Skip to main content
NCBI home page
Physics and Imaging in Radiation Oncology logoLink to Physics and Imaging in Radiation Oncology
. 2026 Apr 17;39:100967. doi: 10.1016/j.phro.2026.100967

Comparison of gantry-less magnetic resonance guided proton therapy with photon therapy for hepatocellular carcinoma patients

Franciska Lebbink a,b, Emilie Alvarez Michael a,c, Anna Kornek a,c, Annika Lattermann a,c, Chiara Valentini a,c, Petra Trnkova a,d, Esther GC Troost a,b,c,e,⁎
PMCID: PMC13137213  PMID: 42088601

Graphical abstract

Unlabelled Image

Open in a new tab

Highlights

  • •

    Realistic setup with a gantry-less proton therapy system in supine position.

  • •

    41/65 fractions required adaptation with magnetic resonance guided proton therapy (MRgPT).

  • •

    Lower mean dose to healthy liver [median 6.1 Gy(RBE)] with MRgPT compared to photons.

Abstract

Background and purpose: With magnetic resonance imaging (MRI), soft-tissue tumors such as hepatocellular carcinoma (HCC) are visualised. MR guided proton therapy (MRgPT) may combine the advantages of proton therapy with soft-tissue image-guidance. Therefore, we assessed the calculated dose differences between MR-guided photon therapy (MR-LINAC) and MRgPT at a horizontal beamline for HCC patients, considering non-adapted and adapted plans.

Materials and Methods: Planning computed tomography and T2-weighted MRI of ten HCC patients with 13 lesions were included. Adaptive hypofractionated MR-LINAC, gantry-based PT as well as MRgPT treatment plans at a horizontal beamline were optimized using RayStation 2023B-R on the daily MRIs. The comparison included various dose-volume-histogram (DVH) parameters of target volumes and normal tissue.

Results: Clinical target volume (CTV) coverage could be fulfilled for 12/13 cases, where plan adaptation was required for 41, 36 and 15 out of 65 fractions for MRgPT, gantry-based PT and MR-LINAC, respectively. With plan adaptation, the D98%,CTV could be restored for 56, 60 and 59 fractions, respectively. The median adapted Davg of the uninvolved liver was reduced from 10.1 Gy for MR-LINAC to 6.1 Gy(RBE) for MRgPT, and 4.5 Gy(RBE) for gantry-based PT. For all adapted plans, the normal tissue constraints were fulfilled in all patients.

Conclusions: More than half of the proton fractions needed plan adaptation for adequate target coverage. MRgPT lead to clinically acceptable treatment plans for HCC patients and reduced the dose to the uninvolved liver compared to MR-LINAC, with similar target coverage. HCC may thus be a promising indication for hypofractionated adaptive MRgPT.

1. Introduction

Advances in the field of radiation therapy (RT; e.g. image guidance, radiotherapy delivery and treatment planning techniques) and consequently reduced dose to healthy liver, allow for treatment of hepatocellular carcinoma (HCC) patient with curative intent [1]. With daily online image guidance visualizing the effect of breathing on tumor and normal tissue motion, the margins compensating for internal motion and thus uncertainty can be reduced leading to less dose in uninvolved tissue [2]. Magnetic resonance (MR)-guided radiotherapy (MRgRT) improves the accuracy of the beam delivery for moving soft-tissue tumors, including HCC, allowing for daily plan adaptation and real-time imaging [2], [3]. MRgRT may be preferred over conventional stereotactic photon therapy if the soft-tissue tumor is in close proximity to organs at risk (OAR), e.g. stomach or bowel, or if the impact of breathing motion needs to be mitigated by applying gating and/or tracking [4]. In particular, promising tumor control has been obtained for hypofractionated MR-LINAC treatment in HCC patients applying 40–60 Gy in 5–10 fractions [3].

Compared to photon therapy, proton therapy (PT) for HCC reduces the dose to the uninvolved liver, thus lowering the incidence of radiation-induced liver disease, while maintaining local control; this combination may be the cause for better overall survival [5], [6]. It may be treatment-of-choice for large tumors with a relatively small uninvolved liver volume, advanced liver disease or re-irradiation [4]. Additionally, PT allows for higher doses to the tumor without significantly increasing the risk of radiation induced liver disease compared to RT [7]. However, PT is highly sensitive to anatomical changes, where small changes can result in a target underdosage and/or overdosage of OARs, and thus the impact of motion, e.g. due to breathing or bowel fillings, has to be considered [8], [9]. This impact can be partially mitigated by selecting appropriate beam angles, e.g. avoiding the bowel, by rescanning or gating.

