4\u2005False hope, capitalising on fear, and cultural obstacles to heeding randomised trial evidence concerning secondary surveillance and metastasectomy in colorectal cancer

Abstract

Objectives The body’s total blood volume is filtered through the lungs about once every minute and circulating ‘seeds’ of carcinoma find a fertile ‘soil’ in the lung’s interstices. Growing there, metastases are easily visualised against the radiolucent air-filled lung and in the course of cancer surveillance. Increasing numbers of people are sent for surgery. What we already know is that patients considered for these operations have no symptoms from their metastases and these low volume deposits do not contribute to symptoms or the terminal course of disease. So there was no palliative role for these operations. The hope is that removing these nodules cure the cancer. Benefit is assumed. Long-term survival is reported anecdotally but is this attributable to the operation or the selection of natural survivors? Is detection for metastasectomy an instance of overdiagnosis leading to unavailing treatment? We embarked on stepwise studies to find answers to these questions. Method Systematic review and quantitative synthesis was conducted with colleagues in the Clinical Operational Research Unit at University College London (UCL). By far the commonest lung metastases operated on are from colorectal cancer so we searched for clinical reports and extracted all available data to build a full picture of practice and outcomes. Systematic review and meta–analysis of randomised trials (RCTs) of surveillance of increasing intensity for recurrence of colorectal cancer was conducted in collaboration with Erasmus University Rotterdam, Cambridge MRC Biostatistics Unit, the Surgical and Interventional Trials Unit (UCL) A ‘Big data’ analysis was conducted to re–examine published NHS data The PulMiCC trial (Pulmonary Metastasectomy in Colorectal Cancer) is an RCT recruiting participants referred for consideration of lung metastasectomy into a registered cohort. Those eligible, and willing to accept randomised treatment assignment, and for whom the clinician could reach equipoise, entered the PulMiCC trial. Results 1. Systematic Review and Quantitative synthesis. There were no RCTs. 51 observational reports 1971–2007 described 3504 patients with consistent features. 60% had a single metastasis, the median interval since primary cancer was 36 months, and 60% were dead within five years. There were no data on symptoms, quality of life or of comparable non-operated patients. (JRSM 2010;103:60–66) 2.Meta-analysis. 16 RCTs. Surveillance advanced the diagnosis of cancer by 1–2 years, providing opportunities for metastasectomy, but with a detrimental (non-significant) effect on survival. (BJS 2016;103:1259–1268) 3.Big data analysis. 114,155 people had colorectal cancer surgery and 3,116 (2.7%) had liver resections. Analysis of the survival curve showed a plateau characteristic of selection and immortal time biases. (Cancer Epidemiology 2017;49:152–155; 2018;52:160–161) 4.PulMiCC: 512 patients were recruited but only 65 were analysable in the RCT. Analysis of reasons for not randomising revealed overwhelming clinician bias. There was no evident benefit from metastasectomy. (Publication pending) Conclusions This methodical succession of studies failed to find evidence for a survival benefit, reasonably attributable to the detection and surgical removal of lung metastases. The reason for a clinical impression of benefit is selection of <3%, i.e the most favourable cases. Similarly selected patients, assigned to a non-operated control arm, had similar survival, much better than assumed, without the predicted burden of anxiety. Timely diagnosis and interventions for local recurrence are of benefit but probably not lung metastasectomy. But recruitment into PulMICC was actively prejudiced by a thoracic surgical journal Editorial opening ‘Surgery for pulmonary metastases is a pillar of modern thoracic Surgery’, decrying the trial, and alluding to the parachute analogy. It will be difficult to roll back this practice. Monitoring for cancer recurrence is generally seen as the right thing to do. There would be financial consequences for imagers and surgeons if a halt were to be called.