**7. Modulation indices**

Modern treatment techniques, such as IMRT and VMAT, have enabled the escalation of target dose with fewer side effects to the surrounding OARs by modulation of the treatment plans to achieve the desired dose distribution. In IMRT, MLCs are moving during treatment, thereby delivering a radiation field with a non-uniform intensity while in VMAT technique, in addition to MLC motions, gantry speed and dose rate are also variable when the radiation beam is continuously on. For patients' protection and safety, pretreatment dosimetric verification is done to provide

*Parameters Affecting Pre-Treatment Dosimetry Verification DOI: http://dx.doi.org/10.5772/intechopen.102517*

#### **Figure 1.**

*Comparisons of standard density vs. high-density modes between different planned grid sizes and energy. (reprinted from Salari, et al., "evaluation of parameters affecting gamma passing rate in patient-specific QA's for multiple brain lesions IMRS treatments using Ray-Station treatment planning system. In print: J Appl Clin med Phys. 2021).*

sufficient data on the safety and reliability of treatment plans and delivery, even though performing pretreatment dosimetric verification is considered an additional workload. Therefore, a retrospective analysis of which parameter (leaf travel, beam aperture and shapes, control point angular separation, dose rate, and gantry variations) can affect the ability of the TPS to calculate a dose may provide important information on the limits of TPSs for IMRT/VMAT plans. The difference between calculated and measured dose distribution may be affected by the accuracy of the TPS calculation and the delivery accuracy. Discriminating between the two causes of errors is not an easy task. Furthermore, the delivery accuracy of IMRT/VMAT plans can be predicted by the score of plan modulation complexity [65]. For this purpose, many authors introduced or evaluated different Modulation Indices (MI)/parameters to find a correlation between plan complexity and GPR.

Nicolini et al. [66] studied the effect of gantry speed (deg/s) and dose rate (MU/min) on the quality of VMAT plans and showed using a higher dose rate improves plan quality and reduces delivery time. They also used dynamic log files generated by linac controllers to evaluate the delivery accuracy of plans and found out accuracy slightly improved in delivery when using a low dose rate. Wu et al. [67] analyzed the results of dose verification of 924 patients including the relationship between gamma pass rates and the location of lesions, the total number of monitor units, and the maximum area of the collective dose. They observed a correlation between the treatment site and GPR plus a strong negative correlation between total MUs and GPR that indicates increasing MU results in lower GPR. Moreover, a weak negative correlation between the largest area of the acquisition dose and GPR was reported [67]. McNiven et al. [68] proposed *Modulation Complexity Score (MCS)* for step-and-shoot IMRT. This score is contribution of variability in the shape of segments and variations in their area. The range of MCS is from 0 to 1. The lower value of the MCS means higher complexity. This metric provides more information about the plan quality than simple metrics such as total MUs and number of segments, but no correlation was observed between GPR and MCS which is in a good agreement with other research [69, 70]. This index was later adapted by Masi et al. [65] for VMAT plans by substituting control points for segments and called it (MCSv). Also, Masi et al. introduced *Leaf Travel (LT)* as the average distance that MLC is traveling over one arc in VMAT and LTMCS index which takes into account both LT and MCSv and has a range between 0 and 1. Zero shows a higher degree of modulation and leaf motion. They reported a moderate correlation between LT, MCSv, LTMCS, and GPR

and a weak correlation between MU and GPR. Hernandez et al. [71] modified LT for multiple arcs or partial arc by dividing LT over arc length (LT/AL). Another index is *Edge Metric (EM)* which was defined by Young et al. [72] and it calculates the complexity as the ratio of MLC side length edge to aperture area. The larger EM index indicates the difference between the positions of adjacent leaves are larger which is closely related to the tongue-and-groove effect. Du et al. in 2014 [73] introduced several MIs to evaluate plan complexity such as *plan averaged beam area (PA), plan averaged beam irregularity (PI), plan averaged beam modulation (PM), and plan normalized MU (PMU).* PA is the average area of beam apertures; PI indicates the noncircularity of the shape of aperture and PM describes to what extent a beam is modulated with multiple smaller apertures. PMU is to compare the total MU among all plans with different prescription dose levels. According to a number of studies [70, 71, 73] MCS, EM, and PI provide similar information. In 2014, Park et al. [74] defined MIs, MIa, and *MItotal* which MItotal unlike previous metrics include both gantry speed and dose rate variations besides MLC motions to quantify the total delivery complexity for VMAT plans. MIs which evaluate MLC speed was originally introduced by Webb [75] to evaluate the modulation degree of IMRT and were modified by Park et al. for VMAT treatment plans and MIa evaluates both speed and acceleration of MLCs. They also studied the MCSv and LTMCS and did not see correlations as high as those found in a previous study (Masi et al) to the pre-treatment VMAT QA results.

In summary, various studies were conducted in this area and revealed different results regarding the correlation between plan complexity indices and QA metrics [65–79]. We believe, these differences may depend on the linac model and its commissioning plus TPS limitations such as beam model, dose engine, and algorithm [71, 80, 81].

### **8. Conclusions**

As described in this chapter, there are a number of sources which may contribute and arise different levels of discrepancy between the computed dose by TPS and measurements. Much effort has been devoted to improve the accuracy of dose calculation algorithms, computing technology and measurements, and through all these developments the accuracy of dose calculation and measurements seems close to our clinical goals. Although, the accuracy of dose calculation in homogenous medium (e.g., water) does not much rely on the algorithm, in heterogeneous media such as lungs or bone, the accuracy of calculation depends strongly on the kernels of calculation algorithms and how well they can simulate the actual scattering of photon and electrons. As mentioned previously in this chapter and noted by authors in various literatures, the accuracy of dose calculation algorithms is rated as principle-based algorithms such as Monte Carlo, and the linear Boltzmann transport as the most accurate, followed by model-based algorithms such as CCC, AAA, and PBC in that order for accuracy; and correction-based algorithms. Another important item to be considered is the beam modeling which will directly affect the accuracy of dose calculation where each TPS has its own features to model beams. Therefore, following the beam data measurements, commissioning of the modeled beams becomes a necessary step typically achieved through end-to-end testing. This is to verify dose distribution and accurate computation under different clinical conditions before any clinical use. Moreover, it is important to understand the response and limitations of each equipment used along with gamma index analysis due to different combinations

*Parameters Affecting Pre-Treatment Dosimetry Verification DOI: http://dx.doi.org/10.5772/intechopen.102517*

of QA devices and software packages, which may result in varying levels of agreement with the predicted gamma analysis for the same pass-rate criteria. Various reasons result in different correlations between GPR and complexity metrics, hence, these correlations are not generic and should be defined for each TPS.