Daily 3D imaging, such as cone beam computed tomography (CBCT), repeated CTs or MRI, and real-time imaging is required to monitor the inter- and intra-fractional anatomical changes during beam delivery, respectively. Recently, daily adaptive PT based on in-room CT was clinically implemented [10]. Integrating MRI with PT would allow for daily adaptive PT based on the excellent soft tissue contrast of the MRI, while taking advantage of the radiation-free imaging offering anatomical and functional information [11], [12], [13]. Different treatment planning studies showed the potential for adaptive MR guided PT (MRgPT) for liver and central lung tumors, albeit not in a realistic setting [14], [15], [16]. Those studies were based on a gantry, with full flexibility for the beam angle selection. However, the integration of gantry-based MRgPT is technically complicated. A whole-body MR scanner has been installed at our horizontal proton pencil beam scanning (PBS) beamline, resulting in a realistic gantry-less MRgPT system. MRgPT with a fixed beamline is thus technically feasible, yet to date, the clinical impact of this MRgPT configuration remains to be determined [17].

The aim of this in silico treatment planning study was therefore to investigate potential differences in delivered clinical target volume (CTV) and OARs doses for treatments based on adaptive MR-LINAC, adaptive gantry-less MRgPT with a horizontal PBS beamline with patients in supine position, and gantry-based PT plans in HCC patients.

2. Materials and methods

2.1. Patient and tumor characteristics

Retrospective data from ten HCC patients with in total 13 lesions treated between January 2022 and December 2024, on the Unity MR-LINAC (Elekta AB, Stockholm, Sweden) were included (Table 1). The Ethics Committee of the TUD Dresden University of Technology, Germany, approved the study (BO-EK-272072024).

Table 1.

Details on the underlying diagnosis, affected liver segment, surrounding critical organ(s) and size of the gross tumor volume. Abbreviations: HCC: hepatocellular carcinoma; S: segment; GTV: gross tumor volume.

Patient Diagnosis Case Segment Critical organ
within 2 cm of the CTV		 GTV
volume [cm3]		 Uninvolved
liver volume [cm3]
1 Liver met 1 S2/3 heart, stomach 122.4 1613
2 Liver met 2 S7 − 6.9 1781
3 S5 − 21.5 1719
3 Liver 4 S8 − 25.3 1395
4 Liver 5 S8 − 48.9 1182
5 Liver 6 S6 small bowel, colon 64.9 1610
6 Liver 7 S5 colon 5.7 1390
8 S6 − 4.9 1391
7 Liver met 9 S5 colon 3.3 1940
8 Liver 10 S4 heart 11.9 1048
9 Liver 11 S3 stomach, heart 41.9 1250
12 S3 − 20.0 1272
10 Liver 13 S6 colon 39.9 1656
Open in a new tab

2.2. Imaging, delineation and workflow

The clinical planningCT (pCT) was acquired on the SOMATOM Definition AS scanner (Siemens Healthineers, Erlangen, Germany) using the clinical abdomen protocol (120 kV, 300 mAs, slice thickness 2 mm). It was 3D reconstructed as average CT from an amplitude-based 4DCT with eight breathing amplitudes. To mitigate the impact of breathing motion, the patients were treated with a pressure belt. 2D cine MRI were acquired to monitor the breathing motion.

T2-weighted MR images (voxel size 2x2x2.4 mm3, repetition time (TR) 2100 ms, echo time (TE) 212.2 ms) were acquired on the 1.5T MR scanner of the MR-LINAC on the same day as the pCT and before daily treatment delivery. All imaging (CT and MR) was performed with the patient in head first supine position.

The CTV was delineated on the pCT encompassing gross tumor volume (GTV) from different 4DCT breathing phases (average, full-inhale and -exhale) following the internal target volume (ITV) concept. The T2-weighted MRI was used to correct anatomical boundaries. The planning target volume (PTV) consisted of the CTV with 5 mm isotropic margin to mitigate set-up uncertainties of the MR-LINAC plans [18]. The GTV was recontoured on every new daily image. The clinically approved structures from the pCT were deformably mapped onto the daily MRIs and corrected if necessary [19].

The daily MRIs were converted to synthetic CTs (sCTs), applying density overrides individually gathered from the respective contours of the pCT (for external contour, ribs, liver, stomach, right lung, vertebra and heart) using the research version of RayStation 2023B-R (RaySearch Laboratories AB, Stockholm, Sweden). The average Hounsfield units (HU) values of the specific organs were extracted and converted to mass density with the HU values to mass density curve from the pCT. An example is shown in Fig. 1.

Fig. 1.

Fig. 1

Open in a new tab

For case 5, dose distributions of MR-guided proton therapy (MRgPT) for the nominal plan (upper left), non-adapted (upper right) and adapted (lower left) fraction on single-energy computed tomography (CT) derived synthetic CT (sCT) with density overrides. On the lower right, the dose volume histograms are shown, with a D98% of 46.1 Gy (RBE) for the nominal plan, 34.8 Gy(RBE) for the non-adapted, and 46.0 Gy(RBE) for the adapted plan.

