Proton therapy has a completely different dose distribution compared with conventional photon beams. Unlike X-ray irradiation, the energy during proton therapy is deposited with depth and produces a maximum peak close to the end of the range
[8]. The maximum peak is well known as the “Bragg peak”, which may be used for dose increment for cancer therapy while reducing the radiation dose to the normal tissue
[10,11][10][11]. Indeed, published dosimetry studies have indicated that proton therapy significantly reduces the dose to normal structures, especially in relation to the lung, heart, and esophagus, when maintaining similar robust target volume coverage to the clinical target volume (CTV) in both early and advanced NSCLC compared with photon therapy. Currently, passive scattered proton therapy (PSPT) and active pencil beam scanning (PBS) are the two forms of proton therapy in use
[12]. The former form uses one or two levels of scatterer to widen the proton beam enough in order to cover the target, while the latter form uses magnets to deflect the proton beams directly, rather than a scatterer. The majority of comparative studies about dosimetry included patients with advanced NSCLC. Studies on the impacts of proton therapy on early-stage cancers were limited, as listed in
Table 1. Those that do exist were mainly conducted in a retrospective manner, and include only two prospective studies
[13,14,15,16,17,18,19,20,21,22,23,24,25,26,27][13][14][15][16][17][18][19][20][21][22][23][24][25][26][27].
2.1. PSPT
Among the limited studies using proton therapy for early-stage NSCLC, PSPT has favorable CTV coverage and distributes lower mean doses to the normal tissues, compared with photon therapy. As reported by Wink et al.
[15] in a retrospective study including 25 patients, CTV doses were more homogenous, and the dose directed to the spinal cord was lowest with PSPT, compared with IMRT, VMAT, and CyberKnife. Wang et al.
[13] reported that in 24 patients with stage I NSCLC, the 95% isodose line of PSPT covered more CTV than that of 3D-CRT (86.4% versus 43.2%), and the mean dose to lung, heart, esophagus, and spinal cord was also lower, as well as V
5Gy and V
20Gy to the lungs. The two studies mentioned above were focused on early-stage patients undergoing a hypo-fractionated radiation therapy regimen (60–66 Gy in 8–10 fractions).
For locally advanced NSCLC, PSPT also reduces the dose to the critical normal tissues and prevent lower-dose target coverage. One of the only two prospective studies indicated that PSPT could keep the dose to the target at 70 Gy for patients with stage IA–IIIB NSCLC, while sparing the lung, compared with 3D-CRT/IMRT (mean lung dose, 13.5 Gy versus 18.9 Gy/16.4 Gy)
[17]. The second prospective study was a phase III trial, reported by Giaddui T et al. comparing the dose parameters for 26 lung IMRT, with 26 proton PSPT plans. As a result, the dose parameters for the IMRT and PSPT plans were very close. However, the PSPT plans led to lower dose values for normal structures (including lung V
5Gy, 34.4% versus 47.2%; maximum spinal cord dose, 31.7 Gy versus 43.5 Gy; heart V
5Gy, 19% versus 47%; and heart V
30Gy, 11% versus 19%)
[23]. The dosimetry comparative studies of PSPT for advanced-stage patients were mostly using conventional regimens (66–74 Gy in 33–37 fractions).
However, two respective comparative studies revealed similar or worse dose distribution to the lung or esophagus for PSPT. Wu et al.
[22] reported that in 33 patients with stage III NSCLC, all of the dose parameters of proton therapy were lower than 3D-CRT, except for the esophageal dose, which was slightly higher than that of the photon plan (V
50Gy, 20.2 versus 16.6%), but the difference was not significant. Another study by Shusharina et al.
[24] with 83 patients (II-IV stage NSCLC), reported that, although higher lung V
5Gy was observed for IMRT, whereas higher V
60Gy for was observed for PSPT, the mean lung dose was similar. However, these two studies were both retrospectives and may have been prone to selection bias.
2.2. PBS
PBS may have advantages compared with PSPT in terms of offering greater dose conformality
[28]. The entry dose of PSPT is often unmodulated, even after using the layer-stacking method
[5]. Meanwhile, the movement of the target during PSPT causes dose distribution disturbances due to interplay and blurring effects, which leads to dose misses and unwanted doses to healthy organs. PBS generates more conformal high-dose volumes than PSPT, with significant sparing of nearby organs, and intensity-modulated proton therapy (IMPT) can be comprehended
[29]. Gjyshi et al.
