Materials/Methods: ITC as part of the Advanced Technology QA Consortium (ATC) developed a modular system for digitaldata submission, queriable archival storage, and web-based remote QA review (“ATC Method 1”), which has been used insupport of RTOG AT protocols. This technology has also played a key role in assisting treatment planning system (TPS)manufacturers in verifying that their RTOG Data Exchange and DICOM implementations (CT, RT Structure Set, RT Dose, RTPlan, and RT Image) match ATC’s digital data exchange conformance statement. ITC and RTOG have developed credentialingcriteria, e.g., on-line Facility Questionnaire and “Dry-Run” test designed to demonstrate participating institution’s ability tosubmit a protocol compliant digital data set prior to placing patients on study. Data are sent to ITC via FTP or media. QA reviewincludes (1) data integrity review by ITC for completeness of protocol required elements, format of data, and possible datacorruption; (2) recalculation of Dose Volume Histograms (DVHs) by ITC; (3) review of target volume and organ at riskcontours compliance by study chair using web-based Remote Review Tool (RRT); and (4) review of dose prescription and doseheterogeneity compliance by RTOG HQ Dosimetry Group using RRT.

Results: To date, 15 TPS vendors have released ATC-compliant RTOG/DICOM export software. ITC has successfullysupported 15 RTOG AT protocols (Phase I-III trials). Over 350 institutions have been credentialed to submit digital data andover 3400 digital data sets have been submitted to ITC. Overall, approximately 1/4 of cases submitted on these trials requiredintervention by ITC to correct data integrity/completeness problems before data could be evaluated by dosimetrists/study chairs.Explicit problems in digital data submission discovered by ITC have been categorized and will be reviewed. Dry Run testexperience varies, e.g., for an IMRT protocol only 1/3 of the credentialed institutions passed on first submission. ITC found thatsubmitted DVHs lack consistency due to algorithmic differences among TPSs. For dose distributions with high gradients (e.g.,brachy, IMRT), discrepancies in excess of 15% were observed between submitted and ITC-recalculated DVHs for volumes �50 cc.

Conclusions: Experience in managing data for AT clinical trials has demonstrated the need for an active data integrity QAprocess to assure completeness and integrity of data submitted from participating institutions prior to review for protocolcompliance by QA reviewers. Re-calculation of DVHs by ITC is necessary for consistent correlation of dosimetry withoutcomes. ITC’s web-based RRT is both an effective tool for QA review of AT clinical trials data by study chairs and RTOGDosimetry Group and an aid to TPS vendors in developing/verifying implementation of digital data export. Future softwaredesign should emphasize use of modular architecture with well-defined interfaces to enable integration of commercial-off-the-shelf, open-source and custom software components.

Supported by NIH U24 grant CA81647 and U10 grant CA21661.

Author Disclosure: J.A. Purdy, None; W.R. Bosch, None; W.L. Straube, None; J.W. Matthews, None; R.J. Haynes, None; J.M.Michalski, None; E.A. Martin, None; K. Winter, None; W.J. Curran, None; J.D. Cox, None.

1012 Defining the Need for Breast Cancer Radiotherapy in the General Population: A Criterion-BasedBenchmarking Approach

M. Kerba, Q. Miao, W. Mackillop

Queen’s Cancer Research Institute, Kingston, ON, Canada

Purpose/Objective(s): Determining the appropriate rate of radiotherapy (RT) for breast cancer (BCa) is important for ensuringoptimal RT utilization, improving accessibility and forecasting needs. A criterion-based benchmark (CBB) was developed asa possible standard for estimating the need for initial RT among incident BCa cases. Our primary objective was to compare anevidence-based estimate (Ebest) of the need for RT in BCa against the CBB. Secondary objectives included comparing observedRT rates in Ontario, Canada and the United States to the Ebest and CBB rates. Lumpectomy (Lx) and mastectomy (Mx) rateswere also examined.

Materials/Methods: RT benchmarks were defined in Ontario as communities in proximity to cancer centres that were withoutlong waiting lists and that had site-specialized radiation oncologists practicing in a multidisciplinary environment. 1997-2001patient data was prospectively collected from provincial RT cancer centers. The Canadian Institute for Health Information(CIHI) database provided surgical information. The public use file of the Surveillance, Epidemiology and End Results Registries(SEER) described RT and surgical rates in the US.

Results: Between 1997-2001 there were 33,869 cases of BCa in Ontario and 84,872 cases in SEER. The median patient agein all populations was 61.0 years. The overall Ebest RT rate for BCa was 64.0% (95% CI: 58.1%,69.8%) compared to the CBBestimate of 60.7% (59.3%,62.1%). Adjuvant RT rates stratified by lumpectomy were 100% in Ebest and 83.6% (82.0%,85.1%)by the CBB. Post-mastectomy RT rates were 21.9% (20.6%,23.3%) by Ebest and 34.6% (32.5%,36.7%) for the CBB.Post-lumpectomy RT rates were 75.1% (74.4%,75.7%) in Ontario and 65.3% (64.9%,65.8%) in SEER. Post-mastectomy RTrates were 29.5%(28.7%,30.3%) in Ontario and 17.0%(16.6%,17.4%) in SEER. Observed Lx and Mx rates were 54.4%(53.8%,54.9%) and 36.0% (35.5%,36.6%) in Ontario and 51.8% (51.4%,52.1%) and 44.2% (43.4%,44.6%) in SEER.

Conclusions: The CBB approach may serve as a standard to estimate the need for RT in BCa. The evidence-based estimatewas concordant with the overall need for RT provided by the CBB. Ebest underestimated the need for post-mastectomy RT asdetermined by the CBB. The estimates of the need for post-lumpectomy RT were greater than the observed rates in Ontario andSEER. Both the CBB and the Ebest suggest a shortfall of RT utilization in the population.

