Robot-assisted rectal resection for rectal cancer: A health technology assessment
Health technology assessment
|Published
The purpose of this health technology assessment (HTA) was to evaluate the clinical effects and to conduct a health economic evaluation of robot-assisted rectal resection for rectal cancer.
Key message
Robot-assisted surgery has been used in rectal cancer surgery in Norway since 2011. Use of the robotic-assisted technique has increased in recent years, and today close to half of the operations are performed with this approach. In this health technology assessment (HTA), we have compared robot-assisted rectal resection with traditional laparoscopy and open surgery in patients with rectal cancer.
Compared with laparoscopic surgery, we found that robot-assisted rectal resection
- led to fewer conversions to open surgery
- probably led to longer operating times, shorter hospital stays, and fewer patients with circumferential margin positivity
- probably did not lead to a difference in bladder function or postoperative complications
- led to small or no difference in long-term survival (overall survival 3 years), but we have low confidence in the results.
The documentation was weaker for the comparison of robot-assisted and open surgery. We found that robot-assisted rectal resection led to shorter hospital stays and fewer patients with circumferential margin positivity than open surgery, but we have low confidence in the results. We did not find sufficient documentation to assess differences in other outcomes.
Costs related to investment, disposable materials and service, were substantially higher for robot-assisted than for laparoscopic and open rectum resection. Based on the current information regarding effect and costs, it seems unlikely that robot-assisted surgery is a cost-effective alternative to laparoscopy. We could not conclude regarding the cost-effectiveness with comparison to open surgery.
Summary
Introduction
In surgical treatment of rectal cancer, both robot-assisted, traditional laparoscopic and open techniques are used. Robot-assisted technique has been used for this indication in Norway since 2011. The proportion of robot-assisted interventions has increased significantly in recent years, and up to half of all rectal resections for rectal cancer were performed robot-assisted in 2022.
Objective
The purpose of this health technology assessment (HTA) was to evaluate the clinical effects and to conduct a health economic evaluation of robot-assisted rectal resection for rectal cancer. Robot-assisted technique was compared to traditional laparoscopic and open surgery. We have also described organizational aspects of robot-assisted surgery.
Method
We conducted a systematic literature search in relevant databases. Two researchers assessed titles, abstracts, and full-text publications of potentially relevant references against the inclusion criteria. Studies of learning curve or first experiences with robotic surgery were excluded. We included randomized trials (RCTs) and non-randomized trials with control groups (non-RCTs) and compiled the results in separate meta-analyses. Two researchers assessed risk of bias and assessed confidence in the results using GRADE. We performed a simplified health economic assessment, where we collected investment costs and costs per hospital stay (KPP) for the surgical techniques. We present examples of organization from three Norwegian hospitals.
Results
For the comparison of robot-assisted and laparoscopic surgery, we included six RCTs (2,459 patients) and 20 non-RCTs (50,237 patients, with some overlap in study population). The patients in the RCTs were mainly from Asia, while patients in non-RCTs were from the USA, Asia, Australia, and several European countries, including Norway.
We found that robot-assisted rectal resection led to fewer conversions to open surgery (Summary Table). Furthermore, robot-assisted surgery probably resulted in longer operating time, shorter length of stay and fewer patients with circumferential margin positivity than traditional laparoscopy. There were probably no differences in short-term complications or bladder function between the groups. There was small or no difference in long-term survival, but we have low confidence in the results. There was also uncertainty related to the results for reoperations and short-term mortality. It was not possible to assess how robot-assisted surgery affected sexual function based on available documentation.
For the comparison of robot-assisted and open surgery, we identified five non-RCTs. All the studies were from one patient registry (USA), and only some of the outcomes of interest were reported. We found that robot-assisted rectal resection led to shorter length of stay and fewer patients with circumferential margin positivity compared to open surgery, but we have low confidence in the results. It was not possible to assess how robot-assisted surgery affected short-term mortality or long-term survival based on the available documentation.
