See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/307157344

In silico comparison of photons versus carbon

ions in single fraction therapy of lung cancer

Article in Physica Medica · August 2016

DOI: 10.1016/j.ejmp.2016.08.014

CITATION READS

1 60

6 authors, including:

Marco Durante Christian Graeff

Trento Institute for Fundamental Physics and… GSI Helmholtzzentrum für Schwerionenforsc…
543 PUBLICATIONS 6,083 CITATIONS 64 PUBLICATIONS 518 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

FOOT (FragmentatiOn Of Target) View project

All content following this page was uploaded by Marco Durante on 06 September 2016.

The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document
and are linked to publications on ResearchGate, letting you access and read them immediately.
 Physica Medica xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Physica Medica
journal homepage: http://www.physicamedica.com

Original paper

In silico comparison of photons versus carbon ions in single fraction

therapy of lung cancer
Kristjan Anderle a,⇑, Joep Stroom b, Nuno Pimentel b, Carlo Greco b, Marco Durante a, Christian Graeff a
a
GSI Helmholtz Centre for Heavy Ion Research, Planckstr. 1, 64291 Darmstadt, Germany
b
Champalimaud Centre for the Unknown, Av. Brasília, 1400-038 Lisbon, Portugal

a r t i c l e i n f o a b s t r a c t

Article history: Purpose: Stereotactic body image guided radiation therapy (SBRT) shows good results for lung cancer
Received 17 April 2016 treatment. Better normal tissue sparing might be achieved with scanned carbon ion therapy (PT).
Received in Revised form 12 August 2016 Therefore an in silico trial was conducted to find potential advantages of and patients suited for PT.
Accepted 18 August 2016
Methods: For 19 patients treated with SBRT, PT plans were calculated on 4D-CTs with simulated breath-
Available online xxxx
ing motion. Prescribed single fraction dose was 24 Gy and OAR constraints used for photon planning were
respected. Motion was mitigated by rescanning and range-adapted ITVs. Doses were compared to the
Keywords:
original SBRT plans.
SDRT
Particle therapy
Results: CTV coverage was the same in SBRT and PT. The field-specific PTV including range margins for PT
Lung cancer was 1.5 (median, 25–75% 1.3–2.1) times larger than for SBRT. Nevertheless, maximum point dose and
Motion mitigation mean dose in OARs were higher in SBRT by 2.8 (1.6–3.7) Gy and 0.7 (0.3–1.6) Gy, respectively. Patients
Single fraction with a CTV >2.5 cc or with multiple lung lesions showed larger differences in OAR doses in favor of PT.
Conclusions: Patients receive less dose in critical OARs such as heart, spinal cord, esophagus, trachea and
aorta with PT, while maintaining the same target coverage. Patients with multiple or larger lesions are
particularly suited for PT.
Ó 2016 Published by Elsevier Ltd on behalf of Associazione Italiana di Fisica Medica.

1. Introduction carbon-ions (around 40%). However, the number of patients trea-

ted with particle therapy was low and they advise caution when
Lung cancer is one of the leading medical problems worldwide interpreting the data. Also different fractionation schemes were
with approximately 1.4 million deaths per year [1]. Surgery is usu- used in the comparison. A more recent review was published by
ally the first choice in treating localized non-small cell lung cancer Kamada et al. [10] where they reported a high 3-year survival rate
(NSCLC). However, in recent years stereotactic body-radiation for single-fraction carbon-ions (76.9%), with no late treatment-
therapy with photons (SBRT) showed very promising results, with related adverse effects. In comparison, SBRT had 55.8% 3-year sur-
high local control-rates of NSCLC [2–7]. vival rate, with 10–27% of patients exhibiting grade 3 treatment-
Scanned particle therapy can produce sharp dose gradients with related adverse effects. [6]. It is important to note that all of these
a finite range of the beam and can thus provide higher healthy studies used passive beam scattering, avoiding the problem of
tissue sparing. This reduces both side effects as well as the risk interplay between organ motion and scanning beam motion. On
of secondary cancer [8]. Treatment of lung tumors with particles the other hand, active beam scanning can provide even better dose
is still challenging due to interplay and radiological path length shaping which becomes essential in high dose single fractionation
changes [9]. The latter can be substantial when dense tissue (e.g. regimes. The effects of motion and motion mitigation techniques
the solid tumor mass) is replaced with low-density tissue (lung) on scanned carbon ion dose distribution therefore need to be con-
due to motion. sidered in a fair comparison of photons and carbon ions.
Grutters et al. have performed a meta-analysis on comparison To evaluate potential advantages of active scanning with carbon
between photon, proton and carbon ions in treating NSCLC [4]. ions (PT), an in silico comparison of simulated PT plans to SBRT
They found similar 5-year survival rates for SBRT, protons and plans actually delivered was conducted. Target coverage and a
wide range of OAR doses were assessed both with and without
simulated motion on time-resolved computed tomographies
⇑ Corresponding author. (4D-CTs).
E-mail address: k.anderle@gsi.de (K. Anderle).

