Reirradiation With Proton Therapy Using Hypofractionated Schedules: A Systematic Review of the Literature

Abstract

Aim

Re-irradiation (re-RT) with hypofractionated proton therapy (PT) may be a treatment option in selected cases. Aim is to provide a critical overview of the state of the art.

Methods

PRISMA-compliant systematic review was performed on 2nd October 2025. All tumor sites were considered, provided that re-RT was performed with dose/fraction>2Gy(RBE) in adults. Oncological and toxicity outcomes were retrieved.

Results

Twelve original research works were identified. The total number of patients treated with hypofractionated PT re-RT was 478. The majority of patients were treated for recurrent tumors of either the gastrointestinal (GI) region (n = 143) or the lung (n = 123). The majority of the published series refers to data derived from retrospective and monoinstitutional experiences. Most patients were treated for recurrent tumors of either the gastrointestinal (GI) region (n = 143) or the lung (n = 123). Median sample size was 32 (interquartile range: 26-46, range: 14-83). Moderate hypofractionation (< 5 Gy(RBE)/fraction) was the most frequent schedule. All included studies had a radical intent. Overall, 2 year overall survival ranged from 8.4% (head & neck) to 87.5% (liver). The rate of severe side effects remained relevant. Of note, clinical outcomes and toxicity profiles were highly heterogeneous among the different tumor sites. Therefore, in the setting of re-RT with PT, conclusive data on the efficacy and toxicity profile of hypofractionated schedules could not be provided yet. Re-RT remains a personalized approach in almost all settings of tumor recurrence.

Conclusion

Evidence supporting the use of hypofractionated PT re-RT remains limited but encouraging. Identifying predictors of toxicity and treatment response is essential to guide future research and to establish this approach as a viable option for selected patients.

Introduction

In recent years, high-precision photon-based irradiation techniques such as stereotactic body radiation therapy have enabled the adoption of hypofractionated schedules (dose per fraction >2 Gy(RBE)) across multiple oncologic indications, including prostate, lung, and breast cancers.1, 2, 3, 4 Compared with conventional 2 Gy-per-fraction regimens, hypofractionation offers several advantages, primarily related to the delivery of a higher radiobiological dose (particularly in ultra-hypofractionated schedules) and a shorter overall treatment time. The former may contribute to overcoming tumor radioresistance, whereas the latter reduces patient inconvenience, optimizes healthcare resources, and facilitates integration with systemic therapies.⁵ Consequently, the use of hypofractionated regimens has progressively expanded, including in the management of recurrent tumors.6, 7

In the re-irradiation (re-RT) setting, however, the main challenge remains treatment-related toxicity. To mitigate the risk of late adverse effects due to cumulative dose exposure in normal tissues adjacent to the recurrence, standard or hyperfractionated schedules have traditionally been preferred.8, 9 Therefore, high-dose-per-fraction treatments are typically reserved for small recurrences located at a safe distance from critical structures.

Proton therapy (PT) has emerged as a promising external beam radiation modality. Its distinctive physical property—the Bragg peak—allows for a steep dose fall-off beyond the target volume, thereby substantially reducing radiation exposure to surrounding healthy tissues. Owing to this favorable dose distribution and the growing global availability of proton facilities, PT is increasingly being incorporated into the treatment of various malignancies. Similar to photon therapy, hypofractionated PT has been explored in several tumor types, including lung, liver, and prostate cancers, with encouraging toxicity profiles also reported for breast cancer and benign intracranial tumors.10, 11, 12, 13, 14 Given its dosimetric advantages, re-RT represents a key area of interest for hypofractionated PT,15, 16 however, the feasibility and safety of hypofractionated PT re-RT remain poorly defined. Accordingly, this systematic review aims to critically evaluate current evidence on clinical outcomes and toxicity of hypofractionated PT in the re-RT setting across tumor sites.

Materials and methods

The applied methodology was compliant with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) recommendations.¹⁷ Full-texts on re-RT with PT in adult patients were considered eligible whenever the delivered dose/fraction was greater than 2 Gy(RBE)/fraction. All tumor sites of origin (namely, brain, head and neck, lung, breast, upper abdomen, and male/female pelvis) were included and analyzed. PubMed (NCBI, National Center for Biotechnology Information), EMBASE (Elsevier), and Scopus (Elsevier) databases were interrogated on 2nd October 2025 with no date restriction by an author with experience in bibliographic searching (SV). Review articles were considered for cross-referencing, as well. Study screening and data extraction were independently performed by 2 authors (V.P. and S.V.). In cases of disagreement, a third author (D.A.) was consulted to reach a consensus decision. Benign tumors, non-PT treatments (i.e., carbon ions, photons), pre-clinical, in silico, and dosimetric analyses were excluded. Studies describing mixed cohorts (protons + other re-RT techniques, or different fractionation schedules, standard and hypofractionated) were also included. In these cases, when feasible, data related to hypofractionated PT patients only were reported. Case reports and case series (sample size ≤ 5 patients) were not considered. The search strategy is provided in the Supplementary Materials. The workflow of article selection is shown in Figure 1.

Figure 1.

PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) diagram for the included studies.