Author Disclosure: M. Kerba, None; Q. Miao, None; W. Mackillop, None.

1013 Older Adolescents and Young Adults With Cancer: Conditional Survival Deficit

S. J. Wang1, C. D. Fuller2, J. Y. Luh2, C. R. Thomas1, W. A. Bleyer3

1Oregon Health & Science University, Portland, OR, 2University of Texas Health Science Center, San Antonio, TX, 3St.Charles Medical Center, Bend, OR

Purpose/Objective(s): Conditional survival expresses change in prognosis for survivors as a function of their time sincediagnosis. When applied to cancer, this matrix estimates the risk of dying after an interval of survival and allows survivors andtheir healthcare providers to know the risks at intervals after diagnosis and to base follow-up accordingly.

S135Proceedings of the 48th Annual ASTRO Meeting

Materials/Methods: The National Cancer Institute Surveillance, Epidemiology and End Results (SEER) database (April 2005release) was used to determine the conditional survival of 15- to 29-year olds diagnosed with cancer during 1975-2000 and tocompare their results with younger and older patients. Patients were subsetted into 4 groups by age: � 15 years old, 15-29 years,30-44 years, and � 45 years of age. We computed the conditional survival for each age group, and measured its change aspatients survived a period of time since diagnosis.

Results: Whereas 15- to 29-year olds diagnosed with cancer during the past quarter century had a better prognosis at diagnosis,their conditional survival thereafter did not increase as rapidly as younger and older patients. Our analysis demonstrated that15- to 29-year olds have had a lower conditional survival improvement than any other age group, with an average annualimprovement of only 3.6% per year, compared to an average annual improvement of 4.4% to 6.2% for other age groups. Whenthe conditional survival at 1, 2, 3 and 5 years after diagnosis are compared (Figure), it is apparent that the deficit among 15-to 29-year olds is apparent at one year and continues to at least five years after diagnosis with no evidence for a narrowing ofthe gap.

Conclusions: These profiles indicate that, during the past 25 years, adolescents and young adults with cancer have not enjoyedthe improvement in prognosis with passage of time since diagnosis to the extent that younger and older patients have. The deficitin survival that has been previously demonstrated among older adolescents and young adults at the time of diagnosis continuesthrough five years after diagnosis. The reasons for this deficit are unknown, but one explanation may be related to the specificmix of cancer types in this age group. Further studies, possibly from pediatric cooperative group tumor-specific datasets, mayprovide more insight into the reasons for this relative lack of improvement in conditional survival for this group when comparedto other age groups. Failure to have a comparable improvement in conditional survival is another deficit faced by olderadolescent and young adult cancer patients.

Author Disclosure: S.J. Wang, None; C.D. Fuller, None; J.Y. Luh, None; C.R. Thomas, None; W.A. Bleyer, None.

1014 Methods to Address Cancer Disparities in Special Populations

D. G. Petereit1,2, M. P. Mehta2, J. S. Kaur3, F. Govern4, C. N. Coleman4, M. A. Ritter2, R. Koscik2, A. R. Moser2,L. Burhansstipanov5, D. Rogers1, et al.1John T. Vucurevich Cancer Center Institute, Rapid City, SD, 2University of Wisconsin School of Medicine and PublicHealth, Madison, WI, 3Mayo Clinic Comprehensive Cancer Center, Rochester, MN, 4Radiation Research Program,National Cancer Institute, Bethesda, MD, 5Native American Cancer Research, Corp, Pine, CO

Purpose/Objective(s): American Indians (AIs) present with more advanced stages of cancer and, therefore, suffer from highercancer mortality rates compared to non-AIs. Under the National Cancer Institute Disparities Research Program, we areresearching methods of improving cancer treatment and outcomes for AIs in Western, South Dakota.

Materials/Methods: This program consists of patient navigation, clinical trials focusing on reducing treatment length tofacilitate enhanced participation, surveys to evaluate barriers to access, and a molecular study to address whether there is aspecific profile that increases toxicity risks. The phase II clinical trials utilize tomotherapy and brachytherapy in order to shortenthe overall treatment duration as AIs live a median of 110 miles from the cancer center. Two surveys in progress investigatepotential barriers to accessing health care. Finally, the ATM gene is being sequenced do determine if there is a molecularexplanation for increased rates of toxicities in this population. The success of enrolling AIs on this program will be presented.

Results: As of April 2006, 123 AIs have undergone patient navigation with a median number of contacts/visits of 5 (range0-85). For AIs who were navigated during radiation the median number of contacts was 19 (range 0-59). For both the AI andnon-AI population, 25 patients have been enrolled in the tomotherapy prostate trial, 14 in the breast brachytherapy trial, and4 in the HDR prostate brachytherapy trial. To date, 10 AIs have participated in a clinical trial, 7 of which have a componentof radiation (12% of the population served by the cancer center are AIs). The general survey has recruited 595 AIs and 16 tothe cancer survey. In the first three months of the ATM trial we have recruited 29 patients, 6 of which are AI. Preliminary resultsbased on the first 289 surveys suggest that an educational program on cancer and the importance of early detection and treatmentcould be effective in enhancing the rate at which AIs seek cancer screening and therapy. A primary endpoint for this projectis entry of AIs on NCI sponsored clinical trials. For those AIs who underwent patient navigation during radiation, 21% weretreated on a clinical trial. Reasons for non-participation in clinical trials included ineligibility due to metastases and/or poor

S136 I. J. Radiation Oncology ● Biology ● Physics Volume 66, Number 3, Supplement, 2006