The reliability of optimization under dose-volume limits

Langer, Mark; Brown, Richard; Kijewski, Peter K.; Ha, Chul S.

doi:10.1016/0360-3016(93)90972-x

Cited by 19 publications

(6 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the theoretical situation, the number of constraint points was limited to 100 in the target and each OAR, and the time limit was set to 15 days to allow for the MILP technique to find the global optimum or a solution quite close to the global optimum. In the clinical situation, the number of constraint points was set to 500 in the target and each OAR, so that dose-volume evaluations had an accuracy of 3% to 4% (Langer et al 1993). The time limit was set to half an hour, which was thought to be within the clinically acceptable range.…”

Section: Data Collectionmentioning

confidence: 99%

“…The third technique, also proposed by Langer et al (1990), is to model the treatment planning optimization accurately in a mixed integer linear programming problem (MILP) that can be solved by solving a sequence of linear programming relaxation problems with the branch-and-bound technique. The number of integer variables (or, more strictly, binary variables) is the number of constraint points in OARs, which must number at least in the hundreds in each OAR to achieve an accuracy of 3% to 4% for dose-volume evaluations (Langer et al 1993). Therefore, there will be hundreds to thousands of these constraint points.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Conversion of dose–volume constraints to dose limits

Dai

Zhu

2003

Phys. Med. Biol.

View full text Add to dashboard Cite

The purpose of this study is to introduce two techniques for converting dose-volume constraints to dose limits for treatment planning optimization, and to evaluate their performance. The first technique, called dose-sorting, is based on the assumption that higher dose limits should be assigned to the constraint points receiving higher doses, and vice versa. The second technique, the hybrid technique, is a hybrid of the dose-sorting technique and the mixed integer linear programming (MILP) technique. Among all constraint points in an organ at risk, the dose limits for the points far from a dose-volume constraint are determined by dose-sorting, while the dose limits for the points close to a dose-volume constraint are determined by MILP. We evaluated the performance of the two new techniques for one treatment geometry by comparing them with the MILP technique. The dose-sorting technique had a high probability of finding the global optimum when no more than three organs at risk have dose-volume constraints. It was much faster than the MILP technique. The hybrid technique always found the global optimum when the MILP percentage (the percentage of constraint points for which the dose limits are determined by the MILP technique) was large enough, but its computation time increased dramatically with the MILP percentage. In conclusion, the dose-sorting technique and the hybrid technique with a low MILP percentage are clinically feasible.

show abstract

Section: Data Collectionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Conversion of dose–volume constraints to dose limits

Dai

Zhu

2003

Phys. Med. Biol.

View full text Add to dashboard Cite

show abstract

“…An important issue in inverse planning is how to formalize the clinical goals to objectively evaluate the figures of merit of different IMRT plans. Despite intense research effort in modelling the clinical decision-making strategies (Amols and Ling 2002, Deasy et al 2002, Earl et al 2003, Hou et al 2003, Lahanas et al 2003, Langer et al 1993, Lee et al 2000, Llacer et al 2001, Mohan et al 1994, Webb 2004, Xing et al 1999, Yan et al 2003, the appropriate form of the objective function remains illusive. Presently, two types of objective functions are widely used: dose or dose-volume histogram (DVH)-based (physical objective functions) (Chen et al 2002, Cho et al 1998, Holmes et al 1995, Hristov et al 2002, Michalski et al 2004, Starkschall et al 2001, Shepard et al 2002, Xing et al 1998 and dose-response-based objective functions (biological objective functions) (Brahme 2001, Kallman et al 1992, Miften et al 2004, Mohan et al 1992, Wang et al 1995, Webb and Nahum 1993.…”

Section: Introductionmentioning

confidence: 99%

Clinical knowledge-based inverse treatment planning

Yang

2004

Phys. Med. Biol.

