Improving Target Coverage and Organ-at-Risk Sparing in Intensity-Modulated Radiotherapy for Cervical Oesophageal Cancer Using a Simple Optimisation Method
Improving Target Coverage and Organ-at- Risk Sparing in Intensity-Modulated Radiotherapy for Cervical Oesophageal Cancer Using a Simple Optimisation Method
Jia-Yang Lu 0 1
Michael Lok-Man Cheung 0 1
Bao-Tian Huang 0 1
Li-Li Wu 0 1
Wen-Jia Xie 0 1
Zhi-Jian Chen 0 1
De-Rui Li 0 1
Liang-Xi Xie 0 1
0 1 Department of Radiation Oncology, Cancer Hospital of Shantou University Medical College , Shantou , China , 2 Department of Clinical Oncology, Prince of Wales Hospital , Shatin, Hong Kong , China
1 Academic Editor: Xuefeng Liu, Georgetown University , UNITED STATES
Funding: This work was sponsored in part by both
National Natural Science Foundation of China (Grant
No. 81171994) and Shantou University Medical
College Clinical Research Enhancement Initiative
(Grant No. 201425). No additional external funding
was received for this study. The funders had no role
in study design, data collection and analysis, decision
to publish, or preparation of the manuscript.
To assess the performance of a simple optimisation method for improving target coverage
and organ-at-risk (OAR) sparing in intensity-modulated radiotherapy (IMRT) for cervical
For 20 selected patients, clinically acceptable original IMRT plans (Original plans) were
created, and two optimisation methods were adopted to improve the plans: 1) a base dose
function (BDF)-based method, in which the treatment plans were re-optimised based on the
original plans, and 2) a dose-controlling structure (DCS)-based method, in which the original
plans were re-optimised by assigning additional constraints for hot and cold spots. The
Original, BDF-based and DCS-based plans were compared with regard to target dose
homogeneity, conformity, OAR sparing, planning time and monitor units (MUs). Dosimetric
verifications were performed and delivery times were recorded for the BDF-based and
The BDF-based plans provided significantly superior dose homogeneity and conformity
compared with both the DCS-based and Original plans. The BDF-based method further
reduced the doses delivered to the OARs by approximately 13%. The re-optimisation time
was reduced by approximately 28%, but the MUs and delivery time were slightly increased.
All verification tests were passed and no significant differences were found.
Competing Interests: The authors have declared
that no competing interests exist.
The BDF-based method for the optimisation of IMRT for cervical oesophageal cancer can
achieve significantly better dose distributions with better planning efficiency at the expense
of slightly more MUs.
Oesophageal cancer is a frequently diagnosed cancer worldwide . To achieve optimal
tumour locoregional control and quality of life, multi-modal treatment strategies including
operation, chemotherapy and radiotherapy are typically applied [2,3]. In fact, it is now standard to
treat locally advanced cervical oesophageal cancer using concurrent chemoradiotherapy
because of the difficulty of achieving a clear margin in surgical resection .
Intensity-modulated radiotherapy (IMRT) is an advanced radiotherapy technique that is
performed using multiple small beams of non-uniform intensity that can generate very steep
dose gradients, resulting in improved tumour control and fewer normal-tissue complications
in general . Many studies have shown that IMRT can minimise the trade-off between target
coverage and organ-at-risk (OAR) sparing for oesophageal cancer . Several clinical trials
 have also reported that IMRT provides promising locoregional control with a low
Cervical oesophageal cancer is typically treated with the IMRT technique because of the
irregular shape of planning target volume (PTV) and the less dosimetric uncertainty caused by
respiratory motion compared with that in the thoracic region. However, it is challenging to
achieve optimal IMRT plans for cervical oesophageal cancer. Common reasons for this
difficulty include the rapidly changing neck-to-shoulder anatomy and the presence of dose-limiting
OARs; another important reason is the dose discrepancy between optimiser plans and finally
calculated plans. This discrepancy is caused by an optimisation-convergence error (OCE) that
originates from the following major sources, as described by Dogan et al.: tissue heterogeneity,
the buildup effect, multi-leaf collimator (MLC) modulation and the optimisation algorithm
[13,14]. The OCE can lead to locally high doses (hot spots) or locally low doses (cold spots) in
the final dose distributions. The OCE is especially significant in the case of cervical oesophageal
cancer because the PTV typically contains air cavities, such as the trachea, as well as lung tissue
and the buildup region. Although selecting the optimal arrangement and number of beams is
an effective approach for improving IMRT plans [15,16], the optimal beam arrangement and
number alone are not able to overcome the OCE because it is a systematic error.