2.3. Treatment planning

This study followed the RATING guidelines [20]. For each patient, the nominal hypofractionated treatment plans with MR-LINAC, gantry-based PT and gantry-less MRgPT plans were created on the first daily MRI of the MR-LINAC system using the research version of RayStation 2023B-R). For the MR-LINAC, an average dose Davg of 48 Gy in 6 fractions was prescribed to the PTV, aiming for D98%≥90%. For both PT plans, relative biological effectiveness (RBE) (RBE = 1.1) weighted dose Davg of 48 Gy(RBE)1 in 6 fractions was prescribed to the CTV and robustly optimized with 5 mm setup and 3% range uncertainty [22]. In the worst-case scenario, the CTV coverage D98% was preferably D98%≥98%, albeit D98%≥95% was acceptable, while maintaining the OARs within the institutional constraints. For all modalities, a helping structure was applied to the target by adding a margin around the OARs and subtracting the overlap from the CTV if the OARs were within 1 cm from the target. A maximum dose limit was added in the optimization for the helping structure to prevent the clinical limit of the heart, stomach and colon could not be fulfilled. Sparing of the OARs was given the highest priority, followed by CTV minimum dose, dose fall-off outside the target and maximum dose within the target. A maximum dose constraint of 52.8 Gy was applied. The following normal tissue dose constraints were applied within the volume of a given OAR receiving x Gy (VxGy): heart V32Gy < 15 cm3, stomach V25Gy < 10 cm3, colon V25Gy < 20 cm3, duodenum V18Gy < 5 cm3, small bowel V19.5Gy < 5 cm3, and at least 700 cm3 of the total non-involved liver (liver-GTV) received less than 21 Gy (i.e. “Total liver volume − V21Gy > 700 cm3″). Most patients had tumors near the heart, stomach, small bowel or colon (Table 1).

A MR-LINAC beam model was created in RayStation 2023B, considering the impact of a homogeneous magnetic field of 1.5 T [23]. Step-and-shoot IMRT plans were optimized on all daily MRs with a Monte-Carlo dose engine (v5.5), a dose grid resolution of 3 mm and a statistical uncertainty of 2%. The selected beam angles of the clinical plans (varying between 7 and 11 fields) were used for the new treatment plans created in RayStation.

Gantry-less MRgPT plans with patients in supine position using 2 intensity-modulated proton therapy (IMPT) beams of 90° and couch rotations of ± 10° were created with RayStation 2023B. The fixed beamline was based on the ProteusPLUS proton therapy system (Ion Beam Applications, Louvain-la-Neuve, Belgium) [24]. The distance between the MR-isocenter and the PBS-isocenter was 144.5 cm along the beam axis. Of note, the magnetic field was not yet considered in the proton plans. Gantry-based PT plans were created for a gantry setup with the clinical beam model for the ProteusPLUS gantry using three patient-specific beam angles. For both PT approaches, a range shifter (6.5 cm thick PMMA) was used for patients with a superficial HCC. The final proton dose was computed for a dose grid resolution of 3 mm and 0.3% statistical uncertainty and using a Monte-Carlo dose engine (v5.5). Additionally, the nominal plan on the first daily MR of the MR-LINAC, gantry-based PT and gantry-less MRgPT plans were recalculated without adaptation on all consecutive daily MRI scans to determine the need for adaptation, further described as the ‘non-adapted’ plans. The consecutive MRIs were rigidly registered onto the first daily MRI focusing on the CTV. For the ‘adapted’ plans, the nominal plans were reoptimized on all consecutive daily MRI scans with the same approach as the nominal plans by correcting the isocenter to the new target and by fulfilling the constraints. All plans were approved by a board-certified medical physicist.

2.4. Evaluation and statistics

The treatment plans were compared in terms of dose volume histogram (DVH) parameters for the GTV, CTV and OARs (i.e., colon, small bowel, stomach, duodenum, uninvolved liver and heart) by calculating the median of all individual fractions of the treatment plans for all HCC patients and reporting them with the respective absolute range. The uninvolved liver was compared in terms of V21Gy and Davg. Robustness evaluation on the nominal and adapted plans was performed by applying shifts of 5 mm in all directions and including range uncertainties of 3% for all proton plans. The loss of the CTV in D98% was defined as the ΔD98% = D98%,adapted plan − D98%,worst case. Statistical significance was assessed using a Wilcoxon signed Rank Test in Python, and a p < 0.05 was considered statistically significant.

3. Results

Clinically acceptable adaptive gantry-less MRgPT, gantry-based PT, and MR-LINAC treatment plans were generated for all ten patients with 13 lesions, resulting in 13 nominal plans with 65 adapted plans and non-adapted recalculations for each modality. A RATING score of 96% was obtained.

The median adapted D98%,CTV was 46.2 Gy(RBE) [35.1;47.3 Gy(RBE)] for gantry-less MRgPT, 46.9 Gy(RBE) [36.8;47.4 Gy(RBE)] for the gantry-based PT, and 46.7 Gy [38.9;46.8 Gy] for MR-LINAC, respectively (Fig. 2). The threshold set for plan adaptation was D98%,CTV > 45.6 Gy, which was reached for 41/65, 36/65 and 15/65 non-adapted fractions for gantry-less MRgPT, gantry-based PT and MR-LINAC, respectively. The median adapted D98%,CTV was significantly higher for gantry-based PT compared to gantry-less MRgPT (p < 1e-3), and MR-LINAC (p < 1e-3). The median D98%,CTV in the non-adapted fractions compared to the corresponding adapted fraction decreased statistically significant by 2.9% [−0.2;24.6%] for MRgPT (p < 1e-3), 0.4% [−0.1;79.2%] for gantry-based PT (p < 1e-3) and non-significant 0.1% [−1.3;20.5%] for MR-LINAC (p = 0.08). With plan adaptation, the D98%,CTV was above 45.6 Gy for 56/65, 60/65 and 59/65 fractions for MRgPT, gantry-based PT and MR-LINAC, respectively. For patient 1, the D98%,CTV of the nominal plan was below 95% for all modalities, which could be increased with plan adaptation by 7.6% for MRgPT, 7.7% for MR-LINAC and 10.3% for gantry-based PT. Similar for the MR-LINAC plan of patient 5, where plan adaptation increased the D98%,CTV by 12%.

