Volumetric modulated arc therapy versus intensity-modulated proton therapy in the postoperative irradiation of thymoma

Abstract

Background

To investigate the role of intensity-modulated proton therapy (IMPT) compared to volumetric modulated arc therapy (VMAT) for the radiation treatment of thymoma cancer.

Methods

Twenty patients were retrospectively planned for IMPT [with (IMPT_R1 or IMPT_R2 according to the approach adopted) and without robust optimization] and VMAT. The results were compared according to dose-volume metrics on the clinical and planning target volumes (CTV and PTV) and the main organs at risk (heart, breasts, lungs, spinal cord and oesophagus). Estimates of the excess absolute risk (EAR) of secondary cancer induction were determined for the oesophagus, the breasts and the composite lungs. For the heart, the relative risk (RR) of chronic heart failure (CHF) was assessed.

Results

IMPT and VMAT plans resulted equivalent in terms of target coverage for both the CTV and the PTV. The CTV homogeneity index resulted in 0.03 ± 0.01 and 0.04 ± 0.01 for VMAT and all IMPT plans, respectively. The conformality index resulted in 1.1 ± 0.1 and 1.2 ± 0.1 for VMAT and all IMPT plans. The mean dose to the breasts resulted in 10.5 ± 5.0, 4.5 ± 3.4, 4.7 ± 3.5 and 4.6 ± 3.4 Gy for VMAT, IMPT, IMPT_R1 and IMPT_R2. For the lungs, the mean dose was 9.6 ± 2.3, 3.5 ± 1.5, 3.6 ± 1.6 and 3.8 ± 1.4 Gy; for the heart: 8.7 ± 4.4, 4.3 ± 1.9, 4.5 ± 2.0 and 4.4 ± 2.4 Gy and for the oesophagus 8.2 ± 3.5, 2.2 ± 3.4, 2.4 ± 3.6 and 2.5 ± 3.5 Gy. The RR for CHF was 1.6 ± 0.3 for VMAT and 1.3 ± 0.2 for IMPT (R1 or R2). The EAR was 3.6 ± 0.v vs 1.0 ± 0.6 or 1.2 ± 0.6 (excess cases/10,000 patients year) for the oesophagus; 17.4 ± 6.5 vs 5.7 ± 3.2 or 6.1 ± 3.8 for the breasts and 24.8 ± 4.3 vs 8.1 ± 2.7 or 8.7 ± 2.3 for the composite lungs for VMAT and IMPT_R, respectively.

Conclusion

The data from this in-silico study suggest that intensity-modulated proton therapy could be significantly advantageous in the treatment of thymoma patients with particular emphasis to a substantial reduction of the risk of cardiac failure and secondary cancer induction. Robust planning is a technical pre-requisite for the safety of the delivery.

Background

Thymoma, a neoplasm arising from epithelial cells of the thymus, is a rare tumour with a yearly incidence of 0.17 per 100,000 in Europe (Siesling et al. 2012). Nevertheless, it is the most common malignancy of the anterior mediastinum, representing 20–50% of all mediastinal tumours (Engels 2010). Surgical resection is currently the standard of care for localized thymoma (Girard et al. 2015), and the extent of resection has been repeatedly shown to be an independent predictor of improved outcome (Maggi et al. 1991; Song and Zhang 2014; Hishida et al. 2016; Ahmad et al. 2015). Due to the proximity of the thymic lodge to critical mediastinal structures, the rate of R0 resection may vary in function of the extent of disease. In particular, clear margins are obtained only in 50 and 25% of cases for Masaoka stage III and IV thymomas, respectively (Detterbeck et al. 2011). Moreover, even in the presence of R0 resection, the risk of local relapse may be increased in case of capsular effraction (Ruffini et al. 1997). For this reason, despite a lack of randomized studies for this rare tumour, postoperative radiotherapy (PORT) has been proposed as adjuvant treatment in resected thymoma Masaoka stage II or higher to increase disease control based on conflicting data from multiple meta-analyses (Jackson et al. 2017; Zhou et al. 2016; Lim et al. 2016; Korst et al. 2009) and is currently integrated into the treatment algorithm proposed by national and international scientific societies, with doses ranging from 50 to 54 up to 60–70 Gy in case of complete resection, microscopically positive margins and gross residual disease, respectively (Girard et al. 2015; Imbimbo et al. 2018; Gomez et al. 2011). The radiation treatment approach is conventionally based on 3D-conformal radiotherapy. Still, advanced techniques including intensity-modulated therapy (IMRT) with image guidance and possibly respiratory motion management are advisable to optimize target definition and reduce healthy tissue involvement (Girard and Mornex 2011; Gomez and Komaki 2010; Komaki and Gomez 2014; Giannopoulou et al. 2013).