Accordingly, we proposed an optimisation method to compensate for the OCE with the
objective of improving the planning quality for cervical oesophageal cancer. To assess the
application of this new method, the original plans were used in a longitudinal comparison to
demonstrate its efficacy, and another common optimisation method was used for a lateral
Materials and Methods
The protocol was approved by the Ethical Commission of the Cancer Hospital of Shantou
University Medical College. Because this was not a treatment-based study, our institutional review
board waived the need for written informed consent from the participants. The patient
information was anonymised and de-identified to protect patient confidentiality.
We retrospectively identified twenty previously untreated patients (median age 58 years, range
4174 years), including 3 females and 17 males, with cervical oesophageal squamous cell
cancers in Stage T3-T4 and N0-N1. Tumour staging was based on the American Joint Committee
on Cancer 2010 7th edition staging criteria. The patients were immobilised in
head-neckshoulder thermoplastic masks in the supine position.
Target delineation and OAR definition
The gross tumour volume (GTV), lymph nodes (LNs), clinical target volumes (CTVs), PTVs
and OARs were contoured on an Eclipse version 10.0 treatment planning system (Varian
Medical Systems, Palo Alto, USA).
The GTV was determined using planning CT, MR, positron emission tomography (PET)
and clinical information. Two CTVs (CTV64 and CTV54) were defined for simultaneous
integrated boost IMRT. The high-risk CTV (CTV64) was contoured with superiorinferior
margins of 34 cm and 1-cm transverse margins around the GTV and with 1-cm margins around
the positive LNs. The high-risk PTV (PTV64), which was generated by adding 0.5-cm margins
to the CTV64, was prescribed a 64-Gy dose (2 Gy/fraction) administered in 32 fractions. The
low-risk CTV (CTV54) covered the CTV64 plus nodal basins at risk of harbouring metastatic
disease, namely, the lymphatic drainage area in the bilateral supraclavicular zone and the
mediastinum. The low-risk PTV (PTV54), which was generated by adding 0.5-cm margins to the
CTV54, was prescribed a 54-Gy dose (1.69 Gy/fraction) administered in 32 fractions. The
mean volumes of the PTV64 and PTV54 were 130.5 72.5 and 321.2 88.9 cubic centimetres
The OARs, including the spinal cord and lungs, were delineated on each image. The
planning OAR volume (PRV) that was generated from the spinal cord plus 5-mm margins was
denoted as PRV spinal cord .
IMRT planning techniques and planning objectives
Five coplanar sliding-window IMRT fields of 6-MV photons from a TrueBeam (Varian
Medical Systems, Palo Alto, USA) linear accelerator were generated for each plan in Eclipse. The
gantry angles were evenly distributed at 216, 288, 0, 72 and 144. Dose-limiting ring
structures  were created to form dose gradients around the PTVs. The Dose Volume Optimiser
(DVO, version 10.0.28) and the Anisotropic Analytical Algorithm (AAA, version 10.0.28) were
employed for optimisation and for final dose calculations, respectively. The plans were
normalised to the 64-Gy prescribed dose which covered 95% of the PTV64.
The optimisation objectives for the inverse planning were to achieve 95% coverage of the
PTVs at the prescribed doses with the PTV64 maximum dose 70.4 Gy while limiting the
doses to the OARs within specified tolerances. The PTV coverage objectives were assigned the
highest priorities, followed by the OAR sparing. The notation Dx represents the dose that was
reached or exceeded in x of the volume. The notation VxGy represents the % volume that
received a dose of at least x Gy. The dose-volume constraints of the OARs were set as follows: the
D0.1cc of the PRV spinal cord was constrained to be < 45 Gy ; the lung volumes were
constrained to be V5Gy < 45%, V10Gy < 35%, V20Gy < 20% and V30Gy < 10%; and the mean lung
dose (MLD) was constrained to be < 15 Gy .
