Dosimetric evaluation of a simple planning method for improving intensity-modulated radiotherapy for stage III lung cancer

Abstract

This study aimed to evaluate the dosimetric outcomes of a base-dose-plan-compensation (BDPC) planning method for improving intensity-modulated radiotherapy (IMRT) for stage III lung cancer. For each of the thirteen included patients, three types of planning methods were applied to obtain clinically acceptable plans: (1) the conventional optimization method (CO); (2) a split-target optimization method (STO), in which the optimization objectives were set higher dose for the target with lung density; (3) the BDPC method, which compensated for the optimization-convergence error by further optimization based on the CO plan. The CO, STO and BDPC methods were then compared regarding conformity index (CI), homogeneity index (HI) of the target, organs at risk (OARs) sparing and monitor units (MUs). The BDPC method provided better HI/CI by 54%/7% on average compared to the CO method and by 38%/3% compared to the STO method. The BDPC method also spared most of the OARs by up to 9%. The average MUs of the CO, STO and BDPC plans were 890, 937 and 1023, respectively. Our results indicated that the BDPC method can effectively improve the dose distribution in IMRT for stage III lung cancer, at the expense of more MUs.

Introduction

Lung cancer constitutes a major source of mortality in the world and the disease is usually diagnosed in advanced stages^1,2. Concurrent chemoradiotherapy (CCRT) represents the standard of care for patients with stage III non-small cell lung cancer (NSCLC)³. Previous studies demonstrated that the use of intensity-modulated radiation therapy (IMRT) improved plan quality, reduced the toxicity and improved local control and survival rates⁴. However, treatment of larger tumors has been reported to be associated with increased risk of severe radiation pneumonitis and esophageal toxicity^3,5,6. Verbakel et al.⁷ implemented a hybrid intensity-modulated radiotherapy technique to achieve lung sparing for stage III lung cancer treatment, however, radiotherapy complications such as radiation esophagitis and radiation-induced heart diseases were not alleviated in their method. Therefore, it is essential for us to develop another method to spare the organs at risk (OARs).

On the other hand, because the optimizers of current treatment planning system use simplified algorithms instead of full volume dose algorithms for fast dose computation, the finally calculated dose was subject to an optimization-convergence error (OCE)^8,9, which leads to the discrepancy between the optimizer and finally calculated dose and thus resulting in suboptimal deliverable treatment plan. Especially in lung cancer cases with low density in lung tissue, significantly lower dose in the target containing lung tissue

Table 2 Summary of the dose constraints for stage III lung cancer.

Plan evaluation

D_98%, D_2%, D_50%, conformity index (CI) and homogeneity index (HI) was evaluated for PTV among the three planning methods. D_x% represents the dose received by x% volume of the organ. For example, D_50% means the dose received by 50% volume of the organ. The CI proposed by Paddick¹⁵ was defined as the location of the prescription isodose volume (PIV) with respect to the target volume (TV). HI is defined by the following formula according to the recommendations of ICRU report 83¹⁶. The CI value was between 0 and 1 with 1 representing ideal conformity. On the contrary, the HI value of 0 represented ideal homogeneity in the target.

The maximum dose and various dose-volume parameters to specific OARs were generated for the plans to assess their effectiveness in OAR sparing. V_x stands for the volume of the organ receiving a dose of ≥x Gy. For example, V₄₀ means the volume of organ receiving a dose of ≥40 Gy. Specifically, the spinal cord was assessed in terms of its maximum dose. The total lung (T-L) was evaluated using V₅, V₁₀, V₂₀, V₃₀ and the mean lung dose (MLD); the contralateral lung (C-L) was evaluated using V₅; the esophagus was appraised with the maximum dose, mean dose, V₃₅, V₅₀ and V₆₀. The heart was assessed in terms of V₃₀, V₄₀ and mean dose. Monitor units (MUs) per fraction were compared among the three planning methods.

Statistical analysis

Data analysis was carried out using the SPSS version 19.0 software (SPSS, Inc., Chicago, IL, USA). The differences among the BDPC, STO and CO plans were evaluated using repeated measures ANOVA. When p of <0.05 was achieved, a further Least Significant Difference (LSD) measurement was performed to compare the difference between groups. Differences were considered to be statistically significant when p was <0.017 due to the adjustment of the observed significance level by one third.

