Robot-assisted prostatectomy for prostate cancer: A health technology assessment
Health technology assessment
|Published
The purpose of this HTA was to evaluate the clinical effects and safety of robot-assisted prostatectomy for prostate cancer. Robot-assisted prostatectomy was compared to open prostatectomy.
Key message
Prostate cancer is the most common cancer among men in Norway, and approximately 5,000 men are diagnosed with prostate cancer each year. The most common curative treatment is surgical removal of the prostate and the seminal vesicles (radical prostatectomy). Prostatectomy may, in principle, be performed by open surgery, conventional laparoscopy, or robot-assisted laparoscopy. Robot-assisted prostatectomy has been the dominating surgical approach for prostate cancer in Norway for several years, but the technique has not been evaluated in a health technology assessment (HTA) within a Norwegian context. In this HTA we compared robot-assisted and open prostatectomy for prostate cancer.
We found that robot-assisted prostatectomy:
- Led to shorter operation time and less blood loss
- Resulted in small or no differences in urine and erectile function
- Probably led to shorter hospital stay and less postoperative pain
- Probably led to improved survival
- Possibly resulted in fewer complications and readmissions, fewer recurrences and reduced the need for additional treatment
The findings suggest that robot-assisted prostatectomy resulted in benefits for surgery-related outcomes as well as improved survival. Robot-assisted and open prostatectomy had comparable results for the important functional outcomes urine and erectile function. The benefits should be evaluated against the costs of the two techniques. A health economic evaluation is being conducted and will be published in a separate report.
Summary
Introduction
Prostate cancer is the most common cancer among men in Norway, and approximately 5,000 men are diagnosed with prostate cancer each year. The progression of the disease varies substantially. In many cases, the cancer develops slowly, and patients die of other causes than prostate cancer. In other cases, the disease develops rapidly and frequently spreads to bones and other organs. Patients that do not have low risk prostate cancer, but have cancer confined to the prostate, can be offered surgical removal of the prostate and the seminal vesicles (radical prostatectomy). Prostatectomy may, in principle, be performed by open surgery, conventional laparoscopy, or robot-assisted laparoscopy. Robot-assisted prostatectomy has been the dominating surgical approach for prostate cancer in Norway for several years, but the technique has not been evaluated in a health technology assessment (HTA) within a Norwegian context.
Objective
The purpose of this HTA was to evaluate the clinical effects and safety of robot-assisted prostatectomy for prostate cancer. Robot-assisted prostatectomy was compared to open prostatectomy. A companion health economic evaluation is being conducted and will be published in a separate report.
Method
We conducted a systematic literature search in relevant databases. Two researchers assessed titles, abstracts, and full texts of potentially relevant references against the inclusion criteria. We used EPPI-Reviewer in the selection process and its machine learning functionalities to streamline the screening process. One researcher extracted and analyzed data from the included studies, and another checked the data extraction. Two researchers independently assessed the methodological quality and assessed our confidence in the results using GRADE. We also assessed the transferability to the Norwegian context. We collected information from clinical experts about the organization of robot-assisted surgery in three Norwegian hospitals (Haukeland universitetssykehus, Universitetssykehuset Nord-Norge - Tromsø, and Akershus universitetssykehus).
Results
We included one RCT with 326 patients and 16 non-randomized studies with 138 162 patients (in total). The studies were conducted in Australia, Sweden, the USA, Germany, France, Austria, the UK, and Canada. All studies compared robot-assisted prostatectomy to open prostatectomy. Robot-assisted prostatectomy resulted in shorter operation time, less blood loss, and small or no differences in urine and erectile function (Summary of findings table). Robot-assisted prostatectomy probably also led to shorter hospital stay, less postoperative pain, and improved survival (Summary of findings table). Additionally, robot-assisted prostatectomy may result in fewer complications, fewer readmissions, lower short-term mortality, fewer anastomotic strictures, fewer recurrences, reduced need of additional treatment, and small or no differences in risk of inguinal hernia, when compared to open prostatectomy. Based on the available documentation, it is not possible to judge whether robot-assisted prostatectomy alters the risk of incisional hernia and positive surgical margins.
