been rising due to advances in imaging [1]. Today, nephron-sparing surgery has become the treatment of choice for solitary kidney tumors smaller than 7 cm in diameter in order to preserve renal function. The oncological outcome has been proved to be comparable to radical tumor nephrectomy [2]. However, surgery may not be applicable in high-risk patients with bad performance status due to multiple comorbidities. In these patients, alternative treatment options are necessary to assure local tumor control [3]. Image-guided local ablative treatments like magnetic resonance (MR)-guided laser-induced thermotherapy (LITT), microwave coagulation, radiofrequency ablation (RFA), or cryoablation are used frequently as well-tolerated renal-sparing treatments with curative potential. The currently most widely used techniques are RFA and cryoablation. Today, laser ablation is routinely used for the treatment of liver tumors with excellent local tumor control. LITT is usually performed under MR guidance, which offers an excellent soft tissue resolution and the possibility to document heat propagation during the ablation with thermo-sensitive sequences [4]. Although it is likely that laser treatment is just as effective in ablating RCCs as in hepatic malignancies, experience in renal ablation is very limited and only few studies have been reported on the treatment of RCC with LITT [1,5,6].
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported. Correspondence to: Dr. Stefan Vallo, MD, University Hospital Frankfurt, Goethe University of Frankfurt, Theodor-Stern-Kai 7, 60590 Frankfurt am Main, Germany. E-mail: [email protected] Accepted 13 May 2014 Published online 6 June 2014 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/lsm.22262
MR-GUIDED LASER-INDUCED THERMOTHERAPY
Generally, the success of thermal ablation depends on tumor size, tumor location, and perfusion of the tumor and its surrounding structures. A main challenge one faces in the kidney is the cooling effect of vessels and the high blood flow, which may influence a sufficient thermal ablation. In this study, we compared different MR sequences for realtime monitoring during MR-guided LITT for renal treatment under standardized conditions.
was measured on an Advantage Workstation (GE Healthcare, Barrington, IL) using standard digital measurement techniques. After ablation, the kidneys were cut along the longitudinal plane passing through the axis of the electrode. On this plane, the longest coagulation diameter (Dx) parallel to the catheter as well as the coagulation diameter Dy perpendicular to Dx were measured with a tape measure. The area of the ablation zone was calculated using the formula of the rotational ellipsoid:
MATERIALS AND METHODS For this study we used 28 kidneys of slaughtered pigs. Measurements were performed not longer than 24 hours after slaughter and kidneys were stored at 48C until treatment. Prior to the experiment, the kidneys were warmed up and fixed in a water basin (378C). For the laser ablation, a Nd:YAG laser (Dornier MedLas 5100) was used which was positioned outside the scanner room. The laser energy was delivered into the MR room through fibers over 10 m in length. Laser light was emitted at an effective distance of 30 mm. A special protective catheter (Somatex, Berlin, Germany, length 43 cm, 9 French) from a laser application kit was placed orthogonally to the organ surface and advanced at least 3.5 cm into the tissue at the upper and lower pole of the kidney. The power laser application system allows permanent cooling with water and prevents carbonization at the tip of the applicator, increasing the volume of coagulation necrosis. During the ablation, permanent cooling was achieved through an integrated roller pump with a flow rate of 60 ml/ minute. MR measurements were performed with a conventional high-field 1.5 T system (Magnetom Espree or Avanto Fit; Siemens, Erlangen, Germany) using the integrated spine coil. In the first step, time-dependent lesion volume growth was analyzed. Laser energy (30 W, continuous pulse) was administered for 7 (n ¼ 9), 10 (n ¼ 16), or 13 (n ¼ 13) minutes. To monitor temperature changes, a thermosensitive T1-weighted gradient-echo sequence (FLASH2D) (repetition time [TR]/echo time [TE]: 100/4.8 milliseconds; flip angle [FA] 908, field of view [FOV] 200 mm; matrix 162 250; slice thickness 5 mm) was used. In the second step, 7-minute ablations (n ¼ 16) were monitored using T2-weighted Haste sequences (TR/TE: 140/12 milliseconds; FA 808, FOV 350 