Due to the proximity of the thymic lodge to critical tissues (such as spinal cord, heart, lungs and oesophagus), there is concern about possible treatment-related toxicity. In particular late cardiovascular sequelae, whose risk may increase over time in consideration of the expected survival in these patients: previous experiences showed a 7% increase in the incidence of severe cardiac events per 1 Gy to the mean dose to the heart (Darby et al. 2013). In a cohort of 74 consecutive thymoma patients (Liao et al. 2018), treated with complete tumour resection and postoperative radiation (74), cardiovascular disease was the leading nonmalignant cause of death (50% of the deaths) with a median time for diagnosis of 101 months after treatment. The mean dose to the heart was found to be an independent risk factor.

It is also noteworthy that, while the incidence of thymoma peaks in late adulthood, increased risk of secondary cancers (in particular oesophagal and lung cancer, Non-Hodgkin lymphoma and soft tissue sarcoma) has been previously reported in patients with thymoma as a possible result of combined immune dysregulation, genetic and environmental factors (Kumar et al. 2018; Travis et al. 2003; Pan et al. 2001). This has important implications, since increased dose exposure to healthy tissue may on one hand increase tumorigenesis due to cumulative genotoxic damage, and on the other hand reduce the possibility to efficiently deliver a second course of radiation therapy in case of treated volume overlap due to increased risk of treatment-related adverse events.

While technical advances such as intensity-modulated radiotherapy (IMRT) allows more conformal dose distribution to the target regions (Willmann and Rimner 2018), there are still limitations related to the ballistic properties of photons that hinders adequate dose sparing of critical structures. Proton therapy has been proposed to improve the therapeutic ratio of PORT. It might reduce the radiation dose to the healthy tissues (Zhu et al. 2018), thus decreasing the risk of acute toxicity (Parikh et al. 2016) and the occurrence of second cancers (Vogel et al. 2017).

A dosimetric planning study was reported by Haefner et al. (2018) with the comparison of 3D-conformal therapy, volumetric modulated arc therapy (VMAT), tomotherapy, proton therapy and carbon ion therapy.

The clinical use of protons for thymoma and thymic carcinoma is still episodically reported. Ten patients were panned for 50 Gy in 25 fractions, and the results showed that particle therapy allowed a significant reduction of the mean doses to the lungs, breasts, heart and oesophagus. No outcome nor complication models were included in the study.

Parikh et al. (2016) compared proton beam therapy (PBT) with photon IMRT over a group of only four patients treated with PBT. Compared to photons, PBT was associated with lower mean doses to the various critical structures (heart, lung and oesophagus) while no difference was observed in the breast.

In 2005, a Swedish initiative (Björk-Eriksson et al. 2005) aiming to determine the number of patients suitable for proton beam therapy identified thymomas as potential candidates given the possibility to decrease the dose to the heart and the lungs. The total number of potential patients resulted, nevertheless quite low. Limited clinical evidence is available about the use of particle therapy and protons in particular for thymomas.

Vogel et al. (2016) reported the outcome of a cohort of 27 patients treated with passive double-scattering proton beam therapy. With a median follow-up of 2 years, 100% local control was achieved, at 3 years, the overall survival rate was of 94%. Modest to moderate toxicity was reported with no grade 3 or higher cases. The same group also reported about the predicted rates of secondary malignancies from PBT versus IMRT (Vogel et al. 2017). Ten patients treated with the passive scattering method were compared dosimetrically to IMRT. Authors concluded that five excess secondary malignancies per 100 patients could be avoided using proton therapy. The calculations were performed with the linear-exponential (the “bell-shaped”) model which includes cell killing but neglects repopulation and repair effects.

