Background: The Brazilian Medical Association (BMA) had started in 2001 the ‘Guidelines Project’ which should be the official recommendations to medical practice in Brazil. The BMA oriented that the guidelines must be developed by Specialty Societies, according to principles of evidence-based medicine. Ideally, recommendations in guidelines should be based on strong evidence as summarized in a Systematic review (SR). However, some case studies have shown that the use of reviews in guidelines is quite low and could be improved. Objectives: To identify the number of systematic reviews, including Cochrane reviews, used to answer the clinical questions in guidelines on cancer developed by Brazilian Medical Association. Methods: The latest version of the BMA’s Guidelines Project was reviewed. We searched for guidelines developed to answer questions in oncology. The guidelines could be about epidemiology, risk factors, prevention, diagnosis and treatment. A search for SRs used to support recommendations was made evaluating the citations in the bibliography of each guideline. Another search for SRs that could be included in each guideline according to the theme was made in MEDLINE and The Cochrane Library. We calculated the number of SRs cited and verified the inclusion of the relevant SRs. Results: We analyzed 30 clinical practice guidelines of the BMA’s project published in 2007. There were recommendations for 14 types of cancer, including breast, endometrial, cervical, colorectal, kidney, bladder, prostate, penis, adrenal, thyroid, skin, head and neck cancer. There were 53 citations of SR, eight of them produced by The Cochrane Collaboration representing an average of 1.8 SR cited for each guideline. Our additional search retrieved an average of 55 SR (two Cochrane reviews) for each guideline until the date of publication. Conclusions: The number of SRs used to support the statements of the Brazilian official guidelines on oncology was extremely low. The medical specialty societies used many expert opinions in the development of recommendations. Some of them clearly were done based on just a few expert opinions. Collaboration between review groups and guideline developers could enhance the effective production and use of strong evidence in clinical guidelines.