A Novel Computed Tomography-Ultrasound Image Fusion Technique for Guiding the Percutaneous Kidney Access Xiaobo Shen1,2,3*, Kaiwen Li1,2,3*, Zhenyu Wu1,2,3, Cheng Liu1,2,3, Hao Yu1,2,3, Cong Lai1,2,3, Zhuang Tang1,2,3, Kui- qing Li1,2,3, Kewei Xu1,2,3 Purpose: To describe the feasibility of computed tomography (CT)-ultrasound image fusion technique on guiding percutaneous kidney access in vitro and vivo. Materials and Methods: we compare CT-ultrasound image fusion technique and ultrasound for percutaneous kid- ney puncture guidance by using an in vitro pig kidney model. The fusion method, fusion time, ultrasound screening time, and success rate of puncture were compared between the groups. Next, patients with kidney stones in our hospital were randomized in the study of simulated puncture guidance. The general condition of patients, fusion method, fusion time, and ultrasound screening time were compared between the groups. Results: A total of 45 pig models were established, including 23 in the CT-ultrasound group and 22 in the ultra- sound group. The ultrasound screening time in the CT-ultrasound group was significantly shorter than that in the ultrasound group (P < .001). In addition, the success rate of puncture in the CT-ultrasound group was significantly higher than that in the ultrasound group (P =.015). Furthermore, in the simulated PCNL puncture study, baseline data including age, BMI, and S.T.O.N.E score between the two groups showed no statistical difference. The ultra- sound screening time of the two groups was (2.60 ± 0.33) min and (3.37 ± 0.51) min respectively, and the differ- ence was statistically significant (P < .001). Conclusion: Our research revealed that the CT-ultrasound image fusion technique was a feasible and safe method to guide PCNL puncture. Compared with traditional ultrasound guidance, the CT-ultrasound image fusion tech- nique can shorten the learning curve of PCNL puncture, improve the success rate of puncture, and shorten the ultrasound screening time. Keywords: kidney stone; percutaneous nephrolithotomy; computed tomography; ultrasound; puncture INTRODUCTION Kidney stone is one of the most common urolog-ical diseases around the world. Due to the high stone free rate (SFR), percutaneous nephrolithotomy (PCNL) has been introduced for the treatment of pa- tients with large kidney stones (> 20 mm)(1). The out- comes of PCNL are highly related to the accuracy of percutaneous kidney puncture in the targeted calyx(2,3). Appropriate percutaneous renal access can guarantee the effectiveness and safety of PCNL and reduce the risk of complications (4). This challenging step can be accomplished by ultra- sound and fluoroscopy, both of which have their short- comings that may increase the risk of complications (5,6). It is well known that computed tomography (CT) scan is the gold-standard modality for the diagnosis of kid- ney stones and preoperative planning of PCNL due to its high sensitivity and specificity, precise stone sizing, and the feasibility of evaluating non-stone pathologies (7). Therefore, we proposed to integrate the preoperative CT images into intraoperative ultrasound images to im- 1Department of Urology, Sun Yat-sen Memorial Hospital, Sun Yat-sen University, Guangzhou, Guangdong, P. R. China. 2Guangdong Provincial Key Laboratory of Malignant Tumor Epigenetics and Gene Regulation, Sun Yat-sen Memorial Hospital, Sun Yat-sen University, Guangzhou, Guangdong, P. R. China. 3Guangdong Provincial Clinical Research Center for Urological Diseases. *Equal contribution *Correspondence: ? Department of Urology, Sun Yat-sen Memorial Hospital, Sun Yat-sen University, Guangzhou, Guangdong, P. R. China. Received October 2022 & Accepted February 2023 prove the accuracy of percutaneous kidney puncture. CT-ultrasound image fusion technique has shown ad- vantages in the treatment of other diseases. Xu et al.