In the current study, we delivered PRT plus CIRT with PBS to 20 patients with primary esophageal SCC. With a median follow-up period of 25.0 months, 2-year OS and PFS rates of 69.2% and 57.4%, respectively, were achieved with mild RT-related side effects in a cohort of patients with stage II–III and IV (limited to patients with metastasis only to paraesophageal supraclavicular and cervical non-regional lymph-node) and a median age of 70 years. PRT and CIRT produced lower doses to the OARs than did photon IMRT, notably as Dmax to the spinal cord and Dmean to the lungs and heart. Dosimetric comparison indicated a rationale for PRT and CIRT in esophageal carcinoma. PRT plus CIRT with PBS appears safe and effective for esophageal carcinoma in a short-term observation.

Concurrent chemoradiotherapy was recommended for patients with locally advanced esophageal carcinoma who were not candidates for resection or could not tolerate surgery, in keeping with its preferred status in such patients [26]. Photon radiotherapy concurrently with chemotherapy has shown efficiency and tolerability compared to photon radiotherapy alone against locally advanced esophageal carcinoma, with 2- and 3-year OS rates of 36% and 30% after combined chemoradiotherapy, compared to 10% and 0% in the radiotherapy-only group, respectively [27]. An RT dose of 50–50.4 Gy was demonstrated to be highly acceptable for patients with locally advanced esophageal carcinoma [5]. The 2-year OS rates were typically 36–55% in randomized studies and the highest, reported recently, was 67% [6, 27,28,29].

In trials of neoadjuvant chemoradiotherapy, [30,31,32] a complete response rate of 30–40% after 40–50 Gy was noted in esophageal carcinoma, indicating that an irradiation dose of 40–50 Gy is sufficient for one-third of esophageal carcinoma patients. However, for the remainder, a higher dose might be indicated if the attendant toxicity was tolerable. A population-based, propensity-score–matched analysis suggested that a higher irradiation dose (≥ 60 Gy vs. 50–50.4 Gy) might improve survival in patients with esophageal SCC [33]. However, a randomized study showed that a total dose of 64.8 Gy did not improve clinical outcomes compared with 50.4 Gy in locally advanced esophageal carcinoma and showed an extended treatment duration due to toxicity breaks and a decreased actual dose of fluorouracil [6]. Furthermore, a recent randomized study (ARTDECO) reported similar OS and LPFS rates between RT doses of 50.4 and 61.6 Gy [28]. Analysis of patients with esophageal carcinoma receiving PRT or IMRT showed that PRT and a Dmean of < 15 Gy to the heart are associated with a decreased incidence of severe (grade 3 or higher) cardiac events, and these events are associated with poorer OS (p < 0.05) [34]. These findings indicate that technical improvements in highly conformal therapy can be expected to improve clinical results in locally advanced esophageal carcinoma if they lead to an increased dose to targets without increased radiation exposure to the OARs (especially the heart).

Similar to our results, the literature also shows that protons and carbon ions can better protect normal tissues than can photons, specifically 3DCRT or IMRT, whether they are passive scattering protons or PBS beams. Compared to IMRT, PRT achieved a significant decrease with the passive scattering technique, especially a decline in the Dmax of the spinal cord, Dmean, V25, V30, V40, and V50 of the heart and Dmean, V5, V10, and V15 of the lungs (Table 4) [9]. CIRT with PBS (i.e. IMCT technique, similar to the technique used in our center) produced better dose homogeneity in the target volume and significantly lower doses to the heart, lungs, spinal cord, and skin in esophageal carcinoma than did photon 3DCRT or volumetric modulated arc therapy (VMAT, a variant of IMRT). Compared to VMAT, significant decreases were achieved with CIRT, especially in the Dmax to the spinal cord; Dmean, V10, V20, V30, and V40 to the heart; and Dmean, V5, V10, V20, V30, and V40 to the lungs [35].

To our knowledge, this is the first study to compare dosimetric parameters among IMPT, IMCT, and photon IMRT in esophageal carcinoma. Similar to Ling et al.’s study using the scattering technique, [9] in our study, PRT using the scanning technique produced a more apparent decline in V20 of the lungs than did IMRT. Compared with PRT using passive scattering, [10] the more advanced IMPT with PBS is associated with further dosimetric benefits in sparing the heart in radiotherapy for esophageal cancer and has also been reported as sparing the lungs and heart in radiotherapy for lung cancer [36]. CIRT using the scanning technique here produced results for esophageal cancer similar to those reported by Takakusagi et al., [35] which included a similar decline in the Dmax of the spinal cord, a decrease in the Dmean and V5–V60 of the heart, and a lesser decline in the Dmean and V5–V20 of the lungs. Those authors used the same PTV for photon and carbon-ion planning; in contrast, in this study, a larger PTV (beam-specific PTV) was applied as in routine clinical practice of particle therapy. Although a trend of better dose distribution to OARs was noted in PRT compared to CIRT planning, the differences were minor in this study.

By taking advantage of dose distribution, PRT led to either a lesser treatment-related toxicity burden than photon radiotherapy or an improvement in survivals for esophageal cancer [10, 11]. Xi et al. conducted a retrospective study of PRT and photon IMRT with a total dose of 50.4 Gy/28Fx concurrently with chemotherapy of fluorouracil and platinum/taxane in esophageal carcinoma [10]. PRT led to improved survival, with higher 5-year OS and PFS rates of 41.6% and 34.9% compared with photon radiotherapy rates of 31.6% and 20.4%, respectively. PRT also induced low severe toxicities compared with photon radiotherapy, with treatment-related severe toxicity of 39% vs. 47% (p > 0.05). Furthermore, a prospective phase-IIb randomized study of PRT mostly using a passive scattering technique vs. photon IMRT with a total dose of 50.4 Gy/28Fx concurrently with fluorouracil/taxane-based chemotherapy showed that PRT maintained similar OS and PFS rates (3-year OS, 51.2% vs. 50.8%; 3-year PFS, 44.5% vs. 44.5%). However, PRT was associated with a lower total toxicity burden [11]. Failure to improve clinical outcomes may be related to the passive scattering technique initially used and a lack of clinician experience with the new technology, similar to the lesson learnt with lung cancer. Compared to PRT, our study on IMPT followed by IMCT with a median total dose of 71 Gy (range: 65–73.1) in esophageal carcinoma produced a promising clinical outcome (2-year OS, 69.2% and 2-year PFS, 57.4%) with mild esophageal toxicities (5%). Severe esophageal and pulmonary toxicity rates of more than 20% were observed after photon RT and PRT with passive scattering in early patients, which included a severe late esophageal toxicity rate of 19–21% after photon RT and severe esophagitis and esophageal stricture rates of 11.4% and 9.8% after proton RT, respectively [6, 10, 27]. The clinical benefit of this treatment modality may be attributed to the dosimetric advantages of IMPT and IMCT when combined with PBS, which can further reduce the dose to normal tissue compared with passive scattering technique, thereby further reducing doses to the heart and thus the incidence of adverse events.

This study had some limitations. First, a definitive conclusion could not be drawn owing to the small sample size and the retrospective nature of the study. Second, our study delivered a wide range of dose prescriptions, attempting to balance toxicity and clinical efficacy. Finally, this study included patients with stage IV esophageal cancer (with metastasis only to paraesophageal non-regional lymph nodes, defined in AJCC, 8th edition), which hampered a definitive conclusion for locally advanced esophageal cancer. Currently, a large-scale prospective study to evaluate the efficacy and safety of PRT plus CIRT in locally advanced esophageal carcinoma is underway. We expect that the results will validate our findings.