IN VITRO STUDY OF ULTRASOUND BASED REAL-TIME TRACKING OF RENAL STONES FOR SHOCK WAVE LITHOTRIPSY: PART 1

Transcription

1 /01/ /0 THE JOURNAL OF UROLOGY Vol. 166, 28 32, July 2001 Copyright 2001 by AMERICAN UROLOGICAL ASSOCIATION, INC. Printed in U.S.A. IN VITRO STUDY OF ULTRASOUND BASED REAL-TIME TRACKING OF RENAL STONES FOR SHOCK WAVE LITHOTRIPSY: PART 1 C. C. CHANG,* S. M. LIANG,* Y. R. PU,* C. H. CHEN,* I. MANOUSAKAS,* T. S. CHEN,* C. L. KUO,* F. M. YU* AND Z. F. CHU* From the Department of Urology, ESWL Laboratory, and Department of Electrical Engineering, Institute of Biomedical Engineering and Institute of Aeronautics and Astronautics, National Cheng Kung University and Hospital, Tainan, Taiwan, Republic of China ABSTRACT Purpose: We developed a real-time tracking system for renal stones that decreases the number of shock waves and treatment time of shock wave lithotripsy. Materials and Methods: Ultrasound images were analyzed to identify the renal stones. A computer software module for ultrasound image processing was developed to monitor stone location instantaneously. Another computer software module controlled generator movement in real time for tracking the stone. We used 13 ultrasound stone trajectories recorded from patients to test the system in vitro. Two tests were established to verify tracking system reliability. One test focused on improvement in the coincidence ratio, which denotes the matching extent of the stone within the effective focal area. The other test focused on improvement in the efficiency ratio, that is a decrease in the number of shocks for stone fragmentation. For each recorded stone trajectory 2 experiments were done under tracking and nontracking conditions. Results: The average coincidence and efficiency ratios plus or minus standard deviation were 79.6% 9.8% and 45.0% 12.7% without tracking, and 97.0% 3.0% and % with tracking, respectively. All tests were statistically significant (paired t test p 0.01). Conclusions: An ultrasound based real-time tracking system proved to be significantly helpful for in vitro lithotripsy. It appears that the tracking system may greatly decrease the number of shocks and treatment time for renal stones. KEY WORDS: kidney, kidney calculi, ultrasonography Since the advent of extracorporeal shock wave lithotripters in the 1980s, the treatment of urinary stones has greatly changed. 1 During lithotripsy for a kidney stone, it is desirable to disintegrate the stone by accurately focusing shock waves on it, resulting in effective disintegration. However, renal stone movement is inevitable and may be as much as 7 cm. due to patient breathing, causing inaccurate shock wave focusing. 2 Only about 30% to 50% of shock waves may be accurately focused on the calculus. 3, 4 Consequently the renal parenchyma may be damaged by peak pressure when shock wave focusing is inaccurate. Although respiration may be used for controlling stone movement, it is inconvenient. To improve these defects we developed an ultrasound based real-time stone tracking system on a lithotripter equipped with an ultrasound system. With this tracking system shock waves may be accurately focused on a moving renal stone, resulting in effective fragmentation. MATERIALS AND METHODS Accepted for publication February 2, Supported by National Science Council Grants NSC E R, NSC E and NSC E and Lite- Med Co. * Financial interest and/or other relationship with LiteMed Co. 28 Beginning in June 1997 patients with 0.6 to 1 cm. renal stones were enrolled in this study. Initially before shock wave lithotripsy was performed without any sedation or anesthesia the renal stone was identified using a 3.5 mhz. Hitachi EUB-315 ultrasound system (Hitachi, Tokyo, Japan). When the stone moved with the kidney, the stone movement was followed in its long axis to maintain it always within the ultrasound images. The images were then recorded continuously on video tape for 8 minutes to document stone movement. These recorded trajectories were used in our experiments. Stone detection technique. An automatic gray level threshold method was developed to detect renal calculi that appeared on ultrasound as a higher gray level intensity. 5 This method was based on gray level histogram entropy. 6 The threshold was applied in a preselected region of interest that moved to follow the stone in real time. Frame matching technique. Often a kidney moved in and out of the image plane. In such cases the ultrasound image of the stone section may have shown not only changes in stone position, but also changes in shape and size, making it difficult to identify the stone. To overcome this difficulty a technique of subsequent frame matching was developed. Stone trajectory recording module. An image processing program to analyze the continuous ultrasound images was developed. From these images the sequences of and changes in stone centers were computed and used as data inputs for an especially designed computer software module. This module was used to control the servomotor that moved the shock wave generator. The computer used had a Pentium II-350 processor (Intel Co., Santa Clara, California). Real-time stone tracking system. Displacement of the stone center in 2 subsequent frames was measured and transmitted to the stone tracking system. Under system directions the shock wave generator was moved from its current position to a new position in real time by the servomotor. However, stone movement was limited to the designated area or region of interest enclosing the calculus. If the stone moved out of the designated area, the tracking system did not function. When this situation occurred, it was necessary to move the designated area to enclose the stone, so that the tracking system would operate. This real-time tracking system was

