Evaluating the learning curve for flexible ureteroscopic thulium fiber laser lithotripsy using CUSUM analysis

Abstract

Objectives:

Effective training methods and learning curve (LC) assessment are crucial for more difficult endoscopic procedures. The present study sought to assess the LC of flexible ureteroscopes (fURS) for thulium fiber laser (TFL) lithotripsy and determine the effectiveness of using a porcine kidney model in training modality, to inform future training.

Methods:

Postgraduate medical students without experience in ureteroscopy were recruited, and surgical data were collected from 10 consecutive flexible ureteroscopic lithotripsy (fURL) procedures performed on our homemade porcine kidney training model. Cumulative sum (CUSUM) analysis and pooled mean CUSUM curves were applied to determine the LC turning points (TPs), and pre-LC and post-LC data were compared with that of an experienced attending physician.

Results:

Data from 110 surgeries were included in the analysis. The mean number of surgical units to overcome the LC for the duration of fURS TFL lithotripsy was 4. The operation time, number of tissue damage, and number of central visual shifts were significantly lower in students after the TP than before. No statistically significant difference in operation time was found between the students after the LC (817.50 [703.75-964.75]) and the surgeon group (732.50 [51.00-822.25]).

Conclusion:

Overcoming the LC of fURS TFL lithotripsy requires four surgeries, and the developed ex vivo porcine kidney is a conveniently accessible and effective clinical training model.

Introduction

Urinary tract stones are one of the most common conditions. The ureteroscopic lithotripsy (URL) has become one of the preferred treatments for urinary tract stones requiring surgical intervention due to the favorable development and maturation of urologic endoscopic techniques and the strong guideline recommendations^[1,2]. The latest European Association of Urology (EAU) and American Urological Association guidelines recognize the potential of flexible ureteroscopes (fURS), expanding its applicability in the treatment of renal stones^[1,2]. Moreover, the continuous improvement in ureteroscopic equipment and laser technology has increased the application of flexible URL^[3]. The latest emergence of thulium fiber laser (TFL) is gaining increasing attention in lithotripsy. Clinical and preclinical studies have demonstrated the advantages of TFL in lithotripsy, particularly in terms of operation time, intraoperative complications, ablation efficiency, stone retropulsion, stone fragmentation, and intraoperative visualization^[4,5]. Furthermore, according to Kronenberg and Traxer^[6] and Traxer and Corrales^[7], the better vision means that TFL is more easily operated by less experienced users, thereby reducing the learning cycle.

Although URL is an endoluminal procedure with relatively few complications, it still presents significant risks such as ureteral perforation, bleeding, and mucosal injury^[8]. The necessity for repeated practical training and standardized learning cannot be overstated^[9]. Nevertheless, during their clinical tenure, young physicians are compelled to master these skills potentially at the cost of patient care. Many hospitals have recently established high-technology surgical skills centers for simulation-based training (SBT) to expedite the acquisition of clinical surgical skills among young physicians and minimize complications. SBT allows novice surgeons to become familiar and proficient in the challenging endoscopic surgery in a low-risk and low-pressure environment. This approach significantly shortens the learning curve (LC) for young surgeons^[10,11]. The current training simulators primarily consist of bench models, virtual reality (VR) models, and animal or cadaveric models. These models differ significantly in their fidelity and their associated costs^[12]. Benchtop simulators are constrained by their reduced tissue haptic fidelity and limited anatomical complexity, rendering them appropriate for the instruction of basic skills^[13]. VR simulators, despite their instant performance feedback, are impeded by their substantial cost and unreliable haptics^[13]. High-fidelity animal models, exemplified by porcine renal and ureteral models, are distinguished by their high face validity. Nonetheless, these models are costly, designed for single use, and may engender ethical considerations^[14].

