Clinical implications in the use of the PBC algorithm versus the AAA by comparison of different NTCP models/parameters.

PURPOSE: Retrospective analysis of 3D clinical treatment plans to investigate qualitative, possible, clinical consequences of the use of PBC versus AAA. METHODS: The 3D dose distributions of 80 treatment plans at four different tumour sites, produced using PBC algorithm, were recalculated using AAA and the same number of monitor units provided by PBC and clinically delivered to each patient; the consequences of the difference on the dose-effect relations for normal tissue injury were studied by comparing different NTCP model/parameters extracted from a review of published studies. In this study the AAA dose calculation is considered as benchmark data. The paired Student t-test was used for statistical comparison of all results obtained from the use of the two algorithms. RESULTS: In the prostate plans, the AAA predicted lower NTCP value (NTCPAAA) for the risk of late rectal bleeding for each of the seven combinations of NTCP parameters, the maximum mean decrease was 2.2%. In the head-and-neck treatments, each combination of parameters used for the risk of xerostemia from irradiation of the parotid glands involved lower NTCPAAA, that varied from 12.8% (sd=3.0%) to 57.5% (sd=4.0%), while when the PBC algorithm was used the NTCPPBC's ranging was from 15.2% (sd=2.7%) to 63.8% (sd=3.8%), according the combination of parameters used; the differences were statistically significant. Also NTCPAAA regarding the risk of radiation pneumonitis in the lung treatments was found to be lower than NTCPPBC for each of the eight sets of NTCP parameters; the maximum mean decrease was 4.5%. A mean increase of 4.3% was found when the NTCPAAA was calculated by the parameters evaluated from dose distribution calculated by a convolution-superposition (CS) algorithm. A markedly different pattern was observed for the risk relating to the development of pneumonitis following breast treatments: the AAA predicted higher NTCP value. The mean NTCPAAA varied from 0.2% (sd = 0.1%) to 2.1% (sd = 0.3%), while the mean NTCPPBC varied from 0.1% (sd = 0.0%) to 1.8% (sd = 0.2%) depending on the chosen parameters set. CONCLUSIONS: When the original PBC treatment plans were recalculated using AAA with the same number of monitor units provided by PBC, the NTCPAAA was lower than the NTCPPBC, except for the breast treatments. The NTCP is strongly affected by the wide-ranging values of radiobiological parameters.

Comparison between the ideal reference dose level and the actual reference dose level from clinical 3D radiotherapy treatment plans.

PURPOSE: Retrospective study of 3D clinical treatment plans based on radiobiological considerations in the choice of the reference dose level from tumor dose-volume histograms. METHODS AND MATERIALS: When a radiation oncologist evaluates the 3D dose distribution calculated by a treatment planning system, a decision must be made on the percentage dose level at which the prescribed dose should be delivered. Much effort is dedicated to deliver a dose as uniform as possible to the tumor volume. However due to the presence of critical organs, the result may be a rather inhomogeneous dose distribution throughout the tumor volume. In this study we use a formulation of tumor control probability (TCP) based on the linear quadratic model and on a parameter, the F factor. The F factor allows one to write TCP, from the heterogeneous dose distribution (TCP{(epsilon(j),D(j))}), as a function of TCP under condition of homogeneous irradiation of tumor volume (V) with dose D (TCP(V,D)). We used the expression of the F factor to calculate the "ideal" percentage dose level (iDL(r)) to be used as reference level for the prescribed dose D delivery, so as to render TCP{(epsilon(j),D(j))} equal to TCP(V,D). The 3D dose distributions of 53 clinical treatment plans were re-evaluated to derive the iDL(r) and to compare it with the one (D(tp)L) to which the dose was actually administered. RESULTS: For the majority of prostate treatments, we observed a low overdosing following the choice of a D(tp)L lower than the iDL(r.) While for the breast and head-and-neck treatments, the method showed that in many cases we underdosed choosing a D(tp)L greater than the iDL(r). The maximum difference between the iDL(r) and the D(tp)L was -3.24% for one of the head-and-neck treatments. CONCLUSIONS: Using the TCP model, the probability of tumor control is compromised following an incorrect choice of D(tp)L; so we conclude that the application of the F factor is an effective tool and clinical aid to derive the optimal reference dose level from the dose-volume histogram (DVH) of each treatment plan.