Outcome Reporting in Neoadjuvant Surgical Trials: A Systematic Review of the Literature and Proposals for New Standards

N. S. Blencowe, P. Chana, R. N. Whistance, D. Stevens, N. A. C. S. Wong, S. J. Falk, J. M. Blazeby
2014 Journal of the National Cancer Institute  
The use of neoadjuvant therapy before surgery for gastrointestinal cancer is increasing; however, patients may not complete both treatment components. Understanding completion rates of each treatment stage is necessary for treatment evaluation and to inform decision-making. This study evaluates reporting for recent neoadjuvant surgical trials, focusing on treatment progression and other key outcomes. Methods Systematic literature searches identified randomized and nonrandomized phase II and III
more » ... studies evaluating neoadjuvant treatment and surgery for esophageal, stomach, and colorectal cancer, and colorectal liver metastases. Rates of reporting of failure to complete neoadjuvant treatment, nonprogression to surgery after neoadjuvant treatment, and nonresection at planned surgery were assessed. For each measure, reporting was categorized as "full, " "partial, " and "absent" according to predefined criteria, and reasons for nonprogression at each stage of treatment were examined to inform proposed standards. Results Of 9854 abstracts, 123 papers were reviewed and 62 articles were included, reporting outcomes for 9126 patients. Details of noncompletion of neoadjuvant treatment and nonprogression to surgery were completely absent in 21 (33.9%) and 19 (30.6%) studies, respectively. Reporting of nonresection at planned surgery was also deficient, with 21 (33.9%) studies providing no information about this outcome. Reasons for noncompletion and nonprogression were similar and included disease progression, treatment toxicity, and patient choice. Common reasons for nonresection were locally advanced disease and the discovery of unsuspected metastases. Conclusions Reports of recent neoadjuvant surgical trials often fail to include treatment progression and other key outcomes. These findings support the need for minimum reporting standards.
doi:10.1093/jnci/dju217 pmid:25106656 fatcat:7lmiaf7tufe4fjia3tncs7v7sy