Breast intensity-modulated radiation therapy (IMRT) improves dose distribution homogeneity within the whole breast. Previous publications report the use of inverse or forward dose optimization algorithms. Because the inverse technique is not widely available in commercial treatment planning systems, it is important to compare the 2 algorithms. The goal of this work is to compare them on a prospective cohort of 30 patients. Dose distributions were evaluated on differential dose-volume histograms using the volumes receiving more than 105% (V(105)) and 110% (V(110)) of the prescribed dose, and on the maximum dose (D(max)) or hot spot and the sagittal dose gradient (SDG) being the gradient between the dose on inframammary crease and the dose prescribed. The data were analyzed using Wilcoxon signed rank test. The inverse planning significantly improves the V(105) (mean value 9.7% vs. 14.5%, p=0.002), and the V(110) (mean value 1.4% vs. 3.2%, p=0.006). However, the SDG is not statistically significantly different for either algorithm. Looking at the potential impact on skin acute reaction, although there is a significant reduction of V(110) using an inverse algorithm, it is unlikely this 1.6% volume reduction will present a significant clinical advantage over a forward algorithm. Both algorithms are equivalent in removing the hot spots on the inframammary fold, where acute skin reactions occur more frequently using a conventional wedge technique. Based on these results, we recommend that both forward and inverse algorithms should be considered for breast IMRT planning.