With early stage primary non-small-cell lung cancer (NSCLC) of T1 or T2 lesion not including metastases, usually the surgical resection was chosen to manage using a lobectomy technique. Unfortunately, significant complication could be associated with lobectomy for those medically inoperable patients [2] -[3] . SBRT has been shown to be an effective treatment option for inoperable patients with lung cancer and metastatic lung lesions. Noticeable local tumor control rates had been reported with SBRT treatment technique but most reports have only one lesion in their studies [4] -[6] . However, most reports in this area documented patients with only one lung lesion instead of multiple lung lesions. Kelly et al. [7] has studied lung patients with up to three multiple metastatic lung lesions and reported no grade 4 - 5 toxicity. Another group by Okunieff et al. [8] also reported dosimetric evaluation with more than five metastatic lesions but without any explicit reports on the outcomes of the SBRT treatments. In order to acquire the most valuable dosimetric information and understand how different radiation schedules is being adopted with fractionated scheme for SBRT in lung, we have also listed the Biological Equivalent Dose (BED) derived from the report of Kavanagh et al. [9] . This simple table has not correctly predicted a linear quadratic correlation between the lung lesions and late responding normal tissues; however, it does support the concept of SBRT which could generate a higher equivalent dose for hypofractionated treatment schemes (see Table 1).

There have been reports which relate the clinical efficacy of SBRT over different fractionations to a BED cutoff. Thus, Onishi, et al. [10] reported that improved local control and survival are associated with SBRT regimens whose BED is >100 Gy. In other words, the purpose of SBRT treatment is to produce a higher tumor control intent and safer outcome with possible minimum complications. The recognition of dose escalation for hypofractionated treatment has gained popularity in recent years, with many clinical implementation, presented the successful local control of metastatic lung disease treatment [11] -[13] .

Traditional SBRT treatment was performed with 3-D conformal therapy. The basic principle is to utilize a set of static beams and to organize the optimized beam and couch angle combinations in order to form the best dose coverage to the target. This process needs tremendous amount of experience and can be system dependent to avoid collision in case they are non-coplanar. The process could be tedious and time consuming. Using IMRT could be a more efficient way to obtain the expected dose coverage, while minimizing the normal structures, either by coplanar or non-coplanar field and beam designs. Tomotherapy is a rotational unit for helical pattern

treatment. With that, the calculated dose distribution could be utilized to compare with Tomotherapy technique.

In this study, we have specifically designed the treatment protocols to reproduce the dosimetry and targeting patients which related to the plans carried out by Pinnacle planning system, either with coplanar and non-co- planar beams entries.

Nine patients diagnosed with NSCLC staged from T1 - T3 with multiple lung metastases were selected (n = 9) for this dosimetric study. There were two patients with two lesions on the right lung, one patient with two lesions on the left lung, five patients with one lesion on the right and one lesion on the left lung, and one patient with two lesions on the right side and one lesion on the left. The range of PTV was from 14.31 to 91.26 cc which presented the mean volume of 37.6 ± 23.4 cc. Plans on the Pinnacle system were programmed with nine non-opposing gantry angles (0, 30, 135, 165, 200, 240, 270, 300 and 330 degrees) for coplanar beams. And with non-coplanar beam settings, the following combination of gantry and couch angles was implemented for individual lung planning (Table 2).

The energy was selected at 6 MV photons; the prescription was 20 Gy per fraction with a total dose of 60 Gy (3 SBRT fractions). Single isocenter on Pinnacle was set to the geometric center of the two or more lesions for planning and prescription purposes. The sorted DICOM images with isocenter and structures such as PTV, lungs (exclude GTV), heart, cord, cord-exp (1 cm expansion of contoured cord structure), trachea, brachial plexus, ribs, chest wall, high dose, low dose (external contour minus 2 cm expansion of PTV), RING structure (external subtract 1 cm of PTV expansion) were transferred from Pinnacle to Tomotherapy. The RING structure was used to avoid any hot spot and to generate a smoother dose distribution among the three planning tools. For both coplanar and non-coplanar plans, Direct Machine Parameters Optimization (DMPO) was used for optimization, max iterations was 50, convolution dose iteration was 20, maximum number of segments were set to 50 which was tested to suit most of the clinical cases, minimum segment area was 4 cm², and minimum segment MUs was designed 5 to avoid small MU delivery. For Tomotherapy plans, the jaw width was set at 1 cm, grid size was set to normal, modulation factor was preset at 1.7 for four cases, and five cases were programmed at 1.4 due to problem with gantry period (Tomotherapy has the speed limit with maximum 6 rotations in 1 minute). Pitch was 0.09 for one case, and other eight cases were set to 0.1. Similar optimization goals were set for all cases. Multivariate analysis using PASW™ (formerly SPSS™) statistical tool version 18 with a significance level of 0.05 were used for statistical analysis (i.e. p < 0.05 indicates clinical significance among the study cases).

Planning outcomes such as V₉₅ (95% prescribed dose to volume, D₉₉ (99% of the target volume receives a minimum of 90% of the prescription dose), D₉₅ (95% of the target volume receives the prescribed dose), D₅, D₁, mean and minimum dose to PTV; V₂₀ (percent volume receiving 20 Gy), V₅ (percent volume receiving 5 Gy), and mean dose to lungs; Maximum dose to heart, esophagus, cord, trachea, brachial plexus, rib, and chest wall; V₃₀ (percent volume receiving 30 Gy) to chest wall, and mean dose to liver were reported for this study.

V₁₀₅ (high dose) was 105% of prescribed dose delivered to volume of body minus PTV, low dose as the falloff gradient was defined as dose to volume of body minus 2 cm expansion of PTV according to RTOG0236 protocol [1] .