Summary Table: Robot-assisted versus laparoscopic rectal resection for rectal cancer | ||||||
Outcome | Anticipated absolute effects (95% CI) | Relative effect | Number of participants | Certainty | Comments | |
Laparoscopy | Robot | |||||
Conversion to open surgery | 49 per 1 000 | 22 fewer per 1 000 (31 fewer to 8 fewer) | RR 0.55 | 2 452 | ⨁⨁⨁⨁ | The finding is supported by findings from non-RCTs (n = 20 410) |
Operation time | 213 min | MD 44 min longer | - | 2 457 | ⨁⨁⨁◯ | The finding is supported by findings from non-RCTs (n = 7 906) |
Complications (30 days) | 248 per 1 000 | 40 fewer per 1 000 | RR 0.84 | 2 123 (4 RCT) | ⨁⨁⨁◯ | The finding is supported by findings from non-RCTs (n = 2 005) |
Hospital stay | 8.49 days | MD 0.86 days (0.01 shorter to 1.71 shorter) | - | 2 452 (6 RCT) | ⨁⨁⨁◯ | The finding is supported by findings from non-RCTs (n = 17 460) |
Bladder function 6 months* | 6.1 | MD 0,56 lower | - | 2 675 (2 RCT) | ⨁⨁⨁◯ |
|
Circumferential margin positivity | 67 per 1 000 | 24 fewer per 1 000 | RR 0.64 | 2 274 (5 RCT) | ⨁⨁⨁◯ | The finding is not supported by findings from non-RCTs (n = 11 558) |
Abbreviations: CI, confidence interval; RR, relative risk; RCT, randomized controlled trial; MD, mean difference; CRM, circumferential resection margin. *Symptom scale from 0–35, where higher scores indicate worse bladder function. The table shows effect estimates in which we have high or moderate confidence from. We have downgraded the certainty in the effect estimates due to: a, high heterogeneity, b: moderate risk of systematic bias, c: wide confidence interval. |
Investment costs and costs associated with disposable materials and service were substantially higher for robot-assisted surgery compared to traditional laparoscopy and open surgery. Excluding investment costs, the simplified cost assessment also indicated that robot-assisted rectal resection was more expensive than laparoscopic rectal resection. The costs per procedure for robot-assisted surgery was closely related to the volume of operations, where higher volume of operations resulted in lower costs. Although the cost per hospital stay (excluding investment costs), based on the KPP model, showed that robot-assisted rectal resection was lower than open surgery, there is considerable uncertainty associated with these data since they were not adjusted for differences in patient populations.
There is great variation in the organization of robotic surgery in the hospitals, both in terms of distribution between specialist areas, types of interventions, and number of operations. For rectal resections, the composition of the operating team is the same for the three techniques.
Discussion
The knowledge base in this HTA was strongest for the comparison of robot-assisted and laparoscopic rectal resection. We found that robot-assisted resections led to fewer conversions to open surgery than traditional laparoscopy. In Norway, the proportion of rectal resections performed minimally invasively has increased in recent years, and the increase can be attributed to the increasing use of robot-assisted surgery. At the same time, the proportion of conversions has decreased and now stands at approximately 4 %.
There are challenges in evaluating studies of newer technology, such as robot-assisted surgery, against established techniques such as laparoscopy and open surgery. Most of the studies in this HTA were carried out in the early phase after the robot-assisted technique was introduced. However, we excluded studies that examined the learning curve or first experiences with robotic surgery and have thus not evaluated the safety aspect of introducing the method. Different experiences with the techniques may still have influenced the results.
We did not find differences in health effects that were considered sufficient for the development of a model-based health economic analysis, and therefore carried out a simplified cost assessment. There is significant variation in the execution and organization of rectal resections in Norwegian hospitals. Therefore, we reported average national costs per hospital stay (KPP). An important limitation of KPP, however, is the lack of comparability between the patient groups forming the basis for the cost calculations.
There are no national procurements for robotic surgical systems, so individual hospitals or regional health authorities make their own purchases. We have presented prices from historical, confidential procurements from the South-Eastern Norway Regional Health Authority since investment costs are not included in the KPP model. These prices were from a single supplier. From 2024, several suppliers are expected to enter the Norwegian market, and acquisition prices may change.
Conclusion
For patients with rectal cancer, we found that robot-assisted rectal resection led to fewer conversions to open surgery compared to traditional laparoscopy. The operating time was probably longer, and the hospital stay shorter with robot-assisted surgery, and there were probably fewer patients with circumferential margin positivity. There were probably no differences in short-term complications or bladder function. There appeared to be no difference in long-term survival, but these results were inconclusive. For the comparison with open surgery, the results showed possible advantages of robot-assisted surgery, but the evidence was weak.
Investment costs and costs associated with disposable materials and service were substantially higher for robot-assisted than for traditional laparoscopic and open surgery. Excluding investment costs, the simplified cost assessment suggested that the cost per hospital stay was higher for robot-assisted rectal surgery than for laparoscopy. Based on the current information regarding effect and costs, it seems unlikely that robot-assisted surgery is a cost-effective alternative to laparoscopy. However, costs per hospital stay have not been adjusted for differences in patient population. We could not conclude regarding the cost-effectiveness with comparison to open surgery.