Zhu reported a series of 6 patients with stage II–III thymic cancer treated with protons and compared the plans against IMRT (Zhu et al. 2018). Concerning clinical outcome, median follow-up of 2.6 years, three patients demonstrated recurrences and no severe toxicity were observed. The mean dose to the heart was sharply reduced (from 33 to 60%) for the heart, lung and oesophagus in the comparison against IMRT.

More recently, in 2019, Mercado et al. (2019) described the patterns-of-care and the early clinical outcome based on prospective registries for a cohort of 30 patients. Results demonstrated a good toxicity profile (no grade 3 or higher reported) with acceptable rates of recurrence. The short median follow-up of 13 months prevented a more in-depth assessment of the outcome.

In the absence of planning studies and clinical evidence, the primary aim of this in-silico planning study was to investigate the relative figure of merit of IMPT versus VMAT for thymoma cancer patients on a relatively large cohort compared to earlier investigations. The focus was set to the assessment of several appropriate dose-volume metrics for cardiac, abdominal and lung structures. The more novel aspect was the inclusion in the study of the estimation the risk of severe cardiac complications as chronic heart failure (CHF) according to the most recent models (Darby et al. 2013; Nimwegen et al. 2016, 2018) and the estimate of the excess absolute risk (EAR) of secondary cancer induction with the so-called full model (Preston et al. 2007; Schneider et al. 2011a, 2011b) which includes cell killing, repopulation and repair and fractionation effects. Calculations have been done for the oesophagus, the breasts (for female patients) and the lungs.

Materials and methods

Patients selection, contouring and dose prescription

Twenty patients treated with PORT for completely resected (R0) thymoma (Masaoka stage II and III) were selected for this retrospective in-silico study. For all patients, the clinical target volume (CTV) was delineated on a 3 mm-thick slices CT (acquired in free-breathing as defined by the institutional clinical practice). It included the thymic lodge, the tumour bed (using preoperative imaging coregistration), the surgical clips and any potential sites of residual disease (i.e. pericardium or pleural sheet in contact with the tumour bulk). Planning treated volume (PTV) was obtained through 5 mm isotropic expansion of the CTV.

The same CTV was used for both VMAT and IMPT planning. The PTV was used for VMAT planning and reporting purposes for both techniques. For the optimisation of the IMPT plans, the PTV or the relative target volume (RTV) were used as described below.

The dose prescription was set to 50 Gy in 25 fractions for all the patients. The normalization was set to the mean dose to the PTV for all plans for comparison reasons. All the planning aims for the targets and the organs at risk are summarized with the results in Tables 1 and 2. The organs at risk considered in the study were the lungs, the whole heart, the spinal cord, the oesophagus and the breast for the female patients.

Table 1.

Summary of the planning objectives and average results (uncertainty expressed as 1 standard deviation) for the clinical target volume (CTV) and for the planning target volume (PTV)

Parameter	Objective	VMAT	IMPT	IMPT R1	IMPT R2	p
CTV Volume: 126 ± 69 Range: [30–306] cm3
Mean [Gy]	50	50.5 ± 0.5	50.1 ± 0.1	50.1 ± 0.1	50.3 ± 0.7	–
D95% [Gy]	>48.2 (98%)	49.6 ± 0.2	49.2 ± 0.2	49.3 ± 0.2	49.3 ± 0.2	b, c, f
D98% [Gy]	>47.5 (95%)	49.4 ± 0.3	48.8 ± 0.3	49.0 ± 0.3	49.0 ± 0.3	a, b, c, d, f
D1% [Gy]	<53.5 (107%)	51.7 ± 0.4	51.7 ± 0.6	52.2 ± 0.7	52.5 ± 0.9	a, c, d, f
SD [Gy]	Minimize	0.5 ± 0.1	0.6 ± 0.2	0.6 ± 0.2	0.7 ± 0.2	a, b, c, d, f
HI	Minimize	0.03 ± 0.01	0.04 ± 0.01	0.04 ± 0.01	0.04 ± 0.01	b, c, f
PTV Volume: 319 ± 124 Range: [122–556] cm3
Mean [Gy]	50	50.0 ± 0.0	50.0 ± 0.0	50.0 ± 0.0	50.0 ± 0.0	–
D95% [Gy]	>47.5 (95%)	47.6 ± 0.5	48.0 ± 0.7	47.7 ± 0.7	47.1 ± 1.0	b
D98% [Gy]	>45.0 (90%)	46.7 ± 0.6	46.6 ± 1.2	46.2 ± 1.0	45.3 ± 1.6	c, e, f
D2% [Gy]	<53.5 (107%)	51.9 ± 0.3	52.5 ± 0.7	52.8 ± 0.8	52.9 ± 0.9	b, c, f
SD [Gy]	Minimize	1.2 ± 0.2	1.2 ± 0.4	1.4 ± 0.4	1.6 ± 0.5	c, e, f
HI	Minimize	0.08 ± 0.01	0.07 ± 0.01	0.08 ± 0.01	0.09 ± 0.01	e

