Portal dosimetry in radiotherapy repeatability evaluation.

The accuracy of radiotherapy is the subject of continuous discussion, and dosimetry methods, particularly in dynamic techniques, are being developed. At the same time, many oncology centers develop quality procedures, including pretreatment and online dose verification and proper patient tracking methods. This work aims to present the possibility of using portal dosimetry in the assessment of radiotherapy repeatability. The analysis was conducted on 74 cases treated with dynamic techniques. Transit dosimetry was made for each collision-free radiation beam. It allowed the comparison of summary fluence maps, obtained for fractions with the corresponding summary maps from all other treatment fractions. For evaluation of the compatibility in the fluence map pairs (6798), the gamma coefficient was calculated. The results were considered in four groups, depending on the used radiotherapy technique: stereotactic fractionated radiotherapy, breath-hold, free-breathing, and conventionally fractionated other cases. The chi2 or Fisher's exact test was made depending on the size of the analyzed set and also Mann-Whitney U-test was used to compare treatment repeatability of different techniques. The aim was to test whether the null hypothesis of error-free therapy was met. The patient is treated repeatedly if the P-value in all the fluence maps sets is higher than the level of 0.01. The best compatibility between treatment fractions was obtained for the stereotactic technique. The technique with breath-holding gave the lowest percentage of compliance of the analyzed fluence pairs. The results indicate that the repeatability of the treatment is associated with the radiotherapy technique. Treated volume location is also an essential factor found in the evaluation of treatment accuracy. The EPID device is a useful tool in assessing the repeatability of radiotherapy. The proposed method of fluence maps comparison also allows us to assess in which therapeutic session the patient was treated differently from the other fractions.


| INTRODUCTION
Radiotherapy, as a treatment using ionizing radiation, is an effective method of oncology. However, several criteria must be met. One of them is the precision of the radiation dose delivery.
One of the principles in medicine is "first, do no harm." This dictum applied in radiotherapy means to protect the critical organs as much as possible. Dynamic intensity-modulated radiotherapy (IMRT) or volumetric modulated arc therapy (VMAT) techniques allow to  which is lower than the planned one. This situation is not clinically relevant. Unfortunately, at the same time the dose in the PTV volume is lower than planned, which can result in lower local control probability. Figure 1(c) presents irradiation of critical structures with a higher dose. This can lead to complications after the treatment.
In radiotherapy, orthogonal x-ray imaging or cone-beam tomography (CBCT) is performed before the therapeutic session to minimize the differences between the planned and actual patient position.
The acquired two-or three-dimensional images are compared with reference images sent from the treatment planning system. Nevertheless, one should remember that on C-arm linear accelerators, the image verifications are mostly performed before switching to radiation exposure. If there are differences between the planned and the actual patient position, the action is required. Patient position is adjusted to the planned one. However, during the therapeutic session, the patient may intentionally or passively change the position.
In such case, the actual dose distribution in the patient body is unknown. Hence, it seems so important to control these changes occurring during the therapeutic session.
Modern C-arm accelerators are equipped with real-time dosemonitoring systems, that is, portal matrices (EPID -Electronic Portal Imaging Device). Therefore, in addition to the old-fashioned use F I G . 1. Example of planned and changed dose distribution in lymph node radiotherapy for the VMAT technique with 6 MV FFF beam. Visible contours of PTV lymph node with margin (red) and critical structures: bladder (yellow), rectum (brown), femoral heads (blue). Dose distribution ≥90% of the planned total dose; (a) in the planned patient treatment position. (b) after moving the patient 3mm forward in the anteroposterior direction relative to the planned position (c) after moving the patient 3mm backward relative to the planned position.
| 157 of matrices to control the patient treatment position and regular use for pretreatment dosimetry, EPID matrices allow to register a signal that can later be reconstructed, as a dose distribution during the therapeutic session with the patient. [1][2][3][4][5][6] The signal measured by the EPID matrix is called the fluence map.
For physics and computer science, the VMAT technique with an acting EPID is a type of megavolt cone-beam computed tomography.
From that, as in CT, the needful information is given to reconstruct the patient anatomy and dose distribution. There are many published articles on how to calculate the dose distribution based on the realtime EPID image. 7-10 However, the software described in the literature is mostly not commercial. If it is available, its use is time-consuming and difficult in clinical practice. Nowadays, the compatibility of dose distribution between the calculations from treatment planning system (TPS) and measured dose distribution in dynamic techniques is performed without a patient.
Since useful tools to compare the measured fluence map with that calculated during treatment planning are available, one can focus solely on treatment repeatability. Fluence maps measured during the therapeutic session can be compared with each other. It can be assumed that the patient is repeatably treated if these do not differ from each other. A patient who is treated repeatably is highly likely to be treated as planned. Measurement of a dose or fluence map, taking into account the patient's body, is called transit dosimetry. 11 The aforementioned common use of portal matrices allows to The parameter adopted for comparing two data sets, including dose distributions or fluence maps, is the gamma coefficient. 15 Each measurement is characterized by uncertainty influenced by several factors, such as the precision of the measurement system settings.
The gamma coefficient defined in the form of values given in brackets (ΔD = x %, DTA = y mm, 98%) means that the acceptable agreement between two data sets is x %, their offset (DTA) is y mm in 98% of analyzed field points. If these conditions are met, then the gamma value is less than or equal to 1. This method of assessing the compliance of two quantities can be used to compare calculations with measurements (Fig. 2) as a mandatory QA procedure in radiotherapy to compare two calculated data (e.g., a good option for testing calculation algorithms) and to compare two measurement data ( Fig. 3). Figure 2 presents the algorithm of dose distribution verification as the calculated and measured fluence maps comparison. It is the classic QA process. Figure 3 shows the possibility of comparing two fluence maps measured during the therapeutic session. The experience described in the paper of Klimas et al. 16 shows that the repeatability of fluence maps measurements is 2%, 2 mm in 98% of analyzed field points. Therefore, EPID can be used to verify differences between fluence maps. The given compliance criterion, developed for the phantom conditions, was used in this study as the starting point for fluence maps comparison in clinical cases.
The study aims to present the possibilities of EPID dosimetry use in assessing the repeatability of radiotherapy.

