First, The authors did not comment on the category of research which could be particularly relevant to the local population including screening for early detection, correlative translational research for mechanisms of drug resistance and predictive markers of response, pharmacogenomic-based studies to explore variations in risk of toxicity and tumor response, and biomarker led studies to look for sub-populations of maximal treatment benefit. Classifying publications based on tumor sub-type and the cancer incidence could further reflect the research activity in different cancer sites.

Second, the authors used the 2011 Oxford Centre of Evidence-Based Medicine Levels of Evidence (OCEBM-LOE) to grade the quality of publications which is a widely acceptable tool for evaluation. But this system has been questioned due to missing clear definition of study design limitations, and neglecting the value of non-randomized trials in certain circumstances where randomization might not be an ethical or practical approach.  It is also hard to rely on the current grading systems to make a conclusion on significance of the study in respect to novelty of the research and the impact of previous work on our understanding of cancer and quality of patients’ lives. Moreover, observational studies remain the most appropriate method to evaluate rare side effects in post-marketing surveillance, particularly taken into account the genetic variability across populations which could influence treatment response and toxicity.  In another aspect, the authors compared publications of the 2 periods in respect to the journal impact factors, but there have been discussions regarding the merits and reliability of the journal impact factor as a reflection of the publication impact and quality...