JNCI | Editorials 1817 In this issue of the Journal, Bonilla et al. (1) present the results of what the authors call a systematic review and meta-analysis of dosedense chemotherapy in nonmetastatic breast cancer. The rationale for dosing chemotherapy at shorter intervals (“dose-dense” chemotherapy) is based on Gompertzian kinetics, as described by Norton et al. (2). This mathematical model predicts that tumor doubling time decreases with increasing tumor size. Conversely, tumor cells grow faster as the tumor burden decreases with the initiation of chemotherapy (2). Dose-dense chemotherapy is predicted to attack rapidly dividing tumor cells more effectively compared with the conventional dosing schedule. The safety and feasibility of dosedense chemotherapy dosing were first demonstrated in a pilot phase II study that used an anthracycline-based chemotherapy agent with growth factor support (3). The results were validated in a large phase III trial, the cancer and leukemia group B (CALGB) trial 9741, which showed better clinical outcomes in the dose-dense arm when using an anthracycline and taxane-based chemotherapy regimen (4). Since this landmark trial, the dose-dense regimen has been widely adopted in clinical practice. It is important to distinguish the concept of dose density from dose intensity. Dose intensity is defined by the amount of treatment delivered per unit of time (5). For example, 60 mg/m2 of doxorubicin given every 3 weeks results in a dose intensity of 20 mg/m2/wk. Multiple other studies comparing frequent dosing with conventional dosing schedule have been published in an attempt to validate the results published by Citron et al. (4) for CALGB 9741. Some of these studies are included in the article by Bonilla et al. (1) and are discussed. Meta-analysis is a statistical combination of results from two or more separate studies on a specific subject (6). The purpose of the meta-analysis is to estimate the effect size by means of the weighted average. The advantages of meta-analysis include higher statistical power than a small individual study; a meta-analysis also allows for generalizations to the studied population. However, meta-analysis has several potential weaknesses. Even a statistically Meta-analysis: Should It be More Than the Sum of Its Parts?