Results

Target coverage and MUs

All the plans created by the three planning methods fulfilled the specified dose constraints. The BDPC method achieved more homogeneous dose distribution, irrespective of isodose distribution (Fig. 1) and dose volume histogram (DVH) display (Fig. 2a). Table 3 summarized the dose coverage parameters and MUs among the three planning methods. It could be seen from Table 3 that the BDPC method obtained significantly higher D_98% (9.7 ± 0.9% and 1.3 ± 0.9% higher than the CO and STO) of the PTV. Meanwhile, it also achieved lower D_2% (5.8 ± 1.7% and 2.7 ± 0.9% lower than the CO and STO). With regard to the HI, BDPC was significantly superior to CO and STO by 54.4 ± 8.9% and 37.6 ± 9.6%, respectively. With regard to the CI, BDPC was better than CO and STO by 6.9 ± 3.8% and 3.4 ± 2.1%, respectively. However, BDPC required more MUs than the other two methods. The MUs were increased by 15.0 ± 3.2% and 9.3 ± 3.9% compared to CO and STO, respectively.

Table 3 Summary of the target dose coverage parameters and monitor units (MUs) for the three planning methods.

Figure 1

Isodose distribution for the conventional optimization (CO) (a), split-target optimization (STO) (b) and base-dose-plan-compensation (BDPC) planning methods (c) from one representative case.

Figure 2

Dose-volume histograms (DVHs) of the planning target volume (PTV) (a) and organs at risk (OARs) (b–f) for the conventional optimization (CO), split-target optimization (STO) and base-dose-plan-compensation (BDPC) planning methods. T-L = total lung; C-L = contralateral lung.

OARs sparing

All the OARs exhibited lower dose with the BDPC method. The numerical statistics from the DVH analysis were listed in Table 4. The V₃₅, V₅₀, V₆₀ and mean dose to the esophagus were reduced by 3.2%, 3.0%, 9.1% and 7.0% compared to CO and by 2.5%, 1.9%, 5.6%, 4.5% compared to STO. The V₅, V₁₀, V₂₀, V₃₀ and MLD of the T-L were reduced by up to 2.2%. The V₅ of the C-L was 4.0% and 3.4% lower compared to CO and STO, respectively. BDPC also reduced the V₃₀, V₄₀ and mean dose of the heart by 1.8%, 1.1% and 4.4% compared to CO and by 1.0%, 0.7% and 2.1% compared to STO and it reduced the spinal cord dose by 3.1% compared to CO. Additionally, BDPC resulted in the least volume receiving a high dose (107% of the prescribed dose) in the normal tissue. Figure 2b–f displayed the DVH of the OARs for the three planning methods in one representative case. It appeared that most of the DVH curves of the OARs of BDPC shifted to the left, indicating lower dose received.

Table 4 Summary of the dose to organs at risk (OARs) for the three planning methods.

Discussion

To improve the therapeutic ratio and obtain optimal clinical outcomes, it is essential to improve the planning technique to give full scope to the advantages of IMRT for lung cancer, that is, to achieve better target dose homogeneity, conformity and better OARs sparing. Our study demonstrated that the introduced BDPC planning method has the ability to further improve the IMRT for stage III lung cancer.

The main advantage of the BDPC method lies in its homogeneous dose distribution in the target with significantly fewer hot and cold spots with an improvement by 38–54%. For lung radiotherapy treatment, PTV was typically generated to account for position, size and shape caused by respiratory motion and uncertainties during patient positioning and alignment of the therapeutic beams during the treatment¹⁷. Underdosage in the PTV may result in insufficient dose to the tumor and may lead to the likelihood of tumor recurrence¹⁸, because the tumor control probability (TCP) predominately correlates with the minimum dose of tumor¹⁹. And overdosage may result in severe acute reactions in tissues (such as esophagus) or late complications²⁰. Accordingly, the improvement of homogeneity may have potential clinical benefits of lowering the risk of tumor recurrence and decreasing the radiation-induced toxicity due to unnecessary excessive dose.

