Definitive proton therapy for unresectable trunk soft tissue sarcoma: a single-institution prospective registry-based cohort study Authors: Shuou Sudo, Koichiro Nakajima, Hiroyuki Ogino, Yukihiro Yokoyama, Masaki Sunagawa, Hiromitsu Iwata, Yukiko Hattori, Kensuke Hayashi, Toshiyuki Toshito, Akio Hiwatashi Published: 2026 ABSTRACT Highlights•Thirty-three patients with unresectable trunk STS received PT.•The most common histology was dedifferentiated liposarcoma (42%).•The 3-year local control rate was 81%.•No grade 3 or higher PT-related toxicities were observed.•Quality-of-life analysis showed preserved overall QOL with PT. Introduction Soft tissue sarcomas (STS) are rare malignant tumors that can arise in various parts of the body and exhibit diverse histological subtypes [1]. Among STS, retroperitoneal sarcomas are a particularly rare subtype, accounting for about 20% of cases in this anatomical subset [2]. Surgical resection remains the gold standard treatment for STS [3]. However, unlike sarcomas in the extremities, which are generally more accessible for surgery, retroperitoneal and other trunk-located STS often pose significant challenges to complete resection due to their deep anatomical location and proximity to vital organs. These anatomical constraints contribute to a higher risk of local recurrence and a poorer overall prognosis [4]. The establishment of an optimal treatment strategy for unresectable trunk STS remains a major challenge. Although a multidisciplinary approach is ideal, systemic therapy alone is often used throughout the course of treatment, even in cases of locally advanced disease without metastasis. However, the clinical outcomes of systemic therapy are generally unsatisfactory [5] and quality of life (QOL) is negatively impacted [6]. Patients with locally advanced disease typically do not receive radiation therapy (RT) primarily because of the radio-resistant nature of STS and their anatomical proximity to radiation-sensitive organs. RT is generally only used in the perioperative setting for these tumors [7]. With advances in RT, long-term local control (LC) has become more attainable, supporting high-precision, definitive RT even for unresectable localized disease. Recent studies have highlighted the potential of charged particle therapy, which offers superior dose distribution to photon beams as a result of the Bragg peak effect [8]. In this area, carbon-ion radiotherapy (CIRT), as a representative high-linear energy transfer (LET) charged particle therapy, has been the most widely used [9], [10]. Although most studies are retrospective due to the rarity of the disease, favorable treatment outcomes have been reported. In contrast, despite the rapid global expansion of proton therapy (PT) facilities [11], there are few studies on the curative use of PT for unresectable STS [12]. Therefore, in the present study, we retrospectively reviewed cases of definitive PT for unresectable trunk STS treated at our institution. This study focuses on treatment efficacy, safety, and changes in QOL following PT. Materials and methods Study design and patient eligibility We conducted a single-institution retrospective analysis of patients with unresectable STS in the trunk who underwent PT at our institution between December 2013 and September 2022, drawn primarily from a prospective registry‑based cohort with a small number of non‑registry cases included for this study. This study was approved by our institutional review board (approval number: 60-23-0085). This prospective in-house registry was conducted to evaluate the clinical outcomes of PT, primarily delivered in 16 or 26 fractions, for bone and soft tissue sarcomas. In this study, only cases with tumors originating from soft tissue that were histologically diagnosed as sarcomas and deemed to be unresectable by surgeons were included. The “trunk” was defined as regions excluding the extremities and head and neck areas. Patients younger than 18 years old were excluded. Furthermore, cases for which the main purpose of treatment was palliative (such as hemostasis and pain relief) and those on which no imaging tests were performed or the results obtained were not confirmed after PT were also excluded. Prior to treatment, all eligible patients provided written informed consent including agreement to the possible use of their data for research purposes. The primary endpoint was overall survival (OS). We also analyzed LC, progression-free survival (PFS), the incidence of PT-related toxicities, and changes in the QOL score. Simulation, target delineation, and prescription Patients were immobilized in the supine or prone position with an ESFORM immobilization system (Engineering System, Nagano, Japan); 2-mm-thick CT were taken using a 16-row multidetector CT. Magnetic resonance imaging (MRI) and, if available, 18F-deoxyglucose-positron emission tomography/computed tomography (FDG-PET/CT) were used to evaluate tumor invasion. The gross tumor volume (GTV) included radiographically visible lesions. The clinical target volume (CTV) was typically defined as GTV plus a 0 to 5 mm margin, and margins were slightly adjusted. The fractionation regimens were primarily based on those previously adopted in CIRT [13] at the time the registry was initiated, with either 60.8 Gy (relative biological effectiveness [RBE]) in 16 fractions, or 65 or 70.2 Gy (RBE) in 26 fractions. The choice between these regimens was made based on factors such as proximity to organs at risk (OARs), patient age, and general condition. Since April 2016, 26-fractions regimen has been predominant in accordance with the guidelines for advanced medical care in Japan. However, in more recent years, 16-fraction regimen has become the mainstay, reflecting the recent trend in CIRT. Irradiation was performed once a day, 5 days a week. Treatment planning and delivery The treatment systems at our institution were previously described in detail [14], [15]. Briefly, proton treatments were delivered by PROBEAT-III (Hitachi, Ltd., Tokyo, Japan) and planned with VQA (Hitachi, Ltd.). In the present study, a passive scattering technique with mainly 140 to 250 MeV proton beams and the spot scanning technique were employed. In the spot scanning technique, single- or multi-field optimization was used. Beam energies ranged from 71.6 to 220.1 MeV. Two to four beam portals were used for each treatment. The margin strategy for the CTV differed between the passive scattering and spot scanning techniques. To generate robust plans, separate margins to CTV were adopted perpendicular to and along the beam direction instead of a uniform expansion. The idea of margins was previously described in detail [15], [16]. The RBE value for our proton beams was 1.1 [17]. Dose constraints are summarized in Table E1. Gastrointestinal OARs, including the stomach, duodenum, small bowel, and large bowel, were contoured individually whenever identifiable. The prescribed dose was delivered uniformly to the target volume, with dose distribution adjusted as needed to prioritize OAR constraints. Surgical spacer placement (SSP) was conducted before PT for some patients with tumors adjacent to the GI tract in order to create a sufficient distance between the tumor and GI tract [18], [19]. During PT, daily pretreatment orthogonal 2-dimensional kV radiographic images were obtained to ensure precise patient positioning. The setup was continued until proper alignment with the bony landmarks on the digitally reconstructed radiographs was confirmed. Moreover, a verification CT was typically performed at least once during the treatment course, and adaptive replanning was conducted when considered necessary. A representative case is shown in Fig. 1.Fig. 1(A) Shows the dose distribution (70.2 Gy(RBE)/ 26 Fr) of a representative case of retroperitoneal dedifferentiated liposarcoma in a 63-year-old woman. (B) and (C) show magnetic resonance imaging and fluorodeoxyglucose-positron emission tomography images comparing tumor changes before and 5 years after proton therapy, respectively. The tumor size on the left side of the abdomen increased in (B), and the hyperaccumulated area in the same region disappeared in (C). The tumor may be evaluated as metabolic CR due to proton therapy. (A) Shows the dose distribution (70.2 Gy(RBE)/ 26 Fr) of a representative case of retroperitoneal dedifferentiated liposarcoma in a 63-year-old woman. (B) and (C) show magnetic resonance imaging and fluorodeoxyglucose-positron emission tomography images comparing tumor changes before and 5 years after proton therapy, respectively. The tumor size on the left side of the abdomen increased in (B), and the hyperaccumulated area in the same region disappeared in (C). The tumor may be evaluated as metabolic CR due to proton therapy. Evaluation of clinical outcomes and statistical analysis OS, LC, and PFS were estimated using the Kaplan-Meier method and were analyzed from the date of the