mm; matrix 128 200; slice thickness 8 mm, acquisition time 8 seconds), thermo-sensitive T1-weighted gradient-echo sequence (FLASH-2D) (TR/TE: 100/4.8 milliseconds; FA 908, FOV 200 mm; matrix 162 250; slice thickness 5 mm, acquisition time 15 seconds), newly developed T1-weighted radial VIBE sequences (TR/TE: 5/2.05 milliseconds; FA 158, FOV 250 mm; matrix 256 256; slice thickness 4 mm, acquisition time 25 seconds) or Caipirinha DIXON sequences (TR/TE: 6.8/4.8 milliseconds; FA 108, FOV 200 mm; matrix 195 320; slice thickness 3 mm, acquisition time 9 seconds) to detect the most appropriate sequence for real-time MR monitoring of laser ablation ex vivo. Measurements were repeated every 30 seconds during the ablation and after cooling. All images were analyzed and the lesion dimension
559
V¼
4p 3
2 Dx Dy 2 2
For statistical analysis, Student’s t-test was used and a P-value of or below 0.05 was considered to be significant. RESULTS There were no technical problems during the LITT procedure. Due to the brighter color of the ablated tissue, the laser lesions were easy to distinguish macroscopically inside the renal parenchyma. In the renal pelvis, however, the generally light tissue color caused difficulties when trying to differentiate macroscopically between treated and untreated tissue. In the first step, lesion volume showed a time-dependent increase, macroscopically as well as in MR volumetry. The mean macroscopic ablation volumes ranged from 3,784 2,107 mm3 after 7 minutes ablation to 7,682 2,934 mm3 after 13 minutes ablation. The ablation volumes, measured with the FLASH sequence, ranged from 2,107 1,674 mm3 (P ¼ 0.08) after 7 minutes ablation to 2,934 1,549 mm3 (P ¼ 0.04) after 13 minutes ablation (56% of the measured macroscopic size after 7 minutes and 38% after 13 minutes ablation time) (Table 1). Because of the better prediction of lesion size after 7 minutes of ablation compared to 13 minutes, we performed another ablation series for 7 minutes in a second step. On the basis of signal loss during the ablation, a realtime monitoring was possible in three (FLASH, radial VIBE, and Caipirinha DIXON) of the four tested sequences. During the laser ablation, the induced thermal lesions presented as hypointense round lesions with an increasing signal loss in a time-dependent manner. Real-time monitoring with Haste sequences was not feasible since signal changes during laser ablation were not visible (Fig. 1). With the new Caipirinha DIXON sequence, lesion size was significantly overestimated by far which corresponded to
TABLE 1. Mean Lesion Volumes After Laser Ablation Measured Macroscopically vs. MR Volumes (FLASH Sequence)
Time (minute)
Macroscopic volume (mm3)
MR volume (mm3)
7 10 13
3,784 1,525 6,920 3,825 7,682 5,756
2,107 1,674 2,868 1,166 2,934 1,549
560
VALLO ET AL.
down using FLASH, Caipirinha DIXON, and radial VIBE sequences. In contrast, the lesions showed a hypointense signal after cooling in the Haste sequence. In the Haste, FLASH and radial VIBE sequences, the estimated mean lesion sizes measured by MR after cooling were smaller compared to the macroscopic measurements, while they were larger in the Caipirinha DIXON sequence. The measured volume corresponded to 61 30% (P ¼ 0.001) (Haste), 67 28% (P ¼ 0.01) (radial VIBE), 47 14% (P ¼ 0.001) (FLASH), or 142 2% (P ¼ 0.001) (Caipirinha DIXON) of the macroscopic lesion volume (Fig. 3). DISCUSSION
Fig. 1. Comparison of four different MRI sequences during laser ablation of a porcine kidney. The ablation was performed at the lower pole of the kidney. Images were made every minute after beginning of the ablation starting from the top left (A: Caipirinha DIXON, B: FLASH, C: Haste, D: radial VIBE).
323 24% of the macroscopic size (macroscopic lesion size 4,392 651 mm3 vs. MR lesion size 14,224 3,517 mm3) after 7 minutes of ablation (P ¼ 0.003). In contrast, a satisfactory real-time monitoring of the lesion size was possible with the FLASH (132 34%; P ¼ 0.40) (macroscopic lesion size 5,788 980 mm3 vs. MR lesion size 7,651 2,604 mm3) and radial VIBE (120 43%; P ¼ 0.38) (macroscopic lesion size 8,940 2,348 mm3 vs. MR lesion size 10,731 4,627 mm3) sequences, showing a tendency to overestimation in size (Fig. 2). Next, we compared the four different MR sequences after a cooling period of 1 minute. Hereby, we observed a signal change of the ablation zone to hyperintensity after cooling
Fig. 2. Real-time monitoring of lesion size in renal tissue measured with FLASH, radial VIBE, or Caipirinha DIXON MR sequences. P < 0.05.