CI and HI were also evaluated according to RTOG0236 protocol as:

where V_PTV is the volume of PTV, and V_PIV is volume of prescription isodose volume. The ideal CI is <1.2 with a minimum tumor size of 3.5 cm.

HI was also defined by RTOG0236protocol as:

where D₅ is prescribed dose to cover 5% of PTV, and D₉₅ is prescribed dose to cover 95% of PTV.

Treatment plans were executed on the Pinnacle system by setting the same dose constraints as in Tomotherapy with the same contours transferred from Pinnacle. The coplanar was with 40 degrees separation each, and the non-coplanar beams were followed with the same angles but with couch rotations in a pre-set parameters for left and right lung lesions, respectively.

SBRT treatment opens a new era for treating the lung metastases compared to the conventional surgery, which was invasive with higher risks. In North America, SBRT has been the standard choice of treatment for selected

group of patients with superior clinical outcome [14] . Systematic study has shown that SBRT created very comparable results to conventional approach [15] . Dunlap et al. [16] published a comparison of T1 and T2 peripheral lesion SBRT treatment comparison, with emphasis on the tumor size factor based on the Tomotherapy delivery results. The corresponding CI and HI were 1.17 and 1.06, respectively. The societies of ASTRO and AAPM have provided user guidelines of valuable information about how to safely utilizing the imaging technique as well as SBRT clinical practices [17] -[19] . As cross checked and examined in the literatures [20] -[23] , patients with multiple lung metastases presented dosimetric challenges in 3D conformal therapy due to the complexity in designing the treatment ports and to avoid crossing the beam angles for hot spots. With IMRT technique, the plan might generate better coverage and eliminate possible dose spreading to the normal lung tissues. Though chemotherapy maybe an option, but the outcomes are always disappointing in managing those cases [24] . With the fast developments in radiotherapy treatment hardware and software during the last decade, two techniques have become available for the treatment of patients with multiple metastases by using static beam IMRT and/or Volumetric Modulated Arc Therapy (VMAT) technology [25] . The study by Sterzing et al. [26] has already shown that helical Tomotherapy is capable of treating multiple lesions in their early investigation. Recently arc therapy treatment modality has been adopted in clinics to manage multiple lung lesions, Li et al. [27] has studied a frameless SBRT treatment with arc based planning and delivery, which reported the CI and HI with the ranges of (0.513 - 0.562) and (0.0709 - 0.0794) of two plans. However, the CI and HI indexes were not derived from the RTOG0236 protocol, but they have indicated the V₂₀ values fell into 8.46% and 14.39%, respectively. They were utilizing the Elekta™ system VMAT technique to compare the planning results with multiple lung lesions. This article also demonstrated the arc based therapy could be beneficial to create suitable dose distribution with reasonable sparing to normal lung tissues. Monte Carlo calculation also has been proposed to avoid and dose discrepancy in the lung-tissue interface for inhomogeneity corrections. Lax et al. [28] has found that the results of their particular study showed average differences of 9% (minimum, −8%; maximum, 29%), 12% (minimum, 0; maximum, 28%), 7% (minimum, 3%; maximum, 13%), and 18% (minimum, 11%; maximum, 29%) in R_100%, R_50%, D_{2 cm}, and V₂₀, respectively. However, the current calculation algorithm in our study was still based on convolution/superposition which Tomotherapy and Pinnacle systems utilized in RTOG2036 comparisons.

The planning comparison focused on the physics strength from helical delivery, along with simple steps in creating an acceptable plan with better planning quality. Helical Tomotherapy presented with a good option to plan and treat those tough cases (multiple lesions) with very encouraging clinical outcomes. Our study has shown that Tomotherapy has better coverage and less normal tissues doses among those 9 patients, as they only concentrate on the feasibility of planning multiple lung lesions with one isocenter. The treatment time was not compared against coplanar or non-coplanar plans from Pinnacle systems due to the continuous couch movement with Tomotherapy delivery, which the MLC was in fact close in between PTVs. The Tomotherapy delivery time was be longer than the static beams treatment in IMRT. Though TomoEdge™ has been released for clinical usage, but at this moment, our system was not upgraded to execute this option, which can dramatically reduce the treatment delivery time by varying the jaw sizes to reduce the time in between PTVs.

We also noticed that for the total lung planning results, where the V₅ of coplanar plan was the smallest among all the calculated results. One of the reasons was that the size of lung lesions only occupied a small portion within the lung volume, and the coplanar beams would just penetrate through the section of slices where the PTVs possessed. The static beam IMRT delivery tended to minimize the lung doses, as expected, but the tradeoff is losing coverage to cover 95% of PTVs, which Tomotherapy had the highest scores. With all factors considered, Tomotherapy has the overall benefits of better CI and HI, and with less critical structure impact.

SBRT using a helical Tomotherapy delivery unit is a well-tolerated and documented methodology in treating multiple lung metastases of inoperable early-stage NSCLC. From the conventional planning of view, it is always difficult to carry multiple areas treatments which were complex and hard to normalize. Based on our study results, dosimetric analysis of multiple lung lesions has shown that Tomotherapy could still produce higher CI and HI in selected cases. Though the cases were not treated with Linac based arc therapy technique (or VMAT) due to the availability at our center, even compared to the non-coplanar SBRT in the analysis, we found the Tomotherapy tended to create more suitable target and critical structures dose statistics as a possible factor for complications indicated [29] . However, the low dose regions created by Tomotherapy needs to be addressed more due to the spillage characteristics and the volume effect of low dose with large volume coverage may be a critical factor in toxicity analysis for future follow-up of multiple lung metastases patients.

None.