Results

A total of 16 works were retrieved. The most frequently reported cancers were of head and neck origin (n = 518, 19, 20, 21, 22); 3 studies described primary cancers of lung,23, 24, 25 gastro-intestinal,26, 27, 28 and the eye,29, 30, 31 while 2 focused on central nervous system tumors.32, 33 Of these 16 works, 3 studies included a mixed cohort of patients treated with different re-RT technique, and proton-specific outcomes could not be separately extracted.19, 20, 22 Among these, only 1 studies was considered, as patients treated with protons were more than 50%.¹⁹ Finally, 12 non-overlapping patients’ series were considered within the scope of this work. The most relevant characteristics of the selected papers are reported in Table 1. Out of the 12 studies, 9 reported data exclusively on patients treated with PT; 2 included mixed cohorts (patients treated with PT and other re-RT techniques) but allowed extraction of data specifically related to proton-treated patients; and 1 included mixed cohorts in which such extraction was not possible. The majority of the included series had applied moderate hypofractionation. A summary of the total number of patients considered for each sub-site and the prescribed doses is shown in Figure 2.

Table 1.

Summary of included studies.

Author, year	Type (Retrospective/prospective), location, years of enrollment	Site and total number of patients	Previous RT technique	Median time (months) (range) interval between previous course and reRT	ReRT technique (passive scattering vs active scanning)	Median tumor volume (cc)	ReRT Total dose	ReRT Dose/fraction	Site of recurrence (In field/out of field)	Concurrent treatments	Clinical outcome	Median follow-up (months)	grade (G) > 3 reported toxicities Acute/Late
Gordon et al, 2021 18	R Russia 2015-2020	HN 30	Conformal (70%), Non conformal 30%	38	Active	NR	Median 55 Gy EqD2 = 57.6 Gy (range 42.1-68.0)	2 Gy(RBE) 2.4 Gy(RBE) 3 Gy(RBE)	NR	NR	2year LC 21% 2year OS 8.4% Median OS 16 months	21	Acute: 3.3% G3 mucositis Late: One patient died for carotid bleeding 16.6% late G3 (necrosis, chewing trismus)
Badiyan et al, 2019 23	P USA 2010-2016	Lung 79	Conformal RT (32% SBRT)	19.9 (2.2-265.2)	5% Active	NR	EqD2 = 60 Gy(RBE) (range 40-74 Gy) EqD2=62.7 Gy(RBE) (range 60-65.4 Gy) EqD2= 60.4 Gy(RBE) (range 32.5-89.3 Gy)	1.8-2 Gy(RBE) 1.2 Gy(RBE) 2.1-7 Gy(RBE)	least partial overlap of prior	30%	5year PFS 8% Median OS 15.2 months	10.7	Acute: 6% G3, pneumonia, syncope, cellulitis, neck pain Late: 1% G3 (fatigue) 2 patients died of pulmonary artery hemorrhage
Yang et al, 2022 24a	R Korea 2011-2019	Lung 22	8 (36.4%)3D 8 (36.4%) IMRT 1 (4.5%) SBRT 2 (9.1%) 3DPT 3 (13.6%) IMPT	14.0 (5.8-28.9)	14% Passive 86% Active	NR	60-70 Gy(RBE) 50-66 Gy(RBE)	2-2.2 Gy(RBE) 2-8 GyREBE	NR	27.3%	Not specified for the proton cohort	19.7	Acute/Late 18.2% G3 toxicity
Janopaul-Naylor et al, 2023 25	R USA 2019-2021	Lung 22	stereotactic ablative radiotherapy (n = 2) or chemoradiation (n = 20)	28 (9-173)	Active	NR	EqD2=60 Gy(RBE) (range 44-60 Gy)	2-4 Gy(RBE)	NR	36%	2year OS 58% 2year PFS 48%	7	Acute/Late 9% bronchial necrosis and skin desquamation with fibrosis
Murayama et al, 2023 28a	P Japan 2016-2019	GI (Colon-rectum) 33	NR	NR	NR	37.7	60-70 Gy(RBE) 72-75 Gy(RBE)	2 Gy(RBE) 4-3 Gy(RBE)	NR	NR	2year OS 86.5% 3year OS 62.6%	34	Acute/Late 9% G3, neuropathy, Abscess, Rectal ulcer, Cystitis, Skin ulcer 3% G4, rectal perforation
Marucci L et al, 2006–2011 29, 30	R USA 2000	Eye 31	Proton therapy	36 (8-165)	NR	1.46	70 in 5 fractions to 90%, 50 Gy in 5 fractions in 10%	14 Gy CGE	Recurrence in field (23%), border (61%), through the sclera in 2 patients (6%), out of field (13%).	no	5year OS 64% 5year LR 31%	50	Late: 12.9% enucleated eye for intractable pain 45% vision deterioration 12.9% neovascular glaucoma
Riechardt et al, 2014 31	R Germany 2000-2010	Eye 48	43.7% brachytherapy. 39.6% Transupupillar thermotherapy. 8.3% Proton 4.2% multiple laser coagulations, 2.1% with CyberKnife 2.1% photodynamic therapy.	16.4 (5.8-76.8)	NR	0.17	60 Gy CGE	15 Gy CGE	Central margin 16 pz, diffuse growth 4, 1 periferal margin	no	5year OS 64% 10year LC 92.1%	81	Late: 4% enucleation for local recurrence 12.5% Vitrectomy
Gulidov et al, 2021 32	R Russia 2016-2019	CNS 44	75% conformal RT (3D/IMRT) 25 non conformal RT	28 (12-73)	NR	62.7	EqD2= 55 Gy(RBE) (range 46-61.75 Gy)	2-3 Gy(RBE)	In field recurrence	no	2year OS 35% 2year PFS 10%	16	Acute/Late: 6.8% symptomatic radionecrosis
Saeed et al, 2020 33	P USA 2012-2018	CNS 45	Photon RT (technique not specified)	20 (3-77)	56% Active44% Passive	NS	46.2 Gy(RBE) (range 24.9-60.8 Gy)	2.2 Gy(RBE) (range 1.5-4 Gy)	NR	Temozolomide alone 36% Bevacizumab alone 9% Temozolomide+ bevacizumab 22% Vorinostat bevacizumab 2%	2year OS 16% 2year PFS 14%	10.7	Acute/Late 2% G3, ataxia 8.8% G3, neuropathy, cognitive disturbance, optic nerve disorder, seizure
Hashimoto et al, 2006 26	R Japan 1989-2000	GI (HCC) 27	Proton Therapy	24.5 (3.3-79.8)	NS	3	66 Gy(RBE)	4.12 Gy(RBE)	Overlapped volume 81%	Only Protontherapy 11% 89% with mixed local treatments (radiotherapy or other)	5year OS 55.6% 5year LC 87.8%	62	Acute: 14.8% G4 (hyperbilirubinemia, hepatic coma) Late: 11.1% G >3 (rib fracture, infectious biloma, obstructive cholangitis)
Oshiro et al, 2017 27	R Japan 2002-2010	GI (HCC) 83	Proton therapy	NR	NS	28.6	66 Gy(RBE) 72.6 Gy(RBE) 74 Gy(RBE)	6.6 Gy(RBE) 3.3 Gy(RBE) 2.0 Gy(RBE)	NR	NR	2year OS 87.5% 5year OS 49.4%	45	No patients experienced a severe long-term toxicity
Guan et al, 2019 19 $	P China 2014-2018	HN 14	15.4% photon-based RT (gamma-knife = 7, cyber-knife = 2, or conventional RT= 5) 84.6 PT	NR	6 (42.8%) Protons, 8 (57.2%) Carbon ions	NR	57-69 Gy(RBE)	2-3 Gy(RBE)	NR	NR	2year OS 50.3% 2year PFS 41.3%	28	No patients experienced severe long-term toxicity