The cumulative summation (CUSUM) method, first introduced by Page in 1954, has been applied in various fields, including surgical LC assessment, and is currently one of the most commonly used methods for measuring surgical LCs^[15–17]. As a control chart statistical method, the CUSUM curve can accurately judge the changes in the LC by calculating the cumulative total difference between the observed values and the target values of the observed indicators during the skill acquisition process^[18,19]. In addition, the CUSUM curve has good visualization and can monitor surgical performance through asymptotes and curve turning points (TPs)^[20]. Currently, CUSUM is used to monitor the learning process of surgical procedures in robotic surgery^[21], magnetic resonance imaging-targeted biopsy^[22], and laparoscopic single-site robot-assisted radical prostatectomy^[23]. The ex vivo porcine kidney model offers the advantage of closely replicating real-world surgical conditions while streamlining preoperative and postoperative procedures. Porcine kidneys are genetically and anatomically closer to humans, providing a more realistic representation of the upper urinary tract anatomy. In addition, ex vivo porcine kidneys are readily available and cost-effective.

Thus, the present study used disposable ureteroscopes (also known as single-use flexible ureteroscopes) combined with the latest TFL lithotripsy technology to perform stone surgeries on a homemade ex vivo pig kidney model. The primary objective was to investigate the LC of fURS TFL lithotripsy using the CUSUM method. The second objective was to assess whether the ex vivo pig kidney model we used is suitable as an effective and readily available training model. Young doctors were trained in in vitro techniques and effective models to acquire competence in this procedure. Young doctors may prevent surgical errors and complications by gaining the required skills before starting clinical procedures through efficient skills training and LC analysis.

Materials and methods

Participants

Nine postgraduate medical students in the field of urology were voluntarily recruited from Zhongnan Hospital of Wuhan University. The inclusion criteria were as follows: (1) participants who had completed a urology theory course and have a basic understanding of urinary tract anatomy and endoscopic surgery; (2) those with a strong interest in urological surgery, especially ureteroscope (URS) and fiber optic laser lithotripsy; and (3) those who participated in the entire training and evaluation period. The exclusion criteria were as follows: (1) participants with clinical experience in URS or any other endoscopic procedure and (2) those with medical conditions that may interfere with their ability to perform procedures, such as severe hand tremors or visual impairment. In addition, two urology-attending physicians with clinical experience were invited to participate as references. The two urologists could independently execute laser lithotripsy on their own as they had finished their residency training, with more than 3 years of clinical work experience.

Before each student performed the surgery, they received training from experienced urology attending physicians. This training included understanding the use of the fURS, its maneuverability, and the activation of the TFL. In addition, students were allowed to practice repeatedly in vitro. After an initial understanding, the students inserted the URS through ureteral access sheath (UAS) into the renal model. Under the guidance of attending physicians, the macroscopic deflection of the URS was adjusted with the dominant hand to find direction within the kidney, with the ability to autonomously enter the upper, middle, and lower calyces and each calyceal of the porcine kidney as the criterion for the end of the training.

All participants completed 10 fURS TFL lithotripsy surgeries according to standardized protocols. All participants were de-identified. Informed consent was obtained from each participant.

Experimental equipment

Homemade pig kidney model

Before the experiment started, 110 fresh porcine renal units were sequentially obtained from the abattoir, selecting those from ternary pigs weighing between 110 and 120 kg, which were slaughtered for purposes other than this experiment. The fresh porcine renal units were of similar size, and a 12Fr dilator in the minimally invasive dilation and drainage kit was used to retrogradely enter the renal pelvis from the pig ureter. After reaching the upper pole of the kidney, the guide needle was inserted to penetrate the renal pelvic mucosa and renal parenchyma to reach the surface of the pig kidney and then the guide needle was removed. The outer sheath (working channel width 8.4 mm), inner sheath, and obturator of the resectoscope (Wuhan Tangji Technology Limited Corporation, Wuhan, China) were assembled and joined with the 12Fr dilator to enter the renal pelvis. The dilator, the inner sheath, and the obturator of the resectoscope were removed, and the fURS was placed in the ureter. Under direct vision of the fURS, BegoStones with a diameter of 6 mm were placed from the working channel of the outer sheath of the resectoscope into the position of the middle calyx of the kidney. The leak was sutured throughout the entire process to prevent leakage of liquid during the experiment (Fig. 1).

Figure 1.

Establishment of a pig kidney model.

Spherical artificial stones (BegoStones) with a diameter of 6 mm were used for the experiment. The BegoStones were prepared as described in our previous experiments^[24,25]. In addition, an artificial pelvic model was designed and produced. These pelvic models were constructed using relatively inexpensive polystyrene as the primary material, offering good impact resistance and thermal insulation properties. Surgical sutures and hot-melt adhesive were used to fix the simulated human body silicone inside the pelvic model and at the pelvic outlet, allowing the UAS to freely enter and provide stability while preventing any leakage during the procedure, thus ensuring uninterrupted surgical progress.

