Publication and Interpretation of Clinical Trial Results: The Need for Caution

Abstract

When the results of a clinical trial are published, modern information technology ensures that news of the trial outcome is disseminated rapidly to the medical community worldwide, and often to the general public as well, through the media. However, there is always a danger that physicians, patients, and the general public may be inappropriately influenced by the ‘‘spin’’ that is put on what they read or hear, without subjecting the trial report to critical evaluation. The adverse and inappropriate effects that the publication of clinical trial results can have on medical practice were discussed in an editorial by Kaplan, who pointed out the heavy responsibilities that fall on the shoulders of reviewers and journal editors to ensure that trials do not have serious design flaws and that their results are clearly and accurately reported. The publication in February 2011 of the results of a large, randomized study assessing the value of completion lymph node dissection (CLND) in sentinel node (SN)positive breast cancer patients illustrates how problems of misinterpretation can arise. The trial report was published in the Journal of the American Medical Association (JAMA), a leading international medical journal, and it was featured on the front page of the New York Times shortly afterwards. The title of the report in JAMA was ‘‘Axillary dissection versus no axillary dissection in women with invasive breast cancer and sentinel node metastasis: a randomized clinical trial.’’ However, the dramatic headline in the New York Times read, ‘‘Lymph node study shakes pillar of breast cancer care,’’ and there were similar alarmist headlines in other major newspapers around the world. As a result of this widespread press publicity, we have encountered patients with melanoma who are not willing to consider entry into another large clinical trial, this one assessing the value of CLND when metastatic melanoma has been found in a SN (MSLT-II—Clinicaltrials.gov identifier NCT00297895). These patients have become aware that in the trial report in JAMA, Giuliano and colleagues concluded that CLND is not necessary for breast cancer patients found to be SN-positive. The patients assume that the situation for SN-positive melanoma patients will be the same, not understanding that the two tumor types behave in quite different ways, and that in any case all the SN-positive patients with breast cancer enrolled in the trial received radiation therapy to the axilla, and most (96.5 %) also received systemic chemotherapy (known to be effective for breast cancer, but not for melanoma). Another trial, published in a major medical journal in November 2009, also is having a major impact on clinical trial accrual in patients with metastatic melanoma, in this case those with cerebral metastases. The trial was reported by Chang and colleagues in Lancet Oncology and will be referred to hereafter as the Chang trial. It studied patients with cerebral metastases from cancers of various types who were treated with stereotactic radiosurgery (SRS) with or without whole brain radiotherapy (WBRT). The Chang trial was stopped early because one of six neurocognitive function assessment tools showed a detrimental result at 4 months. It is unknown whether this was a temporary or permanent effect of WBRT, because the trial was terminated before sufficient follow-up times had been achieved to assess long-term results. There are many potential pitfalls in the design, conduct, and reporting of clinical trials, and the Chang trial appears to have had several flaws. Some of these have been detailed in published correspondence, but there are a Society of Surgical Oncology 2012