CONKISS: CONFORMAL KIDNEYS SPARING 3D NONCOPLANAR
RADIOTHERAPY TREATMENT FOR PANCREATIC CANCER AS AN
ALTERNATIVE TO IMRT
ZSOLT SEBESTYÉN, M.SC., PÉTER KOVÁCS, M.SC., ÁKOS GULYBÁN, M.SC.,
RÓBERT FARKAS, M.D., SZABOLCS BELLYEI, M.D., PH.D., GÁBOR LIPOSITS, M.D.,
ANDRÁS SZIGETI, M.D., PH.D., OLGA ÉSIK, M.D., D.SC., KATALIN DÉRCZY, M.D.,
and LÁSZLÓ MANGEL, M.D., PH.D.
University of Pécs, Institute of Oncotherapy, Pécs, Hungary; EORTC Headquarters, Brussels, Belgium; and University of
Pécs, Department of Radiology, Pécs, Hungary
(Received 4 June 2009; accepted 10 November 2009)
Abstract—When treating pancreatic cancer using standard (ST) 3D conformal radiotherapy (3D-CRT) beam
arrangements, the kidneys often receive a higher dose than their probable tolerance limit. Our aim was to
elaborate a new planning method that—similarly to IMRT—effectively spares the kidneys without compromis-
ing the target coverage. Conformal kidneys sparing (CONKISS) 5-field, noncoplanar plans were compared with
ST plans for 23 consecutive patients retrospectively. Optimal beam arrangements were used consisting of a left-
and right-wedged beam-pair and an anteroposterior beam inclined in the caudal direction. The wedge direction
determination (WEDDE) algorithm was developed to adjust the adequate direction of wedges. The aimed organs
at risk (OARs) mean dose limits were: kidney <12 Gy, liver <25 Gy, small bowels <30 Gy, and spinal cord
maximum <45 Gy. Conformity and homogeneity indexes with z-test were used to evaluate and compare the
different planning approaches. The mean dose to the kidneys decreased significantly (p < 0.05): left kidney 7.7
vs. 10.7 Gy, right kidney 9.1 vs. 11.7 Gy. Meanwhile the mean dose to the liver increased significantly (18.1 vs.
15.0 Gy). The changes in the conformity, homogeneity, and in the doses to other OARs were not significant. The
CONKISS method balances the load among the OARs and significantly reduces the dose to the kidneys, without
any significant change in the conformity and homogeneity. Using 3D-CRT the CONKISS method can be a smart
alternative to IMRT to enhance the possibility of dose escalation.
© 2011 American Association of Medical
Key Words: Pancreatic cancer, 3D conformal radiotherapy, Noncoplanar fields, Dosimetric comparison, IMRT.
Pancreatic cancer is the fourth leading cause of cancer
mortality in the world.1,2The optimal strategy for treating
curable using the existing treatment techniques. Several
authors have already published the importance of different
chemotherapies used as part of a chemoradiotherapy
(CHRT) treatment of patients who present with unresect-
able, locally advanced pancreatic cancer. Considering
these data, radiotherapy (RT) is widely used as part of
the treatment strategy.1,3–7
Delivering adequate radiation doses to the pan-
creas is limited by the presence of radiation-sensitive
normal structures in the upper abdomen. These include
the kidneys, liver, small bowels, stomach, and spinal
The 5-fluorouracil (FU)–based CHT combined with
the standard (ST) 3D conformal RT treatment (3D-CRT)
technique was used in our department.8The disadvan-
tage of the ST technique is that the kidneys often receive
higher mean dose than their generally accepted tolerance
limit. Our aim was to find a conformal treatment tech-
nique that delivers a lower dose to the kidneys than their
tolerance limit—similar to intensity-modulated radiation
therapy (IMRT),9but taking minimal time and technical
METHODS AND MATERIALS
Between February 2005 and August 2008, 23 con-
secutive patients in our department with locally ad-
vanced, unresectable pancreatic cancer were treated with
ST 3D-CRT technique.8The patient immobilization was
done using individual vacuum cushion in supine posi-
tion. During the RT procedure 10-mm increment com-
puter tomography (CT) scans were taken with a Siemens
Somatom CT (Siemens, Erlangen, Germany) scanner
and transferred to our Precise Plan treatment planning
system (TPS) (Elekta, PrecisePLAN 2.02/2.03, Atlanta
GA). The prescribed dose was 45 Gy to the PTV in 1.8
Gy per fractions. During the planning process we fol-
lowed the ICRU 50, 62 recommendations.10,11
Reprint requests to: Zsolt Sebestyén, M.Sc., H-7624 Pécs, Éde-
sanyák útja 17, Hungary. E-mail: email@example.com
Medical Dosimetry, Vol. 36, No. 1, pp. 35-40, 2011
Copyright © 2011 American Association of Medical Dosimetrists
Printed in the USA. All rights reserved
0958-3947/11/$–see front matter
First, the primary gross tumor volume (GTV) and
the clinical target volume (CTV) were defined. Organ
motion and set-up errors were also considered, thus the
planning target volume (PTV) was defined as CTV with
a uniform margin of 15 mm. The clinically uninvolved
regional lymphatics were not included into any of the
target volumes. As organs at risk (OARs), the kidneys,
liver, small bowels, and spinal cord were contoured on
all CT images.
