Session Item

The dose per fraction is typically constant in radiotherapy. The purpose of this study was to explore the potential of isotoxic temporal modulation of dose per fraction in NSCLC radiotherapy.

We implemented the TCP model of Jeong et al. [1] that dynamically simulates the effect of proliferating tumor cells and radiation on tumors every 15 minutes. It includes compartments for proliferating cells as well as intermediate and highly hypoxic cells. The model accurately predicts local control for a wide range of fractionation schemes. Subsequently we first calculated a constant dose per fraction (d_f^C) yielding a range of TCPs [0.5,0.6,0.7,0.8,0.9] for various number of fractions (F) ([12;17;20;24;30];5 fractions/week). For each fractionation scheme, we then isotoxically optimized TCP by modulating the dose per fraction. Fraction dose was modulated using a 1^st and 2^nd order polynomials while keeping the EQD2 (α/β=3Gy) constant for isotoxicity as follows:

d_f=a*(f/F)²+b*(f/F)+c

∑^F_f=1 d_f*(OAR_RE* d_f +3)/(2+3)=F*d_f^C*(OAR_RE* d_f^C +3)/(2+3),

where OAR_RE is the ratio of the fraction dose that is received by the OAR assuming serial organs for simplicity.  OAR_RE is assumed to be independent of d_f. Improvements in TCP were quantified as a function of TCP, F and OAR_RE.

Figure 1 illustrates the effect of a linear modulation on a treatment of 20 fractions TCPs. It can be observed that negative slopes (d starting high, ending low) reduce TCP, while positive slopes tend to increase TCP. Moreover, the effect of modulation reduces with increasing TCP.

Figure 2a depicts the optimal TCP for quadratic modulation. Improvements in TCP maximize around 60% TCP and then decrease more or less linearly. Similarly, the benefit of quadratic modulation versus linear was also bigger (3-5 pp) at lower TCP than at higher TCP (0-2 pp). Very little dependence on the number of fractions was observed. Figure 2b illustrates the dose modulation, again showing little dependence on the number of fractions.  

As expected, bigger improvements could be obtained in case of lower OAR_RE corresponding to dose limiting toxicity related to OARs further away from the tumor. Changes in TCP reduced by about 30% for OAR_RE=1 and increased by about 60% for OAR_RE=0.25 compared to OAR_RE=0.5.

Considerable improvements in TCP were observed using isotoxic temporal dose per fraction modulation with lower dose per fraction at the start and increased toward the end over a wide range of fractionation regimens. These results suggest that the therapeutic window at the beginning of treatment is limited due to radioresistant hypoxic cells and widens towards the end following re-oxygenation. Tumor regression during treatment may be synergistic with such temporal modulation and should be investigated further.  Future work should also include parallel OAR exposure and spatial heterogeneity of metabolism and hypoxia.