AUA2021 Panel Discussion: Artificial Intelligence Applications in Urology

By: Andrew J. Hung MD; Jaime Landman MD, FRCS; Prokar Dasgupta MD, FRCS(Urol), FKC | Posted on: 03 Sep 2021

The era of surgical artificial intelligence (AI) has commenced, although the concept is not new, going back to the genius of Alan Turing, who with his decoding skills had a major impact on the outcome of World War II. There is renewed interest in AI that may well become the driving force of the future, digitising surgical practice.1

AI is akin to human intelligence with examples like visual perception, speech recognition and language translation. Machine learning (ML) is a subset of AI and is in simple terms the use of data to answer questions. One of its main uses has been in accurate image recognition. It has had a major impact in diagnosing diabetic retinopathy and predicting cardiovascular events in these patients up to 48 hours before they actually occur. Likewise, a prostate recognition algorithm could make the machine learn whether a given image is that of a prostate cancer or not, thus reducing the variability in magnetic resonance imaging (MRI) readings by radiologists. The MRI can also be automatically segmented to produce 3D prostate models that can guide surgeons during robotic surgery. Early data suggest that such 3D printed models can reduce positive margins in robotic assisted radical prostatectomy for T3 tumours.2 3D printing can also be used to produce low-cost laparoscopic skills trainers that are easy to assemble and use.3 Artificial neural networks, unlimited data storage capacities and computing ability have revolutionised modern day ML systems, making the executions faster, cheaper and more powerful than ever.

Figure 1. AI in Urology. Adapted from Chen et al.4

A recent review of AI in urology summarised more than 100 articles,4 two-thirds of which were in diagnostics such as biomarkers in bladder cancer, a third of which were in outcome prediction such as those of PCNL, while the remaining described treatment plans, for example drugs in castration-resistant prostate cancer and a few in surgical skills evaluation (see figure).

Another application of AI would be its role in democratising surgery by combining low latency ultrafast 5G connectivity with augmented reality. The so-called “Internet of Skills” could make remote robotic surgery, teaching and mentorship easily accessible, irrespective of the location of the expert surgeon.5

Tremendous strides have also been made in the cross-section between surgery and AI to prognosticate patient outcomes, to improve surgeon skill assessment and to one day improve surgeon training altogether. Automated performance metrics (APMs), derived from robotic instrument kinematic data collected during surgery and processed through AI algorithms, can now accurately predict urinary continence recovery days to months after robot-assisted radical prostatectomy.6,7 Ongoing efforts to automate technical skills assessment, utilizing deep learning-based computer vision, will make feedback to surgeons truly objective and scalable.8 Such assessment can soon pinpoint specific deficiencies, such as addressing wrist rotation during suturing. In fact, recent work has found that technical skills assessment are stronger predictors of clinical outcomes than APMs, which are largely measures of efficiency of surgery.9 Naturally, the collective direction with the above efforts is to provide meaningful feedback to surgeons and improve patient outcomes. That will be the natural course of surgical AI in the next 5 years. Time will tell whether these efforts may then serve as stepping stones for semi-autonomous and fully autonomous surgery.10

  1. Ghose A, Ghose A and Dasgupta P: New surgical robots on the horizon and the potential role of artificial intelligence. Investig Clin Urol 2018; 59: 221.
  2. Chandak P, Byrne N, Lynch H et al: Three-dimensional printing in robot-assisted radical prostatectomy - an Idea, Development, Exploration, Assessment, Long-term follow-up (IDEAL) Phase 2a study. BJU Int 2018; 122: 360.
  3. Parkhomenko E, Yoon R, Okhunov Z et al: Multi-institutional evaluation of producing and testing a novel 3D-printed laparoscopic trainer. Urology 2019; 124: 297.
  4. Chen J, Remulla D, Nguyen JH et al: Current status of artificial intelligence applications in urology and their potential to influence clinical practice. BJU Int 2019; doi: 10.1111/bju.14852.
  5. Kim S, Dohler M and Dasgupta P. The Internet of Skills: use of fifth-generation telecommunications, haptics and artificial intelligence in robotic surgery. BJU Int 2018; 122: 356.
  6. Hung AJ, Chen J, Ghodoussipour S et al: A deep-learning model using automated performance metrics and clinical features to predict urinary continence recovery after robot-assisted radical prostatectomy. BJU Int 2019; 124: 487.
  7. Hung AJ, Chen J and Gill IS: Automated performance metrics and machine learning algorithms to measure surgeon performance and anticipate clinical outcomes in robotic surgery. JAMA Surg 2018; 153: 770.
  8. Luongo F, Hakim R, Nguyen JH et al: Deep learning-based computer vision to recognize and classify suturing gestures in robot-assisted surgery. Surgery 2021; 169: 1240.
  9. Trinh L, Mingo S, Vanstrum E et al: Survival analysis using surgeon skill metrics and patient factors to predict urinary continence recovery after robot-assisted radical prostatectomy. Eur Urol Focus 2021; doi: 10.1016/j.euf.2021.04.001.
  10. Connor MJ, Dasgupta P, Ahmed HU, Raza A. Autonomous surgery in the era of robotic urology: friend or foe of the future surgeon? Nat Rev Urol. 2020; 17(11): 643-649.
Top 300x250:
Bottom 300x250: