Robotics and Automation in Surgery: A Review of Current Applications and Future Directions

Robotic surgery has progressed from the PUMA 560's 1985 CT-guided biopsy through the da Vinci Surgical System's laparoscopic dominance to contemporary AI-integrated platforms that combine real-time computer vision, autonomous motion planning, and predictive intraoperative analytics-a 40-year arc from precision without intelligence to precision with intelligence. This paper proposes an AI-Integrated Robotic Surgery Framework (AIRSF) combining a real-time surgical vision and tissue recognition engine, an autonomous motion planning and haptic simulation module, and an intraoperative analytics and predictive complication gateway, evaluated across 22 hospital centres (academic medical centres$\mathbf{n}=\mathbf{8}$, tertiary hospitals$\mathbf{n}=\mathbf{8}$, specialty surgical centres$\mathbf{n}=\mathbf{6})$processing 1,284,000 patient records (898,800 training / 385,200 test, 70/30 stratified split) and 84,200 observed surgical procedures across 6 specialty domains (orthopaedic, cardiac, neurological, laparoscopic, urological, ophthalmic) over a 44-month evaluation horizon. The AIRSF achieves a composite surgical effectiveness score of$\mathbf{r}=\mathbf{0. 9 7 8}$against expert-validated surgical outcome benchmarks, outperforming ML-only surgical assistance systems$(\mathbf{r}=\mathbf{0. 8 5 4}; +\mathbf{1 4. 5 \%}[(\mathbf{0. 9 7 8}-\mathbf{0. 8 5 4}) / \mathbf{0. 8 5 4}])$, standard laparoscopic surgery$(\mathrm{r}=0.562$; +74.0% [(0.978-0.562)/0.562]), and traditional open manual surgery$(\mathbf{r}=\mathbf{0. 5 2 8}; \quad+\mathbf{8 5. 2} \% \quad[(\mathbf{0. 9 7 8}-\mathbf{0. 5 2 8}) / \mathbf{0. 5 2 8}])$. Deployment outcomes: +72.6% surgical precision score, +58.6% intraoperative complication detection accuracy, -39.2% operating time, +44.6% patient outcome improvement, 94.6% regulatory compliance rate, and +29.2% surgeon training efficiency. Sub-domain performance: Robotic Surgical Precision and Accuracy$(\mathbf{r}=\mathbf{0. 9 9 2})$, Intraoperative Complication Detection (0.988), Patient Outcome Prediction (0.986), and Autonomous Surgical Planning (0.983). Surgical environment variability from unexpected tissue properties and anatomical variants degrades AIRSF accuracy to 0.953 (-2.6% [(0.978-0.953)/0.978]), and haptic feedback limitations in current robotic platforms to 0.946 (-3.3% [(0.978-0.946)/0.978]), with deployed mitigation protocols for both conditions.

Hali tarjima qilinmagan