(8) have reported that 92 patients with malignant liver tu- mors underwent radiofrequency ablation (RFA) under the guidance of CT-ultrasound fusion image, proving that fusion technique assisted RFA was a safe and ef- fective option. As a result, our research aimed to apply the CT-ultrasound image fusion method to guide per- cutaneous kidney puncture and expected to increase the efficiency of stone removal and reduce the risk of complications. Therefore, the research explored the feasibility of CT-ultrasound image fusion method in establishing an appropriate renal channel. First, we utilized an in vitro pig kidney model to study the feasibility of CT-ultra- sound fusion in establishing a percutaneous kidney ac- cess. Next, we performed the simulated CT-ultrasound image fusion on patients with kidney stones, further ex- ploring the feasibility of CT-ultrasound fusion method in guiding PCNL puncture. Urology Journal/Vol 20 No. 4/ July-August 2023/ pp. 208-214. [DOI:10.22037/uj.v20i.7465] ENDOUROLOGY AND STONE DISEASES MATERIALS AND METHODS In vitro pig kidney model For the in vitro study, a well-established model was ap- plied to modulate the percutaneous kidney puncture(9). The porcine kidneys and chicken carcasses were bought from commercial slaughterhouse. We incised the renal pelvis and inserted 2-3 artificial stones (10-20 mm) into the calyces through the incision. The prepared pig kid- ney was placed inside the eviscerated chicken carcass, and filled the remaining space with agar. After the agar was fixed, all the openings of the chicken carcass were closed with sutures to simulate a confined space (Fig- ure 1). A total of 45 models were established, including 23 in the CT-ultrasound group and 22 in the ultrasound group. Patient selection Patients in our hospital from December 2019 to Febru- ary 2021 were enrolled in the simulated PCNL puncture study, who were randomly divided into CT-ultrasound group and ultrasound group with the help of SAS soft- ware. The inclusion criteria were as follows: (1) renal pelvis and upper or middle calyx stones with a diameter of more than 2 cm; lower renal calyx stones with a diam- eter of more than 1.5 cm; (2) available urinary CT data; (3) patients were voluntarily participated and signed an informed consent. Exclusion criteria were patients with anatomical structure or location of the kidney, such as horseshoe kidney, polycystic kidney, pelvic heterotopic kidney, etc. Informed consent was taken from all eli- gible patients. This study was approved by the ethics committee of our hospital, whose ethical number was 2020-KY-021, and has been registered in ClinicalTri- als. gov (NCT04645472). Equipment and Techniques The urinary CT image data were acquired using the Op- tima 64-multidetecto helical CT scanner (GE Health- care, Waukesha, WI) 1 month or less before the pro- cedure. All images were obtained with 1.25-mm-thick sections and a 1:1 pitch. The CT-ultrasound image fu- sion was achieved by the Real-time Virtual Sonography (RVS) method using the ARIETTA 70 ultrasound sys- tem (Hitachi Aloka Medical Ltd., Tokyo, Japan, (Fig- ure 2). Prior to the CT-ultrasound image fusion, the DI- COM volume data of the kidney were loaded onto the ARIETTA 70 ultrasound system, and the magnetic gen- erator was placed next to the working area. The image fusion of ultrasound and CT was executed sequential- ly by the RVS method by the professional ultrasound technician. In this step, we need to find out an image fusion region, which can be well visualized on both CT and ultrasound images, we can adjust the corresponding ultrasound image to match the CT plane through the im- age fusion region. Establishment of percutaneous kidney access in the pig kidney model Pig kidney models were randomly divided into the CT-ultrasound group and ultrasound group. We found that CT and ultrasound image can be fused using the long axis of the kidney and stones in the CT-ultrasound group. After the fusion was completed, the ultrasound