http://dx.doi.org/10.1016/j.ejmp.2016.08.014
1120-1797/Ó 2016 Published by Elsevier Ltd on behalf of Associazione Italiana di Fisica Medica.

Please cite this article in press as: Anderle K et al. In silico comparison of photons versus carbon ions in single fraction therapy of lung cancer. Phys. Med.
(2016), http://dx.doi.org/10.1016/j.ejmp.2016.08.014
2 K. Anderle et al. / Physica Medica xxx (2016) xxx–xxx

2. Methods

2.1. Patient data

Our study included 19 patients with in total 26 lesions that

were actually treated with SBRT at the Champalimaud Centre for
the Unknown, Lisbon (Portugal). The lesion size was 2.9 cc
(median, 25–75% 1.4–9.7) and peak-to-peak motion was 3.1 mm
(1.6–5.6). Three patients had two targets, one had five and the rest
one. 13 lesions were right-sided, 12 were left-sided and one was
located in right cardiophrenic space. An overview of tumor
characteristics can be found in Table 1.
Two CTs were available for all patients. A planning CT was used
for OAR delineation and SBRT planning. Target motion was
estimated on a 4D-CT, consisting of 10 phases (0%–90%). Clinical
target volumes (CTV) were delineated using a registered positron
emission tomography (PET) scan.
The planning objectives were that 99% of planning target
volume (PTV) must receive at least 24 Gy (D99% P 24 Gy) in a
single fraction, while all OAR constraints as defined in the AAPM
task group 101 report on stereotactic radiotherapy had to be
respected [11].
Fig. 1. Different PTV definitions for SBRT (PTVSBRT) and PT (PTVPT). For PTVSBRT
isotropic margins of 3 mm plus maximum tumor displacement due to breathing
2.2. Definition of target volumes
were used on the CTV; for PTVPT margins of 3 mm laterally and 1 mm in beam’s eye
view were used and then range-ITV was constructed with 2 mm + 2% range margins
To account for range changes relevant for particles only, differ- added for PTVPT in the end-inhale phase.
ent PTV definitions were used for SBRT and PT, as shown in Fig. 1.
Within this paper they will be named PTVSBRT and PTVPT for SBRT
and PT, respectively. of 3 mm laterally and 1 mm in beam’s eyes view (BEV) were
In SBRT, the responsible clinician determined the maximum used on the CTV. Afterwards a water-equivalent path length ITV
breathing motion of the CTV from the 4D-CT, hence creating an (WEPL-ITV) was build, using transformation maps from the
ITV. This ITV plus an additional 3 mm for setup uncertainty yielded B-Spline deformable registration of the 4D-CT data [13]. Additional
the PTVSBRT. 2 mm + 2% proximal and distal margins were added in BEV to
PTVPT was constructed following principles from Graeff et al. account for uncertainty from Hounsfield units to water equivalent
[12]. Each beam has a unique PTVPT. For setup uncertainty margins path length conversion.

Table 1
Patient characteristics showing lesion locations, stages, peak-to-peak motions, and volumes of corresponding CTV, PTVSBRT and PTVPT (given as range).