^§The study included a mixed cohort of patients treated with different re-RT techniques, and proton-specific outcomes could not be separately extracted. However, patients treated with protons were more than 50% of the cohort.

Abbreviations: 3D, three-dimensional radiotherapy; CGE, cobalt gray equivalent; CNS, central nervous system; EqD2, equivalent dose in 2 Gy fractions; GI, gastrointestinal; G, grade; HCC, hepatocellular carcinoma; HN, head and neck; IMPT, intensity-modulated proton therapy; IMRT, intensity-modulated radiation therapy; LC, local control; LR, local recurrence; NR, not reported; NS, not specified; OS, overall survival; P, prospective; PBT, proton beam therapy; PFS, progression-free survival; PT, proton therapy; R, retrospective; re-RT, re-irradiation; RT, radiotherapy; SBRT, stereotactic body radiotherapy.

^aThe study included a mixed cohort of patients treated with different re-RT techniques; however, it was possible to extract and analyze the results specifically related to patients treated with protons. The reported number of patients refers only to patients who underwent proton re-RT.

Figure 2.

Summary of the total number of patients considered for each subsite. List of abbreviations: CNS (Central nervous system); GI (gastrointestinal); HN (Head & Neck); LC (Local control).

The total number of patients treated with hypofractionated PT re-RT was 478. The majority of patients were treated for recurrent tumors of either the gastrointestinal (GI) region (n = 143) or the lung (n = 123). Median sample size was 32 (interquartile range: 26-46, range: 14-83). Moderate hypofractionation (< 5 Gy(RBE)/fraction) was the most frequent schedule. Generally, the RBE value of 1.1 for protons is used in clinical practice. However, none of the included works specified the RBE value in the analyzed cohort. The widest range of fractionation was found among patients treated for lung tumors (from 2.1 to 18 Gy(RBE)/fraction). All included studies had a radical intent and reported both on clinical outcomes and toxicity profiles. A more detailed description of the analyzed papers is reported in the following sections.

Head and neck

Two works reported results of hypofractionated re-RT in patients with recurrent tumors in the head and neck region. Gordon et al¹⁸ reported results of 30 patients treated with PT for recurrent head and neck cancers. The majority (76%) of patients had squamous cell histology, while approximately one-third (33%) had a nasopharyngeal recurrent tumor. The median total volume was 158 cc (range 13-280 cc). A median EqD₂ total dose of 57.6 Gy(RBE) was administered, and 60% of patients received a dose/fraction of 3 Gy(RBE). The median OS was 16 months, with a 2-year OS rate of 8%. Skull base localization was associated with worse clinical outcomes. Five patients experienced severe toxicity, and 1 patient (3%) died due to carotid bleeding.