[30] compared two independent cohorts with locally advanced NSCLC (86 received PSPT and 53 received IMPT) with data extracted from a prospective registry study, and found that lower mean radiation doses to the lungs (16.0 Gy versus 13.0 Gy,
p < 0.001), heart (10.7 Gy versus 6.6 Gy,
p = 0.004), and esophagus (27.4 Gy versus 21.8 Gy,
p = 0.005) resulted in lower rates of pulmonary (28% versus 3%,
p = 0.006) and cardiac (14% versus 0%,
p = 0.05) toxicities for IMPT.
IMPT is also sensitive to uncertainties or target motion. Four-dimensional (4D)-computed tomography (CT) ventilation imaging-guided proton therapy, based on breathing patterns, may be helpful for reducing uncertainties and dosing to the normal tissues
[31,32,33][31][32][33]. IMPT via a deep-inspiration breath-hold, deformable image registration with daily adaptive proton therapy, and liver-ultrasound-based motion modeling may also provide additional benefits
[34,35,36,37][34][35][36][37]. FLASH proton therapy which optimizes tissue-receiving dose rate distribution and dose distribution may also provide substantial improvements, compared to IMPT, for normal tissue sparing
[38].
As displayed in
Table 1, published dosimetry comparative studies with proton and photon therapy for IMPT were all retrospective studies with <30 cases. The only study for early-stage NSCLC (15 patients with centrally/superiorly located stage I NSCLC) was reported by Register et al., which revealed that IMPT and PSPT significantly reduced doses to the surrounding normal tissues while maintaining a high radiation dose focused on the tumor, compared with SBRT (total lung volume receiving 5 Gy, 10 Gy, and 20 Gy, respectively)
[14]. The rest of the dosimetry studies included patients with stage III NSCLC, and consistent results were observed for IMPT with comparable, if not better, CTV dose homogeneity/coverage while sparing the lung, heart, spinal cord, and esophagus to a greater extent. In addition, IMPT allowed for further dose escalation, compared with photon therapy
[16]. Zhang X et al. reported that IMPT might allow further dose escalation (a mean maximum tolerated dose to 83.5 Gy or 84.4 Gy) and prevent lower-dose target coverage for the treatment of stage IIIB NSCLC, while sparing more lung, heart, spinal cord, and esophagus, compared with IMRT, and with similar normal tissue sparing compared with PSPT
[16]. Therefore, PBS, which is gradually replacing PSPT in the clinical practice of proton therapy, may potentially overcome the limitations of PSPT and reduce treatment-related toxicity.
Notably, some studies reported special characters for proton, compared with photon therapy. Palma G et al.
[39] reported that in 178 patients with advanced NSCLC who were treated with PSPT/IMRT (66/74 Gy, conventional fractionation) with concurrent chemotherapy, significant dose differences of the heart and the lower lungs was found in the 40 patients who developed clinically symptomatic pneumonitis, compared with those without pneumonitis, which may substantiate potential factors in the development of pneumonitis. Harris et al.
[40] retrospectively reported that in 160 (78 photons, 82 protons) patients with locally advanced NSCLC who were treated with chemoradiotherapy, among them, 40 (20 photons, 20 protons) patients exhibited grade ≥2 pneumonitis. After multivariate analysis, V
40Gy turns out to be statistically significant for proton and a potential pneumonitis predictor is V
40Gy ≤ 23%, and not V
20Gy or Dmean which are traditionally used in photon therapy. However, the dose-response of proton therapy for normal tissue complications has been validated as similar to that of photon therapy, based on a pneumonitis model
[41]. Xiang et al.
[42] identified 450,373 pediatric and adult patients with cancers (33.5% with 3D-CRT, 65.2% with IMRT, and 1.3% received proton therapy) from the National Cancer Database, and during a median follow-up of 5.1 years, the rate of diagnosed secondary cancer was 1.55% per year, suggesting that proton therapy was associated with lower risk of secondary cancer compared with IMRT (adjusted odds ratio 0.31,
p < 0.0001). Further study with a long follow-up duration is needed.