Fig. 2.

Fig. 2

Open in a new tab

The dose to 98% (D98%) of the clinical target volume (CTV) for all patients is shown for MR-guided photon therapy (MR-LINAC; blue), MR-guided proton therapy (MRGPT; orange) and gantry-based proton therapy (gantry; green), by adding up the non-adapted and adapted fractions for every patient respectively. The red line represents the clinical goal value of D98%>95%. The inset shows a zoomed-in view of the adapted plans for patient 2–13. The boxplot shows the first quartile to the third quartile of the data, with the orange line as the median, the whiskers the inter-quartile range multiplied by 1.5 and the circles the outliers beyond this range. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

The median adapted V21Gy of the uninvolved liver was 202 cm3 [61;573 cm3] for MRgPT, 124 cm3 [31;297 cm3] for gantry-based PT and 235 cm3 [38;499 cm3] for MR-LINAC (Fig. 3). The median adapted Davg of the uninvolved liver was 6.1 Gy(RBE) [1.7;15.8 Gy(RBE)] for MRgPT, 4.5 Gy(RBE) [0.9;12.4 Gy(RBE)] for gantry-based PT and 10.1 Gy [1.7;19.9 Gy] for MR-LINAC. For both parameters, the differences between MR-LINAC and MRgPT as well as gantry-based PT to MRgPT were statistically significant (p < 1e-3).

Fig. 3.

Fig. 3

Open in a new tab

Relevant dose-volume histogram parameters for MR-guided proton therapy (MRgPT) and MR-guided photon therapy (MR-LINAC) treatment plans, assessed for a subset of patients (number of cases; N), in whom the OAR received a clinically relevant dose: These include the volume of the heart receiving a dose of 32 Gy (V32Gy), V25Gy of the colon, V19.5Gy of the small bowel, V25Gy of the stomach and V21Gy of the UninvolvedLiver (Liver-GTV) for MRgPT, gantry-based PT and MR-LINAC for the number of cases (N), in whom the OAR received a clinically relevant dose. The red lines show the clinical goals. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

The V25Gy of the colon for the non-adapted plans exceeded the clinical limit for four patients for MRgPT and MR-LINAC and for one patient for the gantry-based PT (Fig. 3). Plan adaptation reduced median V25Gy from 15.7 cm3 [0;47.5 cm3] to 16.1 cm3 [0;20 cm3] for MR-LINAC, changed from 12.1 cm3 [0;55.1 cm3] to 16.9 cm3 [0.4;20.5 cm3] for MRgPT and from 7.1 cm3 [0;21.9 cm3] to 2.5 cm3 [0;16.8 cm3] for gantry-based PT. Without plan adaptation, the V32Gy of the heart exceeded the clinical limit for three patients for all modalities without adaptation, the V25Gy of the stomach of the non-adapted plans was exceeded in numerous fractions in four patients, and the V19.5Gy of the small bowel was exceeded in two patients (Fig. 3). All the clinical goals for the OARs were fulfilled with plan adaptation.

Robust evaluation of the nominal and adapted plans showed D98% of the CTV to be robust with a median standard deviation within 1 Gy for all modalities (results shown in Supplementary Material A and Fig. S1). The loss of the worst case scenario in D98% of the CTV was 0.8[0.5;10.8] Gy for MR-LINAC, 2.8[0.4;12.5] Gy(RBE) for MRgPT, and 0.7[1.9;5.0] Gy(RBE) for gantry-based PT. The D98%>90% for 84% of the MR-LINAC scenarios, 82% of the MRgPT, and 91% of gantry-based PT.

4. Discussion

This in silico treatment planning study demonstrated (1) the applicability of (adaptive) gantry-less MRgPT and (2) the need for plan adaptation in all three investigated modalities and (3) dosimetric superiority of the gantry-based PT plans underlining the advantages of flexible beam angle selection to exploit the full potential of PT. This is the first dosimetric study comparing hypofractionated MRgRT with clinically realistic gantry-less MRgPT, and gantry-based PT in HCC patients.