Patient Lesion number Lesion location Stage Peak-to-peak motion (mm) Volume (cc)
CTV PTVSBRT PTVPT
1 1 LSL IIa 4.8 36 100 143.0–193.9
2 2 LSL Ia 3.1 1.6 7.7 25.8–32.3
3 3 IRL IV 12 2.3 12 25.2–39.1
3 4 RCS IV 11.8 0.4 6.6 8.4–8.7
4 5 RSL Ia 0.5 6.9 25 28.9–38.2
5 6 ILL IV 4.4 2.4 15 0.0–19.2
6 7 ILL IV 7.5 1.4 7.7 34.3–36.8
7 8 RSL IV 3.9 16 40 62.6–76.5
8 9 ILL IV 0.6 139 261 44.4–251.7
8 10 LSL IV 2 9.2 35 242.3–251.7
9 11 IRL IV 3.4 10 38 54.4–59.9
9 12 ILL IV 2.8 14 46 44.1–47.0
10 13 ILL IV 5.8 3.8 17 23.4–31.2
10 14 RSL IV 0.8 4.3 18 26.9–30.3
10 15 LSL IV 3.4 2.7 15 23.2–27.0
10 16 RSL IV 2.1 3.1 15 31.0–35.9
10 17 LSL IV 0.5 0.5 5.4 6.7–7.4
11 18 ILL IV 7.8 0.8 6.1 23.3–24.2
12 19 LSL IV 0.1 1.7 15 22.4–25.1
13 20 IRL IIIb 11.4 27 137 85.0–121.5
14 21 RSL Ia 2.2 1.7 10 23.4–23.4
15 22 RSL IV 0.2 0.9 3.2 13.9–15.8
16 23 RSL IV 2.2 3.9 22 25.8–29.4
17 24 LSL IV 3.1 9.8 28 49.3–52.9
18 25 RSL IV 8.1 0.6 3.3 7.7–8.5
19 26 LSL IV 1.4 0.8 5.9 12.0–12.4

cc, cubic centimeters; RSL, right superior lung; IRL, inferior right lung; LSL, left superior lung; ILL, inferior left lung; RCS right cardiophrenic space; CTV, clinical target volume;
PTVSBRT, SBRT planning target volume; PTVPT, field-specific target volume;

Please cite this article in press as: Anderle K et al. In silico comparison of photons versus carbon ions in single fraction therapy of lung cancer. Phys. Med.
(2016), http://dx.doi.org/10.1016/j.ejmp.2016.08.014
 K. Anderle et al. / Physica Medica xxx (2016) xxx–xxx 3

If the target overlapped with an OAR (e.g. small airways) then respectively. This led to an RBE of approximately 1 in target tissue
OAR plus a margin of 2–5 mm was subtracted from PTVSBRT or and approximately 1.1 to 3 in OARs.
PTVPT. Dose was calculated on end-inhale (3D-Dose0%) and end-exhale
(3D-Dose50%) phases. 4D dose delivery was simulated as described
2.3. SBRT treatment planning by Richter et al. [16]. Two different breathing periods (3.6 and 5 s)
and two different starting phases (0° and 90°) were used. Simula-
The clinical plans were calculated with the Eclipse v10 planning tions without motion compensation (4D-Doseinterplay) and with
system (Varian Medical Systems, Palo Alto, Ca) using the AAA beam slice-by-slice raster rescanning were performed (4D-Doserescan).
model. They were all VMAT plans generally consisting of 4 overlap- Five rescans were used for the majority of targets (n = 24), whereas
ping partial arcs, 2 in clockwise and 2 anticlockwise direction, with 20 rescans were used for targets where the interplay effects were
a gantry range of typically 200°. For tumor sizes >2.5 cm a calcula- too big to achieve a satisfactory target coverage (n = 2; lesions 3
tion grid of 2.5 mm was used, otherwise this was 1 mm. During and 18 in Table 1).
optimization, a first iteration included the PTVSBRT only, after
which the OARs were added. In order to lower OAR dose and 2.5. Dose metrics and analysis
improve the PTVSBRT homogeneity, we created an artificial shell
of 2 cm around the PTVSBRT and minimized the dose there as well. For comparison between SBRT and PT the following dose
During optimization the fast Multi Resolution Dose Calculation metrics were used – for the target the minimum dose in 99% of
(MRDC) model was used, with one intermediate step using the the volume (D99%), which should be higher than 24 Gy; for OARs,
slower but more accurate AAA model to get an adequate PTVSBRT the maximum point dose (DMax) and the mean dose (DMean). Addi-
coverage after optimization. tionally, the volume receiving 20% of the planned dose (V20%) was
used to assess differences in lung doses. In all cases, absorbed dose
2.4. Carbon ion treatment planning in Gy for SBRT was compared to biologically-equivalent dose in Gy
(RBE) for PT
For PT, state of the art 4D treatment planning software TRiP98 Paired t-tests were performed to compare the dose metrics and
was used [14]. A single field uniform dose plan (SFUD) was opti- for post hoc exploratory analysis between groups a two-sided
mized on the PTVPT in the end-inhale reference phase of the 4D- t-test with Welch correction for different variances was carried
CT. Most targets (n = 20) were planned with two fields. For the out. A p-value <0.05 was considered significant. Dose differences
remaining targets, one (n = 1), three (n = 3) or four (n = 2) fields are always reported such that higher dose levels for SBRT result
were used due to proximity of OARs. A regular grid of beam spots in positive values.
with a spacing of 2 mm, a beam spot full width at half maximum
(FWHM) size of approximately 6 mm and a 3 mm ripple filter were 3. Results
used. To compensate for short particle ranges in lung tissue, a bolus
of 80 mm water-equivalent thickness was added. Examples of two SBRT and 4D-Doserescan PT treatment plans are
The relative biological effectiveness (RBE) was used, following shown in Fig. 2. Patient 9 has two lesions in close proximity to the
the local effect model (LEM) IV [15]. For a conservative estimation, spinal cord. Patient 2 has a small lesion (1.6 cc) in the superior
an alpha-beta ratio of 6 Gy and 2 Gy were used for target and OARs, position of the left lung. D99% is 100% for SBRT and PT in all CTVs