Today, about two-thirds of all detected renal masses are found accidentally during diagnostic imaging such as ultrasonography, computed tomography (CT), or magnetic resonance imaging (MRI) [7]. As a consequence, an approximate threefold increase in the diagnosis of small renal masses (
MR-Guided Laser-Induced Thermotherapy in Ex Vivo Porcine Kidney: Comparison of Four Different Imaging Sequences Stefan Vallo, MD,1 Katrin Eichler, MD, PD,2 Kathrin Kelly,1 Boris Schulz, MD, PD,2 Georg Bartsch, MD, PD,1 Axel Haferkamp, MD,1 Thomas J Vogl, MD,2 and Stephan Zangos, MD2 1 Department of Urology, University Hospital Frankfurt, Frankfurt am Main, Germany 2 Department of Diagnostic and Interventional Radiology, University Hospital Frankfurt, Frankfurt am Main, Germany
Purpose: To evaluate the clinical value of different magnetic resonance imaging (MRI) sequences for a realtime thermo-monitoring during laser-induced thermotherapy (LITT) in kidneys. Methods: Twenty-eight ex vivo pig kidneys were treated with laser ablation under MR guidance in a high-field MR scanner (Magnetom Espree or Avanto Fit, Siemens, Germany). For the thermal ablation of the kidney, a neodymium yttrium-aluminum-garnet (Nd:YAG) laser was used in combination with a special protective catheter (length 43 cm, 4 French) which is sealed at the distal end. First, ablation was performed for 7, 10, and 13 minutes using FLASH sequences for investigation of time-dependent growth of lesion size. In the second step, we evaluated the optimal imaging sequence during a 7 minutes ablation of the kidney and after cooling using four different MR sequences (Haste, FLASH, radial VIBE, and Caipirinha DIXON). Results: Macroscopic lesion volume increased from 3,784 1,525 mm3 to 7,683 5,756 mm3 after the ablation from 7 to 13 minutes and MR volume ranged from 2,107 1,674 mm3 to 2,934 1,549 mm3 after the ablation from 7 to 13 minutes. During ablation, FLASH (132 34%) and radial VIBE (120 43%) sequences displayed lesion volumes most efficiently with a trend to overestimation. The Caipirinha DIXON (323 24%) sequence overestimated the volumes significantly during real-time monitoring. The volumes measured by MRI with FLASH (61 30%), Haste (67 28%), or radial VIBE (48 14%) sequences after cooling of the kidney after ablation were always underestimated. The Caipirinha DIXON (142 2%) sequence still overestimated the lesion volume after cooling of the kidney. Conclusion: LITT is a feasible ablation modality in kidney tissue. Moreover, macroscopic and MR lesion volume increases time-dependently. For online monitoring, radial VIBE and FLASH sequences seem to be most efficient. Lasers Surg. Med. 46:558–562, 2014. ß 2014 Wiley Periodicals, Inc. Key words: laser ablation; LITT; MRI; real-time monitoring; renal cancer; thermometry ß 2014 Wiley Periodicals, Inc.
INTRODUCTION The number of detected renal cell carcinomas (RCCs) has been rising due to advances in imaging [1]. Today, nephron-sparing surgery has become the treatment of choice for solitary kidney tumors smaller than 7 cm in diameter in order to preserve renal function. The oncological outcome has been proved to be comparable to radical tumor nephrectomy [2]. However, surgery may not be applicable in high-risk patients with bad performance status due to multiple comorbidities. In these patients, alternative treatment options are necessary to assure local tumor control [3]. Image-guided local ablative treatments like magnetic resonance (MR)-guided laser-induced thermotherapy (LITT), microwave coagulation, radiofrequency ablation (RFA), or cryoablation are used frequently as well-tolerated renal-sparing treatments with curative potential. The currently most widely used techniques are RFA and cryoablation. Today, laser ablation is routinely used for the treatment of liver tumors with excellent local tumor control. LITT is usually performed under MR guidance, which offers an excellent soft tissue resolution and the possibility to document heat propagation during the ablation with thermo-sensitive sequences [4]. Although it is likely that laser treatment is just as effective in ablating RCCs as in hepatic malignancies, experience in renal ablation is very limited and only few studies have been reported on the treatment of RCC with LITT [1,5,6].
