Physics Contribution
4π Noncoplanar Stereotactic Body Radiation Therapy for Centrally Located or Larger Lung Tumors

https://doi.org/10.1016/j.ijrobp.2013.02.002Get rights and content

Purpose

To investigate the dosimetric improvements in stereotactic body radiation therapy for patients with larger or central lung tumors using a highly noncoplanar 4π planning system.

Methods and Materials

This study involved 12 patients with centrally located or larger lung tumors previously treated with 7- to 9-field static beam intensity modulated radiation therapy to 50 Gy. They were replanned using volumetric modulated arc therapy and 4π plans, in which a column generation method was used to optimize the beam orientation and the fluence map. Maximum doses to the heart, esophagus, trachea/bronchus, and spinal cord, as well as the 50% isodose volume, the lung volumes receiving 20, 10, and 5 Gy were minimized and compared against the clinical plans. A dose escalation study was performed to determine whether a higher prescription dose to the tumor would be achievable using 4π without violating dose limits set by the clinical plans. The deliverability of 4π plans was preliminarily tested.

Results

Using 4π plans, the maximum heart, esophagus, trachea, bronchus and spinal cord doses were reduced by 32%, 72%, 37%, 44%, and 53% (P≤.001), respectively, and R50 was reduced by more than 50%. Lung V20, V10, and V5 were reduced by 64%, 53%, and 32% (P≤.001), respectively. The improved sparing of organs at risk was achieved while also improving planning target volume (PTV) coverage. The minimal PTV doses were increased by the 4π plans by 12% (P=.002). Consequently, escalated PTV doses of 68 to 70 Gy were achieved in all patients.

Conclusions

We have shown that there is a large potential for plan quality improvement and dose escalation for patients with larger or centrally located lung tumors using noncoplanar beams with sufficient quality and quantity. Compared against the clinical volumetric modulated arc therapy and static intensity modulated radiation therapy plans, the 4π plans yielded significantly and consistently improved tumor coverage and critical organ sparing. Given the known challenges in central structure dose constraints in stereotactic body radiation therapy to the lung, 4π planning may increase efficacy and reduce toxicity.

Introduction

Stereotactic body radiation therapy (SBRT) using highly hypofractionated doses, typically 50 to 60 Gy in 3 to 5 fractions, has achieved remarkable success in treating early-stage lung cancer. Excellent local control rates and low toxicity were reported in many clinical trials 1, 2 for non-small cell lung cancer. For early-stage and peripheral lung cancer, SBRT has become the standard of care for medically inoperable early-stage lung cancer with comparable survival rates to those of surgery (1). However, it has also been realized that local control rates decrease with increasing tumor sizes. Dunlap et al reported 90% and 70% local control rates for T1 and T2 tumors, respectively, showing a significant size factor in lung SBRT efficacy (3). Similarly decreased local control rates were observed by other investigators 1, 4. Beitler et al (5) showed much poorer outcomes for gross tumor volume larger than 65 cm3, roughly the volume dividing stage IA and stage IB lung tumors. Larger treatment volumes also lead to higher doses to organs at risk (OARs). Limited normal organ tolerance not only prevents higher doses from being delivered to larger tumors but also demands compromised, lower prescription doses in practice, further reducing local control rates. Another challenge in lung SBRT has come from treating centrally located lung tumors. Timmerman et al (6) reported an increase of 8- to 10-fold in toxicity compared with peripheral tumors from lung SBRT and associated the toxicity to high doses to central organs with serial radiobiology characteristics. Song et al (7) concluded that SBRT should not be given to centrally located tumors because of the significantly higher probability of severe side effects. However, based on the radiation therapy oncology group (RTOG) central tumor definition of a 2-cm volume around segmental bronchi for central tumors, these constitute a significant number of tumors in which an improvement in sparing of OARs is meaningful. In practice, moderate SBRT doses such as 50 Gy in 4 to 5 fractions, as opposed to 54 to 60 Gy in 3 fractions to peripheral lung tumors, are recommended for these tumors that are larger, centrally located, or both to achieve a balance between local control and normal tissue toxicity (8). The compromise has diminished potential gains in tumor control probability for these tumors because of the significantly lower biological equivalent doses (9).

The RTOG lung SBRT protocol 0813 for centrally located lung tumors (10) constrains the maximum esophagus heart, great vessels, and trachea/bronchus doses to between 18 and 32 Gy, corresponding to approximately 30% to 60% of the prescription doses. The RTOG protocols recommend using either noncoplanar static or coplanar arc beams to achieve the R50 dosimetric goal, which is defined as the 50% prescription dose volume divided by the planning target volume (PTV). Lim et al (11) showed that when compared against coplanar plans, a lower R50 could be achieved in 81% of patients when noncoplanar beam arrangements were used. However, Christian et al (12) found no dosimetric improvements when comparing noncoplanar and coplanar IMRT plans and that the noncoplanar plan quality was highly dependent on the dosimetrists' experience and the plans were more difficult to deliver.

We previously showed in liver SBRT (13) that large dosimetric gains could be achieved using optimized noncoplanar IMRT beams when compared against state-of-the-art VMAT plans. In the current study, we applied a modified version of the method to these challenging lung SBRT cases to test whether significant dosimetric gains could be achieved, allowing simultaneous critical organ dose reduction and PTV dose escalation.

Section snippets

Dose matrix calculation

The planning process began by distributing 1162 noncoplanar candidate beams throughout the entire 4π solid angle space with 6° of separation between 2 nearest neighbor beam pairs. From the candidate pool, we eliminated those beams that would cause collisions between the gantry and the couch or patient. Collisions were determined using a precise computer-assisted design models of the linear accelerator (Varian EX) and a human subject and simulating their relative positions for each candidate

Results

The 4π algorithm was evaluated for each clinical case according to plans that used up to 30 coplanar and noncoplanar fields. The numbers of beams were selected based on the R50 comparison between coplanar and noncoplanar plans of a typical patient (Fig. 1b). With fewer than 10 beams, the difference in R50 was insignificant and could be compensated for by using more coplanar beams. This agreed with previous observations 16, 17, 18. However, with more than 20 noncoplanar beams, the R50 of the

Discussion

It is evident that 4π planning can provide definitive dosimetric improvements to lung SBRT. With 4π radiation therapy, the tumor prescription dose to these patients could be escalated to 68 Gy or higher without exceeding the dose limits set by previous clinical plans, which were well tolerated by these patients. The higher PTV dose has been correlated to significantly higher tumor control probabilities 19, 20, which would be particularly important for larger (stage IB and IIA) and centrally

Conclusion

A novel treatment planning method has been proposed incorporating beam orientation and fluence map optimization algorithms on the full 4π noncoplanar solid angle space. Algorithm performance was examined by comparing lung SBRT plans for patients with tumors that were larger, centrally located, or both. Compared against state-of-the-art VMAT plans and 7 to 9 static IMRT beams selected by dosimetrists, the 4π plans yielded significantly and consistently superior performance in tumor coverage and

References (21)

There are more references available in the full text version of this article.

Cited by (0)

Conflict of interest: none.

View full text