Early stage non-small cell lung cancer treated with pencil beam scanning particle therapy: retrospective analysis of early results on safety and efficacy

The patients were positioned in supine or prone position based on the location of the radiation targets using a vacuum bag. A thermoplastic mask was used to fix the position and restrict the breath motion for patients using free breathing or gating.

All patients were evaluated for tumor motion under fluoroscopy before simulation. If the tumor motion in all the directions was less than 5 mm, the patient would be treated under free breathing (FB); if the motion exceeded 5 mm, a breath control technique, either active breathing control (ABC) or respiratory gating, was required to mitigate the residual motion to < 5 mm during treatment.

The simulation CT scanning started from the angle of mandible to adrenal glands to include tumor lesions, entire lungs, whole neck and all the organs/tissues by which the beams probably pass. CT scanning was performed under the same respiratory mode as that would be used for PT treatment. All CTs for planning were plain CT, and a contrast CT afterwards was scanned for image fusion only when needed.

For patients to be treated under FB or ABC, a plain simulation CT was scanned under the same respiratory mode. The gross tumor volume (GTV) was delineated on the plain CT.

For patients using gating, a 4-dimensional (4D) simulation CT was performed. Ten phases of a whole respiratory cycle were reconstructed. The gating window (respiratory phase time when beam is on, usually around the end of exhale) was selected to restrict the tumor motion to < 5 mm. The GTV was delineated on one phase of them (usually the end of exhale), then propagated to all other phases using deformable registration and confirmed by physicians. The GTVs generated on each phase were combined to form the internal gross tumor volume (iGTV).

The clinical target volume (CTV) was defined as a 0.6–0.8 cm expansion from GTV/iGTV. Range uncertainties and set-up errors were counted when creating the planning target volume (PTV). In most instances, it was CTV plus a 0.5–0.7 cm margin in lateral and a 0.7–1.5 cm margin along beam direction.

The dose of PT is defined as equivalent dose to Gy of photon (GyE). Three radiation beam sources were used for this group of patients: PRT only, PRT + CIRT boost, or CIRT only. The selection of radiation sources or their combination depended on the clinical trials patients involved. For patients who did not participate in clinical trials, patients were treated according to our IRB approved institutional treatment protocols. From March 2016 ~ July 2017, 3 prospective phase I/II clinical trials approved by hospital IRB were started to enroll patients with stage I NSCLC. According to the distance from the primary lesion to critical organs at risk (OARs), patients were categorized to have peripheral, median, and central lesions. Overall, dose and fractions used included 5–7.5 GyE * 8–10 Fx, 4–5 GyE * 15–16 Fx and 2.25–3.5 GyE * 20–31 Fx.

At least 99% of GTV covered by 99% of the prescription dose, 99% of CTV covered by 95% of the prescription dose, and 90% of PTV covered by 90% of the prescription dose were required. All treatment plans aimed to meet stringent guidelines for protection of OARs, including lungs, spinal cord, esophagus, trachea, proximal bronchi, and heart/big vessels. The normal tissue dose-volume constraints for conventionally fractionated and ablative radiotherapy of the National Comprehensive Cancer Network guideline as well as published results from M.D. Anderson Cancer Center (MDACC) and the University of Florida Proton Therapy Institution were considered to set up the dose constraints for our patients according to fraction size [6‐10].

The Siemens Syngo® treatment planning system was used for all patients’ PT planning. PRT (beam energy 30~250 MeV) and CIRT (beam energy 85~430 MeV) plans were designed by using 2 to 4 beams with PBS technique. Single beam optimization (SBO) or intensity modulated particle therapy (IMPT) method was decided to be used in treatment by evaluating delivery uncertainties and doses to critical organs during planning. Without the feature of robustness evaluation in the Syngo® system, several steps were applied to reduce the uncertainties. First, similar to the experience of MDACC [5], the threshold of 5 mm was selected for motion control in our center. Based on the physical experiment on Alderson phantom conducted in our center, the deviation of dose distribution was found to be acceptable (< 5%) when respiratory movement was 4.4 mm or less. Second, iGTV of the plan was contoured to include all tumor trajectory in 10 phases, and the density of the intersection of iGTV and GTV will be overridden to a similar density of the primary lesion before planning. Third, when planning, physicists will avoid the beam angles passing through the heart and great vessels and keep the beam path less inhomogeneous. Forth, for hypo-fractionation plans (fraction dose ≥6 GyE), strategies of repeated SBO which is mimicking the volumetric rescanning [11], and specific beam PTV expansion were typically used to increse the plan robustness. IMPT was only used for lesions with minor motions, or large volumes and irregular shapes. Fifth, large sport size (full width at half maximum, FWHM = 10 mm) and decreased iso-energy slice spacing (Grid resolution = 2 mm) were used to increase target dose heterogeneities [12]. Sixth, 4D plan recalculation and accumulated dose also have been used for hypo-fractionation plans to evaluate dose distribution. Last but not least, before the first treatment and every week during treatment, simulation CT and plan recalculation on the latest CT will be conducted for every patient, to check the dose distribution, and replanning was demanded when recalculation revealed poor coverage of targets or overdosing to the OARs. Two orthogonal X-ray images were taken to verify and registrate patient position according to bone structures before daily irradiation. The set-up error of ≤3 mm before the treatment was allowed. Typical cases of CIRT are shown in Fig. 1.

The breath control techniques, beam types and doses of PT used are detailed in Table 2. Most patients (27/31) were treated with breath control techniques. Prescriptions of biological effective dose (BED, α/β = 10) < 100 GyE (using relatively low fraction dose) mostly prescribed to patients with N1 disease. The mean values of doses to thoracic OARs of all patients were listed in Table 3.

Chemotherapy was recommended to 16 patients with larger tumor (> 4 cm in diameter) and/or N1 disease; however, only 6 received some form of chemotherapy due to co-morbidities or patients’ refusal. Chemotherapy regimens and schedules were detailed in Table 2.