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported. Correspondence to: Dr. Stefan Vallo, MD, University Hospital Frankfurt, Goethe University of Frankfurt, Theodor-Stern-Kai 7, 60590 Frankfurt am Main, Germany. E-mail: [email protected] Accepted 13 May 2014 Published online 6 June 2014 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/lsm.22262
MR-GUIDED LASER-INDUCED THERMOTHERAPY
Generally, the success of thermal ablation depends on tumor size, tumor location, and perfusion of the tumor and its surrounding structures. A main challenge one faces in the kidney is the cooling effect of vessels and the high blood flow, which may influence a sufficient thermal ablation. In this study, we compared different MR sequences for realtime monitoring during MR-guided LITT for renal treatment under standardized conditions.
was measured on an Advantage Workstation (GE Healthcare, Barrington, IL) using standard digital measurement techniques. After ablation, the kidneys were cut along the longitudinal plane passing through the axis of the electrode. On this plane, the longest coagulation diameter (Dx) parallel to the catheter as well as the coagulation diameter Dy perpendicular to Dx were measured with a tape measure. The area of the ablation zone was calculated using the formula of the rotational ellipsoid:
MATERIALS AND METHODS For this study we used 28 kidneys of slaughtered pigs. Measurements were performed not longer than 24 hours after slaughter and kidneys were stored at 48C until treatment. Prior to the experiment, the kidneys were warmed up and fixed in a water basin (378C). For the laser ablation, a Nd:YAG laser (Dornier MedLas 5100) was used which was positioned outside the scanner room. The laser energy was delivered into the MR room through fibers over 10 m in length. Laser light was emitted at an effective distance of 30 mm. A special protective catheter (Somatex, Berlin, Germany, length 43 cm, 9 French) from a laser application kit was placed orthogonally to the organ surface and advanced at least 3.5 cm into the tissue at the upper and lower pole of the kidney. The power laser application system allows permanent cooling with water and prevents carbonization at the tip of the applicator, increasing the volume of coagulation necrosis. During the ablation, permanent cooling was achieved through an integrated roller pump with a flow rate of 60 ml/ minute. MR measurements were performed with a conventional high-field 1.5 T system (Magnetom Espree or Avanto Fit; Siemens, Erlangen, Germany) using the integrated spine coil. In the first step, time-dependent lesion volume growth was analyzed. Laser energy (30 W, continuous pulse) was administered for 7 (n ¼ 9), 10 (n ¼ 16), or 13 (n ¼ 13) minutes. To monitor temperature changes, a thermosensitive T1-weighted gradient-echo sequence (FLASH2D) (repetition time [TR]/echo time [TE]: 100/4.8 milliseconds; flip angle [FA] 908, field of view [FOV] 200 mm; matrix 162 250; slice thickness 5 mm) was used. In the second step, 7-minute ablations (n ¼ 16) were monitored using T2-weighted Haste sequences (TR/TE: 140/12 milliseconds; FA 808, FOV 350 mm; matrix 128 200; slice thickness 8 mm, acquisition time 8 seconds), thermo-sensitive T1-weighted gradient-echo sequence (FLASH-2D) (TR/TE: 100/4.8 milliseconds; FA 908, FOV 200 mm; matrix 162 250; slice thickness 5 mm, acquisition time 15 seconds), newly developed T1-weighted radial VIBE sequences (TR/TE: 5/2.05 milliseconds; FA 158, FOV 250 mm; matrix 256 256; slice thickness 4 mm, acquisition time 25 seconds) or Caipirinha DIXON sequences (TR/TE: 6.8/4.8 milliseconds; FA 108, FOV 200 mm; matrix 195 320; slice thickness 3 mm, acquisition time 9 seconds) to detect the most appropriate sequence for real-time MR monitoring of laser ablation ex vivo. Measurements were repeated every 30 seconds during the ablation and after cooling. All images were analyzed and the lesion dimension
559
V¼
4p 3
2 Dx Dy 2 2