start of PT until the event of interest or the last follow-up visit. Events of OS included death due to any cause; LC included local failures; PFS included local failures, lymph node failures, distant metastases, and all deaths. Patients, in principle, underwent MRI or CT at each visit. FDG-PET/CT was performed whenever necessary. Survival curves for OS were compared between the groups in a univariate analysis applying the Log-rank test or Cox regression analysis. PT treatment-related toxicities were assessed with the Common Terminology Criteria for Adverse Events, version 4.0. QOL scores were assessed using the European Organization for Research and Treatment of Cancer (EORTC) core QOL questionnaire (QLQ-C30, version 3.0) at baseline (before PT), and 6 and 12 months after PT. Scores for QLQ-C30 were calculated in accordance with the EORTC Scoring Manual [20], [21]. Raw scores were standardized by linear transformation to range from 0 to 100. A higher score indicated a better level of functioning and global health status/QOL (GH), or more severe symptoms. The minimally important difference (MID) was prospectively defined as an absolute change (increase or decrease) in the score from the baseline of ≥10 points [22]. All statistical tests were two-sided, and a P value < 0.05 was considered to be significant. All statistical analyses were performed with EZR, which is based on R and R-Commander (R Foundation for Statistical Computing, Vienna, Austria) [23]. Results Tumor, patient, and treatment characteristics Thirty-three eligible patients were enrolled in the present study. Patient, tumor, and treatment characteristics are summarized in Table 1. The most common histological type was dedifferentiated liposarcoma (DDLS) (n = 14; 42%), followed by leiomyosarcoma (LMS) (n = 6; 18%). The most frequent irradiation site was the retroperitoneum (n = 22; 67%), followed by the pelvic region (n = 7; 21%). The median maximum diameter was 7.8 cm (range, 2.3–24 cm). The dose fractionation ranged from 57.2 to 70.2 Gy (RBE) in 13 to 33 fractions, with the most common being 70.2 Gy (RBE) in 26 fractions (n = 19; 58%). Sixteen patients (48%) had primary tumor with no prior surgery. Two patients (6%) had previous RT. Seven patients (21%) had SSP. Table 2 lists the clinical characteristics, treatment, and outcomes of each case. Four patients (#11, #21, #26, and #30) were outside the prospective registry. Case #11 received 13 fractions due to social reasons favoring hypofractionation. Case #21 underwent dose escalation to 64 Gy (RBE) enabled by SSP. Cases #26 and #30 received 1.8 or 2.0 Gy (RBE) per fraction due to their younger age.Table 1Patient, tumor, and treatment characteristics (N=33)CharacteristicAge (years), median (range)61 (19-85)Gender, n (%) Male/Female16 (48)/17 (52)ECOG PS, n (%) 0/129 (88)/4 (12)Histology, n (%) Liposarcoma18 (55) D / M / W / U14/2/1/1 Leiomyosarcoma6 (18) UPS2 (6) MPNST2 (6) Others5 (15)Tumor size (cm), median (range)7.8 (2.3-24)GTV volume (cm3), median (range)266 (9.6-3017)Tumor status, n (%)Primary tumor with no prior surgery16 (48)Residual tumor after surgery3 (9)Recurrent or metastatic tumor after surgery14 (42)Irradiation site, n (%) Retroperitoneum22 (67) Pelvis7 (21) Mediastinum3 (9) Abdominal wall1 (3)Previous surgery17 (52)Chemotherapy before PT, n (%)19 (58)Chemotherapy after PT, n (%)15 (45)Proton technique, n (%) Spot scanning / Passive scattering26 (79) /7 (21)Dose and fraction, n (%) 70.2 Gy (RBE)/26 Fr19 (58) 60.8 Gy (RBE)/16 Fr10 (30) Others4 (12) Adaptive replanning8 (24)Surgical spacer placement7 (21)Abbreviations: ECOG = Eastern Cooperative Oncology Group; PS = performance status; D = dedifferentiated; M = myxoid; WDLS = well-differentiated liposarcoma; U = unspecified; UPS = undifferentiated pleomorphic sarcoma; MPNST = malignant peripheral nerve sheath tumor; GTV = gross tumor volume; PT = proton therapy; RBE = relative biological effectiveness.Table 2Individual case list#Age(sex)HistologyTumor statusPrior surgeryPT siteGTV(cm3)Total dose and fractionation (Gy [RBE]/Fr)SpacerPrior systemic therapyTime to local progression (months)Time to distant progression (months)Follow-up time(months)Survival status155 (M)DDLSRMYesRetroperitoneum1960.8/16NoYes4084120Alive252 (M)DDLSPNoRetroperitoneum301760.8/16NoYes--90Alive363 (F)DDLSPNoRetroperitoneum24270.2/26NoNo--63Alive476 (F)DDLSPNoRetroperitoneum13070.2/26NoYes--54Alive558 (M)DDLSSYesRetroperitoneum26670.2/26NoNo--21Alive655 (F)DDLSSYesRetroperitoneum1060.8/16YesNo--17Alive762 (M)DDLSRMYesRetroperitoneum14660.8/16NoYes-314Dead857 (M)DDLSPNoRetroperitoneum32770.2/26NoYes14116Dead983 (F)DDLSPNoPeriiliac76670.2/26NoNo-23Dead1057 (F)DDLSRMYesRetroperitoneum5360.8/16NoNo-1428Dead1183 (M)DDLSRMYesRetroperitoneum42257.2/13NoYes-927Dead1285 (F)DDLSPNoRetroperitoneum93770.2/26NoNo--39Dead1364 (F)DDLSRMYesRetroperitoneum93270.2/26NoYes-18Dead1474 (M)DDLSRMYesRetroperitoneum47970.2/26NoYes-410Dead1557 (M)MLSRMYesGluteus muscle43960.8/16NoNo-1011Alive1629 (F)MLSPNoRetroperitoneum11870.2/26NoYes14-53Dead1768 (M)WDLSRMYesRetroperitoneum4360.8/16NoNo-634Dead1861 (F)LS NOSRMYesAbdominal wall12570.2/26NoNo-849Dead1962 (F)LMSRMYesSacrum (met)3260.8/16NoNo-8104Alive2056 (F)LMSPNoRetroperitoneum5870.2/26YesYes--47Alive2142 (F)LMSRMYesSacrum (met)4764/16YesYes--25Alive2240 (F)LMSPNoPeriiliac56770.2/26NoYes-940Dead2369 (M)LMSPNoRetroperitoneum71670.2/26NoYes7720Dead2455 (M)LMSRMYesRetroperitoneum23070.2/26NoYes-27Dead2545 (M)MPNSTSYesGluteus muscle3270.2/26NoNo--63Alive2624 (M)MPNSTRMYesMediastinum46666/22NoYes--6Dead2764 (M)UPSPNoRetroperitoneum34070.2/26NoNo121219Dead2875 (F)UPSRMYesRetroperitoneum6160.8/16YesNo--19Unavailable2975 (M)SSPNoPeriiliac84370.2/26YesYes-25Dead3019 (M)ESPNoMediastinum4059.4/33NoYes--35Alive3169 (F)IMTPNoRetroperitoneum74660.8/16YesYes-13Unavailable3257 (F)RMSPNoRetroperitoneum66470.2/26YesYes526Dead3376 (F)SFTPNoMediastinum84770.2/26NoNo--72AliveAbbreviations: M = male; F = female; DDLS = dedifferentiated liposarcoma; MLS = myxoid liposarcoma; WDLS = well-differentiated liposarcoma; LS NOS = liposarcoma not otherwise specified; LMS = leiomyosarcoma; MPNST = malignant peripheral nerve sheath tumor; UPS = undifferentiated pleomorphic sarcoma; SS = synovial sarcoma; ES = Ewing’s sarcoma; IMT = inflammatory myofibroblastic tumor; RMS = rhabdomyosarcoma; SFT = solitary fibrous tumor; RM = recurrent or metastatic tumor after surgery; P = primary tumor with no prior surgery; S = residual tumor after surgery; PT = proton therapy; met = metastasis; GTV = gross tumor volume; RBE = relative biological effectiveness. Table 1Patient, tumor, and treatment characteristics (N=33)CharacteristicAge (years), median (range)61 (19-85)Gender, n (%) Male/Female16 (48)/17 (52)ECOG PS, n (%) 0/129 (88)/4 (12)Histology, n (%) Liposarcoma18 (55) D / M / W / U14/2/1/1 Leiomyosarcoma6 (18) UPS2 (6) MPNST2 (6) Others5 (15)Tumor size (cm), median (range)7.8 (2.3-24)GTV volume (cm3), median (range)266 (9.6-3017)Tumor status, n (%)Primary tumor with no prior surgery16 (48)Residual tumor after surgery3 (9)Recurrent or metastatic tumor after surgery14 (42)Irradiation site, n (%) Retroperitoneum22 (67) Pelvis7 (21) Mediastinum3 (9) Abdominal wall1 (3)Previous surgery17 (52)Chemotherapy before PT, n (%)19 (58)Chemotherapy after PT, n (%)15 (45)Proton technique, n (%) Spot scanning / Passive scattering26 (79) /7 (21)Dose and fraction, n (%) 70.2 Gy (RBE)/26 Fr19 (58) 60.8 Gy (RBE)/16 Fr10 (30) Others4 (12) Adaptive replanning8 (24)Surgical spacer placement7 (21)Abbreviations: ECOG = Eastern Cooperative Oncology Group; PS = performance status; D = dedifferentiated; M = myxoid; WDLS = well-differentiated liposarcoma; U = unspecified; UPS = undifferentiated pleomorphic sarcoma; MPNST = malignant peripheral nerve sheath tumor; GTV = gross tumor volume; PT = proton therapy; RBE = relative biological effectiveness. Abbreviations: ECOG = Eastern Cooperative Oncology Group; PS = performance status; D = dedifferentiated; M = myxoid; WDLS = well-differentiated liposarcoma; U = unspecified; UPS = undifferentiated pleomorphic sarcoma; MPNST = malignant peripheral nerve sheath tumor; GTV = gross tumor volume; PT = proton therapy; RBE = relative biological effectiveness. Table 2Individual case list#Age(sex)HistologyTumor statusPrior surgeryPT siteGTV(cm3)Total dose and fractionation (Gy [RBE]/Fr)SpacerPrior systemic therapyTime to local progression (months)Time to distant progression (months)Follow-up time(months)Survival status155 (M)DDLSRMYesRetroperitoneum1960.8/16NoYes4084120Alive252 (M)DDLSPNoRetroperitoneum301760.8/16NoYes--90Alive363 (F)DDLSPNoRetroperitoneum24270.2/26NoNo--63Alive476 (F)DDLSPNoRetroperitoneum13070.2/26NoYes--54Alive558 (M)DDLSSYesRetroperitoneum26670.2/26NoNo--21Alive655 (F)DDLSSYesRetroperitoneum1060.8/16YesNo--17Alive762 (M)DDLSRMYesRetroperitoneum14660.8/16NoYes-314Dead857 (M)DDLSPNoRetroperitoneum32770.2/26NoYes14116Dead983 (F)DDLSPNoPeriiliac76670.2/26NoNo-23Dead1057 (F)DDLSRMYesRetroperitoneum5360.8/16NoNo-1428Dead1183 (M)DDLSRMYesRetroperitoneum42257.2/13NoYes-927Dead1285 (F)DDLSPNoRetroperitoneum93770.2/26NoNo--39Dead1364 (F)DDLSRMYesRetroperitoneum93270.2/26NoYes-18Dead1474 (M)DDLSRMYesRetroperitoneum47970.2/26NoYes-410Dead1557 (M)MLSRMYesGluteus muscle43960.8/16NoNo-1011Alive1629 (F)MLSPNoRetroperitoneum11870.2/26NoYes14-53Dead1768 (M)WDLSRMYesRetroperitoneum4360.8/16NoNo-634Dead1861 (F)LS NOSRMYesAbdominal wall12570.2/26NoNo-849Dead1962 (F)LMSRMYesSacrum (met)3260.8/16NoNo-8104Alive2056 (F)LMSPNoRetroperitoneum5870.2/26YesYes--47Alive2142 (F)LMSRMYesSacrum (met)4764/16YesYes--25Alive2240 (F)LMSPNoPeriiliac56770.2/26NoYes-940Dead2369 (M)LMSPNoRetroperitoneum71670.2/26NoYes7720Dead2455 (M)LMSRMYesRetroperitoneum23070.2/26NoYes-27Dead2545 (M)MPNSTSYesGluteus muscle3270.2/26NoNo--63Alive2624 (M)MPNSTRMYesMediastinum46666/22NoYes--6Dead2764 (M)UPSPNoRetroperitoneum34070.2/26NoNo121219Dead2875 (F)UPSRMYesRetroperitoneum6160.8/16YesNo--19Unavailable2975 (M)SSPNoPeriiliac84370.2/26YesYes-25Dead3019 (M)ESPNoMediastinum4059.4/33NoYes--35Alive3169 (F)IMTPNoRetroperitoneum74660.8/16YesYes-13Unavailable3257 (F)RMSPNoRetroperitoneum66470.2/26YesYes526Dead3376 (F)SFTPNoMediastinum84770.2/26NoNo--72AliveAbbreviations: M = male; F = female; DDLS = dedifferentiated liposarcoma; MLS = myxoid liposarcoma; WDLS = well-differentiated liposarcoma; LS NOS = liposarcoma not otherwise specified; LMS = leiomyosarcoma; MPNST = malignant peripheral nerve sheath tumor; UPS = undifferentiated pleomorphic sarcoma; SS = synovial sarcoma; ES = Ewing’s sarcoma; IMT = inflammatory myofibroblastic tumor; RMS = rhabdomyosarcoma; SFT = solitary fibrous tumor; RM = recurrent or metastatic tumor after surgery; P = primary tumor with no prior surgery; S = residual tumor after surgery; PT = proton therapy; met = metastasis; GTV = gross tumor volume; RBE = relative biological effectiveness. Individual case list Abbreviations: M = male; F = female; DDLS = dedifferentiated liposarcoma; MLS = myxoid liposarcoma; WDLS = well-differentiated liposarcoma; LS NOS = liposarcoma not otherwise specified; LMS = leiomyosarcoma; MPNST = malignant peripheral nerve sheath tumor; UPS = undifferentiated pleomorphic sarcoma; SS = synovial sarcoma; ES = Ewing’s sarcoma; IMT = inflammatory myofibroblastic tumor; RMS = rhabdomyosarcoma; SFT = solitary fibrous tumor; RM = recurrent or metastatic tumor after surgery; P = primary tumor with no prior surgery; S = residual tumor after surgery; PT = proton therapy; met = metastasis; GTV = gross tumor volume; RBE = relative biological effectiveness. Survival and LC At the time of the data lock, 17 patients had died. The median follow-up was 27 (range, 3–120) months for all patients and 41.5 (range, 8–120) months for living patients (n = 16). OS, LC, and PFS curves are shown in Fig. 2A. Three-year OS, LC, and PFS rates were 51% (95% CI: 31–68%), 81% (60–92%), and 39% (23–55%), respectively. Median OS was 39 months, PFS was 10 months, and LC was not reached. Local failure was observed in six patients between 5 and 40 months after PT (Table 2). Two cases (#1 and #23 in Table 2) showed GTV progression due to limited dose coverage caused by close proximity to the GI tract. Three cases (#8, #27, and #32) showed marginal recurrence, whereas one case (#16) showed central recurrence in a re-irradiation setting. Table 3 shows the results of the univariate analysis. The univariate analysis identified a GTV volume ≥300 cm3 as a significant poor prognostic factor for OS. A multivariate analysis was not performed due to the small sample size.Fig. 2Kaplan-Meier curves in (A) show overall survival (OS), local control (LC), and progression-free survival (PFS) in all 33 cases, while those in (B) show OS, LC, and PFS in cases focusing on dedifferentiated liposarcoma and leiomyosarcoma.Table 3Univariate analysis of prognostic factors for OSOSVariablesn3-yr rateHR95% CIp-valueGenderMale1638%1.540.59-4.060.37Female (ref)1774%Age≥60 years1742%1.910.71-5.180.19<60 years (ref)1672%GTV≥300 cc1635%3.561.30-9.770.009<300 cc (ref)1774%HistologyDDLS and LMS2060%0.650.25-1.700.38Others (ref)1351%Tumor locationRetroperitoneal2248%1.310.46-3.740.61Others (ref)1173%Prior surgeryNo1661%1.020.39-2.680.96Yes (ref)1747%Prior systemic therapyYes1949%1.730.64-4.690.28No (ref)1466%Post systemic therapyYes1547%1.310.50-3.410.58No (ref)1866%Local failureYes642%1.390.45-4.300.56No (ref)2759%PT techniquesSpot