The Smart Tissue Autonomous Robot (STAR): the role of intelligence and autonomy in surgical robotics
Despite the well-recognized benefits of less traumatic approach and the enabling nature of robot assisted surgery (RAS), the spread of the technologies and adoption rates for both RAS and manual minimally invasive surgery (MIS) over the last two to three decades remain questionable for most surgeries other than technically simple procedures. The enabling advantage of RAS is based on superior vision, dexterity and ergonomics, while manual MIS offers more flexible and somewhat versatile access to different anatomic domains. However, the capital and maintenance costs, including consumables, large physical footprint, unshared workspace, lack of haptics and control console isolated from the patient in RAS continue to pose paramount challenges. In contrast, manual MIS, albeit more cost-effective than RAS, continues to be restricted in its applications in non-routine, more complicated surgery, due to tool design, especially pivoting around insertion sites, challenging ergonomics and dexterity. The critical underlying weakness of both technologies remains human factors: surgeon proficiency, practice volume, training and experience critically undermine surgical outcome and access to the techniques and technology. Apart from the benefits attributable to the approach, surgical outcomes related to the techniques and indications have not changed significantly since the open-surgery era. Future surgical techniques and technology therefore must focus not just on trying to see what can be done with the same instrument designs, powered or manual, but on how patient-centered outcomes can be improved and true value as can be created with technology.
The future technology must consider not just variations in tool design but a fundamental shift which enables the surgeon to conduct surgery differently and better, similar to the paradigm shift from open to minimally invasive surgery. The following six areas may be considered:
a shift from a master-slave configuration, with the human surgeon in control, to a collaborative, potentially master-master configuration;
a hybrid approach combining MIS and RAS, using modular tools, where the surgical approach selected would not be restricted by the instrumentation;
a vision system not limited to the human visual spectrum, displaying not just anatomic but also physiologic and functional information;
tools that are more human hand-like, with haptics, dexterity and soft robotics;
intelligent tools, transitioning from a master-slave configuration to supervised autonomy and possibly full autonomy;
accessibility, with optimal surgical techniques and technologies available wherever and whenever needed.
Democratization of surgery
Surgery comprises four phases: incision or access; resection; reconstruction; and closure. The rationale of the Smart Tissue Autonomous Robot (STAR) is that, if a critical part of surgery such as reconstruction (e.g. anastomosis) can benefit from intelligent and possibly autonomous tools, perhaps the remainder of surgery can also be conducted under full or under supervised autonomy in the not too distant future, to improve outcome. By programming optimal surgical techniques, it is possible to imagine a situation in which the best techniques and technology become available to everyone wherever and whenever surgery is needed in the future, thus democratizing surgery. We chose an intestinal anastomosis as proof-of-concept task, since no autonomous surgical task had ever been performed on soft tissue, given its deformability and mobility in the unstructured in-vivo environment.
Smart Tissue Autonomous Robot (STAR)
To accomplish soft tissue autonomous robotic surgery, we improved vision by using 3D plenoptic camera with real-time tissue tracking using near-infrared technology to provide situational awareness ( Figure 12.1 ). We deliberately restricted suturing to a single-hand technique by incorporating a commercially available manual suturing device, manufactured by Endoevolution, and added a force sensor for suture tension. An intelligent suturing algorithm was developed, based on optimal surgical suturing techniques as described in the literature and a consensus of experienced expert surgeons. We then compared the functional outcomes of anastomoses performed by the STAR in ex-vivo porcine intestine and in preclinical porcine models to those done by expert surgeons using open, laparoscopic and robotic-assisted surgery with da Vinci® system ( Science Translational Medicine, May 2016). The results clearly and convincingly demonstrated that the anastomoses performed under full autonomy (60%) and supervised autonomy (40%) were superior in terms of leak pressure, suture consistency (variance) and mistakes and comparable in terms of time and the degree of luminal narrowing due to imbrication of the edge.