Subsequently, the entire fresh porcine renal unit, with BegoStones in place, was appropriately secured to the artificial pelvic model using surgical sutures and immersed in warm normal saline to minimize the compression of the renal pelvis by the upper renal parenchyma and reduce the space of the renal pelvis, which would affect the experimental results (Fig. 2). This experiment did not involve live animals.

Figure 2.

Experimental equipment and surgical procedures.

Laser settings

The TFL treatment device used was Urolase (Raykeen Laser Technology Limited Corporation, Shanghai, China, wavelength 1.9 µm, power up to 40 W, peak power 500 W), and a uniform fiber was used for all procedures, with a diameter of 272 µm (TJ17031; Tangji, Wuhan, China). The lithotripsy parameters were set to 0.3 J and 25 W.

Experimental and surgical procedures

Participants conducted experiments on the homemade pig kidney model. First, UAS was inserted through the entrance of the artificial pelvic model, entering the pelvis into the ureter, with its distal end positioned just below the ureteropelvic junction. The inner diameter of the working channel was 14 French (Fr). Participants inserted a disposable flexible ureteroscope (8.25 Fr; CoralView U10; Shenzhen Honglu Medical Technology Co., Ltd., Shenzhen, China) into the renal pelvis along the ureteroscope sheath and connected it to a peristaltic pump (Kamoer, Shanghai, China). The irrigation fluid flow rate was calibrated to 20 mL/min. Participants were first asked to perform repeated URS to identify the upper, middle, and lower parts of the renal calyx and locate the stones. After becoming proficient in the operation of the URS, the laser was turned on for lithotripsy. The powdered lithotripsy method was used as previously described^[26], and the end point of lithotripsy was when the stone fragments were less than three times the diameter of the optical fiber. Following the lithotripsy procedure, both the operator and the recorder collaboratively inspected the residual stone fragments within the renal pelvis to ensure they were no larger than three times the diameter of the optical fiber. In addition, we conducted a postoperative dissection of the porcine kidneys to examine the remaining stone fragments in the renal pelvis, defining stone clearance success as residual stone fragments less than 3 mm. Each student and the two attending physicians were required to complete 10 fURS TFL lithotripsy surgeries. When the participant performed the procedure for the first time, a recorder was used to provide corrections for any incorrect techniques used by the students during the lithotripsy process (Fig. 2).

Data collection

The operation time (s) was defined as the time from the start of lithotripsy to its completion. The central visual shift was defined as the deviation of the center of the visual field from the intended target area during surgery, including situations where the endoscope was pointed away from the center or misaligned during surgery^[27]. Tissue damage was defined as collateral tissue injuries, which referred to the number of unintended tissue burns caused by the laser during the pursuit of stone fragments. It mainly included collateral damage caused by accidental direct application of laser to the tissue^[28]. The participants assessed stone retropulsion caused by laser activation and intraoperative visibility impairment due to stone powder floating in the water during the lithotripsy process using a three-point Likert scale designed based on a previous study^[29]. The Likert scale consisted of three points: 0 = no retropulsion/clear visibility; 1 = retropulsion/decrease of visibility that does not affect procedure; and 2 = retropulsion/poor visibility that interferes with procedure. Standardized training was conducted on the operation process and observation indicators for participants before the operation to ensure that they had the same understanding of the measurement standards and minimize interobserver variability. All participants completed the operation according to the standardized protocol. Moreover, a standardized form was used during the data collection process to ensure that each step was performed according to the established procedures. Intraoperative recordings were accompanied by a researcher, and any discrepancies between recorder and operator scores were resolved by a consensus discussion with a third researcher. Data were collected from 10 programs for each student as the endpoint for plotting the LC based on a previous study^[30] and our pilot experiment^[31].

