ESTRO 38 Abstract book

S1094 ESTRO 38

Results CTV1 coverage was not affected by the simulated isocenter shifts. For anatomical changes, weight gain and nasal cavity filling deteriorated CTV1 coverage, although V 95 stayed above 95% in all cases (Fig. 1). When isocenter shifts and anatomical variations were combined, CTV1 V 95 dropped to unacceptable levels for a subset of patients. For weight gain combined with 5 mm setup errors, CTV1 V 95 dropped below 90% for 3/5 patients. For filling of the nasal cavity combined with 5 mm setup errors, CTV1 V 95 dropped below 95% for 2/5 patients. Weight loss did not affect CTV1 coverage even when combined with setup errors. The mean dose to the ipsilateral parotid gland was affected by all anatomic changes and setup errors, and the largest increases were seen for weight loss and filling of the cavities (Fig. 2). The maximum dose to the brainstem and the spinal cord had a similar behaviour for all the variations, but only for 1/5 patient the dose was above the clinical constraint.

EP-2000 Dosimetric impact of setup errors and anatomical changes in breast cancer patients E. Costa 1 , T. Bely 1 , M. Laurans 1 , S. Caneva Losa 1 , P. Poortmans 1 , Y.M. Kirova 1 1 Institut Curie Ensemble Hospitalier, Department of Radiation Oncology, Paris cedex 05, France Purpose or Objective To evaluate the dosimetric impact of uncorrected setup errors and anatomical changes in stage I-III breast cancer (BC) patients during loco-regional radiation therapy (RT) with Helical Tomotherapy after adjuvant chemotherapy. Material and Methods We studies the daily MVCT images acquired in patients treated to the breast or chest wall, internal mammary nodes (IMN), and axillary lymph nodes (levels II, III and IV). Setup corrections along the X, Y and Z directions and roll rotation were carried out after rigid registration of the bony anatomy and soft tissues between MVCT and planning CT. Pitch and yaw rotations were recorded only. Target volumes initially delineated on the planning CT were duplicated on every registrated MVCT and adjusted by a physician in order to reflect the anatomical modifications of the day. Similarly, the external body contour was adapted to each MVCT and the volume of this structure was calculated. Using 1 MVCT/patient/week, we calculated the dose distribution on a total of 101 MVCT’s. First, we analyzed the impact of the interval between chemotherapy and RT on the amplitude of anatomical changes. Subsequently, we quantified the changes in target volume coverage and we studied the correlation with the anatomical changes and with pitch and yaw uncorrected setup errors. Results Dosimetric data of 19 patients with whole breast or chest wall RT were studied, of which 17 were treated to the regional lymph nodes as well. Spearman correlation analysis showed a significant correlation between the amplitude of the changes of the external body contour and the period between chemotherapy and RT (r s =0.60; p=0.01).

The analysis of the calculated dose on the MVCT’s showed that in 7% of studied fractions, the near maximum dose (D2%) increased by at least 2% in the breast, IMN and nodes level II and III. In 2% of studied fractions, the D2% increased by at least 5% in nodes level IV. In 5% of studied fractions, the D95% decreased by at least 5% in the breast, IMN and nodes level IV. Pearson correlation analysis showed a significant correlation between the amplitude of external body contour changes and the influence on D2% in breast and IMN, D95% in nodes level IV and mean dose in breast. Pearson correlation analysis showed a significant correlation between pitch residual setup errors and modification of D2% in IMN and nodes level IV and the D95% in nodes level IV. We didn’t note correlation between yaw residual setup errors and target coverage alteration (r<0.3; p>0.01).

Conclusion Target volume dose was not compromised by setup errors for robustly optimized proton plans in nasopharyngeal cancer patients. Weight gain and nasal cavity filling had detrimental effect on target doses, especially in combination with setup errors, reaching unacceptably low levels for several patients. Better robustness methods are needed to account for anatomical variations in order to minimize the need for plan adaption.