For statistical analysis, Student’s t-test was used and a P-value of or below 0.05 was considered to be significant. RESULTS There were no technical problems during the LITT procedure. Due to the brighter color of the ablated tissue, the laser lesions were easy to distinguish macroscopically inside the renal parenchyma. In the renal pelvis, however, the generally light tissue color caused difficulties when trying to differentiate macroscopically between treated and untreated tissue. In the first step, lesion volume showed a time-dependent increase, macroscopically as well as in MR volumetry. The mean macroscopic ablation volumes ranged from 3,784 2,107 mm3 after 7 minutes ablation to 7,682 2,934 mm3 after 13 minutes ablation. The ablation volumes, measured with the FLASH sequence, ranged from 2,107 1,674 mm3 (P ¼ 0.08) after 7 minutes ablation to 2,934 1,549 mm3 (P ¼ 0.04) after 13 minutes ablation (56% of the measured macroscopic size after 7 minutes and 38% after 13 minutes ablation time) (Table 1). Because of the better prediction of lesion size after 7 minutes of ablation compared to 13 minutes, we performed another ablation series for 7 minutes in a second step. On the basis of signal loss during the ablation, a realtime monitoring was possible in three (FLASH, radial VIBE, and Caipirinha DIXON) of the four tested sequences. During the laser ablation, the induced thermal lesions presented as hypointense round lesions with an increasing signal loss in a time-dependent manner. Real-time monitoring with Haste sequences was not feasible since signal changes during laser ablation were not visible (Fig. 1). With the new Caipirinha DIXON sequence, lesion size was significantly overestimated by far which corresponded to
TABLE 1. Mean Lesion Volumes After Laser Ablation Measured Macroscopically vs. MR Volumes (FLASH Sequence)
Time (minute)
Macroscopic volume (mm3)
MR volume (mm3)
7 10 13
3,784 1,525 6,920 3,825 7,682 5,756
2,107 1,674 2,868 1,166 2,934 1,549
560
VALLO ET AL.
down using FLASH, Caipirinha DIXON, and radial VIBE sequences. In contrast, the lesions showed a hypointense signal after cooling in the Haste sequence. In the Haste, FLASH and radial VIBE sequences, the estimated mean lesion sizes measured by MR after cooling were smaller compared to the macroscopic measurements, while they were larger in the Caipirinha DIXON sequence. The measured volume corresponded to 61 30% (P ¼ 0.001) (Haste), 67 28% (P ¼ 0.01) (radial VIBE), 47 14% (P ¼ 0.001) (FLASH), or 142 2% (P ¼ 0.001) (Caipirinha DIXON) of the macroscopic lesion volume (Fig. 3). DISCUSSION
Fig. 1. Comparison of four different MRI sequences during laser ablation of a porcine kidney. The ablation was performed at the lower pole of the kidney. Images were made every minute after beginning of the ablation starting from the top left (A: Caipirinha DIXON, B: FLASH, C: Haste, D: radial VIBE).
323 24% of the macroscopic size (macroscopic lesion size 4,392 651 mm3 vs. MR lesion size 14,224 3,517 mm3) after 7 minutes of ablation (P ¼ 0.003). In contrast, a satisfactory real-time monitoring of the lesion size was possible with the FLASH (132 34%; P ¼ 0.40) (macroscopic lesion size 5,788 980 mm3 vs. MR lesion size 7,651 2,604 mm3) and radial VIBE (120 43%; P ¼ 0.38) (macroscopic lesion size 8,940 2,348 mm3 vs. MR lesion size 10,731 4,627 mm3) sequences, showing a tendency to overestimation in size (Fig. 2). Next, we compared the four different MR sequences after a cooling period of 1 minute. Hereby, we observed a signal change of the ablation zone to hyperintensity after cooling
Fig. 2. Real-time monitoring of lesion size in renal tissue measured with FLASH, radial VIBE, or Caipirinha DIXON MR sequences. P < 0.05.
Today, about two-thirds of all detected renal masses are found accidentally during diagnostic imaging such as ultrasonography, computed tomography (CT), or magnetic resonance imaging (MRI) [7]. As a consequence, an approximate threefold increase in the diagnosis of small renal masses (