probe was used to investigate the sonographic charac- teristics of hydronephrosis, stones, renal cortex, and medulla of the porcine kidney so to determine the tar- geted calyx. We punctured the targeted calyx from the fornix of the calyx. It was defined as a successful punc- ture when the puncture needle passed through the fornix to reach the targeted calyx after the kidney was opened (Figure 3). In the ultrasound group, we used tradition- al ultrasound for screening and puncture. Finally, the puncture results of the CT-ultrasound group and the ul- trasound group were compared. Simulated PCNL puncture in patients We randomly divided the selected patients into the CT-ultrasound group and ultrasound group. In the CT-ultrasound group, according to the previous study (8) and our existing experience, we using the long axis of the kidney and stones as the image fusion region to match the corresponding ultrasound image and the CT plane, since right kidney is adjacent to liver, the liver portal system can also be the image fusion region in right kidney (Figure 4). The targeted renal calyx and simulated puncture site were determined based on the stone location, hydronephrosis and adjacent organs. Notably in the CT-ultrasound group, we first deter- mined the learning curve of the CT-ultrasound fusion technique. After the fusion time and ultrasound screen- ing time reached the plateau, the CT-ultrasound group and ultrasound group were compared. Statistical analysis SPSS 20.0 software was utilized for statistical analysis. The data were expressed as the mean ± standard devi- ation or median ± range, and qualitative variables were expressed as the rate. The independent t test was ap- plied to compare quantitative variables between the two groups. Normality and homogeneity of variance were also checked. The qualitative variables were compared using the Chi-square test, when no expected cell count less than 1 and at most 20% of expected cell counts less CT-ultrasound versus ultrasound for percutaneous kidney access-Shen et al. Variables a CT-ultrasound group Ultrasound group P-value Sample size, n 22 23 - Chicken weight, kg; mean ± SD 3.44 ± 0.23 3.52 ± 0.28 .326 Length of pig kidney, cm; mean ± SD 12.44 ± 0.75 12.45 ± 0.94 .951 Width of pig kidney, cm; mean ± SD 4.75 ± 0.38 4.72 ± 0.43 .846 Number of artificial stones, n 2 (2~3) 2(2~3) .661 Image fusion method, n - Long axis of kidney and kidney stones - 23 Fusion time, min; mean ± SD 4.02 ± 0.54 - - Ultrasound screening time, min; mean ± SD 2.52 ± 0.31 3.38 ± 0.50 < .001 Puncture depth, cm; mean ± SD 5.99 ± 0.71 5.76 ± 0.80 .306 Puncture success rate, n; (%) 19(86.4) 11(47.8) .015 Table 1. Characteristics of CT-ultrasound and ultrasound group in pig kidney model Abbreviations: CT, computed tomography a Continuous variables were compared by independent samples t-test Vol 20 No 4 July-August 2023 209 than 5. A value of p < .05 was considered statistically significant. RESULTS The detailed results of the in vitro study were shown in Table 1. A total of 45 models were established, in- cluding 23 in the CT-ultrasound group and 22 in the ultrasound group. There were no statistically significant differences in chicken weight, the length and width of kidney between the two groups. The ultrasound screen- ing time in the CT-ultrasound group was significantly shorter than that in the ultrasound group (P < .001). In Variables a CT-ultrasound group Ultrasound group P-value Sample size, n 52 53 - Gender, n; (%) .359 Male 34 (65.4) 29 (54.7) Female 18 (34.6) 24 (45.3) Laterality, n; (%) 1 Right 28 (53.8) 29 (54.7) Left 24 (46.2) 24 (45.3) Age, year; mean ± SD (range) 54.15 ± 11.65 (24-81) 52.74 ± 11.49 (43-82) .532 BMI, kg/m²; mean ± SD (range) 23.69 ± 3.98 (16.89-41.32) 23.37 ± 3.33 (18.07-32.47) .653 Previous ipsilateral surgery history, n; (%) 10 (19.2%) 10(18.9%) 1 Stone size, mm; median (IQR) 19.2 (22) 21.1 (17.6) .354 Tract length, mm; mean ± SD 