Planning priorities and OAR tolerance dose limits
The main priority was to deliver the 45-Gy prescribed
mean dose to the PTV homogeneously. The second priority
was to keep the OARs’ mean dose and percentage volume
below their tolerance limits (Table 1).6,9,12,13The kidney
and the spinal cord limit were respected with higher priority
within the OARs.
ST 3D-CRT treatment planning
The ST 3D-CRT plans consisted of 3 fields includ-
ing an open anteroposterior (AP) and two opposed,
wedged lateral 6-MV photon beams.8The isocenter was
defined to the geometrical center of the PTV. For gen-
erating MLC fields, the following shapes were used:
10-mm margin around the PTV from beam’s eye view
(BEV), except near the kidneys and the liver, where they
were manually reduced to 3 and 8 mm, respectively. The
beam-weights were optimized with the IMRT optimizing
module of TPS to achieve 45 Gy mean dose to the PTV.
CONKISS planning method
The baseline of the conformal kidneys sparing
(CONKISS) 5-field beam arrangement was (Fig. 1): 1
AP-like beam with 40° gantry angle and 90° table angle
(G40-T90) and 4 lateral fields: G270-T340, G90-T340,
G270-T20, and G90-T20 followed by individual adjust-
ment. The isocenter was moved from the center of the
PTV anteriorly considering the following:
1. Isocenter should not be closer than 1 cm to the PTV
border, for adequate dose calculation
2. The AP-like beam is not causing gantry-patient/table
Individual beam direction adjustment—equal area
The gantry angles of the lateral fields were adjusted
so that from their BEV the same kidney areas—from
both of the kidneys—were overlapped in the PTV. The
table angle of the AP-like beam was adjusted so that
again the same areas of the kidneys were overlapped in
Wedge direction adjustment
A 60° wedge was used in all 4 lateral beams. In this
article the direction of a wedge is defined as the direction
where the wedge has greater blocking effect. The wedges
of the 2 lateral fields closer to the AP-like beam were
directed to the other lateral beams on the same side. In
the other 2 lateral beams the wedges were directed to the
AP-like beam (Fig. 1).
Wedge direction determination algorithm
(WEDDE) algorithm was made to determine the proper
collimator angle for the required wedge direction (Fig. 2). It
used a polar coordinate system from the table point of
view (POV) (Fig. 2). In this POV the gantry could move
on a unit-radius sphere, limited by the physical exten-
sions of the gantry, table, and patient.
Determination of the required collimator angle
The initial direction of the wedges was always the
upper direction—when the wedge was directed to an AP
beam. The algorithm converted the polar coordinates of
the points on Fig. 2 to Cartesian coordinates. We deter-
mined the equation of the 2 planes defined by points AP,
O, WB and O, WB, B (Fig. 2). Then we determined their
dihedral angle (the angle between these two planes)—
what is the required collimator rotation angle to direct
Fig. 1. The beam arrangement and the wedge directions of the
Table 1. OAR tolerance limits*
PTV CoverageV95–107%as High as Possible
OAR Mean Dose Limit Vx limit
V20 ? 30%
V35 ? 33%
V45 ? 10%
V4 ? 0%
Abbreviations: OAR ? organ at risk; Vx (%) ? percentage of total
volume receiving ? Gy.
*These are mainly institutional guidelines used in the literature.6,9,12,13
Medical DosimetryVolume 36, Number 1, 201136
the wedge to another beam. Using these principles, our
algorithm determined the exact collimator angle in 4
lateral fields. This method can be applied in all similar
treatment planning situations.