To create the original plan (Original plan), the planning objectives from a template were
applied and fine-tuned until the plan was clinically acceptable. With the original planning
objectives unmodified, two independent methods were utilised to improve the original plans,
thereby generating two additional types of plans: 1) re-optimisation utilising the base dose
function (BDF-based plan) and 2) re-optimisation using dose-controlling structures to address
hot and cold spots (DCS-based plan) [17,18].
To generate a BDF-based plan, the number of fractions of the original plan was modified to
50% of the prescribed number of fractions (from 32 to 16, in our cases) to generate a base
dose plan with half of the total prescribed dose. Then, the base dose plan was copied to be a
top dose plan. Afterwards, the top dose plan was re-optimised once based on the base dose
plan using Eclipses base dose function. At this point, the prescribed dose of the plan sum (the
top dose plan plus the base dose plan) was equal to the originally prescribed dose. When the
final dose calculation was complete, the number of fractions of the optimised top dose plan
was changed from 50% (16 fractions) to 100% (32 fractions) of the prescribed number of
fractions, that is, the prescribed dose of the top dose plan was changed from a half dose to the
original dose. The resulting optimised top dose plan was referred to as the BDF-based plan. This
workflow is depicted in Fig. 1. To generate a DCS-based plan, the isodose of 67.2 Gy (105% of
the PTV64 prescription dose) and the 45-Gy isodose in the PRV spinal cord in the original
plan were converted into dose-controlling structures, and a cold-spot dose-controlling
structure was generated from the PTV64 minus the prescription isodose volume (PIV). Then, the
dose-controlling structures for hot and cold spots were assigned new dose objectives. Typically,
for the PTV64, the upper dose objective was set to 2% lower than the prescribed dose for the
PTV64 hot spots, and the lower dose objective was set to 2% higher than the prescribed dose
for the cold spots. The upper dose objective was set to 4045 Gy for the hot spots of the PRV
spinal cord. After one-time re-optimisation and final dose calculation, the DCS-based plan was
complete. A distributed calculation framework (DCF) was applied to accelerate the final dose
calculation. The one-time re-optimisation time was defined as the time from the beginning of
re-optimisation to the completion of the final dose calculation.
According to the International Commission on Radiation Units and Measurements (ICRU)
report 83 , D98% and D2% represent the near-minimal and near-maximal doses for the PTV,
respectively. The homogeneity index (HI), as a measure of the target dose homogeneity, was
defined as follows:
A conformity index (CI)  which takes into account the overlap between target volume
(TV) and PIV, was used to quantify the target dose conformity and was defined as follows:
An HI value of 0 indicated ideal homogeneity, and a CI value of 1 indicated ideal
conformity. With regard to the PTV64, the D98%, D2% and D50% values were used to evaluate the
coldspot, hot-spot and median doses, respectively. For the PTV54, only the CI was used because
the PTV54 was not normalised and included the PTV64. The MLD, V5Gy, V10Gy, V20Gy and
Fig 1. Workflow for generating a BDF-based plan for cervical oesophageal cancer.
V30Gy values were used for the lungs, and D0.1cc was used to evaluate the near-maximum dose
of the PRV spinal cord.
The independent checking software IMSure version 3.4.1 (Standard Imaging, Middleton, USA)
and a Delta4 diode array phantom (Scandidos, Uppsala, Sweden) were used to verify the dose
accuracy of the BDF-based and DCS-based plans. The fluence for each field and the point dose
for the total plan that were re-calculated using IMSure and the 3D delivered dose that was
measured by the Delta4 phantom were compared with the corresponding values calculated in
Eclipse. The fluence discrepancy and the 3D dose discrepancy were evaluated using gamma
analysis with a criterion of 3%/3 mm (3% dose difference and 3 mm distance-to-agreement)
. The acceptable gamma pass rate was 95%, and the acceptable point-dose deviation
calculated using IMSure was within 3%. Moreover, the delivery time was recorded during the
delivery of radiation to the Delta4 phantom.