The study by Guan et al¹⁹ described a cohort of 14 patients treated for recurrent chordoma and chondrosarcoma. Out of the 14 patients, 8 received hypofractionated re-RT with protons up to a total equivalent dose of 57-69 Gy(RBE) in 19-23 fractions. The 2-year OS rate was about 40%. No patients experienced severe long-term toxicity.

Lung

The most robust literature data were provided by Badiyan et al²³ who described results of data collected from 2010 to 2016 in 2 prospective registry trials (Proton Collaborative Group—a clinical research consortium of 12 PT facilities in the United States—and the University of Florida Proton Therapy Institute Registry trial). One hundred and fifty-eight patients, of whom 92% of non-small cell lung cancers histology, were analyzed. Among them, 79/158 received reRT. Different fractionation schedules were adopted, ranging from conventional to hypofractionated regimens, with median cumulative EqD2 doses around 60 Gy(RBE). Clinical outcomes remained modest, with a 5-year progression-free survival of 8% and a median overall survival of 15.2 months after a median follow-up of 10.7 months. With the limitation that results of the hypofractionated cohort were not reported individually, the authors report that 6% of patients experienced severe toxicity. Of these, 2 and 1 treated respectively with a standard and hypofractionated schedule died of massive pulmonary artery hemorrhage, probably due to treatment-related toxicity. The cumulative dose (EqD₂) of the patient treated with hypofractionated re-RT was 207 GyEQD2 (α/β = 3 Gy). Of note, only 4 patients were treated with the pencil beam technique, and dosimetric data of the previous radiation course were not provided.

Yang et al²⁴ compared 41 patients treated with photon-based RT with 22 patients treated with PT for recurrent lung cancer. Different fractionation schedules were used across the whole cohort (ranging from 60 to 70 Gy(RBE) administered with 2-2.2 Gy(RBE)/fraction to 50-66 Gy(RBE) with 2-8 Gy(RBE)/fraction). The authors found that histology (squamous cell carcinoma) and pulmonary function (Diffusing Capacity Of The Lungs For Carbon Monoxide < 60%) were the only factors associated with OS. Treatment technique (photon vs proton) did not impact either clinical outcomes or toxicity, nonetheless, patients who received a smaller number of fractions experienced a high rate of lung and esophageal complications (P = .002 and P = .017, respectively). Tumor location (ultracentral lesions) and concurrent chemotherapy were associated with a higher risk of severe side effects on the esophagus.

Twenty-two patients treated with IMPT re-RT were reviewed at the Department of Radiation Oncology of the Winship Cancer Institute (Atlanta, Georgia, USA).²⁵ The hypofractionated schedule (15 fractions) was used for about one-third of them. Results were not analyzed according to the treatment schedule, but the authors stated that no significant association was found between a higher number of fractions (> 20) and the toxicity profile.

Colon-rectum

Data from 33 patients treated with PT for recurrent rectal cancer between 2016 and 2018 and collected in the multicentric Japanese Proton-Net Registry were presented by Murayama et al.²⁸ Patients were treated with either hypofractionation (72-75 Gy(RBE) in 18-25 fractions) or standard fractionation (60-70 Gy(RBE) in 30-35 fractions). The schedule was chosen based on tumor proximity to the GI tract. The median follow-up was 34.0 months (range: 2.1-53.7), and the estimated 2-year and 3-year OS was 86.5% and 62.6%, respectively. Details on the eventual previous RT history were not reported. Severe toxicity occurred in 9% of patients, including neuropathy (3 cases), abscess (2 cases), rectal ulcer, cystitis, and skin ulcer (1 case), and 1 Grade 4 rectal perforation.

Eye

PT may be an alternative to enucleation for patients with recurrent uveal melanoma. One American²⁹ and one European³¹ series reported results on the use of PT to treat recurrences after a first treatment course performed with either PT²⁹ or other modalities, including brachytherapy and transpupillary thermotherapy.³¹ The 2 cohorts achieved different results both in terms of local control and eye preservation rate.

Indeed, Marucci et al²⁹ reported results of 30 patients treated with a second course and 1 patient with a third course of PT performed at the Massachusetts General Hospital (Boston, Massachusetts, USA). All except 1 patient (who received 48 Cobalt Gray Equivalent (CGE)) received 70 Gy(RBE) in 5 fractions (14 CGE/fraction) after a mean interval of 50.2 months from the first PT treatment administered with the same schedule (cumulative total dose ranging from 118 to 210 CGE). After a median follow-up of 36 months, 5 patients (16%) experienced tumor regrowth. Four patients developed glaucoma and were treated conservatively. Of the 12 patients who had received a cumulative dose of 140 Gy(RBE), 2 lost their eyes due to intractable pain. The overall 5-year cumulative rate of local recurrence after the second treatment was 31% (95% CI, 11.4-68.1). Eye preservation rate was achieved by about two-third of the patients. Eye enucleation was due to either local recurrence (5 patients) or side effects (4 patients). The ciliary body involvement and tumor diameter >15 mm were found in 42% and 58% of the recurrent patients, respectively. This cohort was then compared to a similar cohort of patients treated with enucleation, and results were reported in a subsequent paper.³⁰ This analysis suggested that treating local recurrences with PT did not affect patients’ OS.

