We read with interest the article by Birchard et al1 and commend their efforts to optimize size-based response assessment. However, we do not accept the conclusion that “change in tumor size does not accurately reflect outcome.” A North Central Cancer Treatment Group analysis of 3 trials in lung and colon cancer revealed that response at 12 weeks was associated with survival.2 Yet Birchard et al observed equivalent survival regardless of progression, contrary to these data as well as common clinical experience. This unexpected finding makes us reconsider the validity of their dataset and methods of analysis.
Their patient sample may not have been optimal. Their study had 99 patients with nonsmall cell lung cancer who had a favorable median survival (30 months vs 12 months in clinical trials), which gives the results limited generalizability to patients with a poorer prognosis. Thirteen patients were excluded from the analysis because of the lack of a follow-up computed tomography scan; those patients may have had early symptomatic progression, which would have affected significantly the survival of the progressive disease cohort. In addition, the authors did not describe the clinical characteristics of their cohorts (disease stage, performance status, smoking status), which may confound the survival findings.
Another limitation was several unusual modifications to the Response Evaluation Criteria in Solid Tumors (RECIST). The authors analyzed measurements obtained from a single target lesion despite recommendations to measure up to 5 lesions. It has been demonstrated that the use of a single lesion to measure disease status results in potential miscategorization of up to 67% of patients and can result in response/progression misclassification between 9% and 18% of the time.3, 4 In addition, >33% of patients in their study had their primary cancer irradiated, and the authors noted that the primary frequently was selected as the single target lesion. Measurements of irradiated lesions do not reflect the impact of systemic therapy; for this reason, response guidelines recommend not measuring target lesions in irradiated fields.
We share the authors' concern that RECIST measurements may be inadequate for the accurate determination of response to therapy and that other imaging endpoints will be useful. However, we believe that early change in tumor size continues to be valuable in the assessment of the activity of antineoplastic therapy and has potential as a surrogate for survival.