2 IN VITRO TRACKING OF RENAL STONES FOR SHOCK WAVE LITHOTRIPSY 29 implemented on a LiteMed 9200 lithotripter (LiteMed Co., Taipai, Taiwan) to verify its reliability. Tests. We used the recorded stone trajectories for certain experiments. Coincidence Test: Coincidence tests were performed to evaluate the accuracy of targeting the stone. An effective focal area was considered an area with a radius of 6.5 mm. from the second geometric focus of the shock wave reflector. This area consisted of a 3.5 mm. portion on which pressure was greater than 200 bars and a 3 mm. portion of the stone radius. 7 9 With N coincidence representing the number of images in which the stone center was within the effective focal area and N total representing the total number of processed images of a given stone trajectory we measured the extent to which the stone center was within the effective focal area or coincidence ratio (C) using the formula, C N coincidence /N total 100%. Thus, the coincidence ratio represents the percent of the number of images on which the stone center was within the effective focal area compared with the total number of processed images. 10 A computer controlled simulator was used to imitate stone movement according to the recorded stone trajectories. These trajectories were then used to calculate the coincidence ratios under tracking and nontracking conditions. Stone Fragmentation Efficiency Test: An artificial calculus was made from water and plaster-of-paris into a cylindrical block shape. The weight ratio of water-to-plaster was 1:1.8. This 6 6 mm. artificial stone was placed in the pocket of a filtering mesh with a 2 mm. grid. The mesh was connected to a computer controlled simulator that moved the artificial calculus according to the trajectories. Figure 1 shows the in vitro setup for this efficiency test. A wooden rod was affixed to the other side of the simulator. When the stone was moved by the simulator, the rod also moved synchronously in parallel with the stone. Therefore, the rod movement equaled the stone movement. We performed ultrasound to detect and track the rod movement and, thus, the stone movement. Shock waves were administered to the stone at the rate of 60 shocks per minute, which was maintained throughout each experiment. The number of shock waves needed to disintegrate the calculus was counted. Shock wave firing was stopped when there were no more stone fragments left on the filtering mesh under visual inspection. One artificial stone was used per experiment. The number of shock waves needed to disintegrate the moving stone (N moving ) was almost that of an immobile stone (N fixed ) centrally positioned within the effective focal area. Thus, the efficiency ratio (E) of stone fragmentation was calculated using the formula, E N fixed / N moving 100%, as previously reported. 10 The efficiency ratio was then calculated from recorded stone trajectories under tracking and nontracking conditions. Statistics. Experimental results were analyzed by the paired t test. The null hypothesis states that the mean of the tracking and nontracking pairs would be 0 for the coincidence and efficiency tests. Commercially available computer software was used to perform statistical analysis with 0.05 with 2 tails. We recorded and used 13 stone trajectories in each of the 2 series of tests. RESULTS Coincidence test results. Figure 2 shows representative tracking and nontracking results of 1 stone trajectory. Without tracking many stone center positions were outside of the effective focal area (fig. 2, A). However, with real-time tracking most stone center positions were within the effective focal area (fig. 2, B). Table 1 lists the computed values of the coincidence ratio of the 13 stone trajectories. Trajectory values without tracking were 65.0% to 95.1% (mean plus or minus standard deviation 79.6% 9.8%). Values with tracking were 91% to 100% (mean 97% 3%). In the former group each value was lower than that of the corresponding stone trajectory in the latter group. The lower the coincidence ratio without tracking, the greater the improvement in the corresponding coincidence ratio with tracking. The average value in the tracking group was about 1.2-fold that in the group without tracking. The difference in these 2 groups was statistically significant (t 7.08, p ). Figure 3 shows the coincidence ratios of the 13 stone trajectories. The mean value with and without tracking was 97% and 79.6%, respectively. The slope of the line of about 1.2 represents the average improvement in the coincidence ratio using the tracking system. Efficiency test results. Table 2 lists the results. An average of shocks was needed to disintegrate an immobile stone, while were needed for a moving stone without tracking and were needed for a moving stone with tracking. Efficiency ratios in the nontracking and tracking groups were 17.3% to 63.8% (mean 45.0% 12.7%) and 70.5% to 98.9% (mean 85.5% 6.8%), respectively. For each stone trajectory the efficiency ratio without tracking was less than that