D_x% = dose received by at least (maximum) x% of the volume, HI = (D₅−D₉₅)/D_mean. a: IMPT vs IMPT R1; b: IMPT vs VMAT; c: IMPT R1 vs VMAT; d: IMPT vs IMPT R2; e: IMPT R1 vs IMPT R2; f: IMPT R2 vs VMAT

VMAT volumetric modulated arc therapy (in the RapidArc form), IMPT intensity-modulated proton therapy

Table 2.

Summary of the planning objectives and average results (uncertainty expressed as 1 standard deviation) for the main organs at risk investigated in the study

Parameter	Objective	VMAT	IMPT	IMPT R1	IMPT R2	p
Breasts
Mean [Gy]	<5 Gy	10.5 ± 5.0	4.5 ± 3.4	4.7 ± 3.5	4.6 ± 3.4	a, b, c, d, e, f
D1% [Gy]	Minimize	39.3 ± 12.5	38.5 ± 16.8	38.4 ± 16.7	38.5 ± 16.7	ns
Lungs
Mean [Gy]	<10 Gy	9.6 ± 2.3	3.5 ± 1.5	3.6 ± 1.6	3.8 ± 1.4	a, b, c, d, e, f
V20Gy [%]	<20%	13.8 ± 6.2	6.4 ± 2.9	6.6 ± 3.1	6.5 ± 2.9	b, c, e, f
V5Gy [%]	<55–60%	53.5 ± 10.7	13.7 ± 5.8	14.5 ± 6.1	15.7 ± 4.8	a, b, c, d, e, f
Heart
Mean [Gy]	<5 Gy	8.7 ± 4.4	4.3 ± 1.9	4.5 ± 2.0	4.4 ± 2.4	a, b, c, d, e, f
D1% [Gy]	Minimize	47.4 ± 8.3	48.2 ± 5.6	48.2 ± 5.6	48.1 ± 5.7	ns
Spinal cord
D1% [Gy]	Minimize	9.8 ± 2.6	1.0 ± 2.5	1.0 ± 2.51	1.2 ± 1.8	b, c, e, f
Oesophagus
Mean [Gy]	Minimize	8.2 ± 3.5	2.2 ± 3.4	2.4 ± 3.6	2.5 ± 3.5	a, b, c, d, e, f
D1% [Gy]	Minimize	28.6 ± 13.0	16.2 ± 15.6	16.7 ± 15.8	17.4 ± 15.6	b, c, e, f
Healthy tissue
V10Gy [%]	Minimize	9.8 ± 2.6	7.9 ± 21.7	9.8 ± 2.6	9.7 ± 3.1	a, b, c, d, f
CI95%	1.0	1.1 ± 0.1	1.2 ± 0.1	1.2 ± 0.2	1.2 ± 0.2	a, b, c, d, f

D_x% = dose received by at least (maximum) x% of the volume; V_xGy = volume receiving xGy. a:IMPT vs IMPT R1; b: IMPT vs VMAT; c: IMPT R1 vs VMAT; d: IMPT vs IMPT R2; e: IMPT R1 vs IMPT R2; f: IMPT R2 vs VMAT

VMAT volumetric modulated arc therapy (in the RapidArc form), IMPT intensity-modulated proton therapy, prv planning risk volume, CI conformality index

Photon planning

The photon plans used for the study were designed and optimized according to the VMAT technique in its RapidArc form with flattening filter-free (FFF) photon beams (beam quality of 6MV) from a TrueBeam linear accelerator (Varian Medical Systems, Palo Alto, USA). Optimization was performed using the Eclipse treatment planning system Photon Optimiser algorithm, and the final dose was recalculated with the Acuros-XB engine with a calculation grid of 2.5 mm (version 16.0). All plans were optimised according to a class solution consisting of two partial arcs with collimator angle set to 45–315° and start and stop angles set to 140° and 220° (with clock-wise and counter clock-wise rotation for the arcs). The field size was arranged to not exceed 14.0 cm in the x-direction to avoid loss in modulation power due to the travel limits of the multileaf collimator.

