Comparison of 3D conformal breast radiation treatment plans using the anisotropic analytical algorithm and pencil beam convolution algorithm

Purpose: To investigate (1) dosimetric differences between plans calculated using the anisotropic analytical algorithm (AAA) and pencil beam convolution (PBC) algorithm, (2) the plan quality achieved using AAA compared to PBC based on dosimetric parameters and (3) discrepancies with an independent MU verification calculation for breast treatment planning.

Materials and methods: This study included 10 lumpectomy (Group I) and 10 mastectomy (Group II) cases. Target volumes were defined as breast for Group I and chest-wall for Group II based on the isodose distribution of PBC plans in order to evaluate plans. All plans were initially calculated with PBC. For study aim (1), plans were re-calculated using AAA with the same monitor units (MUs). For study aim (2), plans were calculated using AAA with modifications of wedges, subfields and beam weightings from the original plans to achieve optimal coverage. For study aim (3), independent MU verification was performed. A 3% difference between primary MUs and verification MUs was considered an action level.

Results: (1) Plans using PBC overestimate the dose to the target volume compared to plans using AAA (Group I V(95%)=90.4%:84.4%; Group II V(95%)=83.0%:74.5%; PBC:AAA). (2) The new plans using AAA achieved similar target coverage to the original PBC plans based on dose-volume histograms (DVHs). Yet, the high-dose volume (V(105%)) was significantly larger in AAA plans than PBC plans for Group II (V(105%)=19.5%:24.0%). For both groups, there was a significant increase in the ipsilateral lung volume receiving low dose with AAA plans (Group I V(5 Gy)=23.6%:39.9%; Group II V(5 Gy)=21.2%:33.6%). Isodose distributions of AAA plans displayed insufficient coverage in the superior area. (3) In Group I, all PBC cases passed MU verification versus 30% of AAA cases. In Group II, 80% of PBC plans versus 65% of AAA plans were within the action level.

Conclusions: Plans using AAA calculation can achieve a similar level of target coverage based on DVH as PBC calculation. Nevertheless, the dose distribution shows insufficient coverage in the superior area with AAA plans compared to PBC plans. The lung volume receiving the low-dose (i.e. 5 Gy) is larger and the dose to the skin is greater in AAA plans than PBC plans. Compared with PBC calculation, a larger tolerance in discrepancy between AAA and independent MU verification should be allowed to account for the inadequate heterogeneity corrections in the latter.