with tracking. The lower the efficiency ratio without tracking, the higher the corresponding improvement in the efficiency ratio in with tracking. There was a 1.9-fold average improvement in the efficiency ratio. The difference in these 2 groups was statistically significant (t 11.19, p ). Figure 4 shows the efficiency ratios of 13 stone trajectories. The mean value with and without tracking was 85.5% and 45%, respectively. The slope of the line of about 1.9 represents the average improvement using the tracking system. FIG. 1. In vitro setup for efficiency test. Assembly consisting of artificial stone in pocket of filter mesh was connected to simulator. Wooden rod 3 mm. in diameter was also affixed to simulator. When stone was moved by simulator, rod was also moved synchronously in parallel with stone. Therefore, rod movement equaled stone movement. We used ultrasound to detect rod movement and, thus, stone movement. Considerable amount of air bubble smoke covering whole stone area was produced by shock waves in water, making it impossible to identify stone by ultrasound. Servo-tracking mechanism moved generator with water tank to track and fired shock waves onto moving stone. DISCUSSION An important feature of the real-time tracking system is real-time analysis of ultrasound images. These digitalized data were used as data input into especially developed computer software for controlling the instantaneous movement of the shock wave generator at a maximal range of 4 cm per second. Thus, it was possible to track the moving stone in real time, resulting effective treatment. No matter how the renal calculus moved in each ultrasound frame, the real-time tracking system instantaneously moved the shock wave generator to the new stone position. This type of tracking is different than that of other systems. 3, 10, 11 In our tracking

3 30 IN VITRO TRACKING OF RENAL STONES FOR SHOCK WAVE LITHOTRIPSY FIG. 2. Stone trajectory. A, without real-time tracking. Horizontal lines indicate range of effective focal area. B, with real-time tracking. TABLE 1. Computed coincidence ratios of 13 trajectory files under tracking and nontracking conditions Stone Trajectory No. % Coincidence Ratio Without Tracking With Tracking Mean SD system the shock wave generator was moved at a rate of about 10 times per second, which was the rate of our ultrasound image processing. At this rate the moving stone was tracked in real time. The effective peak pressure of the shock wave was accurately applied to the stone and not to the renal parenchyma. The average coincidence ratio without tracking was 79.6% 9.8%, which was much higher than the previously reported value of 44%. 10, 11 This result was due to the fact that when we used an ultrasound probe to scan the renal stone, it was always necessary to follow the stone. Thus, the average value of the coincidence ratios without tracking was higher than in other reports. However, the average value of 97% 3% with tracking was also higher than the 83% previously reported. 10, 11 The average efficiency ratio without tracking was 45.0% 12.7%, which increased to 85.5% 6.8% after we used the tracking system. Thus, the efficiency ratio improved an average of about 1.9-fold, which was much better than in other studies. 10, 11 Obviously the pressure applied to the stone determines the degree of stone fragmentation. The coincidence ratio of each stone trajectory with tracking was higher than that without tracking, indicating that the stone center was more effectively positioned in the effective pressure region. As a result, the stone was more easily disintegrated because of the improved coincidence ratio. This conclusion is supported by the improved efficiency ratio of each stone trajectory during the tracking trial. Because of the improved coincidence ratio, the renal stone has a higher probability of receiving effective peak pressures. In other words, the renal tissue around the stone may receive more pressure injury from the shock waves than other regions of the kidney. However, with accurate targeting on the stone the renal tissue farther from the stone would have less FIG. 3. Comparison of coincidence ratios of 13 stone trajectories under tracking and nontracking conditions. Each circle represents 1 pair of coincidence values per stone trajectory with and without tracking. Black circle represents average values for no tracking and tracking of 76.9 and 97%, respectively. Diagonal line indicates that values with and without tracking are equal. Circle above diagonal line represents improved targeting by tracking system. Broken line with slope of about 1.2 represents average improvement in coincidence ratio using tracking system. pressure injury from shock waves. Moreover, it is expected that the number of shock waves needed to disintegrate the stone would be greatly decreased. Thus, overall kidney injury and its associated side effects may be greatly decreased. The real-time tracking system functioned at a rate of about 10 times per second, meaning that the tracking system tracks and targets the stone about 10 times per second. With the improved average coincidence ratio of 97% the tracking system has a 97% probability of success for targeting the stone at a rate of about 10 times per second. If a shock wave generator fires about 10 shock waves per second, stone treatment time would be greatly decreased. Ultrasound is a convenient device for detecting renal stones. However, it has some limitations. The ability to detect renal stones by ultrasound depends on stone size and location. 12, 13 When the stone is small, it may be confused with calcification spots and high echo-density kidney structures. Therefore, it is important that a skillful and experienced physician should localize the stone. When a stone is fragmented and these fragments remain in the same calix, ultrasound may localize the stone. 12 If stone fragments became smaller and displaced away

