When prospective randomised trials have unexpected results: where do we go from there?

Abstract

With the modernisation of laparoscopy and subsequent rapid uptake of the robotic platform, minimally invasive surgery (MIS) became the widely accepted standard of care for gynaecological cancer. Virtually all data supporting improved patient outcomes and oncological safety for MIS in gynaecological cancer derive from retrospective studies. Nevertheless, the advantages of MIS have rapidly been accepted, not only because of the perceived benefits to patients. Surgeons appreciate improved optics, magnification, and the facility of performing complex, delicate surgery on an increasingly obese female population. A generation of gynaecological oncology fellows in the USA has been trained in MIS, with a trend towards minimally invasive hysterectomies becoming the norm. Dr Ramirez and colleagues’ strong belief in MIS for patients with cervical cancer led to the Laparoscopic Approach to Cervical Cancer (LACC) study, a well-designed, prospective (WDP) randomised controlled trial (RCT) with the primary hypothesis that MIS was not inferior to laparotomy with respect to overall survival (Ramirez et al. N Engl J Med 2018;20:1895–1904). Unexpectedly, this study required early closure and literally staggered the authors and the gynaecological oncology community. Patients undergoing MIS had a four-fold increased disease recurrence and a six-fold increased rate of death from disease compared with patients undergoing laparotomy. A database analysis published simultaneously in the New England Journal of Medicine supported the RCT results (Melamed et al., N Engl J Med 2018;379:1905–14). In anticipation of the publication of these data, sceptical gynaecological oncologists published retrospective data refuting the study's findings. The current paper is reassuring for patients in the UK, reporting oncological safety in 779 women treated at eight academic centres (Martin-Hirsch et al. BJOG 2019;126:956–9) Unfortunately, it is not rational to challenge results of a WDP RCT with retrospective data and their inherent biases. WDP RCTs are the ‘gold standard’ of medicine providing the most rigorous data upon which to base future clinical care. Why, then, are we so reluctant to accept the results? Even WDP RCTs sometimes fail to influence medical practice (Bothwell et al. N Engl J Med 2016;374:2175–81). WDP studies have been repeated multiple times for gynaecological cancer, including studies of neoadjuvant chemotherapy and intraperitoneal chemotherapy. It is important to note that the LACC study will never be repeated; there will never be the equipoise required for ethical randomisation. The clinical community is therefore left to decide how to interpret and manage the results. Patients with cervical cancer are vulnerable – often disadvantaged socially, economically and educationally – making a shared decision-making strategy challenging. I agree with the authors of the current paper that ‘applying the results of RCTs is not always entirely straightforward’ and applaud their thoughtful analysis of how best to approach the results of the LACC trial. Surgical RCTs have multiple challenges, including establishing appropriate inclusion criteria and standardising surgical interventions. Perhaps, as the authors suggest, there are certain patients (women with very small tumours or women who are very obese) for whom MIS would be appropriate. Unfortunately, the LACC trial cannot answer questions regarding specific groups of patients: the study was not powered to address the safety of MIS for women who had stage-IA or small (<2 cm) stage-IB tumours, nor did the study include a sufficient number of women with a body mass index (BMI) of >30 kg/m2. Exceptions to standards of care can and will always be made in best practice based upon individual, patient-specific factors. But, at this time, the best evidence supports open surgery for the treatment of early cervical cancer. None declared. Completed disclosure of interests form available to view online as supporting information. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.