[en] To improve bolus reliability, we compared three different cavity sizes generated by different bolus thicknesses in patients by introducing intensity-modulated radiation therapy (IMRT) after radical mastectomy and analyzed the differences in dosimetry. Three IMRT plans-1, 2, and 3-were designed for cavity sizes greater than 10 mm, greater than 5 mm and less than 10 mm, and less than 2 mm, respectively. All three plans were compared to Plan 4 (bolus clinging to the skin), which was conventionally used as a standard benchmark case. The present study analyzed the dose-volume histogram (DVH) and three-dimensional dose distribution profiles. The differences in dose distributions between the target area and the organ at risk were evaluated by statistical analysis on statistical package of social sciences (SPSS). The dose parameters of all the four plans satisfied the prescription dose requirements and no statistically significant differences were observed (p < 0.05). However, the dose distribution map showed that the target areas in the cavities had high concentration areas and cold areas nearby the skin for both Plans 1 and 2. The DVH results showed reduced mean dose (D_mean) to the esophagus and trachea, percent volume received by the affected lung (V₃₀ ≥ 30 Gy), and percent volume received of the auxiliary structure ring (V₄₀ ≥ 40 Gy and V₄₅ ≥ 45 Gy) for Plans 3 and 4. The differences were statistically significant (p < 0.05). Comparisons showed that Plan 3 reduced hot and cold spots around the target area, therefore reducing the risk of organ damage and indicating that it was the best plan in the present study. Plan 3 used a superposition method from 2 mm to 3 mm, which reduced the cavity size to less than 2 mm. It satisfies the compensation function of the bolus and avoids dosimetry error caused by the physical characteristics of the bolus, thus improving treatment quality. (authors)