4 Stone Trajectory No. IN VITRO TRACKING OF RENAL STONES FOR SHOCK WAVE LITHOTRIPSY 31 TABLE 2. Efficiency ratios of 13 stone trajectories with and without real-time tracking N fixed N moving Without Tracking % Efficiency Ratio N moving With Tracking % Efficiency Ratio Mean SD from the original calix, ultrasound probably may not be used to localize or track a stone any more. 13 In this case we would change from ultrasound to x-ray localization. Meanwhile, in most cases ultrasound cannot recognize to what extent a stone is fragmented. X-ray imaging is needed for that purpose. Thus, ultrasound has many limitations during extracorporeal shock wave lithotripsy. Nevertheless, the capability of ultrasound would improve the effects of extracorporeal shock wave lithotripsy, we may consider accepting its use. Our real-time tracking system was effective for calculi that were localized and then tracked. Beyond these conditions x-ray imaging should be used to assess treatment progress. At the end of this study we decided to adjust the stone position by moving the table rather than moving the generator. We developed new computer software and incorporated it with the previously developed software. The whole computer software was then equipped into a newly designed LM 9200 Plus lithotriptor (LiteMed Co.), which is a table moving or adjusting model. Thus, the newly developed computer software may easily be equipped on other types of lithotripters to improve the effect of shock wave lithotripsy. CONCLUSIONS The ultrasound based real-time tracking system for renal stones proved to be an effective device for shock wave lithotripsy in our in vitro study. The tracking system increased the accuracy of targeting the renal stone and increased the efficiency of stone fragmentation. Thus, the system may decrease the number of shock waves for treating renal stones. For more evidence to prove the effectiveness of this system we have performed an animal test with the results to be reported in part 2. Drs. Y. H. Yang, Y. C. Tong, G. S. Hsey, U. S. Tzai and Y. H. Liu contributed to this study. REFERENCES FIG. 4. Comparison of efficiency ratios of 13 stone trajectories with and without tracking. Each circle represents 1 pair of efficiency values per stone trajectory with and without tracking. Black circle represents mean values for no tracking and tracking of 45% and 85.5%, respectively. Circle above diagonal line represents improved efficiency of stone fragmentation using tracking system. Broken line with slope of about 1.9 represents average improvement using tracking system. 1. Chaussy, C., Schmiedt, E., Jocham, D. et al: First clinical experience with extracorporeally induced destruction of kidney stones by shock waves. J Urol, 127: 417, Suramo, I., Paivansalo, M. and Myllyla, V.: Cranio-caudal movements of the liver, pancreas and kidneys in respiration. Acta Radiol Diagn, 25: 129, Kuwahara, M., Kambe, K., Taguchi, K. et al: Initial experience using a new type extracorporeal lithotripter with an antimisshot control device. J Lithotr Stone Dis, 3: 141, Devevey, C.: Etude du mouvement sur des sequences d images echographiques: poursuite de cibles rigides et segmentation par une approche connexionniste du champ des vitesses d objets deformables. Ph.D. thesis no. 93 ISAL 0006, INSA de Lyon, France, p. 94, Manousakas, I., Chen, C. H., Chen, I. H. et al: Real time tracking of renal stones in ultrasound images for extracorporeal shock wave lithotripsy. In: Proceedings of Bio-Medical Engineering Symposium, Taipei, Taiwan, Republic of China, December 18-19, p. 83, Kapur, J. N., Sahoo, P. K. and Wong, A. K. C.: A New method for gray-level picture thresholding using the entropy of the histogram. Comput Vis Graphics Image Process, 29: 273, Dai, S. B.: Performance evaluation of extracorporeal shock wave lithotripter. Thesis, Institute of Aeronautics and Astronautics, National Cheng-Kung University, Republic of China, Tainan, Taiwan, p. 20, Bostrom, U.: An overview of different ESWL-systems. Scand J Urol Nephrol, suppl., 122: 73, Coleman, A. J. and Saunders, J. E.: A survey of the acoustic output of commercial extracorporeal shock wave lithotripters.

5 32 IN VITRO TRACKING OF RENAL STONES FOR SHOCK WAVE LITHOTRIPSY Ultrasound Med Biol, 15: 213, Orkisz, M., Farchtchian, T., Saighi, D. et al: Image based renal stone tracking to improve efficacy in extracorporeal lithotripsy. J Urol, 160: 1237, Orkisz, M., Bourlion, M., Gimenez, G. et al: Real-time target tracking applied to improve fragmentation of renal stones in extra-corporeal lithotripsy. Machine Vis App, 11: 138, Middleton, W. D., Dodds, W. J., Lawson, T. L. et al: Renal calculi: sensitivity for detection with US. Radiology, 167: 239, Baumgarter, B. R., Steinberg, H. V., Ambrose, S. S. et al: Sonographic evaluation of renal stones treated by extracorporeal shock-wave lithotripsy.ajramjroentgenol, 149: 183, 1987