The differences among the BDF-based, DCS-based and Original plans were evaluated using
two-sided paired t-tests in which a P-value of < 0.05 was considered to be statistically
significant. SPSS version 19 software (SPSS, Inc., Chicago, IL, USA) was used to analyse the data.
Table 1 summarises the target dose-volume parameters for the 3 plans. The BDF-based plans
provided the best target dose distributions with respect to most parameters, whereas the
DCSbased plans were inferior to the BDF-based plans but superior to the Original plans. Compared
with the Original plans, the BDF-based plans demonstrated significantly improved D2%, D98%,
HI and CI values for the PTV64 and an improved CI for the PTV54 by approximately 4.4%,
0.3%, 50.3%, 11.4% and 3.7%, respectively. Compared with the DCS-based plans, the
BDFbased plans demonstrated better D2%, HI and CI values for the PTV64 and a better CI for the
PTV64 D2% (cGy)
BDF, base dose function; DCS, dose-controlling structure; SD, standard deviation; Dx, dose that is reached or exceeded in x of the volume; HI,
homogeneity index; CI, conformity index; PTV64, planning target volume receiving a prescribed dose of 64 Gy; PTV54, planning target volume receiving a
prescribed dose of 54 Gy.
* Statistical significance.
Fig 2. Dose distributions of the BDF-based, DCS-based and Original plans for one case.
PTV54 by approximately 1.9%, 25.7%, 8.3%, 3.3%, respectively, as well as a comparable D98%
value for the PTV64. The DCS-based plans showed improvements over the Original plans in
all respects except for the comparable CI for the PTV54. In the isodose distributions,
significantly fewer hot spots of 105% (67.2 Gy) of the prescribed dose for the PTV64 were observed
for the BDF-based plans, and the isodose lines appeared more conformal to the PTVs (Fig. 2).
Besides, the dose-volume histogram (DVH) curves of the PTVs seemed far steeper for the
BDF-based plans (Fig. 3).
Table 2 summarises the dose-volume parameters of the OARs for the 3 plans. In terms of the
dose delivered to the PRV spinal cord, the BDF-based plans slightly reduced the D0.1cc value of
the PRV spinal cord by 1.1 1.3% compared with the Original plans and by 2.3 1.8%
compared with the DCS-based plans. Concerning the dose delivered to the lungs, the BDF-based
plans tended to deposit slightly lower doses. The BDF-based plans yielded MLDs which were
lower by 2.7 1.8% compared with the Original plans and by 2.2 1.6% compared with the
DCS-based plans. These results are also illustrated in Fig. 3.
Efficiency of planning, dose delivery and dosimetric verifications
As shown in Table 3, the BDF-based method was more efficient than the DCS-based method
with regard to the planning time. The one-time re-optimisation required 4.06 0.9 and 5.68
1.05 minutes for the BDF- and DCS-based plans, respectively. The BDF-based method reduced
the re-optimisation time by 28.4 25.1%. The MUs of the BDF-based plans were 1.7 2.3%
and 1.2 2.4% higher than those of the DCS-based and Original plans, respectively (Table 1).
Fig 3. Dose-volume histograms (DVHs) of the BDF-based, DCS-based and Original plans for one case.
BDF, base dose function; DCS, dose-controlling structure; SD, standard deviation; Dx, dose that is reached or exceeded in x of the volume; VxGy, %
volume that received a dose of at least x Gy; PRV spinal cord, planning organ-at-risk volume of spinal cord.
* Statistical significance
The average delivery time of the BDF-based plans was 1.3 1.0% more than that of the
DCSbased plans (Table 3).
All verification tests were passed. There were no significant differences observed in terms of
the gamma pass rates indicated by the Delta4 phantom and the point-dose deviations
calculated using IMSure. The gamma pass rates of the BDF-based plans calculated using IMSure were
very slightly lower than those of the DCS-based plans, but statistically significant differences
were observed in only two fields. Nevertheless, these differences were so small as to
To improve the therapeutic ratio and obtain optimal clinical outcomes, it is important to make
full use of the IMRT technique. Our study demonstrated that the introduced BDF-based
optimisation method is capable of further improving target coverage and sparing OARs.
