- Email: [email protected]
Comparison of Prostate Treatment Options using Data Envelopment Analysis (DEA)
Proceedings of the 50th Annual ASTRO Meeting carefully assessed. Eligibility criteria can result in data that may not be applicable to a substantial percentage of patients seen in an academic clinical practice. Author Disclosure: S. Apisarnthanarax, None; R.J. Kimple, None; S.L. Harris, None; D.E. Morris, None; J.E. Tepper, None.
2698
Comparison of Prostate Treatment Options using Data Envelopment Analysis (DEA)
R. Ramer1, A. Holder2, N. Papanikolaou1 1
University of Texas Health Science Center at San Antonio, San Antonio, TX, 2Trinity University, San Antonio, TX
Purpose/Objective(s): To utilize Data Envelopment Analysis (DEA) as a novel method of comparison between treatment options for prostate cancer patients. Materials/Methods: There are many options available for the treatment of prostate cancer. For example, a prostate patient could be treated using low dose rate (LDR) brachytherapy permanent implants, external beam radiotherapy (EBRT), or surgery, among others. Data Envelopment Analysis (DEA) is a linear programming technique designed to compare the relative performance of peer entities.1 It has the potential to be an invaluable tool in radiation therapy in the comparison of various treatment options. DEA handles multiple objectives, including those with different units. It provides a single score for ranking in terms of efficiency, by calculating a weighted sum of the inputs divided by a weighted sum of the outputs. These ranks are normalized so that an efficiency score of 1.0 is the most efficient. These inputs and outputs can be anything; in terms of radiotherapy, one option is for the inputs to be patient cases and the outputs to be the objectives considered clinically, such as the five year biochemical failure free survival rate and the impotence rate. Data regarding outcomes was taken from the literature,2, 3 and analyzed via linear programming techniques using in-house MATLAB 7.5.0. For homogeneity, only favorable (Gleason score \ 6, a PSA of \ 10 mg/L, and stage T1c or T2) patients were considered. Results: Interstitial brachytherapy with permanently implanted I-125 seeds had an efficiency score of E = 1.000, indicating it was most efficient with respect to the objectives considered. This was followed by surgery and EBRT, with efficiencies of 0.975 and 0.914, respectively. Conclusions: Based on the data included in the analysis, LDR brachytherapy permanent implant I-125 seeds gave the best outcome. Both prostatectomy surgery and external beam radiotherapy showed inefficient scores of less than one. To improve the ability of the model to distinguish between DMUs, additional outputs will be added. Furthermore, other prostate treatment options will be added as DMUs. References: 1. Cooper, W.W., Seiford, L.M. and Zhu, J., (2004) Handbook on Data Envelopment Analysis. Kluwer Academic Publishers. Dordrecht. 2. Peschel, R. E. and J. W. Colberg, (2003) ‘‘Surgery, brachytherapy, and external-beam radiotherapy for early prostate cancer,’’ Lancet Oncol 4: pp. 233–41. 3. Roach, M III (2004) ‘‘Reducing the toxicity associated with the use of radiotherapy in men with localized prostate cancer,’’ Urol Clin N Am 31: pp.353–366. Author Disclosure: R. Ramer, None; A. Holder, None; N. Papanikolaou, None.
2699
Conditional Survival in Anal Cancer: Results from the SEER Dataset 1988-2001
1
M. Choi , D. Klish2, C. D. Fuller1, C. R. Thomas3, S. J. Wang3 1 University of Texas Health Science Center at San Antonio, San Antonio, TX, 2University of Kansas Medical Center, Kansas City, KS, 3Oregon Health & Science University, Portland, OR
Purpose/Objective(s): Survival statistics for patients with anal cancer are typically reported in terms of survival from time of initial diagnosis. However, for patients who have survived a given period of time after diagnosis, conditional survival (CS) is a more clinically relevant measure, as it accounts for changes in risk over time. The specific aim of this study was to estimate CS for anal cancer patients through analysis of large-scale cancer registry data. Materials/Methods: Anal cancer cases diagnosed between 1988-2001 were extracted from the Surveillance, Epidemiology, and End Results (SEER 17) database. Five-year relative CS calculations were performed using SEER*Stat 6.3.6, with stratification by SEER historic stage, age, gender, and ethnicity, for patients who had already survived up to 5 years from diagnosis. Results: A total of 5,124 patients with anal cancer were identified in the SEER dataset as meeting inclusion criteria. The 5-year relative CS improved over time for up to 5 years after diagnosis for all anal cancer patients. The largest gains in CS over time were seen for patients with distant disease, age less than 60 years, male gender, and American Indian/Alaskan Native/Asian/Pacific Islander ethnicity. Overall, while patients with more advanced stage disease had lower CS at diagnosis, these patients also saw the greatest increase in CS over time, with CS for patients with distant disease more than tripling over the first 5 years of surveillance (19% to 59%). At the time of diagnosis, patients over 60 years of age had poorer CS (62%) than those under 60 years of age (68%), and CS for the two age cohorts improved comparably for up to 5 years after diagnosis (81% vs. 89%). At diagnosis, CS for male patients was worse than for female patients (59% vs. 69%) but showed greater improvement over time, with CS at 5 years following initial diagnosis equalizing at 85% for both genders. When stratified by ethnicity, White patients initially fared better than Black or other ethnicity patients (66% vs. 56% vs. 57%). By 5 years out from diagnosis, CS for White and Black patients were equal at 84%, while patients of other ethnicity showed an even better CS (94%). Conclusions: Prognosis improves over time for almost all groups of anal cancer patients surviving one or more years after diagnosis. For anal cancer survivors, CS provides a more relevant measure of prognosis than conventional survival estimates, and may provide useful baseline survival estimates for patients with anal cancer who have survived for a given time after diagnosis. Author Disclosure: M. Choi, None; D. Klish, None; C.D. Fuller, None; C.R. Thomas, None; S.J. Wang, None.
