Modern radiotherapy techniques like stereotactic body radiotherapy (SBRT) and stereotactic radiosurgery (SRS) use high fractional doses of radiation. This permits the very precise and accurate delivery of the radiation dose. The approach increases the importance for dosimetric quality assurance (QA) before the treatment of each patient.^1–3) Various methods to verify dose distributions calculated by treatment planning systems (TPS) have been described.⁴⁾ Radiochromic film is commonly used to provide treatment dose verification and measure two-dimensional (2-D) dose distribution in external beam radiotherapy with high spatial resolutions, weak energy dependence, wide dose range, and tissue equivalence.⁵⁾

Gafchromic EBT3 film (International Specialty Products, ISP, Wayne, NJ, released in 2011) was introduced in late 2011 to eliminate measurement orientation effects as well as Newton rings formed during film scanning. It enhances the accuracy of maintaining a dosimetric performance, similar to the EBT2 predecessor.⁶⁾ The details of dosimetry using EBT3 film and a flatbed color scanner have been described.^7,8) EBT3 dosimetry is generally analyzed by the single color channel method for radiotherapy with a fractional dose less than 3 Gy. However, treatment dose verification using high doses (6~13 Gy) is less clear.^9–12) It has been known for a long time that the green channel (GC) method offers good usability in higher doses, although the red channel (RC) has high sensitivity in the conventional dose range.¹³⁾ Use of the RC for doses below 8 Gy and the GC for higher doses is recommended by the manufacture.¹⁴⁾

The GC has not been regarded as a productive approach to verify the treatment dose of SBRT and SRS, due to the low sensitivity of the channel at low dose regions, such as organ at risk (OAR) and other normal tissues. To verify target and OAR doses for high-dose treatment, a dual-channel (DC) compound method combining the RC and GC could have merit, and was the subject of the present study. We investigated the feasibility of the DC compound method in EBT3 film dosimetry for high doses by comparing the single channel method using the RC and GC.

1. Film calibration

Gafchromic EBT3 film was used. Prior to the film dosimetry, the daily output of linear accelerator (LINAC) was checked with a Famer-type ion chamber by applying the AAPM TG 51 protocol on the day of the calibration.¹⁵⁾ Film was arranged in a solid water phantom (30×30×11.5 cm³), 1.5 cm deeper than the phantom surface with a 10 cm solid water layer placed below the film to produce backscattered radiation. The source to skin distance was 100.0 cm. A net-optical density (netOD) curve was obtained by irradiating film with 6-MV photon beam of a TrueBeam LINAC (Varian Medical Systems, Palo Alto, CA) at the 10×10 cm² field and 0° gantry. Doses ranging from 0 to 15 Gy were used to convert the film OD to dose from the same film batch. Films were scanned by an Expression 11000 XL flatbed scanner (Epson America Inc., Long Beach, CA, USA) after 24 hours with an image resolution of a 72 dots per inch. The scanned images were acquired in transmission mode and landscape orientation. The scanner was always warmed up at least 30 min before use and five preliminary scans without film on the scanner bed were performed to eliminate the impacts of scanner noise. The netOD curves (netOD_RC and netOD_GC) for RC and GC were determined as previously described.¹⁶⁾ Each netOD curves were shown in Fig. 1.

2. Measurement and evaluation for film dosimetry

To measure various dose-ranges, the dose distribution was generated in an Eclipse treatment planning system (TPS, version 11.0.34; Varian Medical Systems) using a 6-MV photon beam of 10×20 cm² field with 60° dynamic-wedge. An analysis anisotropic algorithm (AAA, version 11.0.34) was used to calculate the dose distribution. The plan was normalized to deliver a dose of 575 cGy at a depth of 6 cm in the solid-water phantom (30×30×16 cm³). The calculated dose (D_Calc) with a 0.25×0.25 cm² dose resolution was exported in the digital imaging and communications in medicine (DICOM) format. The film was measured under the same phantom setup and beam configuration for planning. The measured dose distribution was acquired by using both netOD curves. To evaluate the difference between D_Calc and dose distributions (D_RC and D_GC) using the netOD_RC and the netOD_GC, the dose-profiles (cross-line and in-line) and gamma analysis were compared. As shown in Fig. 2, the cross-line profile was extracted at depth of the normalization point for central-axis of the film. The in-line profiles were obtained in five positions of −8, −4, 0, 4, and 9 cm (d1, d2, d3, d4, and d5, respectively) from the normalization point. The root-mean-square-error (RMSE) was used to analyze profiles of D_RC and D_GC compared to D_Calc. The gamma analysis described previously^17,18) was performed for dose distributions obtained by using both channels (RC and DC) and D_Calc.

3. Dual-channel compound method

The DC compound method was designed compounding the gamma value for D_RC and D_GC. It was based on the RC gamma analysis (GA_RC) after correcting the GC gamma analysis (GA_GC). The GA_RC was separated the failed-GA_RC (γ>1) and passed-GA_RC (γ<1). The failed-GARC was converted the GA_GC (convert-GA_GC) when the GA_GC passed at the same location. The DC gamma analysis (GA_DC) was defined ultimately as combination of the passed-GA_RC and the convert-GA_GC.