The most obvious advantage of the BDF-based method is that it substantially improves dose
homogeneity. Such improvement may be clinically beneficial for patients with cervical
oesophageal cancer because the PTVs for the treatment of this type of cancer commonly include
such tissues as submucosal tissue, mucosa, and bone, which may suffer complications after
receiving significantly heterogeneous high doses . Werner-Wasik et al.  have stated that a
higher dose to the oesophagus may increase the risk of oesophageal toxicity, which may be
lifethreatening, leading to such potential consequences as perforations and fistulas [23,24]. Our
study demonstrated that the BDF-based method is able to reduce hot spots by approximately
25% and provide excellent uniformity of the dose distribution, with an HI decrease of
approximately 50%. Thus, it may reduce the risk of oesophageal toxicity.
The BDF-based method also demonstrated certain advantages with regard to target
conformity and nearby-OAR sparing. It reduced the dose delivered to the spinal cord by
approximately 13%, thus theoretically reducing the risk of radiation-induced myelitis, especially for
patients with locally persistent or recurrent diseases requiring a second course of treatment.
The BDF-based method also reduced the mean dose delivered to the lungs by approximately
23%, and reduced the V5Gy, V10Gy, V20Gy and V30Gy values of the lungs. It is well known that
an overdose to the lungs may result in radiation-induced pneumonia, which may lead to death
. Many researchers have shown that the MLD, V5Gy, V10Gy and V20Gy values are useful
predictors of pneumonitis [11,26]. Kumar et al.  have also concluded that acute and chronic
pneumonitis are primarily correlated with the V30Gy and V20Gy values, respectively. As such,
reducing all the dose-volume parameters mentioned above may reduce the risk of
The BDF-based optimisation method is efficient in terms of treatment planning time,
because only one parameter, the number of fractions, must be changed and an excellent dose
distribution can be easily achieved via a simple one-time re-optimisation procedure.
Improvement of the planning efficiency is beneficial for reducing the time that patients must
wait until the start of treatment and thus for relieving patients anxieties. By contrast, the
DCSbased method is time-consuming because it always requires multiple re-optimisations to
further improve the plan, and furthermore, it takes time to delineate the dose-controlling
structures and assign new dose constraints.
Traditionally, the base dose function is used for optimising a second plan (top dose plan),
e.g., a boost plan, while taking into consideration the first plan (base dose plan), to achieve an
optimal plan sum in the optimiser but not in the final calculation. However, the base dose
function is used in a new way in the BDF-based method; here, it is employed to achieve an optimal
second plan (top dose plan) but not a plan sum, in the final calculation but not in the optimiser.
In principle, the base dose function is utilised to compensate for the OCE. When the OCE
introduces a hot spot into the final calculated dose in the original plan (base dose plan), the
second plan (top dose plan) will generate a cold spot in the same region to achieve a uniform
summed dose. After the final dose calculation, by the effect of the OCE again, the cold-spot
dose in the optimiser of the second plan (top dose plan) will approach the desired level .