Petersen et al. [25] and Mondlane et al. [26] reported on a reduced dose to the uninvolved liver for IMPT compared to IMRT (Uninvolved liver: V15Gy < 456 cm3 and Davg < 23 Gy), leading to a reduction of the risk for radiation induced liver disease. Previous studies comparing hypofractionated proton and photon RT for HCC demonstrated the superiority of PT based on NTCP models and sparing of the OARs [14], [15]. Nenoff et al. [14] showed a significantly reduced mean liver dose when using IMPT compared to IMRT. Moteabbed et al. [15] showed a decreased mean dose of the uninvolved liver with offline MRgPT compared to CBCT-guided PT with different margins. However, all studies assumed gantry-based PT system. To increase comparability, the PTV margin and CTV robustness settings ought to be identical [27]. Therefore, a 5 mm setup uncertainty was applied for the proton plans in our study, including a 3% range uncertainty.

Our results using a gantry-less PT are in line with previous findings. Even though gantry-based PT may be superior to gantry-less bearing in mind the full flexibility of beam angle selection. Integration of MRI with PT gantry is technically even more challenging, and thus workarounds are sought for [12]. Guerreiro et al. [28] showed that the need for a gantry can be mitigated for pediatric abdominal tumors with varying couch rotations in supine position. This can also be achieved by going for upright MRgPT, where the full flexibility can be preserved with upright positioning in a fixed beamline [29]. The necessity of a gantry was questioned by Yan et al. [30] in a retrospective study. They showed that only 1.3% of the abdomen patients could have the same plan quality with the fixed beamline as with gantry when treated in a supine position, which increased to 68.8% in upright and up to 99.4% with additional 20° chair angle. Noteworthy, the comparison was performed for passive scattering. For PBS, potentially all patients can have comparable plan quality with upright positioning.

Moreover, we demonstrated the need for daily adaptation, especially for MRgPT. MR-guidance would allow for (gated) adaptive treatments to mitigate intra- and inter-fractional changes [11]. Experience from online adaptive MR-LINAC already shows that daily adaptation can significantly reduce the dose for the bowel, duodenum and stomach, while restoring the dose to the target in liver treatments [31], [32]. Larger robustness settings could reduce the need for daily plan adaptation, but leading to increased dose to OAR and may be insufficient for large changes caused by tumor shrinkage, weight-loss, organ motion, organ size, and shape variations [33]. The first clinical implementation of online adaptive PT [10] was based on an in-room CT workflow. The CT based approaches can be implemented with gantry-based PT but have limited soft tissue contrast and add imaging dose to the patient.

The study had several limitations: (1) The first daily MR was considered nominal plan and used for recalculation on consecutive daily MR scans in the non-adapted scenarios. Therefore, the impact was smaller than if the planning CT or MR would be used, due to the longer time between planning phase and treatment. The uncertainty in GTV delineation caused by interobserver variability and imaging artifacts was neglected. If it turns out that online adaptive radiotherapy leads to margin reduction of the CTV, the delineation uncertainty must again be revisited. In some cases, plan adaptation was required, which might have been avoided if additional margins were applied, however a compromise was found between plan robustness and to spare the OARs, in the proximity of the OARs along the proton beam path. Moreover, plan robustness was maintained to result in adequate target coverage. The margins of the iCTV from the breathing motion of the 4DCT were reapplied, which were verified with cineMR imaging during treatment. Noteworthy, other groups have already shown that if the treatment plan is adapted on the first fraction(s) only, the need for a full daily adaptive workflow is reduced [34], [35]. The impact of plan adaptation might be reduced over the whole treatment course compared to the dose of the single fractions. The fractionation effect was not considered, due to the uncertainty of deformable image registration which is applied in dose accumulation (see Nenoff et al. [36]). (2) The impact of breathing motion and the interplay effect between the breathing motion with the spot delivery of the proton beam has not been considered in this in silico planning study. Especially for (hypofractionated) proton therapy, the impact will not average out and needs to be mitigated [37]. Future studies need to focus on the impact of anatomical changes combined with breathing motion, especially in HCC patients, e.g. by implementing gating to mitigate the impact of breathing motion [38]. (3) The integration of the magnetic field effect into the treatment planning process is still pending [12]. From recent work, it has become clear though, that beam deflection may be effectively compensated [39], [40]. (4) Different density overrides were applied to create synthetic CTs, even though the accuracy of synthetic CTs for MR-only based PT planning is still under debate [12], [41].

In conclusion, HCC patients would be suitable candidates for MRgPT using a gantry-less system with patients in supine position, where the dose to the uninvolved liver may be reduced with MRgPT compared to MR-LINAC.

CRediT authorship contribution statement

Franciska Lebbink: Writing – review & editing, Writing – original draft, Software, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Emilie Alvarez Michael: Writing – review & editing, Investigation, Formal analysis. Anna Kornek: Writing – review & editing, Investigation. Annika Lattermann: Writing – review & editing, Formal analysis. Chiara Valentini: Writing – review & editing, Investigation. Petra Trnkova: Writing – review & editing, Supervision, Formal analysis. Esther G.C. Troost: Writing – review & editing, Supervision, Resources, Investigation, Funding acquisition.

Declaration of competing interest

Prof. Petra Trnkova is an Editorial Board Member for Physics and Imaging in Radiation Oncology and was not involved in the editorial review or the decision to publish this article.