Fig. 2. Treatment plans for SBRT (left), PT (middle) and dose volume histogram (right) for SBRT (solid lines) and PT (dashed lines) for two patients. PT curves for OARs without
any dose are not shown. Patient 9 (top row) might be better suited for PT and patient 2 (lower row) for SBRT. Patient 2 has a small lesion (1.6 cc) in a central lung region,
resulting in large PTVPT – up to 32 cc, compared to PTVSBRT 7.7 cc. The CTV contour is outlined in white.

Please cite this article in press as: Anderle K et al. In silico comparison of photons versus carbon ions in single fraction therapy of lung cancer. Phys. Med.
(2016), http://dx.doi.org/10.1016/j.ejmp.2016.08.014
4 K. Anderle et al. / Physica Medica xxx (2016) xxx–xxx

for Patient 9 and 2; average OAR difference between SBRT and PT almost all patients in SBRT and PT. The overall OAR difference for
in DMax is 5.1 Gy and 1.4 Gy and in mean dose 1.4 Gy and 0.7 Gy, patients between SBRT and PT was significant, 2.8 (1.6–3.7) Gy
respectively for patient 9 and 2. for DMax and 0.7 (0.3–1.6) Gy for DMean.
The contralateral lung did not receive any dose in 12 (75%)
3.1. Target coverage patients with PT, of 16 patients affected only in one lung.

Difference in PTV definition resulted in 1.5 (1.3–2.1) times 3.3. Dependence on CTV size
bigger PTVPT than PTVSBRT. All SBRT plans were clinically accept-
able, though in one case the D99% was reduced to 16.8 Gy due to Significant differences were observed between patients with a
proximity of an OAR. 3D-Dose0% and 3D-Dose50%plans provided single CTV smaller (n = 8) or larger (n = 7) than 2.5 cc for DMax
sufficient target coverage in all patients. For 4D-Doseinterplay and and DMean, see Fig. 4. For patients with a smaller CTV, the dosimet-
4D-Doserescan there was 63% and 2% cases of insufficient target ric advantage over SBRT was on average 0.9 Gy and 0.5 Gy lower
coverage, respectively, across all targets and different breathing for DMax and DMean, respectively. This was associated with PTVPT
patterns. Details are shown in Fig. 3. For the patient with reduced definition – the average volume ratio between PTVPT and PTVSBRT
dose in SBRT, PT could increase the D99% from 16.8 Gy in SBRT to was 2.9 (1.6–4.0) and 1.5 (1.3–1.8), for patients with CTV <2.5 cc
20.3 Gy while adhering to OAR constraints. and CTV >2.5 cc, respectively
The 4 patients with multiple lesions were excluded from this
comparison. The DMax and DMean difference were on average higher
3.2. Dose in OARs
in these patients, but the number of patients was too low for sta-
tistical analysis.
There was no significant difference in dose to OAR between the
different PT dose calculations. The DMax and DMean for SBRT and 4D-
Doserescan for OARs heart, spinal cord, smaller airway esophagus, 4. Discussion
trachea, aorta, ipsi- and contralateral lung are presented in Table 2.
There was a significant difference in both DMax and DMean for all This is the first in silico trial directly comparing clinically valid
OARs between SBRT and PT, with PT delivering less dose to OARs. SBRT plans to scanned carbon ion plans using state of the art 4D
Significant difference was also observed for V20% for ipsilateral dose calculation and motion mitigation methods for NSCLC
lung, which was 15.3 (9.6–23.5)% and 10.3 (7.9–13.7)% for SBRT patients. Our study found that PT deposited less dose to OARs com-
and PT, respectively. The V20% for contralateral lung was zero for pared to SBRT. Therefore PT might be considered as an alternative