1. Verification of cross- and in-line profiles

For the positions mentioned in Fig. 2, the cross-line and in-line profiles were acquired from D_RC, D_GC, and D_Calc of the measured film. Fig. 3 shows the comparison of cross-line and in-line profiles obtained in D_RC, D_GC, and D_Calc. The difference of cross-line profile for the D_RC and the D_Calc was increased remarkably in the high doses more than 8 Gy (Fig. 3a). Although the D_GC was also slight difference of cross-line profile with and the DCalc as dose increase, there is no significant difference. The in-line profiles of D_RC, D_GC and D_Calc were similar in four positions, excepted in d₅ of high dose region (Fig. 3b).

Table 1 shows RMSE values in profiles of D_RC and D_GC compared to D_Calc. The RMSE of the cross-line profile for D_RC was roughly twice as high as D_G. For the in-line profiles at positions d₁ to d₅, RMSE values of D_RC smoothly increased (within 21.15) to the d₄ position, with a steep increase to d₅. For the in-line profiles for the D_GC, the RMSE values were constantly increased (from 2.88 to 38.14) from d₁ to d₅. For in-line profiles of D_RC and D_GC, the difference in RMSE values was 5.59, −2.07, −0.98, −14.58, and 77.80 for d₁, d₂, d₃, d₄, and d₅, respectively.

Fig. 4 shows the distributions of GA_RC, GA_GC, GA_DC and convert-GA_GC. All distributions showed that the passing area decreased with increasing doses from 8 Gy. In the distribution of GA_RC, the gamma passing was the fastest decline in higher doses. Gamma values of convert-GA_GC were mainly distributed in the high dose region (Fig. 4d).

2. Gamma analysis in dose distribution for the red, green channel, and dual-channel method

Fig. 5 shows the gamma passing rates which was 3%/3 mm criteria in dose distributions using the RC, GC, and DC compound method. Most of passing rates for GA_RC were higher than those of GA_GC as dose level was up to 10 Gy. For high doses, the passing rates of GA_RC were lower than those of GA_GC. The gamma passing rates of GA_DC were the highest value among the gamma passing rates at overall dose regions, although the gamma passing rate of GA_RC and GA_GC was decreased. In 2%/2 mm criteria, the gamma passing rates of GA_RC, GA_GC, and GA_DC are 55.30%, 53.18% and 70.64%, respectively.

In this study, the DC compound method for film dosimetry was evaluated in comparison to RC and GC. RC was more accurate within 8 Gy than the GC. RC was less accurate in high doses exceeding 8 Gy. The results echo those of prior studies.^13,14) The previous authors also mentioned that the GC for sensitometric curve exceeds the RC in high doses of more than 10 Gy. This was the basis of the DC compound method, which is based on the RC in low dose region after correcting for the GC in high dose region. Therefore, the gamma values in the convert-GA_GC were concentrated in the high dose region.

By using GA_DC, the gamma passing will increase in the low dose region because the failed-GA_RC was re-calculated using GA_GC. This means that the gamma passing of GA_DC in low doses needs to be double-checked using GA_GC. This double-check could decrease the film uncertainty using a single channel and increase the accuracy of the gamma analysis. Therefore, the use of the DC compound method may improve the accuracy for film dosimetry in the whole dose region.

As shown in Fig. 3, the difference between D_GC and the D_Calc was small in overall dose regions, while the difference between D_RC and the D_Calc was obviously in high doses. Borca et al.¹³⁾ reported that the film sensitivity for all color (RGB) channels decreased in high dose level. Our study found that the RC was good sensitivity within 8 Gy compared to RC. However, the RC exceeded sensitivity than GC when used in doses more than the 8 Gy. With depending on these channels trend, the differences between D_Calc and dose distributions (D_RC and D_GC) increased with dose increase, although the D_GC was more accurate than the D_RC in higher doses.

The DC compound method devised in this study will be more suitable in clinical radiation therapy, especially the dosimetric QA (DQA) for SBRT using high doses. In general, the result of DQA for SBRT is evaluated using the RC with the pretreatment QA plan downgraded for prescription dose. However, this downgraded evaluation has difficulty in reflect real clinical practice due to affect in variation of dose distribution and dose rate in pretreatment QA plan by dose degradation. Our DC compound method does not necessary need a dose downgrade for pretreatment QA plan. DQA is practical in the real clinical application. Therefore, we recommend the use of DC compound method in DQA using film dosimetry, especially in high dose treatments, such as SBRT and SRS.

The limitation of the DC compound method is that it was generated by gamma analysis. To perform a more accurate film dosimetry in whole dose region, it is necessary that each netOD curve is merged in a certain dose range. In future study, we will perform application of the merged netOD method for SBRT-DQA using film.

We evaluated the feasibility of DC compound method for the film dosimetry using EBT3. The method is a suitable film dosimetry for QA of treatment using high doses because it can be reduced errors in high dose verifications without the downgrade of prescription dose. Consequently, we recommend the use of DC compound method for film dosimetry of clinical SBRT-DQA.

This research was supported by a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health and Welfare, Republic of Korea (Grant Number: HI15C0638) and by a grant (2014R1A2A1A10050270) from the Mid-career Researcher Program through the National Research Foundation of Korea funded by the Ministry of Science, ICT & Future Planning.

The authors have nothing to disclose.

All relevant data are within the paper and its supporting information files.