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: For the present study, the authors received no financial support involved in the study design or materials used, nor in the collection, analysis and interpretation of data nor in the writing of the publication. Prof. Esther G. C. Troost is a member of the Scientific Advisory Board of IBA and she’s Editorial board member of Radiotherapy and Oncology.

Acknowledgements

The authors would like to thank Sally Abdelhaleem for critical review of the contours and Agnes Angerud from RaySearch Laboratories for providing the MR-LINAC functionality in RayStation including the 1.5 T magnetic field.

Appendix A

Supplementary data to this article can be found online at https://doi.org/10.1016/j.phro.2026.100967.

1

Gy(RBE) according to the ICRU93 definition [21].

Appendix A. Supplementary data

The following are the Supplementary data to this article:

Supplementary Data 1
mmc1.docx (1.9MB, docx)

References

  • 1.Sun Hyun Bae, Seok Joo Chun, Joo Hyun Chung, Eunji Kim, Jin Kyu Kang, and Won Il Jang et al. Stereotactic body radiation therapy for hepatocellular carcinoma: Meta-analysis and international stereotactic radiosurgery society practice guidelines. IJROBP, 2024;18:337-351. doi: 10.1016/j.ijrobp.2023.08.015. [DOI] [PubMed]
  • 2.Prime S., Schiff J.P., Hosni A., Stanescu T., Dawson L.A., Henke L.E. The use of mr-guided radiation therapy for liver cancer. Semin Radiat Oncol. 2024;34:36–44. doi: 10.1016/j.semradonc.2023.10.006. [DOI] [PubMed] [Google Scholar]
  • 3.Bordeau K., Michalet M., Dorion V., Keskes A., Valdenaire S., Debuire P., et al. A prospective registry study of stereotactic magnetic resonance guided radiotherapy (mrgrt) for primary liver tumors. Radiother Oncol. 2023;189 doi: 10.1016/j.radonc.2023.109912. [DOI] [PubMed] [Google Scholar]
  • 4.Zaki P., Chuong M.D., Schaub S.K., Lo S.S., Ibrahim M., Apisarnthanarax S. Proton beam therapy and photon-based magnetic resonance image- guided radiation therapy: the next frontiers of radiation therapy for hepatocellular carcinoma. Technol Cancer Res Treat. 2023;22 doi: 10.1177/15330338231206335. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Nina N. Sanford, Jennifer Pursley, Bridget Noe, Beow Y. Yeap, Lipika Goyal, and Jeffrey W. Clark et al. Protons versus photons for unresectable hepatocellular carcinoma: Liver decompensation and overall survival. IJROBP, 2019;105:64-72. doi: 10.1016/j.ijrobp.2019.01. 076. [DOI] [PubMed]
  • 6.Francesco Dionisi, Alberto Brolese, Benedetta Siniscalchi, Irene Giacomelli, Francesco Frac- chiolla, and Roberto Righetto et al. Clinical results of active scanning proton therapy for primary liver tumors. Tumori, 2021;107:71-79. doi: 10.1177/0300891620937809. [DOI] [PubMed]
  • 7.Yusuke Uchinami, Norio Katoh, Daisuke Abo, Ryo Morita, Hiroshi Taguchi, and Yoshihiro Fu- jita et al. Study of hepatic toxicity in small liver tumors after photon or proton therapy based on factors predicting the benefits of proton. Br J Radiol, 2023;96(1144):20220720. doi: 10.1259/bjr.20220720. [DOI] [PMC free article] [PubMed]
  • 8.Albertini F., Matter M., Nenoff L., Zhang Y., Lomax A. Online daily adaptive proton therapy. Br J Radiol. 2020;93(1107) doi: 10.1259/bjr.20190594. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Bert C., Durante M. Motion in radiotherapy: particle therapy. Phys Med Biol. 2011;56:R113–R144. doi: 10.1088/0031-9155/56/16/R01. [DOI] [PubMed] [Google Scholar]
  • 10.Albertini F., Czerska K., Vazquez M., Andaca I., Bachtiary B., Besson R., et al. First clinical implementation of a highly efficient daily online adapted proton therapy (DAPT) workflow. Phys Med Biol. 2024;69 doi: 10.1088/1361-6560/ad7cbd. [DOI] [PubMed] [Google Scholar]
  • 11.Pham T.T., Whelan B., Oborn B.M., Delaney G.P., Vinod S., Brighi C., et al. Magnetic resonance imaging (mri) guided proton therapy: a review of the clinical challenges, potential benefits and pathway to implementation. Radiother Oncol. 2022;170:37–47. doi: 10.1016/j.radonc.2022.02.031. [DOI] [PubMed] [Google Scholar]