Proton planning

The ProBeam proton system (Varian Medical Systems, Palo Alto, USA) was used as a source of beam data for the optimization and calculation of intensity-modulated proton therapy (IMPT) plans using the beam spot scanning technique. The dose distribution optimization was performed using the Nonlinear Universal Proton Optimiser (NUPO, v16.0) while for the final dose calculation, the Proton Convolution Superposition algorithm (v16.0) was applied using a grid size of 2.5 mm and a constant relative biological effectiveness RBE of 1.1, the only option available today in the treatment planning system utilised.

All patients were planned with a standardised class solution with two anterior oblique fields with gantry angles set to 25° and 315°. Small, individualized gantry angles tuning was allowed, according to the target position, to minimize the healthy tissue involvement. The multifield optimisation technique was applied to all cases.

A standard IMPT set was optimised from an initial cloud of pencil spots defined from the PTV (the same as for VMAT) with a craniocaudal and lateral expansion of 3 mm.

A second set of plans was designed employing the robust optimization technique (IMPT_R1) to account for setup and range uncertainties considering ± 4 mm shifts in the isocentre along each axis and ± 3% in beam range. The robust plans were optimised by applying the uncertainties to the CTVs starting from the same cloud of spots as defined for the standard IMPT. This method is the most straightforward approach to robust optimisation, although relatively crude.

A third, more clinically sound, group of plans was robustly optimised (IMRT_R2) generating the starting set of spots from a relative target volume. The RTV was automatically generated, field by field, from the CTV in the TPS accounting for the range and position uncertainties (the same as above) and a further (smeared) extension of 3 mm in the craniocaudal and lateral directions. This method should be considered as more reliable, particularly when lung tissue is involved.

The robust optimisation engine implemented in the planning system is based on the original investigation of Liu et al. (2012) and expanded, as described in Li et al. (2015).

Three sets of IMPT plans were, therefore, optimised for each patient, and the results aggregated per each set were reported.

Quantitative assessment of dose-volume metrics

The mean dose and a variety of V_x and D_x parameters [V_x represents the volume receiving at least an x level of dose (in % or in Gy), and D_x is the minimum dose that covers an x fraction of volume (in % or in cm³)] (AA.VV 2010) were derived from the dose-volume histograms (DVH) and used as quantitative metrics.

For the CTV and the PTV, the variance of the dose distribution was reported as the standard deviation of the differential histogram and as the homogeneity index (HI) defined as HI = (D_5% − D_95%)/D_mean. The dose conformality was scored with the conformity index, CI 95%, defined as the ratio between the patient volume receiving at least 95% of the prescribed dose and the PTV volume. The average DVHs were computed, for each structure and each cohort, with a dose binning resolution of 0.02 Gy. Proton doses are reported in Cobalt equivalent Gy (corrected for the 1.1 RBE factor).

The significance of the observed differences was determined using the Wilcoxon matched-paired signed-rank test. The SPSS package version 22 (IBM Corporation) was used for the study.

Modelling the risk of toxicity and secondary cancer induction

The relative risk (RR) estimation for the coronaries and for the left ventricle chamber for disease/failure was performed according to the linear model proposed by van Nimwegen et al. (2016, 2018) for Hodgkin lymphoma patients (following the observations of Darby et al. (2013) for breast cancer patients) who correlated coronary heart disease to the mean dose to the heart. The excess relative risk was fixed to 7.4 and 9.0% per Gy, respectively, for the coronaries and for the left ventricle.

