The generalisability of randomised clinical trials: an interim external validity analysis of the ongoing SENOMAC trial in sentinel lymph node-positive breast cancer

Abstract

Abstract Purpose None of the key randomised trials on the omission of axillary lymph node dissection (ALND) in sentinel lymph-positive breast cancer have reported external validity, even though results indicate selection bias. Our aim was to assess the external validity of the ongoing randomised SENOMAC trial by comparing characteristics of Swedish SENOMAC trial participants with non-included eligible patients registered in the Swedish National Breast Cancer Register (NKBC). Methods In the ongoing non-inferiority European SENOMAC trial, clinically node-negative cT1–T3 breast cancer patients with up to two sentinel lymph node macrometastases are randomised to undergo completion ALND or not. Both breast-conserving surgery and mastectomy are eligible interventions. Data from NKBC were extracted for the years 2016 and 2017, and patient and tumour characteristics compared with Swedish trial participants from the same years. Results Overall, 306 NKBC cases from non-participating and 847 NKBC cases from participating sites (excluding SENOMAC participants) were compared with 463 SENOMAC trial participants. Patients belonging to the middle age groups (p = 0.015), with smaller tumours (p = 0.013) treated by breast-conserving therapy (50.3 versus 47.1 versus 65.2%, p < 0.001) and less nodal tumour burden (only 1 macrometastasis in 78.8 versus 79.9 versus 87.3%, p = 0.001) were over-represented in the trial population. Time trends indicated, however, that differences may be mitigated over time. Conclusions This interim external validity analysis specifically addresses selection mechanisms during an ongoing trial, potentially increasing generalisability by the time full accrual is reached. Similar validity checks should be an integral part of prospective clinical trials. Trial registration: NCT 02240472, retrospective registration date September 14, 2015 after trial initiation on January 31, 2015