82.79 ± 15.24 80.15 ± 13.73 .354 Obstruction or hydronephrosis, n; (%) .828 None or mild hydronephrosis 37 (71.2) 37 (69.8) Moderate hydronephrosis 10 (19.2) 9 (17.0) Heavy hydronephrosis 5 (9.6) 7 (13.2) Number of involved calyces, n 2.40 ± 1.16 2.33 ± 0.91 .385 Stone density, HU; mean ± SD 1122.71 ± 371.29 1022.32 ± 323.91 .143 S.T.O.N.E score (range) 7 (5-10) 7 (5-10) .356 Table 2. Demographic characteristics of CT-ultrasound and ultrasound group in simulated puncture guidance Abbreviations: BMI, Body Mass Index; IQR, Interquartile Range a Continuous variables were compared by independent samples t-test Figure 1. Irregular tumor edge of renal cell carcinoma in contrast-enhanced CT (A) A mass with smooth margin and prominent nodules from part of it; (B) A mass with blurred margin; (C) A mass with completely irregular and non-elliptical shape; (D) Renal sinus compression in contrast-enhanced CT CT-ultrasound versus ultrasound for percutaneous kidney access-Shen et al. Endourology & Stone Diseases 210 addition, the success rate of puncture in the CT-ultra- sound group was significantly higher than that in the ultrasound group (P = .015). A total of 150 patients were included in the simulated PCNL puncture study, of which 97 cases in the CT ul- trasound group and 53 cases in the ultrasound group. We first determined the learning curve of the CT-ultra- sound image fusion method, indicating that the screen- ing time of the CT-ultrasound group reached a plateau after 31-45 cases. Therefore, the last 52 patients in the CT-ultrasound group and 53 patients in the ultrasound group were included in the comparison. There was no statistical difference in baseline data between the two groups (Table 2). The fusion time in the CT-ultrasound group was (4.27 ± 0.56) min. Eighteen cases of the right kidney were fused with the portal venous system; 10 cases were fused with the long axis of the kidney and stones; and all cases of the left kidney were fused with the long axis of the kidney and stones. The ultrasound screening time of the two groups was (2.60 ± 0.33) min and (3.37 ± 0.51) min, respectively, and the difference was statistically significant (P < .001, Table 3). DISCUSSION In 1976, Fernstrom et al.(10) first reported the experi- ence of removing kidney stones through percutaneous nephrostomy, which had been widely used since the early 1980s. Currently, PCNL is recommended for the treatment of upper urinary tract stones with a diameter > 2cm and complex kidney stones owing to its high ef- ficiency and minimal invasiveness. Establishing percu- taneous renal access is the most critical step in PCNL. Incorrect puncture can easily damage the blood vessels, thereby increasing the risk of complications such as bleeding and renal function damage(11). Therefore, how to accurately puncture the targeted calyx has become the biggest problem for PCNL. Rassweiler et al.(12) has reported an iPad-assisted tech- nique with a three-dimensional reconstruction and marker tracking method for kidney puncture. Before the surgery, patients underwent a preoperative CT in the prone position with 5 colored metal markers around the puncture site. The CT images were segmented and analyzed to establish a three-dimensional construc- tion and adjacent organ anatomy on an iPad. During Variablesa CT-ultrasound group Ultrasound group P-value Sample size, n 52 53 - Image fusion region, n; (%) - Portal vein 18 (34.6) - Long kidney axis and stones (right kidney) 10 (19.2) - Long kidney axis and stones (left kidney) 24 (46.2) - Fusion time, min; mean ± SD 4.27 ± 0.56 - - Ultrasound screening time, min; mean ± SD (range) 2.60 ± 0.33 (2.0-3.2) 3.37 ± 0.51 (2.2-4.3) < .001 Simulated puncture site, n; (%) .331 Above the 12th rib 15 (28.8) 10 (18.9) Below the 12th rib 37 (71.2) 43 (81.1) Target calyx, n; (%) .818 Upper calyx 3 (5.8) 2 (3.8) Lower calyx 2 (3.8) 3 (5.6) Middle calyx 47 (90.4) 48 (90.6) Table 3. Outcomes of