Furthermore, the BDPC method provided superior conformity by 3.4–6.9%, which could better spare the surrounding normal tissue. All the OARs exhibited 0.7–9.1% dose reduction with the proposed BDPC method. The findings are attractive because a number of studies have reported the association between the dose-volume predictors and the incidence of complications in lung radiotherapy treatments. The common complications include the radiation-induced pneumonitis, radiation esophagitis and radiation-induced heart diseases^21,22,23. Many dosimetric predictors, such as V₅, V₁₀, V₂₀, V₃₀ and MLD of T-L were reported to be associated with radiation-induced pneumonitis^{21,24,25,26,27,28}. Song et al.²⁹ found correlations between fatal pneumonitis and CL-V₅ dose and he suggested that CL-V₅ should be kept to be less than 60%. Radiation esophagitis was another common complication experienced by lung cancer patients receiving radiotherapy treatment³⁰. These complications significantly affected quality-of-life and could negatively impact long term survival. Rose et al.⁶ related V₃₅ and V₆₀ to clinically significant radiation esophagitis. Palma et al.³¹ determined that the esophageal volume receiving ≥60 Gy (V₆₀) alone emerged as the best predictor of grade ≥2 and grade ≥3 radiation esophagitis in patients undergoing concurrent chemoradiation therapy. V₅₀^32,33,34 and the mean dose³⁵ were also associated with the risk of esophageal toxicity. Radiation-induced heart disease had been well documented and was believed to occur during radiotherapy. In a study by Veinot et al.²², patients who received thoracic radiotherapy showed moderate to significant myocardial fibrosis with heart exposure >30 Gy. The heart V₃₀ was reported as a significant predictor³⁶ of radiation-induced pericardial effusion, with a V₃₀ of >46% associated with a 73% rate of pericardial effusion compared with 13% for a V₃₀ of <46%. As all the predictors mentioned above could be further reduced by the BPDC method, the risks of radiation-induced complications may be decreased and the quality of life may be potentially improved for the lung cancer patients.

Traditionally, the “base dose plan” function is usually applied for optimizing a second-course treatment plan (such as a boost plan), while considering the first-course plan, in order to achieve an optimal plan sum (TDP plus BDP) in the optimizer. But the “base dose plan” function is utilized in a different way in our method, because it is used for obtaining a deliverable treatment plan (TDP) with finally calculated dose, not a plan sum in the optimizer. In principle, the “base dose plan” function is adopted to compensate for the OCE. When an OCE introduces a cold spot into the finally calculated dose distribution in the CO plan (BDP), the BDPC plan (TDP) will produce a hot spot in the corresponding location to even out the original cold spot for a uniform summed dose. After calculating the final dose of the optimized BDPC plan (TDP), the OCE introduces a cold spot into the hot-spot region of the BDPC plan (TDP) and finally, the BDPC plan can achieve a uniform dose.

The OCE primarily originated from three major sources including tissue heterogeneity, MLC leaf motion calculation and the optimization algorithm⁹. Possible solutions to the OCE were investigated in several studies. The STO method⁷ is an effective approach to minimize the error arising from tissue heterogeneity, but the errors from the other two sources could not be reduced, so it is not effective enough. By contrast, the BDPC method is able to reduce the overall OCE by compensating for the whole plan and is effective enough. Another planning method proposed by Süss et al.^37,38 corrects the OCE by re-optimization with additional optimization objectives to address hot and cold spots, but it is only locally effective and new hot and cold spots may appear in other region. By contrast, the BDPC method is globally effective throughout the entire treatment region. The Direct Aperture Optimization (DAO) technique^39,40,41 incorporates the deliverable MLC apertures series instead of ideal fluences in the optimizer to eliminate the error arising from the MLC leaf motion calculation. Unfortunately, this technique is not available in the treatment planning systems without DAO, such as Eclipse version 10.0, whereas the BDPC method is commonly available because a “base dose plan” or similar function is a basic feature for treatment planning systems. Zacarias and Mill⁸ also adopted the “base dose plan” function to overcome the OCE, but that method is not the same as ours, because it required a complex process and additional softwares thus leading to significantly increased planning steps and time. On the contrary, our method is much simpler and practical for routine use, as the only required procedure is modifying one parameter (number of fractions of BDP) and the excellent homogeneous dose distribution can be effortlessly achieved through a single further optimization.

However, the introduced method slightly increased the number of MUs, which was reported to be associated with more treatment time and peripheral dose outside the treatment field, leading to a likelihood of intrafraction shifts of tumor position and radiation-induced secondary cancer^42,43,44,45. As shown in Table 3, the MUs were 1023 ± 159, 890 ± 134 and 937 ± 145 for the BDPC, CO and STO method, respectively. We infer that the treatment time with the BDPC method is increased by 13.3 and 8.6 seconds on average compared to the CO and STO methods with the dose rate of 600 MU/min. As to the peripheral dose, we find it is increased by 0.32 cGy and 0.21 cGy on average compared to the CO and STO methods (1 MU generates 2.44 × 10⁻³ cGy peripheral dose at 20 cm away from the isocenter according to the results of our previous measurement). Whether these drawbacks will impact on the clinical treatment needs further investigations.

Conclusions

In this study, we evaluated the dosimetric characteristics of an IMRT planning method, the BDPC method, applied in stage III lung cancer. We found that this method not only improved the conformity and homogeneity of the target but also spared the OARs, thus may increasing the therapeutic ratio. Furthermore, it is simple and effective for routine use. Therefore, the proposed method is recommended for the treatment of stage III lung cancer.