Fig. 3. CTV D99% for SBRT and different PT calculations. Four different breathing
patterns are included for all targets in 4D-interplay and 4D-rescan. The dashed line Fig. 4. Box plots of average OARs max point dose (DMax) and mean dose difference
shows the lower limit for clinical acceptance. One patient was an exception where between SBRT and PT for patients with single CTV smaller (n = 9) or bigger (n = 6)
lower target dose was accepted due to the proximity of a critical structure. Boxes than 2.5 cc. Boxes represent 25%–75%, outliers are shown as whiskers and median is
represent 25%–75%, outliers are shown as whiskers and median is shown with solid shown with solid lines. Values for patients with multiple lesions are shown with
lines. circle symbols.

Table 2
Dose metrics for OARs. First value at each organ is from SBRT and the second from 4D-rescan. All values are shown as median and 25–75% in brackets.

OAR DMax (Gy) DMean (Gy)

Photon Carbon Photon Carbon
Heart 6.0 (0.3–11.6) 0 (0–8.8) 1.3 (0.1–2.2) 0 (0–0.5)
Spinal cord 5.5 (3.3–8.5) 0 (0–0.5) 0.7 (0.3–1.2) 0 (0 0)
Smaller airways 13.0 (9.8–17.1) 10.3 (3.3–19.1) 2.8 (1.5–5.8) 0.5 (0–2.6)
Esophagus 5.8 (3.9–8.4) 0 (0–0.3) 1.1 (0.6–1.5) 0 (0 0)
Trachea 3.9 (1.8–5.4) 0 (0 0) 1 (0.3–1.3) 0 (0 0)
Aorta 8.0 (5.1–21.9) 3.9 (0–18.1) 1.4 (0.7–1.6) 0.1 (0–0.4)
Ipsilateral lung 26.3 (26.0–26.5) 26.3 (25.8–26.5) 1.9 (1.5–3.0) 1.9 (1.4–2.5)
Contralateral lung 5.0 (3.5–9.6) 0 (0–0.9) 0.4 (0.2–0.6) 0 (0 0)

Please cite this article in press as: Anderle K et al. In silico comparison of photons versus carbon ions in single fraction therapy of lung cancer. Phys. Med.
(2016), http://dx.doi.org/10.1016/j.ejmp.2016.08.014
 K. Anderle et al. / Physica Medica xxx (2016) xxx–xxx 5

