Application of Three-dimensional Visualization assisted Ultrasound for Percutaneous Renal Puncture

We present here the three-dimensional (3D) visualization fused with ultrasound and to evaluate its clinical application effect preliminarily. One hundred and eighteen patients with renal calculi in our hospital from September 2017 to December 2019 were prospectively randomized into two groups. The experimental group was treated with percutaneous renal puncture guided by the 3D visualization fused with ultrasound. The control group was treated with percutaneous renal puncture guided by B-ultrasonography (B-US). The puncture time, operation time, and the loss of hemoglobin in the experimental group were lower than those in the control group, and the success rate of establishing the channel at one time, and the coincidence rate between the channel and the longitudinal axis of the target renal calyx were higher. The stone clearance rate in the experimental group was higher, while the postoperative blood transfusion rate and interventional embolization rate in the experimental group were signicantly lower. The difference was statistically signicant (P < 0.05). 3D visualization assisted ultrasound could preliminarily guide precise navigation to puncture tissues, reduced operation time and bleeding, improved stone clearance rate and operation safety.


Introduction
The most important technique during percutaneous nephroscopy is precise percutaneous renal puncture. 1 The ideal route for this intervention entails passing through the axis of the targeted renal calyx, which minimizes vascular injury around the calyx and obtains the largest renal endoscopy angle. 2,3 At present, the main techniques for guiding percutaneous renal puncture are plain radiography and B-ultrasonography (B-US). 4,5 Because radiography and B-US provide only two-dimensional (2D) images, it is impossible to guide the puncture through the axis of the targeted renal calyx that more-precise three-dimensional (3D) imaging could provide, leading to a less than accurate percutaneous renal puncture and an increased risk of bleeding. Medical 3D visualization technology can process 2D computed tomography (CT) images using a computer, converting the 2D images into intuitive, perspective 3D images to display the 3D shapes of human organs and tissues. One of the important functions of 3D visualization is that it can be marked and segmented on digital 3D images, which could be used for navigation during surgery. Therefore, using 3D visualization technology in this study, we preliminarily realized precise navigation for renal puncture through the axis of the targeted renal calyx.

Patients
After obtaining informed consent, the patients were assessed for eligibility. 118 patients (72 men, 46 women; average age 42.4 ± 6.6 years; average body weight 66.5 ± 10.6 kg) with renal calculi > 3 cm in volume were selected during September 2017 to December 2019 to participate in this study. There was no hydronephrosis in 47 patients, mild hydronephrosis in 38, and moderate hydronephrosis in 33. All patients met the criteria for undergoing percutaneous nephrolithotomy. All patients were randomly divided into two groups: experimental group (n = 63) and control group (n = 55). The 3D visual navigation system assisted B-US-guided percutaneous renal puncture was used in the experimental group and conventional B-US-guided puncture was used in the control group. There was no signi cant difference in age, body weight, stone volume and degree of hydronephrosis between the two groups (P > 0.05). All methods were carried out in accordance with relevant guidelines and regulations. This study protocol was approved by the Research Ethics Committee of Jiangsu Taizhou People's Hospital.

Establishment of 3D visualization model
With plain radiography for guidance, a homemade, grid-shaped body surface patch (10.0 × 6.0 cm), with each square 1.0 × 1.0 cm, was place on, and adhered to, the common area for percutaneous renal puncture (Fig. 1)-i.e., in the area of the inferior rib om the affected side and posterior axillary line.
All patients then underwent thin computed tomography scanning in the lateral position. Spiral, 256-slice CT was used to scan the upper and middle abdomen, with a slice thickness of 0.5 mm. Structural images of the kidney were recorded on plain scans in the arterial, venous, and secretory phases of the renal pelvis. The 2D grayscale Digital Imaging and Communication in Medicine image data obtained from CT scans were input into Mimics software for 3D reconstruction. Image segmentation was carried out using threshold segmentation, the region growth method, and Boolean function and was modi ed by a multi-layered manual segmentation method. Renal calculi, spine, ribs, and the gridded body surface patch were segmented and reconstructed according to the plain scanning images. The renal contour was then segmented and reconstructed according to the arterial phase images. The renal pelvis secretory phase images were used to segment and reconstruct the collection system. After registration, the model was de-noised, smoothed, and carefully lled. Finally, the 3D visual model of the spinal column, partial rib, grid-shaped body surface patch, renal collecting system, and renal calculi were reconstructed (Fig. 2).