scanning2658%1.040.33-3.180.95Passive scattering (ref)751%Prescription dose60.8 Gy (RBE)1047%0.730.23-2.290.5870.2 Gy (RBE) (ref)1961%Abbreviations: OS = overall survival; HR = hazard ratio; CI = confidence interval; GTV = gross tumor volume; DDLS = dedifferentiated liposarcoma; LMS = leiomyosarcoma; PT = proton therapy; Gy (RBE) = gray relative biological effectiveness. Kaplan-Meier curves in (A) show overall survival (OS), local control (LC), and progression-free survival (PFS) in all 33 cases, while those in (B) show OS, LC, and PFS in cases focusing on dedifferentiated liposarcoma and leiomyosarcoma. Table 3Univariate analysis of prognostic factors for OSOSVariablesn3-yr rateHR95% CIp-valueGenderMale1638%1.540.59-4.060.37Female (ref)1774%Age≥60 years1742%1.910.71-5.180.19<60 years (ref)1672%GTV≥300 cc1635%3.561.30-9.770.009<300 cc (ref)1774%HistologyDDLS and LMS2060%0.650.25-1.700.38Others (ref)1351%Tumor locationRetroperitoneal2248%1.310.46-3.740.61Others (ref)1173%Prior surgeryNo1661%1.020.39-2.680.96Yes (ref)1747%Prior systemic therapyYes1949%1.730.64-4.690.28No (ref)1466%Post systemic therapyYes1547%1.310.50-3.410.58No (ref)1866%Local failureYes642%1.390.45-4.300.56No (ref)2759%PT techniquesSpot scanning2658%1.040.33-3.180.95Passive scattering (ref)751%Prescription dose60.8 Gy (RBE)1047%0.730.23-2.290.5870.2 Gy (RBE) (ref)1961%Abbreviations: OS = overall survival; HR = hazard ratio; CI = confidence interval; GTV = gross tumor volume; DDLS = dedifferentiated liposarcoma; LMS = leiomyosarcoma; PT = proton therapy; Gy (RBE) = gray relative biological effectiveness. Univariate analysis of prognostic factors for OS Abbreviations: OS = overall survival; HR = hazard ratio; CI = confidence interval; GTV = gross tumor volume; DDLS = dedifferentiated liposarcoma; LMS = leiomyosarcoma; PT = proton therapy; Gy (RBE) = gray relative biological effectiveness. Subgroup analysis of DDLS and LMS Subgroup analyses of DDLS and LMS were conducted because they represented the majority of cases in the present study. The median follow-up was 24 (range, 3–106) months. OS, LC, and PFS curves are shown in Fig. 2B. Three-year OS, LC, and PFS rates were 60% (95% CI: 33–79%), 88% (61–97%), and 45% (23–65%), respectively. Median OS was 40 months, PFS was 11.5 months, and LC was not reached. PT treatment-related toxicities and QOL PT treatment-related toxicities ≥grade 3 were not observed. Four patients (12%) developed grade 2 dermatitis, one patient (3%) a grade 2 gastric ulcer, and one patient (3%) a nervous system disorder (drop foot). The gastric ulcer occurred outside the irradiation field. Drop foot was noted in a patient who received PT for a tumor originating from the sciatic nerve near the piriformis muscle. QOL scores are summarized in Fig. 3 and Table E2. QOL questionnaires were available from 21 (64%) patients at baseline, 14 (42%) at 6 months, and 9 (27%) at 12 months after PT. Mean (SD) GH and physical functioning scores were 49.6 (21.9) and 71.7 (22.4) at baseline, 61.9 (17.2) and 82.4 (18.4) at 6 months, and 67.6 (23.1) and 75.6 (27.9) at 12 months after PT, respectively. A clinically significant deterioration from baseline to the MID threshold (≥10-point decline in functioning/GH or ≥10-point increase in symptoms) was only observed for dyspnea at 6 months.Fig. 3Changes in quality of life (QOL) scores at baseline (before proton therapy), and 6 and 12 months after PT. The items evaluated are in accordance with the European Organization for Research and Treatment of Cancer (EORTC) core QOL questionnaire (QLQ-C30, version 3.0). Changes in quality of life (QOL) scores at baseline (before proton therapy), and 6 and 12 months after PT. The items evaluated are in accordance with the European Organization for Research and Treatment of Cancer (EORTC) core QOL questionnaire (QLQ-C30, version 3.0). Discussion To the best of our knowledge, this is the largest study on definitive PT for unresectable trunk STS. The present results demonstrated that PT achieved favorable LC and OS, with 3-year rates of 81 and 51%, respectively, while limiting severe adverse events. In the subset of patients with available assessments, QOL appeared to be generally maintained after PT. LC not only has the potential to improve OS, but may also play a role in preventing local complications, such as pain and paralysis, thereby contributing to the maintenance of QOL. Prioritizing local therapy is widely accepted in STS management when the tumor is localized. Regardless