Robotic cancer surgery

Robotic surgery has been practised clinically since the late 1990s. With the promise of enhancing surgical capabilities, it continues to gain popularity. Subjectively, many surgeons claim significant benefit from improved visualization, enhanced dexterity, motion scaling and tremor reduction. Despite considerable investment from both industry and clinicians, and over a decade of research, the evidence base supporting robotic assistance remains woefully lacking. The most popular clinical applications, at present, are in urological surgery. In many centres, owing to a combination of factors including marketing and patient expectations, robotic prostatectomy is now the standard of care. Despite the existence of only one randomized controlled trial and considerable publication and selection bias, the most robust evidence available has failed to show any significant improvement in morbidity or cancer outcomes compared with laparoscopic surgery for prostatectomy1. This typifies the lack of highquality randomized controlled trials evaluating robotic surgery for all oncological applications. A recent meta-analysis of short-term outcomes from eight non-randomized controlled trials comparing laparoscopic and robotic rectal cancer surgery showed no differences in duration of surgery, length of hospital stay, time to resume regular diet, postoperative mortality or morbidity, or the oncological accuracy of resection2. Similar outcomes have been reported in non-randomized controlled trials for gastric and oesophageal cancer in terms of numbers of lymph nodes obtained, and overall morbidity and mortality rates3,4. In gynaecological cancer surgery, the quality of the trial evidence is poor, with high risks of bias, precluding meaningful conclusions. A recent Cochrane review and data from randomized controlled trials for benign gynaecological robotic surgery showed no benefit in safety or efficacy5,6. This lack of supporting evidence for robotics in surgery is peculiar when considering other highrisk, high-reliability and technologydriven industries such as aviation and the military, where computer assistance and robotics play crucial roles and have proven to be immensely beneficial. Why, therefore, do robotics and computer assistance not appear to play an equivalent role in surgery? Perhaps the answer to this question lies in the properties of the robotic platform itself. Currently, the only robotic system that is approved and widely used clinically is the da Vinci master–slave robot manufactured by Intuitive Surgical (Sunnyvale, California, USA), an organization that, through patents and business acquisitions, effectively holds a monopoly on the market. This monopolization is at least partially responsible for the comparative lack of progress in surgical robotics. Over the past decade, emerging or competitor companies such as Computer Motion (which manufactured the Zeus TM robot) have been bought and scrapped by Intuitive Surgical, presumably to maintain market control. From an innovation perspective, there is a need for increased market competition. For reasons that are unclear, but based on the evidence available, it should be concluded that the da Vinci interface does not appear to possess the capabilities necessary to translate to an objective clinical benefit. For robotics to improve patient outcomes in cancer surgery, different robotic properties must therefore be explored. This requires radical innovation in both hardware and software. The prospect of flexibleaccess surgery not only allows the surgeon to minimize abdominal wall incisions safely, but can also permit superior navigation and access in complex interventions. This concept has led to the development of the i-Snake surgical robot (Wellcome Trust and Imperial College, London, UK), a hyper-redundant, biologically inspired, flexible-access robot with integrated multimodal and multiscale sensing7, capable of both advanced diagnostic and therapeutic interventions. In vivo trials have yielded promising results, with clinical translation planned in the next couple of years. Real advantages of computer assistance will be seen when software such as dynamic active constraints and inverse realism in augmented reality are implemented to smart robotic platforms, allowing preoperative planning and rehearsal of a prescribed safe operation, and eliminating adverse events or human error, all based on patient-specific data8. By transforming sensory information, for example through gaze behaviour, into physical constraints during surgery9, real-time haptic feedback can be