Riechardt et al³¹ reported the results of a series of 48 patients treated at the PT facility of Charite University Hospital Berlin for recurrent uveal melanoma. In this series, the first treatment was performed with brachytherapy (21 patients), transpupillary thermotherapy (19 patients), cyberknife (1 patient), PT (4 patients), Laser coagulation (2 patients), or photodynamic therapy (1 patient). A total dose of 60 CGE on 4 consecutive days was administered. Three patients (6%) experienced a tumor relapse. Local tumor control was achieved in 92.1% of patients 10 years after PT. The 10-year estimation for eye preservation was 97.7%.

Central nervous system

Primary tumor recurrences are frequent events in the central nervous system, mainly for high-grade gliomas. Gulidov et al³² reported the results of 44 patients treated with a second course of radiation therapy (RT) for primary brain tumor recurrences. The majority of them (63.7%) had recurrent glioblastoma. The median tumor volume was 62.7 cm³. One-fourth of the patients were treated with 3 Gy(RBE)/fraction up to a median equivalent dose in 2 Gy fractions (EqD2) (α/β = 10 Gy) of 55 Gy(RBE) with an accelerated schedule used for patients with a low-performance status (<70) and small tumor volume (gross tumor volume, GTV < 50 cm³). Half of the patients experienced a local recurrence after re-RT. Symptomatic radionecrosis was reported in 3 patients (6.8%), while 6 patients experienced severe chemotherapy-related toxicity. The median survival time was 12 months, and the 3-year OS rate was 13%. Tumor grade was found to be the only prognostic factor associated with OS.

The Proton Collaborative Group (PCG) reported results of 45 patients with recurrent glioblastoma treated with PT coming in 7 facilities.³³ The first course of RT was performed by photons up to a median total dose of 60 Gy. The median Re-RT dose was 46.2 Gy(RBE) with a dose per fraction ranging from 1.2 to 4 Gy(RBE), with the majority of patients (70%) treated with a hypofractionated schedule. The median OS was 14.2 months (95% CI 9.6-20 months). Grade 3 toxicity was recorded for 5 (11%) patients, and in all patients, PT re-RT exceeded the EqD2 of 41 Gy. Grade 4 and 5 toxicities were not observed. Improved OS was associated with surgery prior to treatment (HR 0.087; 95% CI 0.02-0.42, P < .01).

Liver

The University of Tsukuba (Tsukuba, Japan) reported their experience of 2 consecutive series of patients treated for hepatocellular carcinoma (HCC).26, 27 In the first one,²⁶ 27 patients received a median total dose of 66 Gy(RBE) in 16 fractions to a total of 68 lesions. The median target volume was 30 mm³ (range 8-50 mm³). In 23 patients, PT was associated with other local treatments like transarterial chemoembolization, percutaneous ethanol injection, transcatheter arterial chemoinfusion, and radiofrequency ablation. The non-irradiated volume was defined as the liver parenchyma receiving less than 10% of the prescribed dose. Central and peripheral areas of the liver were defined considering the second branch of the portal vein. Overlap of clinical target volumes (CTVs) with the ones of previously delivered plans was found in 22 (81%) patients, of whom 15 were in the peripheral zone and 7 in the central zone. In the majority (70%) of patients, the Child-Pugh class was maintained during their multiple treatment courses. The 5-year local control rate was 87.8%. After a median follow-up of 62 months (range 9-148 months from the beginning of the first course of irradiation), 4 patients were still alive. Patients who underwent a second treatment course showed a lower OS compared to those who underwent their first treatment course (survival rate at 5 years: 25.9% vs 55.6%, respectively). Severe toxicity was recorded in 5 patients (19%), and among them, 2 with Child-Pugh scores of B and C, respectively, experienced acute hepatic failure. Late injuries were observed in 2/7 patients who had received a second PT course in the central zone of the liver. Therefore, the authors concluded that a safe treatment could be proposed for patients with peripheral lesions and Child-Pugh A liver function. In the case of a tumor located in the peripheral region but with a Child-Pugh B function, the treatment could be considered safe only when the amount of non-irradiated liver parenchyma does not exceed 500 ml.

The other series from the University of Tsukuba²⁷ included 83 patients treated with a minimum dose of 50 Gy(RBE). Of these, 26% had been diagnosed with an in-field recurrence. The following doses were employed: 74 Gy(RBE) in 37 fractions (2.0 Gy(RBE)/fraction) for tumors adjacent to the GI tract, 72.6 Gy(RBE) in 22 fractions (3.3 Gy(RBE)/fr) for tumors adjacent to the porta hepatis, 66 Gy(RBE) in ten fractions (6.6 Gy(RBE)/fr) for other centrally located tumors. The median maximal tumor size was 3 cm. The median OS was 61 months with no significant differences between patients with in-field and out-of-field recurrence. No radiation-induced liver dysfunctions were observed. The authors hypothesized that this finding could be explained by the fact that almost half of the liver volume received a dose near zero and that, on average, more than 500 ml of non-involved liver parenchyma was preserved during the re-RT courses. The impact of hypofractionation on either clinical outcomes or toxicity was not explored.