of the use of preoperative RT, surgery for retroperitoneal sarcomas achieves an abdominal recurrence-free survival of approximately 5 years [3]. Furthermore, the Trans-Atlantic Retroperitoneal Sarcoma Working Group has shown that repeated surgeries for recurrent retroperitoneal sarcoma can also lead to favorable outcomes [24]. RT is a representative local therapy modality as an alternative to surgery. Allignet et al. demonstrated that photon-based modern RT achieved 3-year LC, PFS, and OS rates of 56.8, 16.6, and 34%, respectively [25]. More recently, photon-based hypofractionated approaches have also been investigated for unresectable or unresected sarcomas, including hypofractionated RT [26] and hypofractionated accelerated radiation dose-painting [27]. These studies suggest that modern photon-based dose escalation strategies may improve local outcomes. However, in consideration of the importance of dose escalations [28], the use of particle therapy represents a rational treatment option to further improve outcomes. CIRT offers physical and biological advantages over conventional photon therapy. Imai et al. [29] reported the outcomes of CIRT in 128 patients with unresectable axial STS, with 3- (5-) year LC, OS, and disease-free survival rates of 68 (65), 60 (46), and 39 (32)%, respectively. However, whether CIRT is clinically superior to PT for STS remains controversial. In this context, we emphasize the importance of accumulating long-term data on definitive PT. As highlighted by a representative case (Fig. 1), the long-term effects of definitive PT on tumors are unlikely to be predicted based on the short-term tumor responses observed with pre- or postoperative RT. Demizu et al. found no significant differences in outcomes between CIRT and PT for bone and soft tissue sarcomas [12], which is consistent with the present results. Moreover, although PT is classified as low-LET radiation and is considered to be less effective for radioresistant tumors, emerging evidence suggests that it is biologically distinct from photon therapy [30]. Preclinical studies showed marked variability in RBE across histological subtypes for both PT and CIRT, for example, with the RBE of proton beams in MG63 osteosarcoma cells being 17% higher than 1.1 [31]. These findings suggest the need for detailed preclinical and clinical investigations to optimize the use of PT for sarcomas. We observed six cases of local failure in the present study, all of which were retroperitoneal tumors. These results suggest the need to optimize margin settings depending on histological subtypes; however, the application of wide margins remains challenging. SSP is a promising approach to address this challenge, particularly with the use of bioabsorbable polyglycolic acid (PGA) spacers [18], [19]. In the present study, SSP with a PGA spacer was applied in seven patients, and LC was achieved in all but one case, which exhibited rapid tumor progression (#32). In terms of fractionation, local failure occurred in only one case (#1) treated with the 16-fraction regimen. This supports the adoption of 16 fractions as a standard in definitive PT, similar to CIRT. Furthermore, the future clinical implementation of LET-based intensity-modulated PT optimization may allow for more effective local control [32]. To optimize the role of RT, the careful selection of candidates is essential. In the present study, a large tumor size was identified as a poor prognostic factor for OS, consistent with previous RT studies [9]. However, surgical studies have not consistently identified size as a prognostic factor, instead emphasizing the tumor grade and residual disease [33]. This suggests that poor outcomes in RT for large tumors may reflect treatment limitations rather than tumor biology. Notably, we observed long-term survival in a patient with a very large tumor (#2) [34], indicating that tumor size alone should not limit the use of local therapeutic strategies in locally advanced unresectable patients. While SSP may offer a potential solution, it is less suitable for bulky tumors. We are also investigating PT following debulking surgery. In case #6, preoperative imaging suggested that R2 resection was unavoidable; therefore, SSP was intentionally administered at the time of tumor resection, followed by additional PT. The effectiveness of debulking surgery itself has been reported in the management of STS [35]; therefore, this treatment approach may represent a promising