A number of investigators have focused on possible methods or techniques for overcoming
the OCE. The DCS-based optimisation method described by Sss et al.  and used by
Xhaferllari et al.  is useful for compensating for the OCE, but it is only locally effective in the
dose-controlling region, and it is a trial and error approach because the additional constraints
require manual adjustments. By contrast, the BDF-based method is globally effective
throughout the entire treatment region and is a systematic approach. According to the review by
Broderick et al.  and other studies [30,31], the Direct Aperture Optimisation (DAO) technique
incorporates series of deliverable MLC shapes instead of ideal intensity maps in the optimiser
and thus is able to remove the error introduced by MLC modulation. Unfortunately, when it is
applied in cervical oesophageal cancer, the error arising from tissue heterogeneity and the
buildup effect still cannot be removed, and this error will result in hot and cold spots, according
to our experience. Additionally, this technique is not available in non-DAO treatment planning
systems, e.g., Eclipse version 10.0, whereas the BDF-based optimisation method is always
available because a base dose function or similar base dose function is a basic feature provided in
treatment planning systems for the optimisation of a second plan to achieve an optimal plan
sum. Verbakel et al.  have overcome the error originating from tissue heterogeneity by
dividing the PTV into low- and relatively high-density regions and subsequently setting a higher
dose objective for the low-density region in the optimiser. This method is effective but
minimises only one source of the OCE, and its complexity increases when dividing two or more
In addition, because there have been few reports  regarding the BDF-based method to
date, discreet dosimetric verifications should be performed to identify any error originating
from the base dose function. Our verification results indicated that the BDF-based optimisation
method offered adequate dosimetric accuracy, thus confirming the feasibility of this method in
However, the BDF-based method resulted in an increase in the MUs and delivery time by
approximately 12%, which may slightly increase the incidence of secondary cancer . The
attempt to reduce the MUs remains an interesting topic that will be investigated in our
In this study, we evaluated the dosimetric characteristics of a simple optimisation method utilising
the base dose function for cervical oesophageal cancer, and we found that this method can
improve the target dose homogeneity and conformity and reduce the doses to the OARs while
achieving adequate dosimetric accuracy, at the expense of slightly more MUs. Additionally, it
offers improved planning efficiency. Therefore, the proposed optimisation method is recommended
for incorporation into routine clinical practice for the IMRT of cervical oesophageal cancer.
Conceived and designed the experiments: JYL MC. Performed the experiments: JYL MC.
Analyzed the data: JYL MC BTH LLW WJX LXX. Contributed reagents/materials/analysis tools:
JYL MC BTH. Wrote the paper: JYL MC ZJC DRL LXX.
1. Jemal A , Center MM , DeSantis C , Ward EM . Global patterns of cancer incidence and mortality rates and trends . Cancer Epidemiol Biomarkers Prev . 2010 ; 19 : 1893 - 1907 . doi: 10.1158/ 1055 - 9965 . EPI10-0437 PMID: 20647400
2. Shridhar R , Almhanna K , Meredith KL , Biagioli MC , Chuong MD , Cruz A , et al. Radiation therapy and esophageal cancer . Cancer Control . 2013 ; 20 : 97 - 110 . PMID: 23571700
3. Cao CN , Luo JW , Gao L , Xu GZ , Yi JL , Huang XD , et al. Primary radiotherapy compared with primary surgery in cervical esophageal cancer . JAMA Otolaryngol Head Neck Surg . 2014 ; 140 : 918 - 926 . doi: 10.1001/jamaoto. 2014 . 2013 PMID: 25233363
4. Cooper JS , Guo MD , Herskovic A , Macdonald JS , Martenson JA Jr, Al-Sarraf M , et al. Chemoradiotherapy of locally advanced esophageal cancer: long-term follow-up of a prospective randomized trial (RTOG 85-01). Radiation Therapy Oncology Group . JAMA. 1999 ; 281 : 1623 - 1627 . PMID: 10235156
5. Hodapp N. The ICRU Report 83: prescribing, recording and reporting photon-beam intensity-modulated radiation therapy (IMRT) . Strahlenther Onkol . 2012 ; 188 : 97 - 99 .
6. Fenkell L , Kaminsky I , Breen S , Huang S , Van Prooijen M , Ringash J. Dosimetric comparison of IMRT vs. 3D conformal radiotherapy in the treatment of cancer of the cervical esophagus . Radiother Oncol . 2008 ; 89 : 287 - 291 . doi: 10.1016/j.radonc. 2008 . 08.008 PMID: 18789828
7. Nicolini G , Ghosh-Laskar S , Shrivastava SK , Banerjee S , Chaudhary S , Agarwal JP , et al. Volumetric modulation arc radiotherapy with flattening filter-free beams compared with static gantry IMRT and 3D conformal radiotherapy for advanced esophageal cancer: a feasibility study . Int J Radiat Oncol Biol Phys . 2012 ; 84 : 553 - 560 . doi: 10.1016/j.ijrobp. 2011 . 12.041 PMID: 22386376
8. Nutting CM , Bedford JL , Cosgrove VP , Tait DM , Dearnaley DP , Webb S. A comparison of conformal and intensity-modulated techniques for oesophageal radiotherapy . Radiother Oncol . 2001 ; 61 : 157 - 163 . PMID: 11690681
9. Shridhar R , Chuong M , Weber J , Freilich J , Almhanna K , Coppola D , et al. Outcomes of definitive or preoperative IMRT chemoradiation for esophageal cancer . J Radiat Oncol . 2012 ; 1 : 347 - 354 .