CT-ultrasound and ultrasound group in simulated puncture guidance aContinuous variables were compared by independent samples t-test Figure 1. Establishment of percutaneous kidney access with the in vitro pig kidney model. (A) Preparation of in vitro pig kidney model. (B) Establishment of in vitro pig kidney model. (C) Suture of the openings. (D) CT scan of the model. (E) Ultrasound screening before puncture. (F) Percutaneous kidney puncture with the model. CT-ultrasound versus ultrasound for percutaneous kidney access-Shen et al. Vol 20 No 4 July-August 2023 211 the operation, the integrated image was matched with the real-time puncture guidance through the iPad, and the simulated image of the puncture pathway was displayed on the screen. The limitations of this meth- od were the additional CT examination and operation time. The Sonix-GPS system was another technology designed to track needle positioning under ultrasound guidance(13). This technology applied electromagnetic tracking to identify the needle location and display the planned puncture access. A prospective comparative study reported that the Sonix-GPS system can improve the success rate of a single puncture without increasing the risk of complications(14). However, the ultrasound image was greatly affected in obese patients. In addi- tion, the Uro-Dyna-CT system reported by Ritter et al. (15) was a modified angiography device that allowed 3D reconstruction of CT images and showed the exact pathway chosen for puncture. This device offered a 3D anatomical image with the characteristics of safe, fast, and accurate. The disadvantages included high radia- tion exposure, high costs, and difficulty in learning. In the research, we demonstrated an innovative tech- nique for kidney puncture. The RVS technology com- bined the images of ultrasound and CT, allowing the surgeon to puncture with both ultrasound and CT im- ages simultaneously. CT images can provide the infor- mation including the diameter and location of kidney stones, the anatomy of the renal pelvis and calyxes, and the refined construction of surrounding organs. During the puncture process, the surgeon can refer to the ultra- sound and CT images at the same time, thereby increas- ing the accuracy of the percutaneous kidney access. The image fusion technology can improve the effectiveness of stone removal of the percutaneous access and reduce the damage to adjacent organs without increasing the risk of radiation exposure and complications. The im- age fusion technology can be of great help to beginners because of the high success rate of the CT-ultrasound group, which was operated by a urological intern in the whole study. In addition, our results verified the effec- tiveness of this technology through an in vitro pig kid- ney model and simulated CT-ultrasound fusion study. However, there are some limitations in our research. First, our study lacked a description of the intraoperative application of this technique. At present, our center has tested this technique during surgery, whose results will be published in another manuscript. Second, after the fusion is completed, slight movements of the patient's body, such as breathing activity, can easily reduce the accuracy of the fusion image. During the puncture pro- cess, the anesthesiologist can decrease the patient's tidal volume to reduce the impact of breathing activity on the fusion image. Finally, the fusion image cannot be used for the second puncture, as the first puncture changes the structure of the kidney and stones. CONCLUSIONS Our research demonstrated an innovative CT-ultra- Figure 2. The overview of the ARIETTA 70 ultrasound system. (A) ARIETTA 70 ultrasound system. (B) Probe CT-ultrasound versus ultrasound for percutaneous kidney access-Shen et al. Endourology & Stone Diseases 212 sound image fusion technique for percutaneous kidney puncture. The in vitro study revealed that compared with the traditional