strategy and warrants further investigation in prospective studies. Systemic therapy remains the standard first-line treatment for advanced STS; however, its efficacy is often limited and frequently results in a decline in QOL. Although recent randomized trials, including TH CR-406/SARC021 [36], evaluated new systemic strategies, median OS for patients with locally advanced disease remains around 17–23 months, with long-term survival beyond five years being extremely rare. Moreover, although data are limited [37], the ANNOUNCE trial reported QOL deterioration within two months of initiating doxorubicin-based therapy [6]. In contrast, local therapies, such as RT and surgery, may preserve QOL more effectively. The present results suggest that PT does not significantly impair QOL, and previous studies showed only a mild QOL decline even after multivisceral surgery [38]. Systemic therapy is the only option in the metastatic phase; however, when the disease is still localized, effective local therapy may allow for the avoidance of or delay in systemic therapy, providing a meaningful strategy to preserve patient QOL. Nevertheless, the high incidence of distant recurrence in the present cohort underscores the need for appropriate integration of systemic therapy with local treatment to further improve survival outcomes. This study has several limitations that need to be addressed. This was a retrospective analysis of a small sample size, which may have introduced a bias. The heterogeneous nature of the disease also makes direct comparisons with other studies difficult. To aid clinical interpretations, we included detailed individual case data to assist in practical decision-making. In addition, the definition of unresectability was not based on universally standardized criteria. Surgical decision-making in STS is center-dependent and may vary according to surgical expertise. Therefore, some tumors deemed unresectable in the present study might have been considered surgically approachable at other specialized sarcoma centers. Furthermore, competing risks were not considered in the calculation of LC rates, potentially leading to an overestimation. However, we used the Kaplan–Meier method for comparability with previous studies and provided a detailed case list for an alternative analysis if needed. Moreover, tumor grade information was not included in the present study, as it was not mandated by the prospective registry and was therefore unavailable for the majority of cases. Even in cases where data were accessible, grading was primarily based on needle biopsy specimens, which are often inadequate for an accurate evaluation using the FNCLCC system. Given these limitations, along with the known subjectivity and interobserver variability of FNCLCC grading [39], future studies may benefit from molecular-based approaches, such as gene expression profiling, which may provide more reliable prognostic information. In conclusion, the treatment of unresectable trunk STS remains a significant clinical challenge. This study suggests the potential of PT as a safe and effective treatment option and definitive RT for STS may not need to rely solely on CIRT. However, numerous issues remain to be addressed, including optimal irradiation techniques, patient selection, and integration with surgery and systemic therapy. PT may play an important role in the management of this disease, and we hope that the present results will contribute to future advances in this field. CRediT authorship contribution statement Shuou Sudo: Data curation, Writing – original draft, Writing – review & editing. Koichiro Nakajima: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Writing – original draft, Writing – review & editing. Hiroyuki Ogino: Investigation, Project administration, Writing – review & editing. Yukihiro Yokoyama: Investigation, Project administration, Writing – review & editing. Masaki Sunagawa: Investigation, Writing – review & editing. Hiromitsu Iwata: Investigation, Writing – review & editing. Yukiko Hattori: Investigation, Writing – review & editing. Kensuke Hayashi: Resources, Writing – review & editing. Toshiyuki Toshito: Resources, Writing – review & editing. Akio Hiwatashi: Supervision, Writing – review & editing. Funding This work was supported by the JSPS KAKENHI Grant Number 24K18835, and Yong Investigator Clinical Research Grant in Nagoya City University Grant Number 60-23-0085. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.