1.A | MATERIALS AND METHODS
During radiotherapeutic sessions, transit dosimetry was made for each collision-free radiation beam. It means that the EPID matrix In Table 1, a C value equal to the number of all fluence maps comparisons and a D value equal to zero represents the null hypothesis.
If we consider ten fractions of radiation therapy and EPID mea-  . Fluence maps are measured using the EPID matrix (c). In the next step, the calculated fluence map (b) is compared with the measured one (c). If gamma is smaller than 1 in 98% analyzed points, comparisons: calculated vs. measured fluence mapspass.
This way of comparing fluence maps also allows us to assess in which therapeutic session the patient was treated differently from the other fractions.

| RESULTS
The performed analysis indicates that the highest percentage of repeatable treatment is obtained in stereotactic radiotherapy and the lowest in the BH technique (Table 2). It means that statistical significance (P-value higher than 0.01) estimated in chi 2 or Fisher's exact test shows the compliance with measured fluence maps for given gamma criteria of (2%, 2 mm, 98%). Table 2 indicate that all FSRT patients were treated repeatedly. However, in two cases of abdominal nodes, despite statistically repeatable treatment, in three of six fractions, the difference in the dose was higher than 2%. The place of its deposition differed from the planned one by more than 2 mm. Except these two cases, 98% of the analyzed fluence maps pairs meet the gamma criterion.

