S608 I. J. Radiation Oncology d Biology d Physics Volume 75, Number 3, Supplement, 2009

Materials/Methods: A Virtual Water phantom (30x30x5cm3) and its two inserts (8cm diameter) were custom made with groovesin the inserts to accommodate dosimeters in parallel with and orthogonal to the transponders, respectively. In the transponder lo-calization accuracy test, 30 measurements each were performed for control (no dosimeter) and for dosimeter located at 0, 1, and2cm from transponders for parallel as well as orthogonal configurations. A multivariate analysis (Wilk’s Lambda) of the data wasperformed to test the mean localized positions of the transponders among the control and test datasets. In the dosimeter dose read-ings test, a relative dose acquisition and analysis method was adopted to reduce uncertainties. In one day, a total of 7 fractions of1Gy/fraction were sequentially delivered to the dosimeters inside the phantom in a fashion that only fraction 2, 4 and 6 had tran-sponders present. Dose readings from fractions 1, 3, 5 and 7 were fit using a 2nd order polynomial. The dose differences betweenfit-interpolated and corresponding measured values of fraction 2, 4, 6 were calculated. A separate dose accuracy test was performedwith 2Gy/fraction delivered in 5 consecutive days to 2 dosimeters in a Virtual Water phantom with no transponder present. Theaccelerator output variations were removed from dosimeter reading by daily TG51 output monitoring.

Results: For the localization test, the multivariate analysis indicated that the presence of the dosimeter did affect the transponderlocalization accuracy. However, 95% confidence intervals of localization coordinate differences between control and test datasetswere well below 100mm, which is insignificant compared to Calypso clinical display resolution, 0.5mm. For the dosimeter readingtest, the dose differences between polynomial fitted and measured dose values at fractions 2, 4, and 6 is within 1.4%. In the absolutedose test, the temperature corrected 5 day average dose readings were 99.5% and 102.8% of 2Gy for the 2 dosimeters, respectively.

Conclusions: In this study, the results of transponder localization tests proved that, clinically, the transponder localization wasnegligibly affected by the presence of wireless MOSFET dosimeters; the results of dosimeter reading tests demonstrated that tran-sponders will not alter dose readings. In summary, there is no interference between the transponder and dosimeter systems whenboth are placed in the same volume. Using both devices in the same patient for the same treatment course is feasible.

Author Disclosure: Z. Su, Provide phantom and machining service, C. Other Research Support; L. Zhang, None; V. Ramakrishnan,None; M. Hagan, None; M. Anscher, None.

2952 Geometric and Dosimetric Effects of Megavoltage Conebeam-guided Treatment Shifts in Head and Neck

Cancer

A. M. Morris, J. E. Bayouth, J. M. Buatti

The University of Iowa, Iowa City, IA

Purpose/Objective(s): Great attention has been paid to interfraction variation and set-up error in head and neck cancer. This sup-ports the potential role of daily IGRT. However, whether daily translational corrections have a beneficial effect upon the dosimetryof target and organs at risk remains unanswered. In this study, we investigate both the geometric and dosimetric impact of dailymegavoltage conebeam-guided treatment shifts on head and neck cancer patients treated definitively with IMRT.

Materials/Methods: Treatment plans from ten patients with head and neck cancer were retrospectively reviewed. A planning CTwas acquired for each of the ten patients, and 33-35 daily megavoltage conebeam CT scans were also acquired over the course oftreatment. Conebeam CT images were aligned to the planning CT using a translation algorithm designed to simulate the patient’sposition before and after the treatment shift. CTV1 and spinal cord contours were copied from the planning CT to each CBCTdataset and deformed using model-based segmentation. Daily dose distributions were then recomputed on the daily CBCTs usingthe original IMRT optimization. CTV1 V95, CTV1 D95, and spinal cord D0.5cc were noted. The location of vertebral bodies C1-T1 was also recorded before and after the treatment shift.

Results: At a population level, daily MVCB-guided treatment shifts had no significant effect upon the dosimetry of the patientsstudied (V95: p = 0.198, D95: p = 0.291, D0.5cc: 0.314). At an individual level, this generally held true as well; only three of the tenpatients demonstrated a significant shift-induced change (p\0.001) in at least one dosimetric parameter: CTV V95, CTV D95 orspinal cord D0.5cc. On a daily basis, however, 15.3%, 4.6% and 63.3% of the treatment shifts resulted in a significant change inCTV V95, CTV D95, and spinal cord D0.5cc, respectively. Although significant, these changes were still modest and within treat-ment guidelines. There was also a trend towards increasing consistency in dose delivery after the treatment shift. From a geometricperspective, all ten patients underwent a median translational shift of 0.34 cm. Seven demonstrated rotation in at least one axis(mean: X 1.59� [range: 0.45-2.65�] Y 0.96� [range: 0.07-3.74�]) before the treatment shift, and all ten patients qualitatively dem-onstrated non-rigid body deformation and residual set-up error after the shift (mean X: 0.13 cm, mean Y: 0.09 cm, Z: 0.30 cm).

Conclusions: Although insignificant at a population level, MVCB-guided treatment shifts are capable of eliciting significant dailychanges in dose to the target and OAR. Additionally, the presence of rotation, non-rigid body deformation, and residual setup errorindicates that further change might be masked by treatment margins and/or imprecise plan conformity.

Author Disclosure: A.M. Morris, None; J.E. Bayouth, Siemens Medical Solutions, F. Consultant/Advisory Board; J.M. Buatti,None.

2953 Fine-needle Marker for IGRT, a New Fiducial Gold Anchor for High-precision Radiotherapy

I. Naslund, P. Wersall, E. Castellanos, C. Beskow, S. Nyren

Karolinska University Hospital, Stockholm, Sweden

Purpose/Objective(s): Implantation and tracking of gold markers is a direct method of tumor visualization during radiotherapy(4D-RT). Fifty years experience of fine-needle aspirations for cytology specimens using spinal-type narrow needles almost any-where in the body with minimal risk of bleeding or infections were the basis for our development of this new gold marker instru-ment.

Materials/Methods: This fine-needle cannula GA120 mm in length (Gauge 25, 0.53 mm in diameter) and GA200 mm in length(Gauge 22, 0.70 mm in diameter) contains a 20 mm long gold-wire (0,28 mm in diameter) that can be implanted at almost any site inthe body. Small scoring cuts or notches in the gold wire every 2.0 mm force the wire to fold to a dense marker as it enters the tumortissue where it automatically anchors itself, making further delocalization impossible. Patients with various malignancies referred