treatment option to SBRT. The finite range of the beam permits a even better OAR sparing than protons. Our results are in agreement
small number of fields and thus a narrow entry channel, so that with several in silico studies comparing SBRT and proton therapy
critical OARs such as spinal cord, heart, esophagus, and the con- for NSCLC [20,27,28]. Furthermore, a study made by Kadoya
tralateral lung could be effectively spared using PT, with typically reached the same conclusion as our study, that patients with larger
low or even zero dose. PT could be thus highly beneficial to CTV and/or multiple CTVs would receive less dose from proton
patients with impaired contralateral lung function, because PT therapy [27]. A recent phase II trial for patients with multiple sites
deposited no dose in the contralateral lung in 12 patients, while of extracranial disease showed good results for photons [29], how-
SBRT irradiated the contralateral lung in all patients. Being an ever, based on the findings of Kadoya et al. and our study, proton
intensity-modulated arc therapy, SBRT had an advantage in some and/or carbon-ion therapy might result in even better outcome.
patients where the smaller airways were in a close proximity to
CTV; SBRT could shape the dose distribution to reduce dose to
4.3. Study limitations
the smaller airways, compensating PT’s advantageous physical
dose characteristics.
The 4D dose calculations were based on a regular breathing pat-
Further increase in OAR sparing could be achieved by using
tern, which typically varies during patient treatment and/or
intensity modulated particle therapy (IMPT) instead of SFUD.
between 4D-CT acquisition and actual treatment [30,31]. A possi-
While IMPT could lead to less dose in the OARs, it would make
ble solution was proposed by Boye et al. to get motion information
the plans less robust against setup errors due to additional dose
from 4D magnetic resonance imaging (4DMRI) and use it in 4D
gradients between the fields. These gradients can be controlled
dose calculations [32].
by employing robust optimization to account for range, motion
Furthermore, SBRT treatment plans were done on a static case
and setup uncertainties, which we will implement in a future 4D
in contrast to a 4D dose calculation done for PT. This should not
treatment planning study [17,18].
influence the results of our study, since motion has a smaller
impact on photon dose distributions [33], whereas it is imperative
4.1. Range margins and motion mitigation
in PT dose calculations [9].
There were also differences in treatment planning. PT plans
Since conventional geometric margins are not suitable for PT
were done by a single person in a research setting, whereas SBRT
[19], margins based on range changes were used. Another trial
plans were made by different people under clinical conditions with
comparing photon to proton therapy in NSCLC patients also used
the requirement to finish the plans on time.
different PTV definitions to incorporate range changes [20]. As
Slight changes also existed between the planning CT, used for
shown in our study, inclusion of range changes leads to increase
SBRT treatment plans and 4D-CT used for PT treatment plans, even
in PTVPT, up to 4.7 times compared to PTVSBRT. Furthermore, the
though 4D-CT was usually acquired right after the planning CT. The
difference between PTVs is bigger for smaller tumor sizes. Patients
propagation of contours from the planning CT to the 4D-CT and
with bigger tumor volumes (CTV >2.5 cc) are therefore better sui-
also for the 4D dose calculation rely on deformable image registra-
ted for treatment with PT.
tion (DIR), where even small changes can effect 4D dose distribu-
Our results confirm previously published results that interplay
tion [34]. Results from DIR were therefore thoroughly checked
can lead to a dose degradation in treating moving targets with
with different approaches, such as inspection of the warped image,
active scanned beam [21]. Fig. 3 shows the importance of using
evaluation of the Jacobian values and the inverse consistency of the
4D dose calculation and motion mitigation techniques in treating
vector field. However, the transformation of the dose with DIR is a
moving targets with particles. Even small motion amplitude can
debated topic and might jeopardize the simulated results, espe-
lead to underdosage in CTV without proper motion mitigation.
cially with respect to the 4D target coverage. On the other hand,
Considering the average over the 4 simulated motion patterns,
dose differences in OARs were large and should be robust against
15 patients showed a D99% < 24 Gy under interplay conditions, as
vector field errors in the order a few mm. Nevertheless, further
opposed to none when using rescanning (excluding the one patient
studies are warranted, possibly using advanced moving phantoms
with reduced target dose). Rescanning proved to be a strong miti-
for an experimental validation [35] and finally also clinical trials.
gation technique, with robust results across all targets and differ-
First patients are being treated in thoracic and abdominal regions
ent breathing patterns.
with an active beam scanning at the National Institute for Radio-
Recent studies suggest that some patients require phase-
logical Sciences (NIRS) in Japan [36].
controlled layer or volumetric rescanning for sufficiently robust
target coverage [22,23]. The advantage of simple slice-by-slice res-
canning is that no motion monitoring or assumptions on the 4.4. Application
breathing frequency are necessary [9], but the higher required
number of rescans might increase treatment times due to reduced Scanned carbon ion therapy is available only in a limited num-
beam intensities. Another possibility is to combine rescanning ber of clinics, mainly due to the considerably higher cost in com-
with gating, which was already successfully implemented in clinic parison to photon linacs. Therefore a careful patient selection
[24]. appears sensible. Patients with larger and multiple lesions where
SBRT might be limited due to OAR constraints could be referred
4.2. RBE and proton therapy to carbon centers. In this study, already lesions >2.5 cc were found
to benefit significantly stronger from PT.
Carbon ions exhibit a radiobiological advantage, especially in
the Bragg peak region. However, for high doses as used here the
effect of RBE is not well documented and is subject to ongoing 5. Conclusion
research [25]. For these high doses RBE for carbon ions should
approach a value between 1 and 2 [26], which is in agreement with SBRT and PT both achieved satisfactory target dose. In most
values in our study (1.1). patients PT deposited less dose in all OARs (including heart, spinal
Coincidently, RBE values in the target at high doses are similar cord, esophagus, trachea and aorta). Patients with multiple lesions
to those used clinically in proton therapy. Carbon-ions show con- and/or with large target volumes might be preferentially selected
siderably lower lateral scattering though, which should result in for particle therapy.