10. Yu WW , Zhu ZF , Fu XL , Zhao KL , Mao JF , Wu KL , et al. Simultaneous integrated boost intensity-modulated radiotherapy in esophageal carcinoma: early results of a phase II study . Strahlenther Onkol . 2014 ; 190 : 979 - 986 . doi: 10.1007/s00066- 014 - 0636 -y PMID : 24609941
11. Lin SH , Wang L , Myles B , Thall PF , Hofstetter WL , Swisher SG , et al. Propensity score-based comparison of long-term outcomes with 3-dimensional conformal radiotherapy vs intensity-modulated radiotherapy for esophageal cancer . Int J Radiat Oncol Biol Phys . 2012 ; 84 : 1078 - 1085 . doi: 10.1016/j. ijrobp. 2012 . 02.015 PMID: 22867894
12. Freilich J , Hoffe SE , Almhanna K , Dinwoodie W , Yue B , Fulp W , et al. Comparative outcomes for threedimensional conformal versus intensity-modulated radiation therapy for esophageal cancer . Dis Esophagus . 2014 ; In press.
13. Jeraj R , Keall PJ , Siebers JV . The effect of dose calculation accuracy on inverse treatment planning . Phys Med Biol . 2002 ; 47 : 391 - 407 . PMID: 11848119
14. Dogan N , Siebers JV , Keall PJ , Lerma F , Wu Y , Fatyga M , et al. Improving IMRT dose accuracy via deliverable Monte Carlo optimization for the treatment of head and neck cancer patients . Med Phys . 2006 ; 33 : 4033 - 4043 . PMID: 17153383
15. Budrukkar AN , Hope G , Cramb J , Corry J , Peters LJ . Dosimetric study of optimal beam number and arrangement for treatment of nasopharyngeal carcinoma with intensity-modulated radiation therapy . Australas Radiol . 2004 ; 48 : 45 - 50 . PMID: 15027920
16. Cheng MC , Hu YW , Liu CS , Lee JS , Huang PI , Yen SH , et al. Optimal beam design on intensity-modulated radiation therapy with simultaneous integrated boost in nasopharyngeal cancer . Med Dosim . 2014 ; 39 : 246 - 250 . doi: 10.1016/j.meddos. 2014 . 03.003 PMID: 24857279
17. Xhaferllari I , Wong E , Bzdusek K , Lock M , Chen J. Automated IMRT planning with regional optimization using planning scripts . J Appl Clin Med Phys . 2013 ; 14 : 4052. doi: 10.1120/jacmp. v14i1.4052 PMID: 23318393
18. Suss P , Bortz M , Kufer KH , Thieke C. The critical spot eraser-a method to interactively control the correction of local hot and cold spots in IMRT planning . Phys Med Biol . 2013 ; 58 : 1855 - 1867 . doi: 10. 1088/ 0031 - 9155 /58/6/1855 PMID: 23442519
19. Paddick I. A simple scoring ratio to index the conformity of radiosurgical treatment plans . Technical note. J Neurosurg . 2000 ; 93 : 219 - 222 . PMID: 11143252
20. Ezzell GA , Burmeister JW , Dogan N , LoSasso TJ , Mechalakos JG , Mihailidis D , et al. IMRT commissioning: multiple institution planning and dosimetry comparisons, a report from AAPM Task Group 119 . Med Phys . 2009 ; 36 : 5359 - 5373 . PMID: 19994544
21. Vineberg KA , Eisbruch A , Coselmon MM , McShan DL , Kessler ML , Fraass BA . Is uniform target dose possible in IMRT plans in the head and neck? Int J Radiat Oncol Biol Phys . 2002 ; 52 : 1159 - 1172 . PMID: 11955726