  • 12.Hoffmann A., Oborn B., Moteabbed M., Yan S., Bortfeld T., Knopf A., et al. Mr-guided proton therapy: a review and a preview. Radiat Oncol. 2020;15(1):129. doi: 10.1186/s13014-020-01571-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Keall P.J., Brighi C., Glide-Hurst C., Liney G., Liu P.Z.Y., Ly- S., et al. Integrated mri-guided radiotherapy — opportunities and challenges. Nat Rev Clin Oncol. 2022;19:458–470. doi: 10.1038/s41571-022-00631-3. [DOI] [PubMed] [Google Scholar]
  • 14.Nenoff L., Sudhyadhom A., Lau J., Sharp G.C., Paganetti H., Pursley Jennifer. Comparing predicted toxicities between hypofractionated proton and photon radiotherapy of liver cancer patients with different adaptive schemes. Cancers. 2023;15(18):4592. doi: 10.3390/cancers15184592. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Moteabbed M., Smeets J., Hong T.S., Janssens G., Labarbe R., Wolfgang J.A., et al. Toward MR-integrated proton therapy: modeling the potential benefits for liver tumors. Phys Med Biol. 2021;66(19) doi: 10.1088/1361-6560/ac1ef2. [DOI] [PubMed] [Google Scholar]
  • 16.Rabe M., Palacios M.A., van John R., de Koste Sörnsen, Eze Chukwuka, Hillbrand Martin, Belka Claus, et al. Comparison of mr-guided radiotherapy accumulated doses for central lung tumors with non-adaptive and online adaptive proton therapy. Med Phys. 2023;50:2625–2636. doi: 10.1002/mp.16319. [DOI] [PubMed] [Google Scholar]
  • 17.Schellhammer S.M., Hoffmann A.L., Gantz S., Smeets Julien, van der Kraaij Erik, Quets Sébastien, et al. Integrating a low-field open mr scanner with a static proton research beam line: proof of concept. Phys Med Biol. 2018;63 doi: 10.1088/1361-6560/aaece8. [DOI] [PubMed] [Google Scholar]
  • 18.van Herk M. Errors and margins in radiotherapy. Semin Radiat Oncol. 2004;14:52–64. doi: 10.1053/j.semradonc.2003.10.003. [DOI] [PubMed] [Google Scholar]
  • 19.Winkel D., Bol G.H., Kroon P.S., van Asselen B., Hackett S.S., Werensteijn-Honingh A.M., et al. Adaptive radiotherapy: the elekta unity mr-linac concept. Clin Transl Radiat Oncol. 2019;18:54–59. doi: 10.1016/j.ctro.2019.04.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Christian Rønn Hansen, Wouter Crijns, Mohammad Hussein, Linda Rossi, Pedro Gallego, and Wilko Verbakel et al. Radiotherapy treatment planning study guidelines (rating): A framework for setting up and reporting on scientific treatment planning studies. Radiother Oncol, 2020;153:67-78. doi: 10.1016/j.radonc.2020.09.033. [DOI] [PubMed]
  • 21.ICRU93. Icru report 93: Prescribing, recording, and reporting light ion beam therapy. J ICRU, 2020;16. doi: 10.1093/jicru_ndy025.
  • 22.Janson M., Glimelius L., Fredriksson A., Traneus E., Engwall E. Treatment planning of scanned proton beams in RayStation. Med Dosim. 2024;49:2–12. doi: 10.1016/j.meddos.2023.10.009. [DOI] [PubMed] [Google Scholar]
  • 23.Kierkels R.G.J., Hilgers G.C., Angerud A., Fälth C., Minken A. Beam modeling and validation for a 1.5 t mr-linac in an alternative treatment planning system. Phys Med Biol. 2025;70 doi: 10.1088/1361-6560/adcfed. [DOI] [PubMed] [Google Scholar]
  • 24.Horst F., Bodenstein E., Baumann M., Beyreuther E., Bokor J., Enghardt W., et al. The research beamlines at the Dresden proton therapy facility: available infrastructure and experimental capabilities. Front Oncol. 2025;15:1–10. doi: 10.3389/fonc.2025.1594973. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Jørgen B.B. Petersen, Yasmin Lassen, Anders T. Hansen, Ludvig P. Muren, Cai Grau, and Morten Høyer. Normal liver tissue sparing by intensity-modulated proton stereotactic body radiotherapy for solitary liver tumours. Acta Oncol, 2011;50:823–828. doi: 10.3109/0284186X.2011.590526. [DOI] [PubMed]
  • 26.Mondlane G., Ureba A., Michael P.A., Lind Gubanski, Siegbahn Albert. Estimation of the risk for radiation-induced liver disease following photon- or proton-beam radio-surgery of liver metastases. Radiat Oncol. 2018;13:1–9. doi: 10.1186/s13014-018-1151-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Korevaar E.W., Habraken S.J.M., Scandurra D., Kierkels R.G.J., Unipan M., Eenink M.G.C., et al. Practical robustness evaluation in radiotherapy – a photon and proton-proof alternative to PTV-based plan evaluation. Radiother Oncol. 2019;141:267–274. doi: 10.1016/j.radonc.2019.08.005. [DOI] [PubMed] [Google Scholar]
  • 28.Guerreiro F., Svensson S., Seravalli E., Traneus E., Raaymakers B.W. Intra-fractional per-beam adaptive workflow to mitigate the need for a rotating gantry during mri-guided proton therapy. Phys Med Biol. 2021;66(18) doi: 10.1088/1361-6560/ac176f. [DOI] [PubMed] [Google Scholar]