Discussion

The present analysis showed that only a limited number of patients have received hypofractionated re-RT with PT. The majority of the published series refers to head and neck, HCC, and lung tumors, with most data derived from retrospective and monoinstitutional experiences. Selection criteria, treatment schedules, and reported outcomes are heterogeneous, and the overall level of reported data is low. Prognostic and predictive factors, when explored, were found to be different across the analyzed studies. Therefore, in the setting of re-RT with PT, conclusive data on the efficacy and toxicity profile of hypofractionated schedules could not be provided yet.

Re-RT remains a personalized approach in almost all settings of tumor recurrence. A recent consensus of the European Organisation for Research and Treatment of Cancer (EORTC) group³⁴ classified re-RT into 2 groups: type I in case of a new course of RT that has geometrical overlap with the irradiated volume of previous courses; type II in case of a new course with concerns of toxicity from the cumulative doses without overlap of irradiated volumes. As expected, in the setting of type I re-RT reported in the present analysis (mainly for head and neck and lung tumors), prognostic factors reported in cohorts of patients treated with photons were confirmed when a PT approach was offered.

Head and neck

With regards to head and neck cancers, the low number of treated patients and the retrospective nature of data collection, as well as the heterogeneity of patients, tumors, and treatment characteristics, did not allow for drawing definitive conclusions from the above-reported studies.

However, some considerations can be offered: 1) outcomes of patients treated with recurrent head and neck tumors remain poor; 2) carotid hemorrhage represented the most frequent life-threatening side effects, and it is mainly related to tumor factors and total accumulated dose; and 3) patients’ selection (mainly taking into account both tumor histology and volume) still represents a crucial issue, both in terms of treatment efficacy and risk of radiation-related side effects.

Dose constraints for organs at risk were reported only by a few authors (Supplementary material S2). A cumulative dose of < 120% to 125% from QUANTEC tolerance doses³⁵ for healthy tissues (considering α/β ratio of 3 Gy) was accepted by Gordon et al¹⁸ and Gulidov et al³² for tumors of the central nervous system. Conversely, a recovery from a previous RT course set at 70% (regardless of the interval time between the 2).

Carotid hemorrhage represented the main cause of death for patients treated with a photon-based hypofractionated schedule.36, 37, 38 Indeed, carotid involvement has been considered a contraindication for proposing hypofractionated re-RT, as well as skin infiltration.³⁹ Moreover, recurrences of nasopharyngeal tumors are commonly considered at higher risk of developing carotid bleeding due to several factors, including high prescription doses of the primary irradiation course, the proximity to the carotid artery, the osteomyelitis that often occurs after skull base re-RT, and the generally high radiosensitivity of nasopharyngeal tumors, which leads to a potentially rapid tumor shrinkage.40, 41, 42

The series reported in our analysis showed that massive bleeding remains the most relevant fatal side effect in head and neck cancer patients.

Conclusion: Hypofractionated re-RT in the setting of recurrent/second primary head and neck remained a personalized approach with a high risk of severe toxicity, also in the setting of PT.

Lung

In patients with lung cancers, ultracentral lesions are considered at higher risk for severe toxicity compared to peripheral tumors. The thorax “no-fly zone” (in which ultracentral lesions are located) was defined as the area encompassing the 2 cm radius of the trachea and proximal bronchial tree, as well as mediastinal structures.⁴³ A recent meta-analysis on stereotactic body radiotherapy applied for ultracentral masses showed a very low (4%) pooled incidence of treatment-related deaths, with the most common being hemoptysis.⁴⁴ Comparable results were found in the case of hypofractionated re-RT.⁴⁵ As expected, tumor location (ultracentral lesions) was associated with a higher risk of severe side effects in cohorts of patients treated with proton-based hypofractionated re-RT.²⁴

Another meta-analysis included 595 patients treated with photon-based stereotactic radiotherapy for recurrent primary tumors of the lung.⁴⁶ The 2-year OS and local control were 0.54 (95% confidence interval: 0.48-0.61) and 0.73 (95% CI: 0.66-0.80), respectively. The rate of toxicity of grade ≥3 was 0.098 (95% CI: 0.06-13.6), with 1.5% of patients dying due to treatment-related side effects. Ultracental lesions remain a challenge when a hypofractionated schedule is used.⁴⁴ Results of nearly 200 patients treated with hypofractionated re-RT for lung cancers have been reported in the current analysis. Unfortunately, all patients were included in mixed series treated with heterogeneous techniques (photons and protons) and fractionation schedules (standard fractionation, hyper, and hypofractionated approaches).

Conclusion: Overall, risk factors for developing severe toxicities highlighted in the setting of patients treated with photons should be carefully considered also in the case of proton-based re-RT. Accurate patient selection as well as a precise dosimetric analysis of cumulative dose absorbed by critical organs at risk remain a cornerstone for maximizing the cost/benefit ratio of re-RT independently from the treatment technique. Ultracentral lesion seems to be associated with a high risk of esophageal side effects. The impact of fraction size on the toxicity profile has not been defined yet.

