This project aims to develop an augmented reality (AR) tool to enhance skill acquisition for endoscopic kidney stone surgery. Of the 100,000 patients who undergo an endoscopic kidney stone treatment annually in the United States, 25% will require a repeat stone surgery within 20 months of their index surgery. The repeat stone surgery rate is almost completely driven by postoperative residual stone fragments, which lead to ureteral obstruction, causing pain, urinary tract infection, and kidney injury. One significant factor that contributes to residual stone fragments is limited visualization of the entire collecting system - a skill directly associated with surgeon experience. This leads to novice surgeons having a much higher recurrence rate than experienced ones. As the incidence of kidney stone disease continues to increase (prevalence of 10%, incidence of 1116 per 100,000), improved endoscopic surgical training is required to improve outcomes of stone surgeries and minimize complications by improving stone-free rate.

Currently, skill assessment during endoscopic stone surgery is limited. There are no objective metrics for endoscopic surgery to assess skill. The only feedback trainees get is in the form of verbal communication from expert surgeons, usually after the conclusion of surgery. Thus, most feedback is synoptic and limited in facilitating skill acquisition. Operative time and patient safety concerns restrict the amount of active, real-time feedback given during a case for skill acquisition. Endoscopic kidney stone surgery is uniquely challenging given the small depth and field of view of current endoscopes, which complicate the complete visualization of the entire collecting system.

Navigation of the collecting system relies on mentally mapping preoperative imaging to the endoscopic surgical field. Success in mapping relies on hand-eye coordination, memory, and spatial reasoning, which are gained through practice. Thus, there is a need for tools that facilitate endoscopic surgical skill acquisition.

The overarching hypothesis for this research is that surgical skill acquisition and outcomes for endoscopic kidney stone surgery can be improved by analyzing eye gaze data and using expert gaze to guide surgical trainees intraoperatively. Eye gaze guidance has been shown to lead to better skill acquisition in virtual reality surgical tasks compared with motion guidance alone. The proposed system would provide real-time education for trainees during endoscopic stone surgery, such as through head-mounted displays (i.e., the Microsoft HoloLens 2). The investigators have previously demonstrated eye gaze sharing in phantoms. By implementing this system in the operating room (OR), the investigators would be able to instill durable skill acquisition in trainees. The investigators will also implement the NASA-task load index for the trainees to gauge the usability of the system.

We will perform a clinical trial to quantitatively evaluate the benefits of gaze sharing for stone localization in patients. The trainees and experts surgeons will be prospectively randomized into the control and experimental groups. Specifically, each trainee (N=10) will be tasked with stone localization as is standard of care during ten endoscopic stone surgery under the guidance of an expert surgeon. All trainees and experts will wear HoloLens 2s to track their eye gaze. surgeons will review preoperative patient images and the surgical plan to ensure all agree on the surgical plan and no abnormal anatomy is missed. Each trainee and expert surgeon will be tasked with visualizing the entire intrarenal collecting system, including the major branches and renal stones. Each surgeon will visualize the entire collecting system in a retrograde fashion from the proximal ureter successively. To avoid bias, surgeons will be blinded to the surgical monitor while the other surgeon is performing the visualization. While each surgeon is localizing stones, the HoloLens 2 will collect eye gaze data. After completion of stone localization, the remainder of the case will proceed as is standard of care. The trainees and experts in the control group's HoloLenses will not show any augmentation, but trainees in the control group will receive verbal feedback from an expert surgeon as is standard procedure. Additionally, we will validate eye gaze metrics in the control group (i.e., the primary outcome being the gaze area of interest, as well as the number of gaze fixations, fixation times, and fixation points outside the area of interest, as well as total localization time). In the experimental group, the trainees' HoloLenses will project the AR display in addition to verbal feedback from the expert surgeon. Each trainee will perform stone localization in six separate patients with either solely verbal (control group) or verbal and gaze guidance to test skill acquisition (expert group). Then, the trainees will perform stone identification in four additional patients without guidance. After localization, the remainder of the case will continue as is standard of care. After the first six patients, each trainee in both groups will perform the same visualization task on four new patients with only verbal guidance.

The intervention is only during the kidney exploration phase of the surgery. The augmented reality is turned off during the rest of the surgery, which proceeds per standard of care. The study is designed to measure the total percentage of intrarenal collecting system surface visualized during all ten surgeries to assess skill acquisition in addition to gaze data. Additionally, we will measure if there is a significant difference between the performance of each group in the last four surgeries.

Patients will undergo endoscopic stone surgery by urologic surgeons via standard of care. In the current paradigm, expert surgeons teach trainees in real-time during the surgery to develop surgical competency. An augmented reality-based gaze-sharing system deployed on Microsoft HoloLens 2 will be used to evaluate surgical competency and provide a novel method of surgical education. As the data will be passively collected, the expert surgeon will not be affected by the system and the current standard of practice will not change. Furthermore, validation of eye gaze data and guidance will be done post hoc, so there is no impact on operative time. All surgeries will be performed with a Storz Flex XC ureteroscope. Videos will be recorded via Karl Storz StreamConnect--a centralized, cloud-based, secure surgical video platform. No clinical information regarding patients will be recorded.

The goal is to analyze at least 100 datasets of trainee and expert's eye-gaze sharing data in the OR. If patients need operation on both kidneys, it will count as 2 datasets of surgeon's eye-gaze sharing data for data analysis purposes. Thus, we may end up enrolling slightly fewer than 100 patients. All endoscopic videos will be recorded and analyzed postoperatively. Further validation of eye tracking will be done post hoc, to minimize the impact on operative time.

SUMMARY: In this clinical trial, the investigators are asking surgical trainees and experts to wear the HoloLens 2 in the operating room while conducting kidney stone surgery on consenting patients (~N of 100) enrolled for this study. The investigators measure the eye gaze data of the trainee and expert using the HoloLens. In the experimental arm of the trial, the trainees get visual guidance through a hologram showing the expert's gaze in addition to the verbal guidance as their standard of care. The intervention is only on the surgeons during the kidney exploration phase of the surgery. The augmented reality is turned off during the rest of the surgery, which proceeds per standard of care.

Inclusion Criteria:

• Surgeons who are "adult" residents and attending urologists at Vanderbilt University Medical Center in the age range 18 to 65
• Patients who are adults with kidney stones scheduled for endoscopic surgery as clinically indicated.

Exclusion Criteria:

• Vulnerable populations, including incarceration status, pregnancy, inability to give informed consent, and serious illness likely to cause death within 5 years

We will share the endoscope video and eye-gaze information from the study, and associated metadata, including the study protocol.

The data will be uploaded to the NIDDK repository by Aug 2026. We foresee the data being available indefinitely.

Access to individual data will be controlled through the NIDDK repository and available only for research purposes.