Breast Cancer TomoTherapy Positioning and Verification

Patients with breast cancer involving internal mammary chain and clavicular nodes are routinely treated with TomoTherapy in our center. Positioning is started with aligning the patient in the TomoTherapy virtual isocenter using lasers or Catalyst. The couch is displaced 70 cm towards the real isocenter in the cranial direction (+Y) and a daily MVCT image is taken. We use coarse slices and fine reconstruction leading to 3 mm slice thickness. After this MVCT image, the couch is displaced back to the virtual isocenter, after which the MVCT image fusion with the planning kVCT is performed. Finally, the position is corrected for in X, Y, Z and roll (left–right, craniocaudal, ant–post, roll) and displaced back towards the gantry for treatment.
All treatments were performed in 25 fractions. Boost volumes, when present, were treated by delivering 2.4 Gy/fraction as integrated boost. Treatment planning is performed in the helical mode with TomoEdge⁽²⁰⁾ (Accuray) using a 5 cm field width and using details described in the previous work by Crop et al.⁽²¹⁾ This results in a treatment beam‐on time of approximately 5–7 min. Typical volumes of the different PTVs were in the order of 40 cm³ for internal mammary chain PTV, 1000 cm³ for the breast, 100 cm³ for infraclavicular node, and 100 cm³ for the supraclavicular node.
The patient is positioned with both arms up on a breast board with an inclination of 7°. The TomoTherapy couch is positioned as low as possible: 21–22 cm below the virtual isocenter. This allows for the highest amount of liberty, low thread effect,⁽²²⁾ and better Catalyst camera view. The difference between the initial position and positioning result after MVCT fusion was evaluated in order to compare laser‐based and Catalyst‐based positioning. Forty patients with Catalyst‐based setup and 55 patients with laser‐based setup where included, resulting in, respectively, 810 Catalyst‐only sessions and 666 laser positioning‐only sessions. Thirty‐one of these patients had both laser‐based and Catalyst‐based setup on different days. Couch sag for TomoTherapy is different in comparison with the initial kVCT image. This leads to a different height position of the patient in the real isocenter, 70 cm further located, compared to the position in the virtual isocenter. This couch sag thus depends on the patient weight and the treatment location. Couch sag was evaluated with two methods: by adding weight to a MVCT‐scanned phantom and as the uncorrected mean Catalyst‐image bias between patient position outside the bore and inside the bore.