Background: We studied the use of Lexicomp®, an online drug information database, for adequate identification of drug-drug interactions (DDIs) within Comprehensive Geriatric Assessment (CGA) in cancer patients. Materials and Methods: Data of 149 onco-geriatric patients were reviewed. Sixty-three percent participated in an observational study recruiting head and neck cancer patients (H&N-group), 37% in a registry recruiting general oncology patients (GO-group). Baseline drug information was collected by a health professional, through the medical interview within CGA. Drug class usage was quantified and potential DDIs were assessed and categorized (risk rating "C": monitor therapy, "D": consider therapy modification, "X": avoid combination) with Lexicomp®. Results: On average, H&N and GO-patients took 5 and 8 prescription drugs at presentation, respectively. An average of 4 drugs were added in both groups as part of their proposed therapy. Potential DDIs (n=211 H&N; n=247 GO) were detected by Lexicomp® in 64.9% (85.3% "C", 14.7% "D", 0% "X") and 83.6% (83.4% "C", 15.8% "D", 0.8% "X") of H&N and GO patients, respectively, at therapy start. Administration of cancer-therapy-related drugs lead to additional DDIs (n=75 H&N; n=68 GO) in 73.7% and 58.3% of H&N and GO cases, respectively. DDIs occurred mainly with supportive drugs (100% H&N and 83.8% GO). Sixteen percent of potential DDIs were identified with anti-neoplastic drugs in the GO-group. In 28.7% and 60.0% of H&N and GO patients, respectively, at least one drug was not recognized by Lexicomp®. Conclusions: Use of Lexicomp® drug database within CGA is feasible. It could reduce the administration of inappropriate drugs, and in that way improve the quality of patient-individualized therapy.