Concerning the risk of secondary malignancy induction, the excess absolute risk (EAR) for a whole specific organ (org) was estimated according to the methods described in Preston et al. (2007). In brief, it is expressed by:

$${\mathrm{E}\mathrm{A}\mathrm{R}}^{\mathrm{o}\mathrm{r}\mathrm{g}}=\mu \frac{1}{V_{\mathrm{T}}}\sum _{i}V\left(D_i\right)\mathrm{R}\mathrm{E}\mathrm{D}\left(D_i\right),$$

where V_T is the total organ volume. The sum is over all the bins of the differential DVH, V(D_i) is the absolute volume receiving a dose D_i. µ is the slope of cancer induction based on the atomic bomb survivors’ data (Schneider et al. 2011a) corrected for the age distribution. The values used in this analysis were: 5.0, 3.8, 3.2 and 0.4 for breast, lung, oesophagus and thyroid. RED(D) is the dose–response, which has been modelled using different approaches to fit the Hodgkin’s patient’s data group (Schneider et al. 2011a). The organ equivalent dose (OED): $\mathrm{O}\mathrm{E}\mathrm{D}=\frac{1}{V_{\mathrm{T}}}{\sum }_{i}V\left(D_i\right)\mathrm{R}\mathrm{E}\mathrm{D}\left(D_i\right)$ was introduced as the dose in gray, which, when uniformly distributed across the organ, causes the same radiation-induced cancer incidence.

Different models were published to calculate OED. In the present study, the so-called full model was adopted (Schneider et al. 2011b).

$$\mathrm{O}\mathrm{E}\mathrm{D}=\frac{1}{V_{\mathrm{T}}}\sum _{i}V\left(D_i\right)\frac{e^{-{\alpha }^{\prime }{D}_i}}{{\alpha }^{\prime }R}\left(1,-,2,R,+,R^2,e^{{\alpha }^{\prime }{D}_i},-,{\left(1,-,R\right)}^2,e^{-\frac{{\alpha }^{\prime }R}{1-R}{D}_i}\right),$$

where R is the parameter accounting for repopulation and/or repair and models the ability of the tissue to recover between two dose fractions (R = 0 means no recovery, R = 1 full recovery). This model fully includes all the biological aspects of cell killing, repopulation/repair, and fractionation. The used value for R was 0.62, 0.83, 0.85 for the breasts, lungs and oesophagus; the used value for α was 0.067, 0.041 and 0.46 Gy⁻¹, respectively. α/β is the standard LQ parameter: α/β = 3 Gy.

Results

The dose-volume histograms averaged over all the patients are shown in Fig. 1 for the four sets of plans and the various structures. Both VMAT and IMPT plans resulted equivalent for the CTV and the PTV while for all the OARs a macroscopic sparing was observed in the low-medium dose ranges up to the 20–25 Gy range. Above this band, the dose distributions between photons and protons resulted more similar, although significant differences could be observed for the oesophagus. Due to the geometry of the beam arrangement, the spinal cord resulted in uninvolved in the case of protons. No remarkable difference was associated with the use of robust optimization for IMPT.

Fig. 1.

Average dose-volume histograms for the target volumes and the main organs at risk investigated

The summary of the numerical analysis for the target volumes is presented in Table 1. Both VMAT and IMPT plans met the planning aims for all the metrics. No difference, although some statistical significance was detected, shall be outlined as potentially relevant from a clinical perspective.

The variance of the dose distributions resulted in 3–4% for the CTV and of 7–8% for the PTV. Also, the conformality resulted remarkably good (1.1 for photons and 1.2 for protons). The better result for VMAT is linked to the sharper fall-off achievable with FFF photon beams compared to protons.

Table 2 reports the results of the dosimetric analysis carried out on the various OARs. As anticipated from Fig. 1, for all structures, the mean doses resulted significantly and largely improved with IMPT compared to VMAT as well as the V_5Gy and the V_20Gy for the lungs. The near to maximum dose for the breasts and the heart did not differ between the techniques while IMPT allowed for a further sparing of ~ 12 Gy compared to VMAT for the oesophagus near-to-maximum D_1% metric.

The RR estimates for CHF are reported in Table 3 as mean values (with uncertainty at 1 standard deviation) and median value. The comparison is presented only between VMAT and IMPT_R1 and IMPT_R2 (similar results would be achieved for the non robustly optimised plans). IMPT (equivalently R1 or R2) resulted in a 0.3 net reduction of the RR, which corresponds to ~ 19% reduction compared to the VMAT baseline. The difference resulted statistically highly significant.

Table 3.