CUSUM analyses

The CUSUM technique was applied to define the LC as previously described^[32], seeking the relationship between the number of surgeries and the operation time. The number of surgical units was sorted in chronological order and CUSUM was defined as the cumulative total difference between each operation time and the mean of all operation times, namely CUSUM-OT. The CUSUM was calculated using the following formula:

$$\mathrm{C}\mathrm{U}\mathrm{S}\mathrm{U}{\mathrm{M}}_n=\sum _{i=1}^n\left(x_i-\mu \right),$$

where x is the operation time of the ith surgery and μ is the target operation time (the average operation time in our study).

The pooled mean CUSUM curve represents the average of the individual CUSUM of the same serial number group under each identical case^[33]. IBM SPSS Statistics 25 software was used for curve fitting with surgical sequence numbers as the x-axis and CUSUM-OT values as the y-axis; the closer the R2 was to 1, the better the fit. The curve was successfully fitted when P-value <0.05. The abscissa at the peak of the fitting curve was taken as the turning points (TPs). The renal model processing was divided into two groups based on these TPs: Group A (≤TPs) representing the pre-LC group, and Group B (>TPs) representing the post-LC group.

Statistical analysis

Non-normally distributed data were expressed as median and interquartile range and normally distributed data were expressed as mean ± standard deviation. Normal distribution data were statistically compared using the Student’s t-test, whereas non-normally distributed data were compared using the Mann–Whitney U-test or the Kruskal–Wallis test. All experimental data were analyzed using the IBM SPSS Statistics 25 statistical software, with statistical significance set at P-value <0.05.

Results

All participating students and attending physicians completed the required 10 surgeries each, and data from 110 surgeries were included in the analysis. No statistically significant differences in stone quality were found between the groups (P = 0.844, Table 1). Similarly, no statistically significant differences in intraoperative visibility and stone retropulsion were found between the groups, which may affect the procedure.

Table 1.

Comparison of variables between groups

Variables	Number	Intraoperative visibility, (IQR)	Stone retropulsion, (IQR)	Stone quality, g (IQR)
Student 1	10	1.00 (0.00–1.00)	0.50 (0.00–1.00)	0.09240 (0.090375–0.093425)
Student 2	10	0.00 (0.00–1.00)	0.00 (0.00–1.00)	0.09140 (0.090975–0.092875)
Student 3	10	0.50 (0.00–1.00)	1.00 (0.75–1.00)	0.09195 (0.090375–0.095450)
Student 4	10	1.00 (0.00–1.00)	0.00 (0.00–1.00)	0.09260 (0.092175–0.093800)
Student 5	10	0.00 (0.00–1.00)	0.00 (0.00–1.00)	0.09155 (0.090878–0.095900)
Student 6	10	0.00 (0.00–0.25)	1.00 (0.00–1.00)	0.09205 (0.090750–0.092325)
Student 7	10	1.00 (0.00–1.00)	0.00 (0.00–1.00)	0.09230 (0.090750–0.093375)
Student 8	10	0.00 (0.00–1.00)	0.00 (0.00–1.00)	0.09315 (0.091000–0.093875)
Student 9	10	0.00 (0.00–1.00)	0.00 (0.00–1.00)	0.09175 (0.090775–0.093250)
Surgeon 1	10	0.00 (0.00–1.00)	0.00 (0.00–0.00)	0.09230 (0.090975–0.096150)
Surgeon 2	10	0.00 (0.00–1.00)	0.00 (0.00–1.00)	0.09270 (0.091200–0.093250)
P		0.086	0.064	0.844

Learning curve

The CUSUM analysis was conducted using operation time as the main indicator. First, the CUSUM curve of each student was examined (Fig. 3a-i). The TPs were not consistent. However, all CUSUM curves first showed an upward trend until they reached the TP and then a downward and stable trend. Concurrently, a CUSUM analysis was conducted on the pooled mean (R2 = 0.991, P < 0.05). R2 = 0.991 showed that there was a very high match between the fitted model and the data, suggesting that the model did a good job of explaining the variations in operation time. The data indicated that LC peaked at the fourth surgical case. The curve rose smoothly until four examples, at which point it began to progressively drop. The LC for flexible ureteroscope TFL lithotripsy surgery was split into two phases using the fourth case as the TP. As shown in Figure 3j, there were two stages: the learning and improvement stage and the proficiency stage. After grouping based on TPs, significant decreases were observed in the total operation time (1229.00 [1045.75-1425.00] vs 817.50 [703.75-964.75], P < 0.001), the total number of tissue damage (7.83 ± 3.1288 vs 3.72 ± 1.6168, P < 0.001), and the total number of central visual shift (9.00 [6.00-13.50] vs 3.00 [2.00-4.00], P < 0.001) for each student before and after the LC (Table 2).