3D visual puncture navigation technology
First, using the segmentation tool of Magics software, the line between the vertex of the fornix and the midpoint of the calyx was drawn from any side of the model of the targeted renal calyx, and section A (perpendicular to the side of the renal calyx) was drawn along that line ( Second, we identi ed the posterior targeted calyx and the adjacent posterior calyces on the 3D visual model of the renal collection system. The axial section C, segregated along the axis of the targeted renal calyx and as close as possible to the adjacent posterior calyx, was the optimal simulated US-guided section of the targeted calyx that could maximally show the relation between the targeted calyx and the targeted pelvis and their surrounding calyces (Fig. 5). Section C crossed the grid to form a tangent that passed through the puncture point. According to the tangent and the position of the puncture point on the grid, the line and puncture point could be accurately restored on the patient's skin ( Fig. 6A-C). The guiding lines on both sides of the puncture point are the ideal position of the ultrasonic probe during the puncture (Fig. 6D). Finally, a ureteral catheter was retrogradely inserted into the affected renal pelvis of the patient through the ureter inferior to the calculus position, and arti cial hydronephrosis was created by slowly injecting normal saline through the ureteral catheter. In the lateral position on the healthy side, the head or tail of the B-US probe was pressed on the puncture point, swinging the probe after placing it along the location line. The probe was then xed when the 2D US image of the renal pelvis and renal calyx was the same or similar to the image of section C (Fig. 7A,B). The needle was then inserted from the puncture point of the head or tail of the B-US probe. The puncture line of the needle that passed through the central line of the targeted calyx in the 2D US image was the axis of the targeted renal calyx (Fig. 7C).
Each patient underwent the CT examination and surgery in a standard lateral position, and each was examined or punctured at the end of inspiration to avoid complications due to kidney displacement.
The puncture time was recorded, and the changes in hemoglobin levels before and after the operation were compared. On the third postoperative day, contrast agent was injected into the percutaneous renal drainage tube, and the patient underwent CT scanning in the lateral position, and the 3D structures of the renal pelvis, renal calyx, and drainage tube were reconstructed. The drainage tube represented the puncture route. Whether the puncture route passed through the axis of the targeted renal calyx was observed. The puncture passage passes through the target renal calyceal nipple and the skin puncture point is judged to be consistent with the target renal calyx longitudinal axis in the 12 mm around the intersection of the renal calyx longitudinal axis and the skin.

Observation indicators
The operation time, puncture time, hemoglobin loss, the success rate of one-time establishment of the channel, the coincidence rate between the channel and the longitudinal axis of the target renal calyx were compared between the two groups, and the stone clearance rate, blood transfusion rate and interventional embolization hemostasis rate were observed.
Statistical methods SPSS 17.0 statistical software was used for data processing, measurement data were expressed by x±S, T test was used for comparison between groups, percentage was used for counting data, x2 test was used for comparison between groups. P < 0.05 indicates that the difference is statistically signi cant.

Results
Under the guidance of the 3D visualization technique, the targeted renal calyces of the patients were punctured quickly and successfully in accordance with the puncture points and routes located before operation. Intraoperative observation showed that the puncture route entered the targeted renal calyx from the center of the papilla of the targeted renal calyx, and the puncture route faced the renal calyx opening directly. Postoperative 3D reconstruction showed that the puncture channels passed through the axis of the targeted renal calyx (Fig. 8).
The puncture time, operation time and the loss of hemoglobin in the experimental group were lower than those in the control group, and the success rate of establishing the channel at one time and the coincidence rate between the channel and the longitudinal axis of the target renal calyx were higher than those in the control group, and the difference was statistically signi cant (P < 0.05, Table 1). The stone clearance rate in the experimental group was higher than that in the control group, while the postoperative blood transfusion rate and interventional embolization rate in the experimental group were signi cantly lower than those in the control group, and the difference is statistically signi cant. (P < 0.05, Table 2).