Estimates of the relative risk of cardiac failure for VMAT and IMPT R1 and IMPT R2 datasets

Organ	VMAT	IMPT R1	IMPT R2	Δ (VMAT-IMPT R1)	Δ (VMAT-IMPT R2)	p (VMAT-IMPT R1)	p (VMAT-IMPT R2)
Heart	1.6 ± 0.3 1.7	1.3 ± 0.2 1.3	1.3 ± 0.1 1.3	0.3 ± 0.2 0.3	0.3 ± 0.3 0.3	< 0.001	< 0.001

Results are shown as averages (with uncertainty expressed as 1 standard deviation) and median value (second line). The p value is relative to the Wilcoxon’s signed-rank paired test

RR relative risk, VMAT volumetric modulated arc therapy (in the RapidArc form), IMPT R1 (R2) intensity-modulated proton therapy with robust optimization

The estimates of the EAR per 10,000 patients-years of secondary cancer induction are presented graphically in Fig. 2 and numerically in Table 4. In all cases, IMPT resulted in a greatly reduced risk, as reported in the table. The risk reduction resulted in about 72–67%, 67–64% and 67–65% for the oesophagus, the breasts and the lungs and IMPT_R1 or IMPT_R2, respectively. All differences were statistically highly significant.

Fig. 2.

Estimates of the Excess absolute risk (EAR) (per 10,000 patient-years) of secondary cancer induction estimated with the full model for the oesophagus, the breasts and the combined lungs

Table 4.

Estimates of the excess absolute risk (EAR) (per 10,000 patient-years) of secondary cancer induction estimated with the full model for the oesophagus, the breasts and the combined lungs

	VMAT	IMPT R1	IMPT R2	Δ (VMAT-IMPT R1)	Δ (VMAT-IMPT R2)	P (VMAT-IMPT R1)	P (VMAT-IMPT R2)
Oesophagus	3.6 ± 0.4 3.7	1.0 ± 0.6 1.1	1.2 ± 0.6 1.3	2.6 ± 0.7 2.6	2.4 ± 0.7 2.3	<0.001	<0.001
Breasts	17.4 ± 6.5 19.4	5.7 ± 3.2 7.1	6.1 ± 3.8 7.1	11.7 ± 4.7 10.8	11.2 ± 4.4 10.8	<0.001	<0.001
Lungs	24.8 ± 4.3 25.9	8.1 ± 2.7 7.5	8.7 ± 2.3 8.6	16.7 ± 2.7 17.0	16.1 ± 2.7 16.7	<0.001	<0.001

Results are shown as averages (with uncertainty expressed as 1 standard deviation). The p value is relative to the Wilcoxon’s signed-rank paired test

VMAT volumetric modulated arc therapy (in the RapidArc form), IMPT R1 (R2) intensity-modulated proton therapy with robust optimization

Discussion

The present study analysed the role of VMAT and IMPT for advanced stage postoperative thymoma cancer patients. The sample size (20 cases) is adequate for an in-silico planning study and, given the relatively low incidence, it is comparable to some of the published clinical studies.

From a dosimetric point of view, this study expands the current knowledge in the field by comparing two techniques not investigated in parallel so far. The dosimetric advantage that protons usually offers for the sparing of organs at risk was also confirmed in this investigation without any detriment to the coverage of the target. Further factors of innovation of the present report are the inclusion of model-based predictions for radiation-induced cardiac failure and the estimate of the risk of secondary cancer induction. From a practical point of view, for both VMAT and IMPT it was possible to achieve high-quality plans with a limited inter-patient variance with the application of simple class solutions for the beam geometries and standardized clinical aims for the optimization phase. The relevance of this result is in the possibility to streamline the pattern of care and harmonise the treatment approach among the entire patient’s population also for these highly sophisticated treatment techniques.

An issue in comparative studies between photons and protons is the possible bias derived from the use of a constant RBE in the proton calculations. A variable RBE would be more realistic and might impact on both tumour control (TCP) and normal tissue complication (NTCP) probabilities. As pointed out by Jones (2017), RBE could be in the range of 1.1–1.5 or higher depending on the type of tumour or normal tissues involved in the irradiation. Similarly, McNamara et al. (2020), reviewing the data and the models for variable RBE, concluded that the notion of fixed RBE is too simplistic and might raise concerns for treatment planning decisions. We acknowledge this as one of the limiting elements of our present study, and we would consider the inclusion of RBE effects as a fundamental future improvement in proton treatment planning systems.