Table 3.

Comparison of outcomes for data pre-LC and post-LC between student and surgeon groups

	Surgeon group	Student group		Student group
Variables	(n = 20)	pre-LC (n = 40)	P	post-LC (n = 50)	P
Operation time (s)	732.50 (651.00–822.25)	1229.00 (1045.75–1425.00)	0.000	817.5 (703.75–964.75)	0.063
Number of tissue damage	2.00 (1.00–2.75)	8.00 (6.00–10.00)	< 0.001	3.50 (3.00–5.00)	0.000
Number of central visual shift	1.00 (1.00–2.00)	9.00 (6.00–13.50)	0.000	3.00 (2.00–4.00)	0.001

Figure 3.

(a-i) CUSUM curves for individual students. (j) Pooled mean CUSUM curve.

Table 2.

Outcomes before and after the turning point for flexible ureteroscopic TFL lithotripsy

Variables	Number		Operation time, s, mean ± SD/IQR		P	Number of tissue damage, number, mean ± SD/IQR		P	Number of central visual shift, number, mean ± SD/IQR		P
	A	B	A	B		A	B		A	B
Student 1	4	6	1311.75 ± 285.3073	853.83 ± 74.4726	0.005	6.25 ± 1.7078	4.00 ± 0.6325	0.017	9.00 (6.75-12.75)	2.50 (2.00-3.75)	0.012
Student 2	4	6	1592.00 ± 284.03634	1058.33 ± 220.1133	0.010	6.50 ± 1.2910	3.83 ± 1.9408	0.044	9.00 ± 4.2426	4.50 ± 1.5166	0.041
Student 3	4	6	1195.00 ± 110.8061	899.83 ± 124.2359	0.005	7.25 ± 3.7749	2.83 ± 1.3292	0.027	9.50 ± 3.3166	4.50 ± 2.5100	0.026
Student 4	5	5	1109.60 ± 272.9200	719.80 ± 65.5225	0.031	8.80 ± 3.8341	2.60 ± 1.1402	0.008	7.00 (4.00-12.50)	2.00 (2.00-2.00)	0.005
Student 5	3	7	1619.00 ± 315.5677	878.71 ± 136.0964	0.046	10.67 ± 2.5166	6.00 ± 1.6330	0.007	12.00 (12.00-16.00)	4.00 (3.00-7.00)	0.016
Student 6	4	6	1198.25 ± 281.8716	855.50 ± 151.9826	0.036	8.75 ± 3.7749	3.50 ± 1.0488	0.011	10.00 ± 5.2281	1.50 ± 1.0488	0.046
Student 7	6	4	1127.50 ± 207.8613	708.75 ± 157.2289	0.009	5.83 ± 1.9408	3.00 ± 1.4142	0.037	5.00 (3.75-10.00)	0.50 (0.00-1.75)	0.010
Student 8	4	6	1431.00 ± 350.6546	851.17 ± 237.6623	0.014	10.00 ± 3.7417	3.83 ± 0.7528	0.004	12.75 ± 5.9090	3.17 ± 1.3292	0.004
Student 9	6	4	1062.17 ± 193.4781	595.50 ± 98.8619	0.002	7.83 ± 3.3715	2.75 ± 1.7078	0.025	10.50 ± 4.5935	2.75 ± 1.2583	0.012
Total	40	50	1229.00 (1045.75-1425.00)	817.50 (703.75-964.75)	0.000	8 (6.00-10.00)	3.5 (3.00-5.00)	0.000	9.00 (6.00-13.50)	3.00 (2.00-4.00)	0.000

Comparison of outcomes for data pre-LC and post-LC between student and surgeon groups

