Synthetic CT for single-fraction neoadjuvant partial breast irradiation on an MRI-linac

Area of Science:

• Medical Physics
• Radiotherapy Physics
• Image-Guided Therapy

Background:

• Synthetic computed tomography (sCT) is essential for optimizing radiotherapy plans on MRI-linacs.
• Limited data exists on the dosimetric accuracy of sCT for breast cancer radiotherapy.

Purpose of the Study:

• To evaluate the dosimetric accuracy of single-fraction neoadjuvant partial breast irradiation (PBI) plans on a 1.5 T MRI-linac using both bulk-density and deep learning-generated sCTs.
• To determine the optimal number of bulk-density levels for accurate dose calculations.

Main Methods:

• Created three bulk-density sCTs (sCT_BD1, sCT_BD2, sCT_BD3) and one deep learning sCT (sCT_DL) from planning CTs for ten breast cancer patients.
• Optimized and recalculated single-fraction neoadjuvant PBI plans on each sCT and the planning CT.
• Evaluated image accuracy using Mean Absolute Error (MAE) and Mean Error (ME) in Hounsfield Units (HU).
• Assessed dosimetric accuracy via dose differences, gamma pass rates, and dose-volume histogram (DVH) differences.

Main Results:

• MAE within the body contour was 106/104/104/75 HU for sCT_BD1/sCT_BD2/sCT_BD3/sCT_DL, respectively.
• Mean dose difference in the PTV_GTV was 0.15/0.00/0.00/-0.07 Gy.
• Median gamma pass rates (2%/2 mm) were 98.9/98.9/98.7/99.4%.
• DVH parameter differences were below 2% for most structures, except for the skin with sCT_DL.

Conclusions:

• Accurate dose calculations for single-fraction neoadjuvant PBI are feasible on both bulk-density and deep learning sCTs for MRI-linac breast radiotherapy.
• sCT_BD2 represents an optimal bulk-density approach, balancing simplicity and accuracy.
• These findings support the implementation of MRI-guided radiotherapy for breast cancer treatment.