22. Werner-Wasik M , Yorke E , Deasy J , Nam J , Marks LB . Radiation dose-volume effects in the esophagus . Int J Radiat Oncol Biol Phys . 2010 ; 76 : S86 - 93 . doi: 10.1016/j.ijrobp. 2009 . 05.070 PMID: 20171523
23. Kwint M , Uyterlinde W , Nijkamp J , Chen C , de Bois J , Sonke JJ , et al. Acute esophagus toxicity in lung cancer patients after intensity modulated radiation therapy and concurrent chemotherapy . Int J Radiat Oncol Biol Phys . 2012 ; 84 : e223 - 228 . doi: 10.1016/j.ijrobp. 2012 . 03.027 PMID: 22560551
24. Ahn SJ , Kahn D , Zhou S , Yu X , Hollis D , Shafman TD , et al. Dosimetric and clinical predictors for radiation-induced esophageal injury . Int J Radiat Oncol Biol Phys . 2005 ; 61 : 335 - 347 . PMID: 15667951
25. Guckenberger M , Baier K , Polat B , Richter A , Krieger T , Wilbert J , et al. Dose-response relationship for radiation-induced pneumonitis after pulmonary stereotactic body radiotherapy . Radiother Oncol . 2010 ; 97 : 65 - 70 . doi: 10.1016/j.radonc. 2010 . 04.027 PMID: 20605245
26. Tucker SL , Liu HH , Wang S , Wei X , Liao Z , Komaki R , et al. Dose-volume modeling of the risk of postoperative pulmonary complications among esophageal cancer patients treated with concurrent chemoradiotherapy followed by surgery . Int J Radiat Oncol Biol Phys . 2006 ; 66 : 754 - 761 . PMID: 16965865
27. Kumar G , Rawat S , Puri A , Sharma MK , Chadha P , Babu AG , et al. Analysis of dose-volume parameters predicting radiation pneumonitis in patients with esophageal cancer treated with 3D-conformal radiation therapy or IMRT . Jpn J Radiol . 2012 ; 30 : 18 - 24 . doi: 10.1007/s11604- 011 - 0002 - 2 PMID: 22160648
28. Lu JY , Wu LL , Zhang JY , Zheng J , Cheung ML , Ma CC , et al. Improving target dose coverage and organ-at-risk sparing in intensity-modulated radiotherapy of advanced laryngeal cancer by a simple optimization technique . Br J Radiol . 2015 ; 88 : 20140654. doi: 10.1259/bjr.20140654 PMID: 25494885
29. Broderick M , Leech M , Coffey M. Direct aperture optimization as a means of reducing the complexity of Intensity Modulated Radiation Therapy plans . Radiat Oncol . 2009 ; 4 : 8. doi: 10.1186/ 1748 - 717X -4-8 PMID: 19220906
30. Jones S , Williams M. Clinical evaluation of direct aperture optimization when applied to head-and-neck IMRT . Med Dosim . 2008 ; 33 : 86 - 92 . doi: 10.1016/j.meddos. 2007 . 04.002 PMID: 18262129
31. Dobler B , Pohl F , Bogner L , Koelbl O. Comparison of direct machine parameter optimization versus fluence optimization with sequential sequencing in IMRT of hypopharyngeal carcinoma . Radiat Oncol . 2007 ; 2 : 33 . PMID: 17822529
32. Verbakel WF , van Reij E , Ladenius-Lischer I , Cuijpers JP , Slotman BJ , Senan S. Clinical application of a novel hybrid intensity-modulated radiotherapy technique for stage III lung cancer and dosimetric comparison with four other techniques . Int J Radiat Oncol Biol Phys . 2012 ; 83 : e297 - 303 . doi: 10.1016/j. ijrobp. 2011 . 12.059 PMID: 22579380
33. Hall EJ , Wuu CS. Radiation-induced second cancers: the impact of 3D-CRT and IMRT . Int J Radiat Oncol Biol Phys . 2003 ; 56 : 83 - 88 . PMID: 12694826