AI-enabled Kidney Surgery Makes it Easier for Novice Surgeons
Percutaneous nephrolithotomy (PCNL) is an efficient surgical intervention for removing large kidney stones. However, it is a challenging procedure that requires years of training to perform. To address this, a group of scientists from the Nagoya City University, developed and trialled an artificial intelligence (AI)-enabled robotic device for assisting surgeons in PCNL.
Creating a renal access from the skin on the back to the kidney is a crucial yet challenging step in PCNL. A poorly performed renal access can lead to severe complications including massive bleeding, thoracis and bowel injuries, renal pelvis perforation, or even sepsis. This procedure takes years of training to master. The two main renal access methods for PCNL – X-ray guidance and ultrasound (US) guidance deliver similar postoperative outcomes but require experience-based expertise.
Many technologies are being developed to bridge this skill gap. This inspired a Nagoya University research team to question if artificial intelligence (AI)-powered robotic devices could be used for improved guidance compared with conventional US guidance. Specifically, they wanted to see if the AI-powered device called the Automated Needle Targeting with X-ray (ANT-X), which was developed by the Singaporean medical start-up, NDR Medical Technology, offers better precision in percutaneous renal access along with automated needle trajectory.
The team performed a randomised, single-blind, controlled trial comparing their robotic-assisted fluoroscopic-guided (RAF) method with US-guided PCNL. The results of this trial were detailed in the The Journal of Urology. “This was the first human study comparing RAF with conventional ultrasound guidance for renal access during PCNL, and the first clinical application of the ANT-X ,” said Dr Kazumi Taguchi, Assistant Professor at NCU.
The trial was conducted with 71 patients—36 in the RAF group and 35 in the US group. The primary outcome of the study was single puncture success, with stone-free rate (SFR), complication rate, parameters measured during renal access, and fluoroscopy time as secondary outcomes.
The single puncture success rate was ~34 and 50% in the US and RAF groups, respectively. Fewer needle punctures were needed in the RAF group (1.82 times) as opposed to the US group (2.51 times). In 14.3% of US-guided cases, the resident was unable to obtain renal access due to procedural difficulty, prompting a surgeon change. However, none of the RAF cases faced this issue. The median needle puncture duration was also significantly shorter in the RAF group (5.5 minutes vs 8.0 minutes). No significant differences in the other secondary outcomes was found.
Multiple renal accesses during PCNL are directly linked to postoperative complications including, decreased renal function. Therefore, the low needle puncture frequency and shorter puncture duration, as demonstrated by the ANT-X, may provide better long-term outcome for patients. While the actual PCNL was performed by residents in both RAF and US groups, the renal access was created by a single, novice surgeon in the RAF group, using ANT-X. This demonstrates the safety and convenience of the novel robotic device, which could possibly reduce surgeons’ training load and allow more hospitals to offer PCNL procedures.
Dr Taguchi commented, “The ANT-X simplifies a complex procedure, like PCNL, making it easier for more doctors to perform it and help more number of patients in the process. Being an AI-powered robotic technology, this technique may pave the way for automating similar interventional surgeries that could shorten the procedure time, relieve the burden off of senior doctors, and perhaps reduce the occurrence of complications.”
Source: Nagoya City University