ultrasound guidance, the CT-ultra- sound fusion imaging technique can shorten the ultra- sound screening time, and improve the success rate of puncture. The simulated CT-ultrasound fusion research suggested that this technology can shorten the learning curve and ultrasound screening time. In summary, the image fusion assisted percutaneous kidney puncture ap- peared to be safe and effective for PCNL. CONFLICT OF INTEREST The authors declare that they have no competing inter- ests. This work was funded by grants from the National Natural Science Foundation of China (Grant numbers: 81572511, 81702525, 81702528), Guangzhou Science Figure 3. The CT-ultrasound fusion images and puncture outcomes with the in vitro pig kidney model Figure 4. Simulated CT-ultrasound image fusion. (A) Simulated CT-ultrasound fusion on the right kidney; (B) Simulated CT-ultrasound fusion on the left kidney; (C) CT-ultrasound fusion image on the right kidney; (D) CT-ultrasound fusion image on the left kidney CT-ultrasound versus ultrasound for percutaneous kidney access-Shen et al. Vol 20 No 4 July-August 2023 213 and Technology Program key projects (Grant numbers: 201803010029), Natural Science Foundation of Guang- dong Province (Grant numbers: 2016A030313317), Medical Scientific Research Foundation of Guangdong Province (Grant numbers: C2018060), and Yixian Clin- ical Research Project of Sun Yat-sen Memorial Hospi- tal (Grant numbers: sys-c-201802). REFERENCES 1. Turk C, Petrik A, Sarica K, et al. EAU Guidelines on Interventional Treatment for Urolithiasis. Eur Urol. 2016;69:475-82. 2. Michel MS, Trojan L, Rassweiler JJ. Complications in percutaneous nephrolithotomy. Eur Urol. 2007;51:899-906; discussion 3. Patel RM, Okhunov Z, Clayman RV, Landman J. Prone Versus Supine Percutaneous Nephrolithotomy: What Is Your Position? Curr Urol Rep. 2017;18:26. 4. de la Rosette JJ, Laguna MP, Rassweiler JJ, Conort P. Training in percutaneous nephrolithotomy--a critical review. Eur Urol. 2008;54:994-1001. 5. Rodrigues PL, Rodrigues NF, Fonseca J, Lima E, Vilaça JL. Kidney targeting and puncturing during percutaneous nephrolithotomy: recent advances and future perspectives. J Endourol. 2013;27:826-34. 6. de la Rosette J, Kashi AH, Farshid S. Newer Advances in Access. In: Agrawal MS, Mishra DK, Somani B, eds. Minimally Invasive Percutaneous Nephrolithotomy. Singapore: Springer Singapore; 2022:145-55. 7. Brisbane W, Bailey MR, Sorensen MD. An overview of kidney stone imaging techniques. Nat Rev Urol. 2016;13:654-62. 8. Xu ZF, Xie XY, Kuang M, et al. Percutaneous radiofrequency ablation of malignant liver tumors with ultrasound and CT fusion imaging guidance. J Clin Ultrasound. 2014;42:321-30. 9. Hacker A, Wendt-Nordahl G, Honeck P, Michel MS, Alken P, Knoll T. A biological model to teach percutaneous nephrolithotomy technique with ultrasound- and fluoroscopy- guided access. J Endourol. 2007;21:545-50. 10. Fernstrom I, Johansson B. Percutaneous pyelolithotomy. A new extraction technique. Scand J Urol Nephrol. 1976;10:257-9. 11. Hajiha M, Baldwin DD. New Technologies to Aid in Percutaneous Access. Urol Clin North Am. 2019;46:225-43. 12. Rassweiler JJ, Müller M, Fangerau M, et al. iPad-assisted percutaneous access to the kidney using marker-based navigation: initial clinical experience. Eur Urol. 2012;61:628- 31. 13. Li R, Li T, Qian X, Qi J, Wu D, Liu J. Real- time ultrasonography-guided percutaneous nephrolithotomy using SonixGPS navigation: clinical experience and practice in a single center in China. J Endourol. 2015;29:158-61. 14. Li X, Long Q, Chen X, He D, He H. Real- time ultrasound-guided PCNL using a novel SonixGPS needle tracking system. Urolithiasis. 2014;42:341-6. 15. Ritter M, Rassweiler MC, Michel MS. The Uro Dyna-CT Enables Three-dimensional Planned Laser-guided Complex Punctures. Eur Urol. 2015;68:880-4. CT-ultrasound versus ultrasound for percutaneous kidney access-Shen et al. Endourology & Stone Diseases 214