We categorized all students into pre-LC group and post-LC group according to the TPs of LC, and the two attending physicians as the surgeon group (Table 3). We compared the above two sets of data of the student group with those of the surgeon group. The results showed that in the student pre-LC group, the student group was significantly higher than the surgeon group in terms of operation time (1229.00 [1045.75-1425.00] vs 732.50 [651.00-822.25], P < 0.001), the number of tissue damage (8.00 [6.00-10.00] vs 2.00 [1.00-2.75], P < 0.001), and the number of central visual shift (9.00 [6.00-13.50] vs 1.00 [1.00-2.00], P < 0.001). In the student post-LC group, the number of tissue damage (3.50 [3.00-5.00] vs 2.00 [1.00-2.75], P < 0.001), number of central visual shift (3.00 [2.00-4.00] vs 1.00 [1.00-2.00], P = 0.001) was still significantly higher than that of the surgeon group. However, there was no statistically significant difference in operation time between the two groups (817.50 [703.75-964.75] vs 732.50 [651.00-822.25], P = 0.063).

Discussion

Wright introduced the concept of the “learning curve” in 1936 to calculate the cost of aircraft production in the airplane manufacturing industry^[34]. It was later discovered that productivity increases significantly with time and experience^[35]. The theory of the LC is equally applicable in the medical field and is becoming increasingly important with the advent of new technologies. Page^[16] provided a detailed description of the CUSUM technique for sequential analysis testing. This technique is widely used in analyzing the LC of surgical procedures due to its ability to continuously and rapidly assess data, independent of sample size, effectively identifying subtle changes. Although it may reflect variations due to patient heterogeneity rather than changes in surgical performance, CUSUM analysis is still considered one of the best tools for quality control in the healthcare field for monitoring surgical performance and patient outcomes^[36,37]. Other analytical methods applied in LC studies are not suitable for this research, such as cumulative failure rate analysis, exponentially weighted moving average (EWMA), or risk-adjusted cumulative sums. Cumulative failure rate analysis focuses on changes in failure rates, requiring a large amount of failure rate data, whereas dichotomous data do not provide additional information^[38]. The EWMA method is somewhat lacking in the rapid detection of small changes^[39,40]. The risk-adjusted cumulative sums method focuses on adjusting risk factors related to patient outcome indicators^[41,42], which is not suitable for our study design.

This CUSUM analysis can reflect the trend and outcomes of successively conducted surgeries^[43,44]. However, studies specifically using the CUSUM method to analyze the LC of flexible URL in the field of urology are limited. By comparing the surgical index levels of postgraduate medical students before and after training with the relevant index levels of surgeons, this study applied the CUSUM method to evaluate the LC and the process of young doctors mastering in vitro fURS TFL lithotripsy. This not only provides an effective in vitro surgical training method and equipment for young doctors but also ensures that trainees can master the required surgical skills before clinical practice through in-depth skill training and LC analysis. This approach also reduces the risk of errors and complications in clinical surgery, thereby improving the overall quality of surgery and enhancing patient safety and satisfaction.

The present study found significant improvement in operative time, tissue damage, and central visual shift after approximately four training surgeries. The CUSUM value increased steadily during the learning and improvement phase, and a longer operation time reflected the learning process of the students as they mastered basic surgical skills. The long and fluctuating operation time during this phase revealed the difficulties students faced at this stage, including unfamiliarity with the surgical steps, unpredictable complications encountered during the operation, and the inability to predict and avoid complications that might increase the difficulty of the next step, such as stone migration to the lower pole. fURS could be challenging to treat lower pole stones, given the anatomical structure of the lower pole^[45–47]. The CUSUM scatter plot showed that the operation time was significantly prolonged in the third operation of Student 3 and fifth operation Student 8, which was related to the displacement of the stone during the operation. Larger stone fragments became trapped in the space between the renal pelvic mucosa and the renal papilla during lithotripsy because intervention during the early stages of stone displacement was not feasible. Due to this, the laser cannot effectively reach the stone during lithotripsy, and a lot of time must be spent adjusting the stone’s position with fURS. The CUSUM value showed a downward trend as the students gradually transitioned to the proficiency stage, which marked the students’ in-depth understanding and mastery of surgical skills. Students started to effectively plan surgical steps and reduce unnecessary complex operations, thereby improving surgical efficiency. At this stage, students not only mastered the basic skills of surgery but also learned how to predict and avoid situations that require complex operations. The CUSUM scatter plot showed that the operation time was significantly reduced in the fifth operation of Student 2 and third operation of Student 6. There were more renal pelvic mucosal folds in these two operations and the stone position was relatively fixed, which reduced the time for repeated adjustment of the stone position and focus during the operation. Furthermore, a comparison of these results with those of clinical attending physicians further validated the effectiveness of our training method.