Discussion
Puncture along the axis of the targeted renal calyx is based on the same principle used to establish a percutaneous renal passage during percutaneous nephrolithotomy, which is derived from the anatomical structure of the kidney and blood vessels. 3 Cone-like pyramidal structures are found in the medulla of the kidney. The bottom of the pyramid is toward the cortex, and the tip of the pyramid forms the papilla. which inserts into the bottom of the conical calyx. The renal segmental artery and interlobar artery are shaped around the renal calyx and renal vertebrae. The axis passing through the targeted renal calyx also passes through the axis of the adjacent renal pyramidal body, which is the optimal anatomical route with less vascular distribution. 6 Meanwhile, the maximum angle and operating range of renal enzoscopy can be reached throughout the puncture route in the axis of the renal calyx. 6 In clinical practice, ensuring that the percutaneous renal puncture is routed precisely through the axis of the targeted renal calyx is di cult. 7 It requires an accurate location of the central axis of the 3D conical renal calyx and precise, simultaneous navigation in three dimensions (X, Y, Z). At present, the commonly used 2D guidance methods (i.e., plain radiography, B-US) cannot accurately locate the skin puncture point or route, [8][9][10] which is the most signi cant technical obstruction to using percutaneous renal puncture.
The radiography-guided puncture route and the axis of the targeted renal calyx are always in the same longitudinal plane, although these two lines cannot be guaranteed to overlap in this plane. It often takes multiple puncture points, multiple angles, and repeated attempts to penetrate the targeted renal calyx, and it often deviates from the axis of the targeted renal calyx, resulting in injury and bleeding of the blood vessels around the renal vertebrae and renal calyx, which limits the scope of nephroscopy and affects the e ciency of stone removal.
Although US guidance can show the targeted renal calyx at multiple angles and monitor the puncture needle through the center line of the 2D section of the targeted renal calyx, 7 it is di cult to determine whether the 2D section selected by US contains the axis of the targeted renal calyx, which is also the geometric axial section of the pyramidal targeted renal calyx. Especially for a targeted renal calyx with hydronephrosis, which deforms into an oval body or sphere, it is even more di cult to locate the axial section of the targeted renal calyx using US. Thus, it could be concluded that, with US guidance, it is often impossible to pass the axis of the targeted renal calyx accurately, resulting in injury to, and bleeding from, the blood vessels around the renal vertebrae and renal calyx, as well as allowing only limited intraoperative nephroscopic movement.
Accurate target percutaneous renal calyx longitudinal puncture has always been the biggest challenge in percutaneous renal surgery. In recent years, a variety of percutaneous renal puncture navigation systems have emerged, including electromagnetic positioning navigation 11 and iPad augmented reality navigation 12 and so on. The above techniques can achieve fast and safe puncture in the operation to some extent, but their disadvantages are also signi cant. IPad augmented reality auxiliary system needs complex equipment, software and professional personnel. In addition, due to the movement of renal position caused by respiratory activity during the operation, there is a certain deviation between the augmented reality image and the actual position of the patient, while the iPad augmented reality auxiliary system does not have real-time position tracking of the puncture needle, which may cause the puncture needle to deviate from the target puncture line. Electromagnetic positioning and navigation technology also requires complex electromagnetic equipment and professionals, and the exible ureteroscope needs to be placed into the target renal calyx in advance. Ureteral stricture and kidney stone obstruction will limit its use. In addition, because there is no real-time monitoring, it is di cult to ensure that the adjacent organs of kidney are not damaged during the puncture. The above shortcomings limit the clinical application of these techniques, resulting in their exploratory application in only a small number of patients.
The fusion of virtual 3D images and real-time ultrasound images may be the development direction of percutaneous renal puncture navigation in the future. Based on the above assumption, we designed 3D visualization technology combined with real-time ultrasonic monitoring to navigate percutaneous renal puncture more precisely via three standardized steps for the rst time. First, we calculated and marked the geometric axis of the targeted renal calyx and then marked the optimal puncture point on the skin using the grid surface patch. Second, we measured and segmented the optimal simulated guided section image and matched the tangent precisely on the skin using the grid location patch. This tangent line is the optimal placement position of the US probe on skin. Third, we moved the US probe around the puncture point and along the localization tangent. By fusing and comparing the observed sections of the collecting system with the optimal simulated guiding cut section segmented using 3D visualization technology, we determined the optimal direction of the US probe. Throughout these steps, the navigational information gained by 3D visualization technology is fused with real-time guidance of B-US, thereby successfully taking advantage of the rapid, stereoscopic, accurate percutaneous renal puncture navigation.
In this study, because we accurately combine the positioning function of 3D visual image with the real-time ultrasonic monitoring function, we overcome the shortcomings of 3D image and ultrasonic localization, and realize the accurate location of skin puncture point and the accurate guidance of puncture path. First of all, the postoperative puncture channel is highly in line with the longitudinal axis of the target renal calyx, which meets the requirements of accurate percutaneous renal puncture. Secondly, on the basis of accurate puncture, compared with the standard B-U-guided puncture, the loss of hemoglobin, postoperative blood transfusion rate and interventional embolization rate in the 3D visualization navigation puncture group decreased signi cantly, and the stone clearance rate increased to 90.48%. In addition, although the steps of 3D visual navigation puncture are complex, it does not take a long time for each patient after skilled operation, and most of the work is completed before the operation, which greatly shortens the puncture time and operation time. Also, this technology only needs simple software operation learning, and does not need complex equipment. It is very easy to popularize and apply.

Conclusions
In summary, 3D visual navigation fused with ultrasound can help overcome the limitations of current conventional 2D guidance methods, as it preliminarily realized accurate navigation of puncture through the axis of the targeted renal calyx. The clinical application effect was good, and its further clinical use will verify its accuracy.

Declarations
Competing interests: