2893 The Value of Credentialing

J. Leif, J. Roll, D. Followill, G. Ibbott

M.D. Anderson Cancer Center, Houston, TX

Purpose/Objective(s): The Radiological Physics Center (RPC) has performed credentialing of institutions to participate in anumber of protocol studies. The purpose of credentialing is to verify that the radiation oncologist and other personnel involvedare familiar with the protocol prior to enrolling patients with the goal of limiting the number of deviations. Credentialing hasbeen performed for high and low dose rate brachytherapy studies, stereotactic studies and IMRT studies. Credentialing has alsobeen done for several different study groups and for a variety of disease sites including eye, brain, breast, and cervix.

Materials/Methods: Over the years, the RPC’s credentialing procedures have required one or more of the following: reviewof the institution’s physics and QA procedures, submission of knowledge assessment and facility questionnaires, calculation ofgeometric benchmark cases, importing and planning on a predetermined CT image set, irradiating an anthropomorphicphantom, and submitting the first 2 patients for review. The credentialing process not only evaluated the quality of the specificclinical procedure at the institution, but also assured that the institution and participating radiation oncologist had experiencein the procedure. The credentialing process enabled us to offer feedback to the institution to correct mistakes that may occurwith a protocol patient. A retrospective review of radiotherapy of patients entered on the studies was performed. A recalculationof patient dose and a review of the records and all planning and verification films were performed. Deviations from protocolguidelines were assessed according to predefined criteria.

Results: Three protocols for which credentialing was required of all participants had rates of deviation on the order of 0% to4% and two protocols that had limited credentialing requirements had rates of deviation on the order of 7% to 17%. In anotherstudy some institutions were credentialed for one technique but not for another. Those institutions that were credentialedreceived no deviations on the protocol, whether they utilized the technique for which they were credentialed or not, while thoseinstitutions not credentialed had a deviation rate of 16%.

Conclusions: It appears that institutions that go through a credentialing process were better prepared to comply with therequirements of the protocol. This may be because the credentialing process provides feedback on how to better comply withthe treatment protocol prior to submitting a patient onto the study. Therefore, the frequency of deviations can be reduced forinstitutions that go through the credentialing process. QA is an important component of clinical trials. A credentialing procedurecan be a valuable component of clinical trial QA.

This work was supported by PHS grants CA 10953 and CA081647 awarded by NCI, DHHS.

Author Disclosure: J. Leif, None; J. Roll, None; D. Followill, None; G. Ibbott, None.

2894 Lessons From the In-Vivo TLD Dosimetry Practice for Patients Treated With Total Skin ElectronRadiation Therapy (TSEB) at M.D. Anderson Cancer Center

R. C. Tailor, N. B. Tolani, G. S. Ibbott, C. S. Ha

UT MD Anderson Cancer Center, Houston, TX

Purpose/Objective(s): At institutions offering TSEB, it is often customary to determine the actual radiation dose at multiplesites of patients with thermoluminescent dosimetry (TLD). These measurements facilitate adjustments to the patient positionand orientation to achieve dose uniformity and help determine the need for boost treatment. This practice is time consumingand sometimes uncomfortable for the patient, and requires considerable resources. This report describes statistics of TLD resultsfor a large population of patients treated on a single linac, and an evaluation of the rationale for the number of in-vivo TLDmeasurements.

Materials/Methods: Results of TLD measurements at 22 body sites of each of 68 patients treated with TSEB were evaluated.Patients were treated with a modified Stanford 6-field technique using 9 MeV electron beam. For each measured body site, themean dose and associated standard deviation were determined.

Results: The table below presents statistical results of in-vivo TLD on 68 patients at MDACC. The results are consistent withcommissioning data and are compared to those from Chen et. al. (IJROBP 59, P 872–885, 2004). Despite variations in patientsize and uncertainties in positioning, TLD results at main body sites (1–6) show good reproducibility: one standard deviation 7%. Several other sites show considerable variation from one patient to the next: inner thighs, buttocks, axillae, vertex of scalp,and feet.

Conclusions: In-vivo dosimetry for treatments such as TSEB continues to play an important role by providing data for makingclinical decisions on the adequacy of dose delivery. Sites 1–4 are valuable for verifying the reference dose. Upper thorax sites(5, 6) show consistency among patients and could be dropped. Since the TLD results at sites 7–12 show large variations amongpatients and within the same patient (data not shown), decisions to boost or shield are primarily based on clinical judgment. Thefeet almost always receive a higher dose than prescribed, and therefore clinical experience dictates that shielding will berequired. Consequently, the use of TLD at sites 7–12 and 21–22 may be discontinued. Site such as the back of hand may beuseful because differences in left and right may indicate incorrect patient orientation. A careful reduction in the number ofin-vivo TLD measurement sites could not only reduce the time and effort required, but more importantly lessen discomfort tothese patients.

S716 I. J. Radiation Oncology ● Biology ● Physics Volume 66, Number 3, Supplement, 2006

Author Disclosure: R.C. Tailor, None; N.B. Tolani, None; G.S. Ibbott, None; C.S. Ha, None.

2895 Clinical Impact of F-18 Fluorodeoxyglucose (FDG) Positron Emission Tomography (PET) on theManagement of Primary Tumors of the Thymus

J. W. Lee1, M. P. MacManus2, D. L. Ball2, R. J. Hicks3, A. Hogg2

1Department of Radiation Oncology, University of Toronto, Toronto, ON, Canada, 2Department of Radiation Oncology,Peter MacCallum Cancer Institute, Melbourne, Australia, 3Department of Molecular Imaging, Peter MacCallum CancerInstitute, Melbourne, Australia

Purpose/Objective(s): To evaluate the impact of PET on the management of primary tumors of the thymus.

Materials/Methods: Patients with a primary tumor of the thymus who underwent PET were identified from a prospectivedatabase. Prior to PET, referring physicians recorded the scan indication, disease extent by conventional imaging and pathology,and the proposed treatment plan. The impact of PET on clinical management was assessed. Follow-up data were used tomeasure the accuracy of conventional imaging and PET findings.

Results: From 1997 to 2005, a total of 43 PET scans were performed on 26 patients with primary thymic tumors (15malignant/invasive thymoma, 10 thymic carcinoma, 1 benign thymoma). The median age was 53 and median follow-up fromfirst PET scan was 2.7 years. Indications for PET were surveillance(n�20), suspected recurrence(n�12), therapeutic moni-toring(n�6) and staging(n�5). PET accuracy could not be determined in 4 cases. For all assessable scans, the negativepredictive value of PET was 64% and the positive predictive value was 100%.

Conventional imaging and PET findings were discordant in 10 cases(23%), of which PET findings were either truepositive(n�3), true negative(n�2), false negative(n�3) or could not be validated or refuted by clinical outcomes(n�2). Whenconventional and PET findings differed, there was a 30% incidence of false negative PET results.

PET appropriately changed patient management in 4 cases(9.3%) based on accurate results that differed from pre-PETimaging. The treatment modality changed from surgery to observation(n�1) and from chemotherapy to observation(n�1). Theradiotherapy dose(n�1) and target volume(n�1) were also influenced by PET results. Treatment was inappropriately changedby PET in 1 case(2.3%); chemotherapy was withheld based on a false-negative PET result. In 5 cases, discordant PET resultswere ignored and did not change management due to patient co-morbidity(n�2), pre-planned surgery(n�2), or lack of treatmentoptions(n�1).

Conclusions: The majority of PET scans performed(88%) did not influence clinical management. Non-visualization of residualdisease sites presumably reflects low FDG avidity in some forms of thymic tumor, resulting in false negative PET findings.Therefore, in the setting of a residual mass with a negative PET, a policy of surgery/biopsy or ongoing surveillance with CTis warranted. However, the positive predictive value of PET is high and enables appropriate modification of management in asmall but potentially important subset of patients.

Author Disclosure: J.W. Lee, None; M.P. MacManus, None; D.L. Ball, None; R.J. Hicks, None; A. Hogg, None.

2896 Design of Non-Uniform Dose Distribution Based on Spatially Inhomogeneous Radiosensitivity ExtractedFrom Biological Images

G. Chen, E. Ahunbay, C. Schultz, X. A. Li

Department of Radiation Oncology, Medical College of Wisconsin, Milwaukee, WI

Purpose/Objective(s): Inhomogeneous radiobiology exists within tumor volume of malignant gliomas as shown by physio-logical MRI. An inverse planning technique which generates non-uniform dose is developed to account for the inhomogeneousradiosensitivity.

Materials/Methods: Relative cerebral blood volume (rCBV) maps obtained with a dynamic susceptibility contrast-enhancedMRI technique was used to delineate spatial radiosensitivity distributions. With reported data on correlation between rCBV,tumor grade, and radiosensitivity, the voxel rCBV was converted to voxel radiosensitivity parameters (e.g., � and �/�). With

Site # Site

% Av. Dose % Std. Dev

MDACC Refa MDACC Refa

1 Umbilicus (AP) 100 106 9 32 Abdomen: umb. level (PA) 93 104 8 63,4 Abdomen: umb. level (Lat.) 96 100 8 105,6 Upper thorax (AP, PA) 95 — 7 —7,8 Axillae (Rt, Lft) 61 90 27 289,10 Inner Thighs (Rt, Lft) 55 68 26 3411,12 Buttocks (Rt, Lft) 60 — 25 —13 Vertex of scalp 90 75 25 1114 Back of head 103 — 16 —15,16 Top of shoulders (Rt, Lft) 70 89 13 1117,18 Back of hands (Rt, Lft) 87 87 7 1119,20 Elbows (Rt, Lft) 90 — 14 —21,22 Exposed feet (Rt, Lft) 115 85 13 13

S717Proceedings of the 48th Annual ASTRO Meeting