P-values presented in
The results of the analysis performed for the DC group in the mediastinum, abdominal cavity, and pelvis with conventional fractionation demonstrate that for an average of 75% of the fluence maps pairs, the gamma coefficient is less than or equal to one. The statistical analysis shows that the patients in this group are not irradiated in a repeatable way. Since it is a rectal treatment, significant changes in anatomic shape and volume of patient's organs are likely to affect the result achieved. Additionally, there is one patient in this group, who is irradiated in two fractions that differ considerably from other fractions. The analysis presented in Fig. 4 shows that on November 13 and December 11, the compatibility of the compared fluence maps with other maps was much worse.
In the FB techniques for 58% of pairs, the gamma criteria are met ( Table 2). A statistical test indicates that no patient was irradiated repeatedly. For all cases, the p-value is less than 0.01.
The analysis of the fluence maps pairs in BH cases showed that the average compatibility of the analyzed fluence pairs is 20% (Table 2). From a statistical point of view, no patient in BH group was irradiated repeatedly. For all cases, the P-value is less than 0.01.
In the chest wall area, the 2 mm shifts are unavoidable. The radiation dose is related to the energy deposited in the matter. Therefore, the lung volume changes due to respiratory movements affect the dose distribution and the measured fluence map.  techniques. Figure 5 shows the results of comparing three pretreatment fluence map measurements.
Comparison was performed on first vs. second fraction, first vs.
third fraction, and second vs. third fraction. The gamma coefficient for 0.5% and 0.5mm is less than or equal to 1 in 100% of the analyzed field. This means that the repeatability of measurements (EPID) is much lower than 2% and 2mm. It can be assumed that the differences between measurements higher than these values are related to the mobility of the patient.
Previous experiments indicate that such a high degree of maps similarity for (0.5%, 0.5 mm, 99%) gamma criteria are not always obtained and are challenging to achieve. 16 Therefore, the values of (2%, 2 mm, 98%) were adopted. They are sufficient to assure the accuracy of measurement system. Higher differences can be caused by the presence of patient between the radiation source and mea- There is no doubt that they should be associated with the location of the tumor. For brain and craniofacial tumors, a tolerance of 2% and 2 mm is optimal. The analysis shows that all patients in this group were treated reproducibly. In the mediastinum and pelvis tumor location (DC group), the analysis shows that no patient was irradiated repeatably for the (2%, 2 mm, 98%) conditions. Shifts in anatomical organs position in the abdominal cavity are undoubtedly higher than 2 mm, so the values of 3% and 3 mm were adopted, and the analysis was performed again. The reanalysis in DC group shows that approximately 88% of the analyzed pairs of fluence maps meet the gamma condition of (3%, 3 mm, 98%). The value of P = 0.0147 means that there are no differences between the theoretical and analyzed groups. The patients were treated in the repeatable way.
In the FB and BH groups, very little compatibility between pairs of fluence maps was obtained. Therefore the analysis was redone for 3% and 5 mm conditions, which allowed to achieve 98% of the analyzed surface. In irradiation of chest wall after mastectomy at FB, assuming new gamma criteria, 94% of patients were treated repeatedly (P = 0.018), except two patients, whose results differed considerably from the rest. Therefore, it can be assumed that patients treated with this technique are irradiated repeatedly within the dose range of 3% and 5mm shift.
In the group of breast cancer patients irradiated with VMAT on BH technique, the reanalysis was performed for (4%, 5 mm, 98%) condition. Only 75% of the analyzed pairs of fluence maps meet this criterion, and the P-value lower than 0.01 indicates that the patients were treated in a repeated manner. For this technique further analysis should be performed in order to find out whether the value of dose offset should be enlarged to 6-7 mm or the surface percentage should be reduced to, for example, 95%.
Resource data confirm that the use of EPID in radiotherapy verification reduces errors, which undoubtedly improves treatment results. [20][21][22][23] Transit dosimetry with EPID matrices can be used based on two methodscomparing calculated fluence maps with measured ones or by comparing all measured fluence maps with one another. We chose the latter, and we believe it the right one to assess the repeatability. 3. The calculations show that various radiation techniques and various tumor locations require different criteria, in order to optimize the treatment repeatability assessment.

All additional actions aimed at improving the radiotherapy results
and/or patient's safety is extremely important. Fluence maps assessment is an example of such action, especially given the fact that it does not extend the treatment duration, and does not expose patient to additional dose.

5.
Further research is required in order to define the optimum repeatability criteria for various techniques and tumor locations.

Krzysztof
Ślosarek conceived and designed the analysis, contributed data or analysis tools, and wrote the paper.
Dominika Plaza, Aleksandra Nas, and Marta Reudelsdorf collected the data.
Jacek Wendykier and Barbara Bekman performed the analysis and other contribution: language corrections.
Aleksandra Grządziel performed the analysis, wrote the paper, and other contribution: language corrections.