S481
2698
Comparison of Prostate Treatment Options using Data Envelopment Analysis (DEA)
R. Ramer1, A. Holder2, N. Papanikolaou1 1
University of Texas Health Science Center at San Antonio, San Antonio, TX, 2Trinity University, San Antonio, TX
Purpose/Objective(s): To utilize Data Envelopment Analysis (DEA) as a novel method of comparison between treatment options for prostate cancer patients. Materials/Methods: There are many options available for the treatment of prostate cancer. For example, a prostate patient could be treated using low dose rate (LDR) brachytherapy permanent implants, external beam radiotherapy (EBRT), or surgery, among others. Data Envelopment Analysis (DEA) is a linear programming technique designed to compare the relative performance of peer entities.1 It has the potential to be an invaluable tool in radiation therapy in the comparison of various treatment options. DEA handles multiple objectives, including those with different units. It provides a single score for ranking in terms of efficiency, by calculating a weighted sum of the inputs divided by a weighted sum of the outputs. These ranks are normalized so that an efficiency score of 1.0 is the most efficient. These inputs and outputs can be anything; in terms of radiotherapy, one option is for the inputs to be patient cases and the outputs to be the objectives considered clinically, such as the five year biochemical failure free survival rate and the impotence rate. Data regarding outcomes was taken from the literature,2, 3 and analyzed via linear programming techniques using in-house MATLAB 7.5.0. For homogeneity, only favorable (Gleason score \ 6, a PSA of \ 10 mg/L, and stage T1c or T2) patients were considered. Results: Interstitial brachytherapy with permanently implanted I-125 seeds had an efficiency score of E = 1.000, indicating it was most efficient with respect to the objectives considered. This was followed by surgery and EBRT, with efficiencies of 0.975 and 0.914, respectively. Conclusions: Based on the data included in the analysis, LDR brachytherapy permanent implant I-125 seeds gave the best outcome. Both prostatectomy surgery and external beam radiotherapy showed inefficient scores of less than one. To improve the ability of the model to distinguish between DMUs, additional outputs will be added. Furthermore, other prostate treatment options will be added as DMUs. References: 1. Cooper, W.W., Seiford, L.M. and Zhu, J., (2004) Handbook on Data Envelopment Analysis. Kluwer Academic Publishers. Dordrecht. 2. Peschel, R. E. and J. W. Colberg, (2003) ‘‘Surgery, brachytherapy, and external-beam radiotherapy for early prostate cancer,’’ Lancet Oncol 4: pp. 233–41. 3. Roach, M III (2004) ‘‘Reducing the toxicity associated with the use of radiotherapy in men with localized prostate cancer,’’ Urol Clin N Am 31: pp.353–366. Author Disclosure: R. Ramer, None; A. Holder, None; N. Papanikolaou, None.
2699
Conditional Survival in Anal Cancer: Results from the SEER Dataset 1988-2001
1
M. Choi , D. Klish2, C. D. Fuller1, C. R. Thomas3, S. J. Wang3 1 University of Texas Health Science Center at San Antonio, San Antonio, TX, 2University of Kansas Medical Center, Kansas City, KS, 3Oregon Health & Science University, Portland, OR
Purpose/Objective(s): Survival statistics for patients with anal cancer are typically reported in terms of survival from time of initial diagnosis. However, for patients who have survived a given period of time after diagnosis, conditional survival (CS) is a more clinically relevant measure, as it accounts for changes in risk over time. The specific aim of this study was to estimate CS for anal cancer patients through analysis of large-scale cancer registry data. Materials/Methods: Anal cancer cases diagnosed between 1988-2001 were extracted from the Surveillance, Epidemiology, and End Results (SEER 17) database. Five-year relative CS calculations were performed using SEER*Stat 6.3.6, with stratification by SEER historic stage, age, gender, and ethnicity, for patients who had already survived up to 5 years from diagnosis. Results: A total of 5,124 patients with anal cancer were identified in the SEER dataset as meeting inclusion criteria. The 5-year relative CS improved over time for up to 5 years after diagnosis for all anal cancer patients. The largest gains in CS over time were seen for patients with distant disease, age less than 60 years, male gender, and American Indian/Alaskan Native/Asian/Pacific Islander ethnicity. Overall, while patients with more advanced stage disease had lower CS at diagnosis, these patients also saw the greatest increase in CS over time, with CS for patients with distant disease more than tripling over the first 5 years of surveillance (19% to 59%). At the time of diagnosis, patients over 60 years of age had poorer CS (62%) than those under 60 years of age (68%), and CS for the two age cohorts improved comparably for up to 5 years after diagnosis (81% vs. 89%). At diagnosis, CS for male patients was worse than for female patients (59% vs. 69%) but showed greater improvement over time, with CS at 5 years following initial diagnosis equalizing at 85% for both genders. When stratified by ethnicity, White patients initially fared better than Black or other ethnicity patients (66% vs. 56% vs. 57%). By 5 years out from diagnosis, CS for White and Black patients were equal at 84%, while patients of other ethnicity showed an even better CS (94%). Conclusions: Prognosis improves over time for almost all groups of anal cancer patients surviving one or more years after diagnosis. For anal cancer survivors, CS provides a more relevant measure of prognosis than conventional survival estimates, and may provide useful baseline survival estimates for patients with anal cancer who have survived for a given time after diagnosis. Author Disclosure: M. Choi, None; D. Klish, None; C.D. Fuller, None; C.R. Thomas, None; S.J. Wang, None.
S481