IMRT

96

Breast IMRT Cone-beam imaging during breast IMRT

Table 1 Patient translational setup errors. Patient L/R cancer

Fig. 5. Spatial location of low-dose PTV regions at planning (a) and throughout subsequent treatment days (b–i) for P7 (worst-case).

Individual systematic error (cm)

Individual random error (cm)

X = lateral

Y = vertical

Z = longitudinal

X = lateral

Y = vertical

Z = longitudinal

95

iotherapy and Oncology 90 (2009) 93–98

0.58

0.50 0.77 0.83 0.26 0.79 0.41 0.21 0.30 0.57 0.61

1 2 3 4 5 6 7 8 9

L L

! 0.23 ! 0.11 ! 0.34 ! 0.07 ! 0.02 ! 0.33 0.30 ! 0.05 0.20 ! 0.65

0.29 0.36 0.43 0.39 0.21 0.28 0.39 0.33 0.44 0.50 0.39

0.30 0.34 0.37 0.38 0.20 0.46 0.40 0.25 0.34 0.28 0.35

0.27 0.37 0.39 0.19 0.38 0.37 0.30 0.17 0.24 0.28 0.32

r each ot vol- 5% and

! 0.02 0.10 ! 0.68 0.12

R

L L

0.53 1.24 0.71 0.87

R R

L

of dai- nt sys- n and tic and f indi- d devi- ted by he vol- on was or and

R

10

L

! 0.20

Population error

0.57

0.28

0.23

Discussion

volume increases (>15%) were observed in patients 2 and 4; the largest percentage reduction (>15%) was observed in patient 9. Daily PTV contours overlaid onto the planning CT are illustrated for an example patient ( Fig. 4 ). Variation in patient positioning Table 3 Impact of daily PTV variations upon IMRT and conventional RT plans. Isodose level IMRT plan No patients exceeded conventional verification tolerance limits. The mean systematic and random components of setup error determined for each patient using CBCT during the treatment course are given in Table 1 . Values greater than 5 mm are shown in bold. All patients demonstrated a mean systematic error >5 mm in at least one axis, mainly in the lateral ( X ) and longitudi- nal ( Z ) axes. The population lateral systematic error was also >5 mm, where shifts tended towards the left. All longitudinal shifts were in the inferior direction. Mean individual and population ran- dom errors were all 6 5 mm. Table 2 shows the frequency and size of rotations recorded for all patients. Rotations >2 ! were most frequent around the lateral ( X ) axis, and all occurred in the positive (backwards) direction. Y Mean planned volume (% PTV) Mean delivered volume (% PTV) >107% 0.1 0.3 ± 0.6 4.0 >105% 0.5 1.8 ± 2.0 15.7 >95% <105% 96.7 89.5 ± 5.1 82.5 <95% 2.7 8.6 ± 4.0 1.8 <90% 0.3 2.7 ± 2.9 0.2 Fig. 4. Example RT planning image with original (pale blue) and superimposed daily PTV contours. Table 2 Total rotational setup errors. Rotations Lateral ( X ) axis (%) Vertical ( Y ) axis (%) Longitudinal ( Z ) axis (%) 0 6 2 ! 70.8 84.9 98.1 2 ! 4 ! 22.6 12.3 1.9 >4 ! 6.6 2.8 – PTV position Standard tangential RT plan Mean planned volume (% PTV) Mean delivered volume (% PTV) 4.4 ± 1.9 15.6 ± 3.7 79.3 ± 5.1 5.1 ± 2.6 1.9 ± 2.0

P. Jain et al./Radiotherapy and Oncology 90 (2009) 93–98

High dose

Fig. 6. Spatial location of high-dose PTV regions at planning (a), and throughout subsequent treatment days (b–i) for P3 (worst-case). The grey dot represents the PTV isocentre.

P. Jain et al. / Radiotherapy and Oncology 90 (2009) 93–98 Despite meeting conventional setup verification tolerances, a mean systematic shift >5 mm was measured using CBCT in at least one axis in all women. It is possible that variations may be exagger- ated, as patientswere required to lie in treatment position for longer, and the required couchmovesmay have introduced some errors. Im- age registration for patients with limited data CT scans was also Fig. 5. Spatial location of low-dose PTV regions at planning (a) and throughout subsequent treatment days (b–i) for P7 (worst-c This study is the first to use daily on-treatment CBCT to assess 3D inter-fraction patient/organ motion throughout a course of breast RT using IMRT, and to assess the consequences upon both IMRT and standard RT dose homogeneity. Breast depth and daily PTV varied from planning and fluctuated throughout the treatment course for all patients [12,13,23] . It is likely that the breast depth variations were due to short-term oe- dema and vascular effects as a result of prior treatments, showing that independent organ motion remains an important consider- ation for improved verification methods. Unlike a previous study [23] , we found larger PTV changes (>15%) in some patients, and no apparent pattern or regression trend ( P = 0.366), indicating ran- dom organ/patient motion over the RT course. Low Dose

ily EPI Fig. 2 ). and in- neated ations entage

Pt 1 Pt 2 Pt 3 Pt 4 Pt 5 Pt 6 Pt 7 Pt 8 Pt 9 Pt 10