  • 29.Lennart Volz, James Korte, Maria Chiara Martire, Ye Zhang, Nicholas Hardcastle, and Marco Durante et al. Opportunities and challenges of upright patient positioning in radio- therapy. Phys Med Biol, 2024;69. doi: 10.1088/1361-6560/ad70ee. [DOI] [PubMed]
  • 30.Yan S., Lu H.M., Flanz J., Adams J., Trofimov A., Bortfeld T. Reassessment of the necessity of the proton gantry: Analysis of beam orientations from 4332 treatments at the massachusetts general hospital proton center over the past 10 years. Int J Radiat Oncol Biol Phys. 2016;95(1):224–233. doi: 10.1016/j.ijrobp.2015.09.033. [DOI] [PubMed] [Google Scholar]
  • 31.Lukas Nierer, Chukwuka Eze, Vanessa da Silva Mendes, Juliane Braun, Patrick Thum, and Rieke von Bestenbostel et al. Dosimetric benefit of MR-guided online adaptive radiotherapy in different tumor entities: liver, lung, abdominal lymph nodes, pancreas and prostate. Radiat Oncol, 2022;17:1-14. doi: 10.1186/s13014-022-02021-6. [DOI] [PMC free article] [PubMed]
  • 32.Mayinger M., Ludwig R., Christ S.M., Dal Bello R., Ryu A., Weitkamp N., et al. Benefit of replanning in MR-guided online adaptive radiation therapy in the treatment of liver metastasis. Radiat Oncol. 2021;16:1–8. doi: 10.1186/s13014-021-01813-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Petra Trnkova, Ye Zhang, Toshiyuki Toshito, Ben Heijmen, Christian Richter, and Marianne C. Aznar et al. A survey of practice patterns for adaptive particle therapy for interfractional changes. Phys Imaging Radiat Oncol, page 100442, 2023;26:100442. doi: 10.1016/j.phro.2023. 100442. [DOI] [PMC free article] [PubMed]
  • 34.Jeremy de Leon, Tania Twentyman, Madeline Carr, Michael Jameson, and Vikneswary Ba- tumalai. Optimising the MR-Linac as a standard treatment modality. J Med Radiat Sci, 2023;70:491–497. doi: 10.1002/jmrs.712. [DOI] [PMC free article] [PubMed]
  • 35.Van Timmeren J.E., Chamberlain M., Krayenbuehl J., Wilke L., Ehrbar S., Bogowicz M., et al. Treatment plan quality during online adaptive re-planning. Radiat Oncol. 2020;15(1):203. doi: 10.1186/s13014-020-01641-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Nenoff L., Ribeiro C.O., Matter M., Hafner L., Josipovic M., Langendijk Johannes A., et al. Deformable image registration uncertainty for inter-fractional dose accumulation of lung cancer proton therapy. Radiother Oncol. 2020;147:178–185. doi: 10.1016/j.radonc.2020.04.046. [DOI] [PubMed] [Google Scholar]
  • 37.Lebbink F., Stock M., Georg D., Knäusl B. The influence of motion on the delivery accuracy when comparing actively scanned carbon ions versus protons at a synchrotron-based radiotherapy facility. Cancers. 2022;14:1788. doi: 10.3390/cancers14071788. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Ye Zhang, Petra Trnkova, Toshiyuki Toshito, Ben Heijmen, Christian Richter, and Mar- ianne Aznar et al. A survey of practice patterns for real-time intrafractional motion- management in particle therapy. Phys Imaging Radiat Oncol, 2023;26:100439. doi: 10.1016/ j.phro.2023.100439. [DOI] [PMC free article] [PubMed]
  • 39.Padre K.G., Lebbink F., Cobanaj M., Traneus E., Schürer M., Fuchs H., et al. validated energy layer-wise beam deflection compensation method for scanned proton beams in the static magnetic field of an in-beam mr scanner. Radiother Oncol. 1880;2025(206):S3478–S3480. doi: 10.1016/S0167-8140(25)00707-8. ESTRO 2025, 2-6 May 2025, Vienna, Austria. [DOI] [Google Scholar]
  • 40.Marisa Cobanaj, Erik Traneus, Krishna Godino Padre, Felix Horst, Aswin Hoffmann, and Franciska Lebbink. O081 / 155 - spot-wise compensation of proton beam deflection in the magnetic field of an in-beam mr scanner using raystation treatment planning system. Int J Part Ther, 2025;17:100837. doi: 10.1016/j.ijpt.2025.100837. PTCOG 63.
  • 41.Guerreiro F., Koivula L., Seravalli E., Janssens G.O., Maduro J.H., Brouwer C.L., et al. Feasibility of mri-only photon and proton dose calculations for pediatric patients with abdominal tumors. Phys Med Biol. 2019;64 doi: 10.1088/1361-6560/ab0095. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Data 1
mmc1.docx (1.9MB, docx)

Articles from Physics and Imaging in Radiation Oncology are provided here courtesy of Elsevier

RESOURCES