“If you want to continue to pursue the path of knowledge,
you should not revere the master,
study and develop his teachings
you should look for his errors.”
Compared in 1926 to a comic opera,1 blamed in 1996 to be misconceived, described as a waste in 2014, surgery has been the whipping boy of clinical research.2 A very few randomized controlled trials (RCTs) are published by surgeons and despite continuously adapted specific CONSORT criteria,3 most of the published RCTs are of poor quality. A fifth of surgical RCTs are abandoned and less than half are published. Only 5% of funding in oncology goes to surgical trials.4 Unfortunately, it appears that editorial boards and peer reviewers assert that the RCT is the only road to salvation in surgical research.5 In some countries such as Japan or Netherlands, better organization among the surgical communities produced successful trials. However, in majority of Western countries, the situation is less favorable. It is time to propose a new complementary approach based on dedicated, comprehensive research platforms focusing on feasible, prospective, and pragmatic surgical studies. This initiative by the European Society of Surgical Oncology (ESSO), the Japan Clinical Oncology Group (JCOG), and the European Organization for Research and Treatment of Cancer (EORTC) will be successful only if there is a collective awareness in the academic and editorial community, at whom this open letter is aimed.
LIMITATIONS OF RANDOMIZATION IN SURGERY
Randomization was introduced into agriculture in the 19th century applied in medicine in 1948 and belatedly applied, 50 years ago, to surgery. Limitations to conduct RCTs in surgery are well known. First, a lack of equipoise for both surgeon and patients, as exemplified by the famous Buxton law—always too early to randomize, suddenly too late. Second, the accrual of highly selected patients, which of course is one of the major reasons for abandoning RCTs. Third, the lack of funding—only a third of surgical trials receive funding—is a significant limiting factor. Finally and most importantly, the quality and the reproducibility of the procedures, a direct consequence of aforementioned factors, are a major concern.2 In the case of pharmaceutical trials, randomizing drug A versus drug B can be rationalized as the quality of both components is known and guaranteed. However, randomization in surgery (for instance, minor vs major lymphadenectomy in gastric cancers) may lead to major risks associated with re-evaluation by external audit.6 This is also the case in other regional treatments such as radiotherapy and interventional radiology.
QUALITY CONTROL OF SURGERY DURING A TRIAL IS THE CORNERSTONE
Quality assurance is now the main challenge for surgeons. This is also the case for randomized trials involving neoadjuvant or adjuvant therapies. We need a robust change in how randomized trials are conducted in surgery because nonstandardized surgical practices could distort results. Moreover, an RCT is regarded as evidence of the quality of the scientific evidence. Assigning this interpretation to a biased study risks misinterpretation of misleading results. The consequences of RCTs that were performed and reported without meeting the minimum CONSORT requirements3 need to be addressed. A major task of SURCARE will be to elaborate on the certification process of quality in prospective surgical trials.
ALTERNATIVES TO RCTs IN SURGERY
Although RCTs have a robust internal validity, they also suffer from low levels of generalizability7: the gap between clinical research and real life.8 To ensure that a new treatment will be effective on a frail population—different ages, advanced tumors with metastases, comorbidities—when its efficacy was demonstrated on a highly selected population without these risk factors is challenging and unrealistic. Moreover, there is little evidence that the estimates of treatment effects are overinflated in observational studies compared with RCT,9 especially prospective observational studies.10 Finally, there is a need to change the “RCT or nothing” policy and keywords such as prospective, pragmatic, feasible, and cost effective are needed for this new proposal.
RETHINKING CLINICAL RESEARCH IN SURGERY
Most of the tools to improve scientific reporting in surgery do exist but remain untapped, we propose to restore the interest of both the surgical and methodology communities in them.
Phase 0 is an early “first in human” evaluation for drugs. Very low doses are given to a small number of patients for less than 7 days before the tumors are removed by the surgeon as a standard procedure. Between the preclinical proof of concept and the phase 1, phase 0 can boost the evaluation process and open collaborations between surgeons and biologists in the translational context.
Phase 1 in surgery needs to be introduced as it is the only field in medicine where a practitioner can test a “first in human” procedure or a new device without proper controls in place. It is important to note that the Conforme aux Exigences (CE) label is just an industrial standard and does not provide guidelines on medical procedures. A clinical surgical researcher is like a test pilot, but in this case there is a passenger without a parachute enrolled on the flight. Some minimal solutions, such as the adoption of early prospective databases (in which every new case might be recorded), can continue to be used. “Natural orifice transluminal endoscopic surgery (NOTES)” procedures, “associating liver partition and portal vein ligation for staged hepatectomy (ALLPS)” have on-going prospective databases, awaiting political decisions from both the US medical device post market surveillance system and the European Commission.
