skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: SU-C-BRD-01: A Statistical Modeling Method for Quality Control of Intensity- Modulated Radiation Therapy Planning

Abstract

Purpose: To apply a statistical modeling approach, threshold modeling (TM), for quality control of intensity-modulated radiation therapy (IMRT) treatment plans. Methods: A quantitative measure, which was the weighted sum of violations of dose/dose-volume constraints, was first developed to represent the quality of each IMRT plan. Threshold modeling approach, which is is an extension of extreme value theory in statistics and is an effect way to model extreme values, was then applied to analyze the quality of the plans summarized by our quantitative measures. Our approach modeled the plans generated by planners as a series of independent and identically distributed random variables and described the behaviors of them if the plan quality was controlled below certain threshold. We tested our approach with five locally advanced head and neck cancer patients retrospectively. Two statistics were incorporated for numerical analysis: probability of quality improvement (PQI) of the plans and expected amount of improvement on the quantitative measure (EQI). Results: After clinical planners generated 15 plans for each patient, we applied our approach to obtain the PQI and EQI as if planners would generate additional 15 plans. For two of the patients, the PQI was significantly higher than the other three (0.17 and 0.18more » comparing to 0.08, 0.01 and 0.01). The actual percentage of the additional 15 plans that outperformed the best of initial 15 plans was 20% and 27% comparing to 11%, 0% and 0%. EQI for the two potential patients were 34.5 and 32.9 and the rest of three patients were 9.9, 1.4 and 6.6. The actual improvements obtained were 28.3 and 20.5 comparing to 6.2, 0 and 0. Conclusion: TM is capable of reliably identifying the potential quality improvement of IMRT plans. It provides clinicians an effective tool to assess the trade-off between extra planning effort and achievable plan quality. This work was supported in part by NIH/NCI grant CA130814.« less

Authors:
; ;  [1]; ;  [2]
  1. University of Wisconsin-Madison, Madison, WI (United States)
  2. University of Maryland School of Medicine, Baltimore, MD (United States)
Publication Date:
OSTI Identifier:
22412439
Resource Type:
Journal Article
Journal Name:
Medical Physics
Additional Journal Information:
Journal Volume: 41; Journal Issue: 6; Other Information: (c) 2014 American Association of Physicists in Medicine; Country of input: International Atomic Energy Agency (IAEA); Journal ID: ISSN 0094-2405
Country of Publication:
United States
Language:
English
Subject:
60 APPLIED LIFE SCIENCES; HEAD; NECK; NEOPLASMS; NUMERICAL ANALYSIS; PATIENTS; PLANNING; QUALITY CONTROL; RADIOTHERAPY

Citation Formats

Gao, S, Meyer, R, Shi, L, D'Souza, W, and Zhang, H. SU-C-BRD-01: A Statistical Modeling Method for Quality Control of Intensity- Modulated Radiation Therapy Planning. United States: N. p., 2014. Web. doi:10.1118/1.4889714.
Gao, S, Meyer, R, Shi, L, D'Souza, W, & Zhang, H. SU-C-BRD-01: A Statistical Modeling Method for Quality Control of Intensity- Modulated Radiation Therapy Planning. United States. https://doi.org/10.1118/1.4889714
Gao, S, Meyer, R, Shi, L, D'Souza, W, and Zhang, H. 2014. "SU-C-BRD-01: A Statistical Modeling Method for Quality Control of Intensity- Modulated Radiation Therapy Planning". United States. https://doi.org/10.1118/1.4889714.
@article{osti_22412439,
title = {SU-C-BRD-01: A Statistical Modeling Method for Quality Control of Intensity- Modulated Radiation Therapy Planning},
author = {Gao, S and Meyer, R and Shi, L and D'Souza, W and Zhang, H},
abstractNote = {Purpose: To apply a statistical modeling approach, threshold modeling (TM), for quality control of intensity-modulated radiation therapy (IMRT) treatment plans. Methods: A quantitative measure, which was the weighted sum of violations of dose/dose-volume constraints, was first developed to represent the quality of each IMRT plan. Threshold modeling approach, which is is an extension of extreme value theory in statistics and is an effect way to model extreme values, was then applied to analyze the quality of the plans summarized by our quantitative measures. Our approach modeled the plans generated by planners as a series of independent and identically distributed random variables and described the behaviors of them if the plan quality was controlled below certain threshold. We tested our approach with five locally advanced head and neck cancer patients retrospectively. Two statistics were incorporated for numerical analysis: probability of quality improvement (PQI) of the plans and expected amount of improvement on the quantitative measure (EQI). Results: After clinical planners generated 15 plans for each patient, we applied our approach to obtain the PQI and EQI as if planners would generate additional 15 plans. For two of the patients, the PQI was significantly higher than the other three (0.17 and 0.18 comparing to 0.08, 0.01 and 0.01). The actual percentage of the additional 15 plans that outperformed the best of initial 15 plans was 20% and 27% comparing to 11%, 0% and 0%. EQI for the two potential patients were 34.5 and 32.9 and the rest of three patients were 9.9, 1.4 and 6.6. The actual improvements obtained were 28.3 and 20.5 comparing to 6.2, 0 and 0. Conclusion: TM is capable of reliably identifying the potential quality improvement of IMRT plans. It provides clinicians an effective tool to assess the trade-off between extra planning effort and achievable plan quality. This work was supported in part by NIH/NCI grant CA130814.},
doi = {10.1118/1.4889714},
url = {https://www.osti.gov/biblio/22412439}, journal = {Medical Physics},
issn = {0094-2405},
number = 6,
volume = 41,
place = {United States},
year = {Sun Jun 15 00:00:00 EDT 2014},
month = {Sun Jun 15 00:00:00 EDT 2014}
}