Please cite this article in press as: Anderle K et al. In silico comparison of photons versus carbon ions in single fraction therapy of lung cancer. Phys. Med.
(2016), http://dx.doi.org/10.1016/j.ejmp.2016.08.014
6 K. Anderle et al. / Physica Medica xxx (2016) xxx–xxx

References [19] Park PC, Zhu XR, Lee AK, Sahoo N, Melancon AD, Zhang L, et al. A beam-specific
planning target volume (PTV) design for proton therapy to account for setup
and range uncertainties. Int J Radiat Oncol Biol Phys 2012;82:e329–36.
[1] Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer
[20] Roelofs E, Engelsman M, Rasch C, Persoon L, Qamhiyeh S, de Ruysscher D, et al.
statistics. CA Cancer J Clin 2011;61:69–90.
Results of a multicentric in silico clinical trial (ROCOCO): comparing
[2] Baumann P, Nyman J, Hoyer M, Wennberg B, Gagliardi G, Lax I, et al. Outcome
radiotherapy with photons and protons for non-small cell lung cancer. J
in a prospective phase II trial of medically inoperable stage I non-small-cell
Thoracic Oncol 2012;7:165–76.
lung cancer patients treated with stereotactic body radiotherapy. J Clin Oncol
[21] Bert C, Grözinger SO, Rietzel E. Quantification of interplay effects of scanned
2009;27:3290–6.
particle beams and moving targets. Phys Med Biol 2008;53:2253–65.
[3] Fakiris AJ, McGarry RC, Yiannoutsos CT, Papiez L, Willams M, Henderson MA,
[22] Mori S, Furukawa T, Inaniwa T, Zenklusen S, Nakao M, Shirai T, et al. Systematic
et al. Stereotactic body radiation therapy for early-stage non-small-cell lung
evaluation of four-dimensional hybrid depth scanning for carbon-ion lung
carcinoma: four-year results of a prospective phase II study. Int J Radiat Oncol
therapy. Med Phys 2013;40:031720.
2009;75:677–82.
[23] Takahashi W, Mori S, Nakajima M, Yamamoto N, Inaniwa T, Furukawa T, et al.
[4] Grutters JPC, Kessels AGH, Pijls-Johannesma M, Ruysscher D, Joore MA, Lambin
Carbon-ion scanning lung treatment planning with respiratory-gated phase-
P. Comparison of the effectiveness of radiotherapy with photons, protons and
controlled rescanning: simulation study using 4-dimensional CT data. Radiat
carbon-ions for non-small cell lung cancer: a meta-analysis. Radiother Oncol
Oncol 2014;9:238.
2010;95:32–40.
[24] Rossi S. The national centre for oncological hadrontherapy (CNAO): status and
[5] Ricardi U, Filippi AR, Guarneri A, Giglioli FR, Ciammella P, Franco P, et al.
perspectives. Phys Med 2015;31:333–51.
Stereotactic body radiation therapy for early stage non-small cell lung cancer:
[25] Friedrich T, Sholz U, Durante M, Sholz M. RBE of ion beams in hypofractionated
results of a prospective trial. Lung Cancer 2010;68:72–7.
radiotherapy (SBRT). Phys Med 2014;30:588–91.
[6] Timmerman R, Paulus R, Galvin J, Michalski J, Straube W, Bradley J, et al.
[26] Carabe-Fernandez A, Dale RG, Jones B. The incorporation of the concept of
Stereotactic body radiation therapy for inoperable early stage lung cancer.
minimum RBE (RBEmin) into the linear-quadratic model and the potential for
JAMA 2010;303:1070–6.
improved radiobiological analysis of high-LET treatments. Int J Radiat Biol
[7] Greco C, Zelefsky MJ, Lovelock M, Fuks Z, Hunt M, Rosenzweig K, et al.
2007;83:27–9.
Predictors of local control after single-dose stereotactic image-guided
[27] Kadoya N, Obata Y, Kato T, Kagiya M, Nakamura T, Tomoda T, et al. Dose-
intensity-modulated radiotherapy for extracranial metastases. Int J Radiat
volume comparison of proton radiotherapy and stereotactic body radiotherapy
Oncol Biol Phys 2011;79:1151–7.
for non-small-cell lung cancer. Int J Radiat Oncol Biol Phys 2010;79:1225–31.
[8] Newhauser WD, Durante M. Assessing the risk of second malignancies after
[28] Register SP, Zhang X, Mohan R, Chang JY. Proton stereotactic body radiation
modern radiotherapy. Nat Rev Cancer 2011;11:434–48.