According to current international guidelines^[1,2], fURS has become the first-line treatment option for kidney stones under 2 cm in size. With the advancement of fURS and laser technology, fURS remains effective even in larger stones with lower risks of serious complications^[48]. In addition, the advent of TFL technology has led to higher stone ablation rates and reduced stone retropulsion^[49]. Moreover, TFLs, with their smaller diameter delivery fibers, decrease the probability of fiber bending and consequent laser failure, reduce congestion in the working channel, and improve irrigation rates, enhancing visibility during laser lithotripsy^[50]. This trend has shifted clinical practice toward favoring fURS over PCNL in treating kidney stones. However, a lack of proficiency in the use of fURS in clinical practice can lead to prolonged surgical times, surgical failure, increased risk of complications, and endoscope damage^[51]. Therefore, the safe and effective performance of flexible URL requires the accumulation of practical experience and continuous training. To date, there is no standardized or clinically relevant training model for flexible URL.

The use of simulator models in medical education provides a low-stress, low-risk nonsurgical environment while offering realistic settings and training opportunities^[52]. Similar to the conclusions of Grober et al^[53], we believe that high-fidelity animal models should be the preferred choice for trainees. The use of live animals as training models necessitates complex procedures including intraoperative anesthesia and vital sign monitoring, rendering it impractical for widespread implementation^[54]. Previous research has shown that training with inanimate models and animal tissue is beneficial for learning progression^[55]. fURS has several disadvantages, such as high purchase and maintenance costs and limited reuse, which limit their use in both clinical practice and training^[56]. Disposable ureteroscopes are relatively cheap, readily available, and replaceable^[57]. Furthermore, EAU guidelines suggest that disposable ureteroscopes have the same safety and clinical effectiveness as reusable fURS^[1].

Previous studies have shown that surgeons need at least about 56-60 cases of experience to gain the necessary experience and confidence in retrograde intrarenal surgery (RIRS) and reach a plateau on the LC^[58]. Kezer and Ozgor^[59] retrospectively analyzed 80 patients with renal stones who underwent fURS and compared them based on operative time, success rate, and complication rate. The results showed that the success rate of fURS reached a satisfactory level after the 20th case. In addition, 40 cases are sufficient to improve surgical proficiency and reduce preparation time, operative time, and fluoroscopy time. Recently, Koo et al^[30] conducted an RIRS LC analysis of novice surgeons and retrospectively analyzed the data of a mentor surgeon and three novice surgeons. The results showed that the LC reached a plateau after 12-15 RIRS cases of medium-sized stones. The present study analyzed the LC of fURS TFL lithotripsy and found that the LC reached a plateau after four operations. The CUSUM curve of novice surgeons reached a plateau earlier in our study than in previous studies. This might be related to the following reasons. First, TFL lithotripsy was used for the first time in our study. TFL, a new laser, has the advantages of smaller stone fragments and improved visualization^[5]. Kronenberg and Traxer^[6] and Traxer and Corrales^[7] argued that the better vision means that TFL is more easily operated by less experienced users, thus reducing the LC. Second, our study was conducted in a model rather than a surgical environment, and this low-pressure environment may be more conducive to the operator’s operation.

Villa et al^[27] found that students trained in the use of fURS demonstrated greater proficiency with the device and reduced operation times. Consistently, our data showed that the operative time for each student was significantly lower after than before the LC. In terms of proficiency with fURS, the number of tissue damage and the number of central visual shift for each student after the LC was significantly reduced, indicating the effectiveness of our training approach. Furthermore, we found that post-LC operation times (817.50 s) of all students reached the level of clinical attending physicians (732.50 s) (P = 0.063) using our model training and the same lithotripsy method. However, the students and attending physicians exhibited significant differences in terms of the proficiency indicators of tissue damage and central visual shift with fURS. This might be attributed to the limited training and insufficient clinical experience among the students.