Summary of findings
|
Summary of findings |
||||||
Outcome |
Anticipated absolute effects (95% CI) |
Relative effect |
Number of patients (studies) |
Certainty |
|
|
|
Open prost. |
Robot-assisted prost. |
Comments |
||||
|
Surgery-related outcomes |
||||||
|
Operation time
|
234 min |
MD 32.3 min lower |
- |
308 (1 RCT) |
⨁⨁⨁⨁ |
|
|
Blood loss |
1338 ml |
MD 894 ml lower |
- |
308 (1 RCT) |
⨁⨁⨁⨁ |
2 non-RCT (n = 10610) reported the outcome, both agreed with RCT |
|
Length of hospital stay |
3.27 days |
MD 1.72 days lower |
- |
308 (1 RCT) |
⨁⨁⨁◯ |
2 non-RCT (n = 29275) reported the outcome, both agreed with RCT |
|
Functional outcomes |
||||||
|
Urine function (12 months) |
90.8 |
MD 0.77 lower |
- |
281 (1 RCT) |
⨁⨁⨁⨁ |
3 non-RCT (n = 13270) reported the outcome, none found effect |
|
Erectile funct. (12 months) |
42.3 |
MD 0.23 higher |
- |
281 (1 RCT) |
⨁⨁⨁⨁ |
3 non-RCT (n = 9870) reported the outcome, none found effect |
|
Postoperative pain (24 h) |
5.83 |
MD 1.23 lower |
- |
303 (1 RCT) |
⨁⨁⨁◯ |
|
|
Oncological outcomes |
||||||
|
Survival (risk of death) |
57 per 1 000 |
15 fewer per 1 000 (20 fewer to 11 fewer) |
HR 0.73 |
26 944 (2 non-RCT) |
⨁⨁⨁◯ |
|
|
CI: confidence interval; prost.: prostatectomy; MD: mean difference; RCT: randomized controlled trial; h: hours; HR: hazard ratio. We have downgraded the certainty in effect estimates due to: a: wide CI; b: moderate risik of bias Urine function and erectile function were assessed with the questionary Expanded Prostate Cancer Index Composite (EPIC). EPIC has a 0-100 scale where high values indicate good function. Pain was assessed by Surgical Pain Score, which has a 0-10 scale where 10 indicates the worst thinkable pain. For survival, the anticipated absolute effect is the risk of death which was calculated from the risk in the control group (unadjusted) and the hazard ratio (adjusted). |
||||||
The robotic surgery units were used for multiple indications at all three hospitals. The operating time assigned for the various indications and the number of procedures varied between the hospitals. Two hospitals used the same type of surgical team for robot-assisted and traditional surgery, whereas the third hospital replaced one of two surgeons with a specialized nurse for robot-assisted prostatectomy and hysterectomy. This approach produced cost savings and released surgeon capacity.
Discussion
We have probably identified all studies that meet our inclusion criteria by our systematic literature search and selection. Only one RCT was included, but we also included 16 non-randomized studies that had adjusted their results for important confounders. The RCT and the non-randomized studies were analysed separately (in separate meta-analyses), and the results from the different study types were consistent and reinforced each other. We excluded non-randomized studies without sufficient adjustment for important confounders. This suggests that this HTA is based on the most reliable available results. Nevertheless, it remains a weakness that only one RCT that met the inclusion criteria was identified.
Surgical results may be substantially influenced by the surgeons’ experience and skills. For both robot-assisted and open prostatectomy, the surgeons experience a learning curve which complicates any comparison of the techniques. We excluded studies where surgeons obviously were in a learning phase, but several of the non-randomized studies did not describe the surgeons’ experience levels. In these studies, it is fair to assume that both experienced and inexperienced surgeons were included. Because most studies were conducted within the first decade after the introduction of robotic surgery, surgeons may have been less experienced with robot-assisted than with open surgery. This suggests that possible gains with robotic surgery may have been underestimated in this HTA.
Robot-assisted prostatectomy has been the dominating surgical approach for prostate cancer in Norway for several years. Our HTA found that this approach led to improvements in both surgery-related and oncological outcomes, and yielded comparable results to open prostatectomy for important functional outcomes. The improvements may have great impact for individual patients and may reduce resources and time spent on patient follow-up in the health care service.
Robot-assisted is continuously introduced to new indications in Norway and internationally. Our HTA of robot-assisted prostatectomy for prostate cancer may be compared with the evidence available for other indications. Such comparisons may be useful when deciding whether robot-assisted surgery should be introduced for new indications, and to prioritize access to robotic surgery systems between approved indications.
Conclusion
Robot-assisted prostatectomy led to benefits for both surgery-related and oncological outcomes, and yielded comparable results for important functional outcomes when compared to open prostatectomy.
The benefits of robot-assisted prostatectomy should be evaluated against its costs. We are conducting a companion health economic evaluation that will be published in a separate report.