Phase 2 is a completely underutilized tool in surgery that could raise the level of scientific reporting. The threshold for efficacy and nonefficacy are defined and with statistical power (risks α and β), the cohort size is calculated (usually, between 30 and 60 patients). The use of interim stopping rules allows reduction of the required number of patients (eg, Simon design). There are several types of phase 2: single phase 2, twin parallel phase 2, randomized phase 2, and so on. Phase 2 tend to produce over optimistic results, hence their sometimes difficult reputation in medicine. However, surgical phase 2 have a strong added value because of their rigorous prospective design that is not often not recognized by editorial teams. For example, the Journal of Clinical Oncology has a restricted policy in phase 2 publishing that completely ignores surgery.11 The journal sometimes rejects a prospective surgical phase 2 on the grounds that it is too small but at the same time might publish a 20-year retrospective series. This restrictive policy must change.
Phase 3 studies and population-based observational research are partners in the evolution of medical evidence and should not be considered to be in opposition.7 For some of us, RCTs in surgery are a niche in clinical research because of the precise need for the reproducibility and quality assurance of the procedures being studied. For others, especially from Japan, RCTs have still to be considered as the best option, whenever possible. In any case, only a large networking system can endorse such trials. Young surgeons should be encouraged to start clinical research by designing nonrandomized, easier, prospective studies before proceeding to consider risky and more intensive randomized phase 3 studies.
Pragmatic and prospective observational studies7 are other tools to revitalize surgical research in the real world. Several types of studies exist using prospective designs with the possibility of regulating the inclusion and exclusion criteria: broad criteria will lead to real life descriptions in outcome studies and narrow criteria will produce quasiexperimental studies that compete with the randomization process. Such reports can use modern approaches such as the Dindo-Clavien classification for complications of surgery or the core outcome measures for effectiveness trials system.2
THE SURCARE PLATFORM
ESSO, EORTC, and JCOG have built a comprehensive platform with all the necessary competencies and facilities to organize pragmatic clinical research in surgical oncology. Specific funding has been obtained with a sustainable commitment over several years from different national anticancer leagues and charitable organizations. Two EORTC surgical fellows are employed full-time and collaborate directly with ESSO's Clinical Research Committee. A collaborative network comprising expert surgical centers in Europe and Japan has been established. The main focus of the project is quality assurance. The first project is CLIMB-EORTC where 1409 enrolling patients with initially inoperable liver-limited colorectal cancer metastases (CLM) with the first quality assessment after enrollment of the first hundred patients converted to resectability with curative intent. The second project is DREAM-EORTC where 1527 patients devoted to the management of CLM by diffusion-weighted MRI in the context of conversion chemotherapy and disappearing CLM. We intend to launch a new project every year, covering all aspects of surgical oncology (breast, sarcoma, elderly patients, and so on).
Much progress can be made by improving surgical techniques. We also know that postoperative complications decrease overall postoperative survival. Consequently, this quality assurance (QA) program may have a strong impact in the clinical context. With SURCARE, we aspire to enroll young surgeons in pragmatic, feasible, high-quality prospective research. The choice can no longer be between retrospective series and RCTs; alternative ways must be found when RCTs are not feasible. SURCARE is one way to unlock this problem for surgeons and produce high-level data with a rapid return on investment for funding organizations. Comparable initiatives do exist in the US and joining our forces together would be beneficial. Lastly, the scientific community has to endorse this new approach.
The authors would like to thank Dr. Ravi Nookala for the medical writer service.
1. Horton R. Surgical research or comic opera: questions, but few answers. Lancet
2. Evrard S, McKelvie-Sebileau P, van de Velde, et al. What can we learn from oncology surgical trials? Nat Rev Clin Oncol
3. Clavien PA, Lillemoe KD. A new policy to implement CONSORT guidelines for surgical randomized controlled trials. Ann Surg
4. Naredi P, La Quaglia MP. The future of trials in surgical oncology. Nat Rev Clin Oncol
5. Buchwald H. Surgical procedures and devices should be evaluated in the same way as medical therapy. Control Clin Trials
6. Bonenkamp JJ, Hermans J, Sasako M, et al. Extended lymph-node dissection for gastric cancer. Dutch Gastric Cancer Group. N Engl J Med
7. Booth CM, Tannock IF. Randomised controlled trials and population-based observational research: partners in the evolution of medical evidence. Br J Cancer
8. Neuman HB, Greenberg CC. Comparative effectiveness research: opportunities in surgical oncology. Semin Radiat Oncol
9. Concato J, Shah N, Horwitz RI. Randomized, controlled trials, observational studies, and the hierarchy of research designs. N Engl J Med
10. Benson K, Hartz AJ. A comparison of observational studies and randomized, controlled trials. N Engl J Med
11. Cannistra SA. Phase II trials in Journal of Clinical Oncology
. J Clin Oncol