Objective Co‐morbidity of depressive symptomatology is a common indication for use of mental health services in oncology. In this regard, screening instruments are useful for prompt identification of mental disorders in cancer. This study is set to evaluate the diagnostic validity of Centre for Epidemiological Studies Depression Scale Revised (CES‐DR) for depression screening in cancer. Methods The CES‐DR and the Schedule for Clinical Assessment in Neuropsychiatry (SCAN) were administered by the researchers on 200 attendees of a Nigerian hospital with histological diagnoses of cancer. Subsequently, the diagnostic validity of CES‐DR was compared with SCAN. Results Ninety‐eight (49.0%) participants had significant depressive symptomatology (CES‐DR scores of ≥16) as against the diagnosis of depression in 55 (27.5%) participants following SCAN interview. Furthermore, of these 55 (27.5%) depressed participants, two (3.6%) participants had CES‐DR scores <16 (non‐cases). The Cronbach's alpha reliability of CES‐DR was 0.86, and sensitivity and specificity of CES‐DR were 96.4% and 68.7%, respectively, whereas positive and negative predictive values of CES‐DR were found to be 0.54 and 0.98, respectively, in this study. The average administration time of CES‐DR was 6 (±2) min, and an inter‐rater reliability of 93.7% was observed. Conclusions The CES‐DR was found in this study to be a useful tool for screening for depression in cancer but with diagnostic limitation when compared with SCAN. The development as well as popularization of screening instrument(s) with improved diagnostic and administration property for prompt identification of mental disorders to improve consultation–liaison psychiatry services in cancer care is recommended. Furthermore, replication of similar research is warranted. Copyright © 2012 John Wiley & Sons, Ltd.