Improve dosimetric outcome in stage III non-small-cell lung cancer treatment using spot-scanning proton arc (SPArc) therapy

Lung cancer remains the leading cause of cancer deaths for both men and women in the United States [1]. For patients with locally advanced lung cancer, radiotherapy alone or concurrent chemoradiotherapy has been widely used. Many trials have shown the benefits of concurrent chemoradiotherapy to improve the prognosis of patients, in terms of local control and survival [2‐6]. However, the associated toxicities to the surrounding normal tissues can be significant, which could limit the ability of target dose escalation to possibly improve the outcome [7‐12].

Proton beam therapy, unlike photon radiotherapy, has inherent physical characteristics which are advantageous to treat disease sites such as locally advanced lung cancer, due to its sharp distal dose fall-off. Compared to passive scatter proton therapy (PSPT) [13], intensity modulated proton therapy (IMPT) has shown potentials to further reduce the dose to the adjacent normal tissues while maintaining similar or superior target coverage [14‐17] in a more efficient way without having the need to use beam specific blocks or compensators. In IMPT, thousands of mono-energetic narrow beamlets (“spots”) are optimized simultaneously and magnetically scanned to superpose a desired dose distribution. Although such active scanning delivery system provides the greatest flexibility to shape the target dose pattern, it can also be degraded by different uncertainties, including setup, range uncertainties [18‐20], and the motion induced uncertainties [21‐25].

Many studies have shown that robust optimization can effectively reduce the negative dosimetric impact from setup and range uncertainties compared to traditional margin approach in IMPT planning [26‐32]. All studies on robust optimization, however, are still limited to few beam angles, due to low calculation speed and delivery efficiency. These obstacles limit our ability to further exploit the benefits of IMPT. With the recent development of spot-scanning arc therapy (SPArc) [33], which generates robust and delivery efficient spot scanning proton arc plans, we could potentially overcome the current dosimetric limitations [34, 35]. Herein, we evaluate the potential benefits of SPArc for treating locally advanced non-small-cell lung cancer (NSCLC) relative to the conventional multi-field robust optimized IMPT (RO-IMPT) plans.

This study evaluated the dosimetric benefits of using SPArc in treatment of stage III NSCLC patients. These results suggest that SPArc is able to significantly reduce the dose to all critical structures evaluated, while maintaining similar or superior target volume coverage and plan robustness compared to RO-IMPT plans. It is important to note that such additional organ dose sparing could potentially grant the clinician the ability to escalate the dose to the target volume safely. Although RTOG 0617 and a recent meta-analysis of randomized trials [40, 41] did not show a benefit of dose escalation in the chemoradiotherapy setting, the negative results could be partially related to high levels of toxicity, as dose escalation still appears to improve survival in the radiation alone setting. RTOG 0617 showed a strong association between overall survival and higher heart dose, especially the heart volume receiving more than 40 Gy [42]. Furthermore, the secondary analysis showed using IMRT planning also resulted in a significant reduction in grade 3 or higher pneumonitis. With techniques such as SPArc, shown here to further reduce lung and other normal tissues doses while concurrently reducing interplay effects related to respiratory tumor motion which may be critical in proton delivery, there may be further potential to dose escalate the target volume without increasing toxicity or mortality but potentially improving control or survival. As national cooperative groups compare proton to photon irradiation with survival endpoints, such as in the on-going RTOG 1308 and protons become more readily available in the community, these issues will become increasingly important.

One counterintuitive result of our study is that SPArc plans could significantly decrease the integral body dose compared to RO-IMPT plans despite that SPArc generally uses hundreds of beams entering from different angles. One explanation is that the iteratively energy layer selection algorithm of SPArc optimization tends to filter the highest energy of each beam angle and superimpose the dose contributions from other beam angles to cover the distal extent of the target volume to produce a most robust plan. Figure 5a plots the MUs distributions versus energy of both SPArc and RO-IMPT plans for patient 6. As it is shown, the SPArc plan only used energies higher than 82.6 MeV and lower than 164.5 MeV while the RO-IMPT delivered a portion of the plan using energies higher than 164.5 MeV. In order to quantitatively illustrate this concept in more details, the calculated total energy depositions (CTED) were estimated as follows to evaluate the ratios between RO-IMPT and SPArc plans:where MUi, Ei, S(Ei) are the monitor units, energy, and the corresponding proton stopping power for the energy in air for ith energy layer. Such CTED ratios roughly estimate the total energy depositions between these two modalities. Figure 5b shows the ratios among integral dose, total MUs, and CTED. The integral dose ratios match very well with the CTED ratios, which ultimately explain the reduction of integral body dose for the SPArc plan.

One of the major challenges with scanning beam proton therapy in treatment of lung cancer patients is the interplay effect between the spot scanning delivering sequence and the patient specific respiratory motion. Such interplay effect could lead to severe dose distortion especially in cases of large tumor motion, and could possibly be diminished by rescanning [23, 25, 43, 44]. In this study for a fair comparison, we only compared the single-fraction 4D dynamic dose calculation without taking rescanning into account. The results demonstrated the ability of SPArc to reduce the interplay effect as shown in Fig. 4. This improvement could be explained by the washout effect of SPArc’s hundreds of beam angles. Further studies are needed, however, evaluating different tumor motions and ELST simulating different patients and machine parameters to evaluate the interplay effect and SPArc more comprehensively. In routine clinical practice of using IMPT for motion target, rescanning technique is normally used in order to mitigate such interplay effects.

Moreover, it is important to mention that another potential benefit of using SPArc over the RO-IMPT plan could be the relative biological effectiveness (RBE) optimization. The RBE of 1.1 has been used historically and conservatively to ensure target coverage without taking into account the variation of RBE along the beam path. Paganetti et al. [45] has systemically analyzed the published experimental data and supported that there is an increase of RBE to 1.7 towards the distal falloff of the spread-out Bragg peak. Unlike traditional passive scatter proton therapy, IMPT has the potential to intentionally place the higher RBE around the Bragg peak inside the target volume to avoid any extra biological damage to the surrounding normal tissue. Furthermore, in case of SPArc, using hundreds of beam angles along with great energy layer selection magnetism would provide an even greater flexibility to locate the high RBE points inside the target volume. As indicated in Fig. 5a, even without biological optimization, the SPArc plan tends to avoid using very low and high energy beams, and therefore allows a greater probability of placing Bragg Peak into the target volume.

SPArc, as demonstrated here, shows significant potential to further improve the dosimetric outcome and therapeutic ratio in treating locally advanced NSCLC compared to RO-IMPT. Due to the limitations of current proton scanning beam systems, continuous gantry rotation with scanning delivery is still not available. Issues that need to be addressed in order to efficiently and safely implement SPArc for routine clinical use are: (1) developing a new rotational gantry which is capable of rotating around the iso-center while delivering pencil beam therapy with sub-millimeter accuracy; (2) designing new beamline control system to adjust and monitor the beam energy spectrum, spot position and spot profile during the gantry rotation, and (3) implementing a comprehensive SPArc quality assurance program.

SPArc is a robust and delivery-efficient proton spot-scanning arc therapy technique which could potentially be implemented into routine clinical practice to further improve treatment outcomes in patients with locally advanced NSCLC.