therapy for clinically challenging cases of centrally and superiorly located
[9] Bert C, Durante M. Motion in radiotherapy: particle therapy. Phys Med Biol
stage I non-small-cell lung cancer. Int J Radiat Oncol Biol Phys
2011;56:R113–44.
2010;80:1015–22.
[10] Kamada T, Tsujii H, Blakely EA, Debus J, Neve WD, Durante M, et al. Carbon ion
[29] Ivengar P, Kavanagh BD, Wardak Z, Smith I, Ahn C, Gerber DE, et al. Phase II
radiotherapy in Japan: an assessment of 20 years of clinical experience. Lancet
trial of stereotactic body radiation therapy combined with erlotinib for
Oncol 2016;16:e93–e100.
patients with limited but progressive metastatic non-small-cell lung cance. J
[11] Benedict SH, Yenice KM, Followill D, Galvin JM, Hinson W, Kavanagh B, et al.
Clin Oncol 2014;32:3824–54.
Stereotactic body radiation therapy: the report of AAPM Task Group 101. Med
[30] Verma P, Huanmei W, Langer M, Das I, Sandison G. Survey: real-time tumor
Phys 2010;37:4078–101.
motion prediction for image-guided radiation treatment. Prog Nucl Sci Technol
[12] Graeff C, Durante M, Bert C. Motion mitigation in intensity modulated particle
2010;13:24–35.
therapy by internal target volumes covering range changes. Med Phys
[31] Malinowski K, McAvoy TJ, George R, Dietrich S, D’Souza WD. Incidence of
2012;39:6004–13.
changes in respiration-induced tumor motion and its relationship with
[13] Shackleford JA, Kandasamy N, Sharp GC. On developing B-spline registration
respiratory surrogates during individual treatment fractions. Int J Radiat
algorithms for multi-core processors. Phys Med Biol 2010;55:6329–51.
Oncol 2011;82:1665–73.
[14] Richter D, Schwarzkopf A, Trautmann J, Kramer M, Durante M, Jakel O, et al.
[32] Boye D, Lomax T, Knopf A. Mapping motion from 4D-MRI to 3D-CT for use in
Upgrade and benchmarking of a 4D treatment planning system for scanned
4D dose calculations: a technical feasibility study. Med Phys 2013;40:061702.
ion beam therapy. Med Phys 2013;40:051722.
[33] Zou W, Yin L, Shen J, Corradeti MN, Kirk M, Munbodh R, et al. Dynamic
[15] Elsässer T, Weyrather WK, Friedrich T, Durante M, Iancu G, Kramer M, et al.
simulation of motion effects in IMAT lung SBRT. Radiat Oncol 2014;9:225.
Quantification of the relative biological effectiveness for ion beam
[34] Kashani R, Hub M, Balter JM, Kessler ML, Dong L, Zhang L, et al. Objective
radiotherapy: direct experimental comparison of proton and carbon ion
assessment of deformable image registration in radiotherapy: a multi-
beams and a novel approach for treatment planning. Int J Radiat Oncol Biol
institution study. Med Phys 2008;35:5944–53.
Phys 2010;78:1177–83.
[35] Perrin R, Peroni M, Bernatowicz K, Zakova M, Knopf A, Safai S, et al. A realistic
[16] Richter D, Graeff C, Jakel O, Combs SE, Durante M, Bert C. Residual motion
breathing phantom of the thorax for testing new motion mitigation techniques
mitigation in scanned carbon ion beam therapy of liver tumors using enlarged
with scanning proton therapy. Med Phys 2014;41:111.
pencil beam overlap. Radiother Oncol 2014;113:290–5.
[36] Mori S, Karube M, Tajiri M, Takekoshi T, Miki K, Shiraishi Y, et al. Carbon-ion
[17] Chen W, Unkelbach J, Trofimov A, Madden T, Kooy H, Bortfeld T, et al. Including
pencil beam scanning treatment with gated markerless tumor tracking: an
robustness in multi-criteria optimization for intensity-modulated proton
analysis of positional accuracy. Int J Radiat Biol 2016;95:258–66.
therapy. Phys Med Biol 2012;57:591–608.
[18] Graeff C. Motion mitigation in scanned ion beam therapy through 4D-
optimization. Phys Med 2014;30:570–7.

Please cite this article in press as: Anderle K et al. In silico comparison of photons versus carbon ions in single fraction therapy of lung cancer. Phys. Med.
(2016), http://dx.doi.org/10.1016/j.ejmp.2016.08.014
View publication stats