The dosimetric advantage of protons was translated into a remarkable reduction of the risk of toxicity.

Before discussing the potential improvements, it is essential to mention that those studies were based on photon treatments and, therefore, the models might not be strictly applicable also to protons. In the absence of clear evidence about this, we opted for the application of the same models but acknowledged the potential bias.

The determination of the EAR was performed by means of the full model which incorporates repair, repopulation ad fractionation mechanism in an attempt to provide the most comprehensive modelling of the risk. The absolute value of the predictions might still be impacted by the uncertainty on the computational parameters, but the relative difference between techniques should be free from this potential bias.

The estimates of EAR in our study are hardly comparable with the work of Vogel et al. (2017) because of the different model used (linear-exponential versus full) and the sample size. Nevertheless, within the broad uncertainty in the accuracy of these models, the present study reports the first assessment of EAR for lungs, breasts and oesophagus in thymoma patients for VMAT and IMPT with spot scanning technique. Beside eventual bias in the absolute values, the significant difference between photons and protons is suggestive of the relevance of IMPT. Concerning cardiac toxicity, the work of Liao et al. (2018) demonstrated that cardiovascular disease was the leading cause of non-cancer related death and that the mean dose to the heart was a clear predictor. In our study, we reinforced the assessment with the estimation of the RR according to the van Nimwegen and Darby models (Darby et al. 2013; Nimwegen et al. 2016, 2018) and, as for EAR, the definite relative difference reported between VMAT and IMPT enhances the decisive potential clinical role of the latter technique. The possibility to segment cardiac sub-structures like the coronaries or the left-ventricle would be possible either with the use of dedicated cardiac scans (not available for the present patient population) or employing atlas-based contouring (Feng et al. 2011) and done in other investigations (Levis et al. 2019). We acknowledge this as a potential limitation of the study, but the results from the whole heart RR estimate are already strongly supportive of IMPT.

Robust optimization was adopted in this study (and directly compared against plans optimized without it). Concerning the robust optimisation, no clinically alarming differences were reported between IMPT_R1 and IMPT_R2 (also with respect to IMPT). The choice of 3% uncertainty in the range calibration and 4 mm in the isocenter position is somehow arbitrary but seems realistic from routine clinical practice. Robust optimization should be an essential element in the IMPT planning process. The RTV method should be considered as a reference.

Concerning motion management, no further elements were introduced in this planning study but should be taken into account for clinical treatments. In particular, deep inspiration breath-hold (DIBH) is frequently used in clinical practice (e.g. in the treatment of breast) to reduce the dose to the heart. DIBH was not included in the present study for two main reasons. Firstly, patients with thymic tumours patients are expected to display lower compliance with this procedure due to older median age and prior major chest surgery. Secondly, due to anatomical superposition of the CTV with the pericardium in the majority of patients, no expected additional benefit of DIBH was anticipated in our cohort. As a third remark, it is clear that any eventual benefit derived from DIBH would reflect in both VMAT and IMPT data with substantial preservation of the relative merits of the two techniques. In any case, whenever possible, methods for respiratory gating (particularly with breath-hold) could further improve the absolute sparing of the heart against the results presented in this study. Therefore, it is advisable to implement them.

As an early confirmation of the relevance of respiratory motion management, Fracchiolla et al. (2019) proposed a respiratory-gated delivery approach based on optical tracking for patients with limited intrafraction motion and applied it to two patients affected by thymoma. The management of the movements induced by the heartbeat is still a challenging factor which cant be fully modelled at present.

Conclusion

The data from this in-silico study suggest that intensity-modulated proton therapy could be significantly advantageous in the treatment of thymic cancer patients with particular emphasis to a substantial reduction of the risk of cardiac failure and secondary cancer induction. Robust planning is a technical pre-requisite for the safety of the delivery.

Funding

None.

Compliance with ethical standards

Conflict of interest

L. Cozzi acts as Scientific Advisor to Varian Medical Systems and is Clinical Research Scientist at Humanitas Cancer Center. All other co-authors declare that they have no conflict interests.

Ethical approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. The Humanitas research hospital ethical committee approved by notification this retrospective study (study 10.20).

Informed consent

Informed consent was obtained from all individual participants included in the study.