Colon-rectum

Locoregional relapses after treatments occurred in about 4%-8% of patients affected by colon-rectal cancer.47, 48, 49 Surgical resection achieving negative margins represents the most important predictor of disease-specific survival.⁵⁰ The role of re-RT has been defined either in the curative setting (by increasing the rate of negative margins surgery) or for palliative purposes.51, 52, 53, 54 Neoadjuvant treatments could both lead to a higher number of R0 resections, ranging from 39% to 89%, by downsizing the tumor.⁵⁵ In this scenario, photon-based hypofractionation RT showed a median PFS rate of 12.1 months and an OS rate ranging from 28.3 to 38.7 months with an acceptable toxicity profile.56, 57

The use of hypofractionated re-RT with PT was reported in only 1 series. A multicentric prospective data collection enrolled patients treated with different fractionation schedules. Promising data in terms of clinical outcomes (2-year OS of 86.5%) and toxicity profile (9% of severe toxicity) were reported. A prospective trial currently ongoing in Italy, allowing the use of particle therapies.⁵⁸ Results would provide useful data for further analysis.

Conclusion: Hypofractionated re-RT in the setting of recurrent rectal cancer represents a promising field of application for PT. Further studies are required to better define which radiation technique could be applied in different clinical settings.

Eye

Re-RT of malignant eye tumors has been reported by different authors. Among the different RT techniques, a second course of brachytherapy may offer a high probability of tumor control and organ preservation, although visual acuity may deteriorate during the follow-up period.59, 60 Two PT facilities (Massachusetts General Hospital, Boston, Massachusetts, USA, and Charité Universitatsmedizin Berlin, Berlin, Germany) reported their experience in hypofractionated re-RT of patients with uveal melanoma. After treatment, visual acuity was found to be comparable between the 2 cohorts. Nevertheless, a significantly lower rate of enucleation was reported in the European series (4%) compared to that described by Marucci et al (29%).³⁰

Hypothesis for the selection of patients with a more favorable organ preservation rate could be 1) the smaller tumor size, 2) the different treatment of the primary tumors (patients who received radiotherapy as the first treatment are more prone to develop a radioresistant recurrence), 3) the more conservative approach used to treat treatment-related side effects, and 4) the use of different radiation fields in the re-RT course. Overall, these preliminary data on small cohorts of patients showed encouraging results in terms of either efficacy and toxicity of PT in the setting of patients treated with recurrent uveal melanoma after a first course of radiotherapy, mainly performed by PT or brachytherapy.

Conclusion: Hypofractionated re-RT could be offered to selected patients in case of recurrent uveal melanoma as a conservative treatment approach. A proper patient selection, the radiation technique, and the supportive care to manage treatment-related toxicity are of paramount importance to maximize the rate of organ preservation.

Central nervous system

The rate of tumor recurrence in the field of primary tumors of the central nervous system varies according to the tumor histology, dimension, and treatment approaches. Re-RT represents one of the treatment options, and it is increasingly used in this setting, with hypofractionation often used in selected clinical settings.61, 62, 63 The present review reported 2 series of patients treated for high-grade glioma. As expected, the clinical outcome was poor with a median OS of 12 to 14 months. However, an acceptable rate (7%-11%) of relevant side effects was reported.

Conclusion: Literature data showed that the prognosis of patients with recurrent high-grade gliomas treated with hypofractionated PT remains poor. However, the favorable dosimetric characteristics of PT allow for a reasonable rate of side effects.

Liver

HCC is a multifocal disease, and the appearance of concurrent/sequential lesions is expected during the follow-up period. Local treatments are generally employed to achieve additional local control. The compromised function of the liver (due to the baseline cirrhosis) observed in a great proportion of patients with HCC makes the dose absorbed to the uninvolved parenchyma the main concern when a radiation approach is proposed (type II re-RT according to the EORTC consensus definition). Therefore, preserving the not-involved parenchyma represents the primary treatment planning objective in repeated irradiation of primary tumors of the liver. Different tolerance doses for the whole liver have been reported. In an RTOG dose-escalating protocol, Russel et al considered 30 Gy(RBE) in 15 fractions or 36 Gy in 30 fractions as the cut-off value in hyperfractionation schedules.⁶⁴ Emami et al reported 30 to 35 and 50 Gy of tolerance dose when three-thirds, two-thirds, and one-third of the liver is irradiated. Of note, the previous 2 studies referred to patients treated with 2D/3D radiation techniques. Other authors reported a tolerance dose of 90 Gy for one-third of the liver.⁶⁵ In this setting, photon-based IMRT has played a limited role due to the well-known bath of low doses, while stereotactic body radiotherapy has been demonstrated to be a safe and effective approach in the case of multiple treatments, also.66, 67, 68, 69

In line with this philosophy, PT re-RT might represent a new treatment option. Multiple irradiation courses (up to 7) on liver recurrences have been delivered with encouraging results of either tumor local control or toxicity profile at the PT facility of Tsukuba University.26, 27 Indeed, the 2 reported series suggested that PT could be a safe and effective approach. In particular, PT achieved the best results in the case of: 1) lesions located in the peripheral zone, 2) with a Child Pugh A liver function, and 3) with a non-irradiated volume > 500 ml. On the contrary, relative contraindications seem to be: 1) tumor located close to the gastrointestinal tract, 2) the presence of ascites, 3) metastases outside the liver, and 4) a disseminated disease with multiple lesions. With regards to hepatocellular carcinoma, Oshiro et al²⁷ defined a maximum dose to the alimentary tract of 50 GyE (EqD₂), while an out-of-target liver volume was maintained below 500 ml.