View full text Add to dashboard Cite

Clinical IMRT treatment plans are currently made using dose-based optimization algorithms, which do not consider the nonlinear dose-volume effects for tumours and normal structures. The choice of structure specific importance factors represents an additional degree of freedom of the system and makes rigorous optimization intractable. The purpose of this work is to circumvent the two problems by developing a biologically more sensible yet clinically practical inverse planning framework. To implement this, the dose-volume status of a structure was characterized by using the effective volume in the voxel domain. A new objective function was constructed with the incorporation of the volumetric information of the system so that the figure of merit of a given IMRT plan depends not only on the dose deviation from the desired distribution but also the dose-volume status of the involved organs. The conventional importance factor of an organ was written into a product of two components: (i) a generic importance that parametrizes the relative importance of the organs in the ideal situation when the goals for all the organs are met; (ii) a dose-dependent factor that quantifies our level of clinical/dosimetric satisfaction for a given plan. The generic importance can be determined a priori, and in most circumstances, does not need adjustment, whereas the second one, which is responsible for the intractable behaviour of the trade-off seen in conventional inverse planning, was determined automatically. An inverse planning module based on the proposed formalism was implemented and applied to a prostate case and a head-neck case. A comparison with the conventional inverse planning technique indicated that, for the same target dose coverage, the critical structure sparing was substantially improved for both cases. The incorporation of clinical knowledge allows us to obtain better IMRT plans and makes it possible to auto-select the importance factors, greatly facilitating the inverse planning process. The new formalism proposed also reveals the relationship between different inverse planning schemes and gives important insight into the problem of therapeutic plan optimization. In particular, we show that the EUD-based optimization is a special case of the general inverse planning formalism described in this paper.

show abstract

“…[7][8][9] Among the treatment planning optimization techniques based on physical objective functions, the model of partial dose-volume constraints may be the most clinically relevant one. [7][8][9][10][11][12][13] Within this model, various objectives are pursued under constraints demanding that fractional volumes of critical structures do not receive more than certain dose limits, e.g., no more than 30% of the lung volume may receive more than 20 Gy. A maximum dose constraint to a critical organ can be expressed by a partial dose volume constraint through the requirement that no more than 0% of the volume can get more than the specified maximum dose.…”

Section: Introductionmentioning

confidence: 99%

“…Several minimization techniques have been proposed as optimization engines for problems involving dose-volume objective functions and constraints. [8][9][10][11][12][13][14][15][16][17][18][19] While stochastic algorithms [15][16][17][18]20 have the theoretical advantage of being able to escape from local minima present in optimization problems with dose-volume objective functions, 21 gradient algorithms are a few orders of magnitude faster. Furthermore, as argued by Bortfeld, 7 gradient algorithms are generally adequate for inverse planning because ͑i͒ different local minima can be promptly explored from different starting points and ͑ii͒ the value of the objective function at the local minima is often not too different from its value at the global minimum.…”

Section: Introductionmentioning

confidence: 99%

On the implementation of dose‐volume objectives in gradient algorithms for inverse treatment planning

et al. 2002

View full text Add to dashboard Cite

A method that allows a straightforward implementation of dose-volume constraints in gradient algorithms for inverse treatment planning is presented. The method is consistent with the penalty function approach, which requires the formulation of an objective function with penalty terms proportional to the magnitudes of constraint violations. Dose constraints with respect to minimum and maximum target dose levels are incorporated in quadratic, dose-penalty terms. Analogously, quadratic volume-penalty terms in the objective function reflect the violation of dose-volume constraints imposing limits on the fractions of healthy organ volumes that can be irradiated above specified dose levels. It has been demonstrated that within the framework of this formulation neither modified objective functions nor finite difference gradient calculations are necessary for the incorporation of gradient minimization algorithms. As an example, a simple steepest descent algorithm is presented along with its application to illustrate prostate and lung cases.

show abstract

The reliability of optimization under dose-volume limits

Cited by 19 publications

References 19 publications

Conversion of dose–volume constraints to dose limits

Conversion of dose–volume constraints to dose limits

Clinical knowledge-based inverse treatment planning

On the implementation of dose‐volume objectives in gradient algorithms for inverse treatment planning

Contact Info

Product

Resources

About