Strengths and limitations

Currently, surgical simulators have become a mature and effective training method^[60–62]. Using simulation training models to obtain the LC of most procedures may improve operating room performance^[63]. Our porcine kidney model is designed to enhance urologists’ skills and improve patient outcomes. The model accurately simulates the anatomy and physiology of the human kidney and ureter. Previous research has validated the effectiveness of porcine kidney tissue as a training tool for URS^[13]. In a previous study by our team, 20 urologists were trained to conduct fURS using fresh porcine kidney and ureteral tissue, and the results showed that urologists completed the task in a shorter time and their ability was significantly improved^[31]. In addition, our device is simple, readily available, and inexpensive. As a simulation training object, ex vivo pig kidneys were cheaper and easier to obtain than live pig kidneys, which reduced the cost of training and did not pose ethical issues. To the best of our knowledge, this is the first study to report providing data on the LC of fURS TFL lithotripsy. TFL lithotripsy has currently attracted widespread attention due to its advantages compared with the Ho: YAG laser. This study provides a valuable reference for novice surgeons who want to actively gain experience in URS surgery.

This study has certain limitations. First, although participants achieved surgical proficiency comparable to attending physicians after training with our model, it is important to note that clinical scenarios can be highly variable. Factors such as complex stone composition and deeper stone locations can influence the surgical procedures. Second, while our model closely resembles real-world surgical conditions, it does not fully account for the impact of patient respiration during surgery. In addition, our model simplifies the process of UAS and URS insertion, which is extremely difficult to gain experience in model simulation due to the complexity and individual differences of the urinary system and the dynamic changes of the ureter due to its compliance and peristalsis. Moreover, research incorporating data from actual patient procedures is necessary to address this limitation. Furthermore, the surgical procedure scoring in the study did not use double-blind or video play-back mechanisms, which may have resulted in outcomes susceptible to subjective bias. Nevertheless, our findings provide compelling evidence supporting the efficacy of our training method.

Given the potential limitations of our sample size, expanding the participant pool in subsequent research would enhance the generalizability of the findings. This study investigated the short-term effects, and future studies should perform long-term studies to evaluate skill retention. In addition, the applicability of the porcine kidney-ureter model used in our study in various settings needs to be validated in the future. We acknowledge the limitations of the present study design, as model training differs from real-world clinical practice. To address this, we plan to collect surgical data from young physicians who have and have not undergone training with our device. By comparing relevant performance indicators between these two groups, we aim to establish a LC and evaluate the efficacy of our device and training methodology.

Conclusion

Analysis of the LC for fURS TFL lithotripsy shows that it takes approximately four surgeries to reach the learning TP, with significant improvements in operative time, tissue damage, and central visual shift. In addition, the established ex vivo porcine kidney model performed well in fURS laser lithotripsy and is a conveniently accessible and effective clinical training model.

Ethical approval

This study was approved by the Medical Ethics Committee of Wuhan University Zhongnan Hospital (Approval No. 2024023K).

Consent

Written informed consent was obtained from the participant. A copy of the written consent is available for review by the Editor-in-Chief of this journal on request.

Sources of funding

The Key Research and Development Program of Hubei province (2023BCB001) and an Exploration of Teaching and Learning Based on Cognitive Psychology Incorporating Virtual Patients (230901449082231)and the National Key Research and Development Plan of China (Grant No. 2024YFC2421400).

Author’s contribution

X.H.W., S.L., T.Z.L., Z.L.L. and X.Y.T.: conception, interpretation of data, manuscript writing, and editing. Z.L.L., S.J.W., D.W. and T.X.Y.: analysis and interpretation of data, editing of manuscript. S.J.W. and Y.W.L.: Collection of data. X.H.W, S.L., and T.Z.L.: conception, interpretation of data, editing of manuscript, supervision and administrative support. All authors read and approved the final manuscript.

Conflicts of interest disclosure

Not applicable.

Research registration unique identifying number (UIN)

Not applicable.

Guarantor

Prof. Xinghuan Wang.

Provenance and peer review

Not commissioned, externally peer-reviewed.

Data availability statement

The data that support the findings of this study are available from the corresponding author on reasonable request.

Assistance with the study

Not applicable.

Presentation

Not applicable.

Acknowledgments

Not applicable.