Conclusion: The preliminary results suggested that the use of protons might open a new scenario in the setting of HCC, allowing sparing of a high volume of not-involved tissue. This dosimetric characteristic could be particularly useful either in patients with a worsening liver function or in the case of multiple cancer recurrences. The risk-based approach (i.e., personalized doses according to the tumor location and baseline liver function) leads to optimizing the treatment toxicity profile. Nevertheless, larger series are required to confirm these preliminary findings.

Limitations

We are aware of the several limitations of the present study. The most relevant limitation is that most of the studies (9 out of 12) presented data from mixed fractionation schedules cohort (both hypofractionated and conventional), and results regarding the hypofractionated group only could not be separated.

Moreover, 3 out of the 12 included studies reported cohorts treated with different re-RT techniques. Among them, in 2 studies it was possible to extract and report results specifically related to proton-treated patients, while in the remaining study it was not possible to separate outcomes by treatment modality. In addition, the majority of manuscripts were single-center experiences describing a very limited cohort of patients.

In addition, a wide range of hypofractionated schedules has been reported. It is well known that mild hypofractionation schedules (ie, 2.1-2.2 Gy(RBE)/fraction) have different biological paradigms compared to ablative doses (i.e., 15 Gy(RBE)/fraction). Therefore, results provided by these studies cannot be compared either in terms of clinical outcomes or side effects profiles. Along with the fractionation, when different radiation techniques have been compared, a high-quality statistical analysis was not used. Treatment techniques, as well as the constraints applied for the surrounding organs at risk, have been rarely reported. Similarly, the included studies also reported highly heterogeneous toxicity reporting (in terms of scales, timing, and rate). Overall, the available literature is characterized by substantial heterogeneity, predominantly retrospective study designs, limited sample sizes, and variability in treatment and outcome reporting. These factors did not allow a reliable quantitative synthesis; therefore, no statistical analysis or meta-analysis was performed. In this context, another important limitation of the present review is the absence of a registered review protocol and a formal structured quality or risk-of-bias assessment of the included studies, although the review methodology itself complied with PRISMA recommendations for systematic reviews.

Therefore, further research is needed. Particularly, our work highlights the need for future studies to report outcomes separately for hypofractionated and conventionally fractionated cohorts, even when both treatment regimens are included within the same study, to enable more accurate subgroup analyses and strengthen the available evidence.

Considering the rarity of the clinical setting and the still low number of PT facilities worldwide, we advocate cooperative multicenter studies, which would allow to collect standardized prospective data. The classification of Type I and Type II re-RT proposed by Andratschke et al³⁴ represents an interesting and potentially unifying framework. All studies included in this review were published before the publication of Andratschke et al, and therefore, this classification was not available at the time. The adoption of internationally validated classification systems could enable the generation of more robust and comparable data, improving consistency in patient selection, treatment techniques, and reporting of clinical outcomes. Such standardization would facilitate meaningful comparisons across studies, reduce heterogeneity in data interpretation, and ultimately strengthen the evidence base in this clinical setting.

Finally, PT hypofractionated re-RT should be considered as a personalized approach carefully tailored to each patient and ideally performed within the context of prospective data collection. Of note, in some cases, hypofractionated re-RT PT has been associated with a high toxicity rate, underscoring the importance of careful patient selection, appropriate treatment techniques, and dose constraints. Thus, we strongly recommend that data on fundamental steps such as 1) inclusion criteria, 2) equivalent dose at 2 Gy (EqD₂) plan sum, 3) constraints applied for healthy tissues, and 4) treatment technique choice (passive scattering versus active scanning) should be provided in future studies optimizing the quality assurance of data reporting according to published recommendations.34, 70, 71, 72 Moreover, the availability of photon-based high-precision techniques deserves to be taken into account. Therefore, treatment plan comparison should be considered in order to define patients who could benefit more from a PT approach.

Despite the abovementioned limitations, the current analysis represents a comprehensive and updated overview of hypofractionated PT re-RT. Findings could provide a useful tool for daily clinical practice, while more solid results are awaited.

Funding

None.

CRediT authorship contribution statement

Methodology: DA, VEP, SV, MGV: Formal analysis: DA, VEP, SV, MGV; Data Curation: DA, VEP, SV, MGV; Writing - Original Draft: DA; Writing - Review & Editing: All authors; Supervision: DA, BAJF, RO.

Declaration of Conflicts of Interest

The authors declare the following financial interests/personal relationships, which may be considered as potential competing interests: Barbara Alicja Jereczek-Fossa reports a relationship with Roche, Bayer, Janssen, Carl Zeiss, Ipsen, Accuray, Astellas, Elekta, and IBA Astra Zeneca that includes: speaking and lecture fees. Division of Radiotherapy IEO received research funding from AIRC (Italian Association for Cancer Research) and Fondazione IEO-CCM (Istituto Europeo di Oncologia-Centro Cardiologico Monzino) all outside the current project. BAJF received speakers fees from Roche, Bayer, Janssen, Carl Zeiss, Ipsen, Accuray, Astellas, Elekta, and IBA Astra Zeneca, all outside the current project. IM declares occasional fees for advisory boards supported by Eli Lilly, Novartis, Pfizer, SeaGen, Daiichi Sankyo, Astra Zeneca, Gilead, and Menarini StemLine. The remaining authors declare no conflicts of interest that are relevant to the content of this article. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary material

Supplementary material