MLC setting adjustment
The generation of MLC fields and the beam weight
optimization was done in the same way as in case of the
ST technique. When the mean dose to the kidneys was
less than 50% of their tolerance limit, the previously
reduced margins were increased either until the mean
kidney dose reached 66% of the tolerance limit or until it
reached the original value. The procedure was named the
To further increase PTV homogeneity and to reduce
the maximum dose value, a second segment was used in
the AP-like beam that excluded the highest 2–3% dose
cloud from BEV, similarly to Gulybán et al.14
Figure 3 shows the workflow of the whole CONKISS
Plan evaluation and comparison
The conformity of the plans was evaluated using the
conformation number (CN) in the following formula:
where VT,PIis the volume of PTV receiving at least the
prescription dose, VPIis the volume enclosed by the
prescription isodose, and VTis the PTV.15,16
The homogeneity was evaluated in 2 different ways
using the cumulative dose volume histogram (DVH):
First according to ICRU 50, 62,10,11where the V95–107%
represents the percentage of PTV between 95–107% of
the prescribed dose; second, using the D95–5%
where D5%, and D95%were the doses received by 5% and
95% of the PTV volume, and PI is the prescribed isodose.
With regard to the OARs, we evaluated the mean
dose to the kidneys, liver, and small bowel; the maxi-
mum dose to spinal cord; and the percentage of kidneys
and total kidney volumes receiving 20 Gy (V20), liver
V35, and small bowel V45.6,9To compare the 2 tech-
niques, relative evaluation was performed using the per-
centage OAR dose reduction values.18
All data are presented in mean dose ? standard
deviation and as percentage of tolerance limit. Two-
tailed t significance tests were performed to compare the
results of the 2 techniques. The 5% probability level (p ?
0.05) was considered to be statistically significant.
Fig. 3. The workflow of the CONKISS method.
Fig. 2. The model used in the WEDDE algorithm, where WB
represents the gantry position of the wedged beam, B represents
the gantry position of the beam where the wedge in WB will
direct, O is the place of isocentre, and AP represents the gantry
position of the AP beam.
CONKISS 3D-CRT pancreas treatment method ● Z. SEBESTYÉN et al.
The mean PTV volume was 657.8 cm3(range 296–
1080). The CONKISS plans resulted in a better V95–107%
and D95–5%homogeneity and in a slightly worse confor-
mity (Table 2). None of these differences was statisti-
Dose to OARs
With the ST plans, the mean dose to the right kidney
exceeded its defined tolerance limit in 10 cases, for the
left kidney in 8 cases, and for the total kidney in 9 caes.
With the CONKISS method, this number was reduced to
4, 2, and 3, respectively. All of the other OARs’ mean
doses—liver V35, small bowel V45—and the spinal cord
maximum doses were for both of the techniques under
their tolerance limits.
Comparison of the OARs’ mean doses and percent-
age volumes is shown in Table 2. With the CONKISS
technique, the mean left, right, and total kidney doses
were significantly reduced. The mean dose to the liver
significantly increased, whereas the liver V35 decreased.
The differences between the other mean doses and per-
centage volumes were not statistically significant.
With the CONKISS method, the following mean
dose reductions were achieved: left kidney 28.0%, right
kidney 22.2%, total kidney 24.3%. The mean dose to the
liver increased by 20.7%. Concerning the percentage
volumes, the reduction was 26.1, 24.2, 25.0, and 12.3%,
for the left, right, and total kidney and for the liver,
respectively (Table 2). For the CONKISS plans, the
mean doses to the kidneys and to the liver in percentages
of their tolerance limits were similar: left kidney 64%,
right kidney 76%, total kidney 70%, and liver 72%. The
CONKISS method allowed balancing the doses to the
kidneys and to the liver compared wth the ST technique,
where these percentages were 89, 98, 93, and 60%,
respectively (Fig. 4). The doses to the other OARs re-
mained under ?50% of their tolerance limits and none of
these changes were statistically significant.
While developing the CONKISS method, we ap-
plied retrospectively more than 30 different 3-field to
7-field, mainly noncoplanar beam arrangements with dif-
ferent photon energies. Some of them were better only
Table 2. ST—CONKISS comparison10,11,19–21
ST ? SDCONKISS ? SD
Reduction in % (CONKISS/ST)
Mean dose (Gy)
Mean dose (Gy)
Mean dose (Gy)
Mean dose (Gy)
Mean dose (Gy)
95.5 ? 2.6
8.4 ? 2.7
0.656 ? 0.06
96.4 ? 2.1
7.6 ? 2.1
0.636 ? 0.06
10.7 ? 4.2
11.5 ? 10.0
7.7 ? 2.8
8.5 ? 6.7
11.7 ? 5.0
12.8 ? 12.6
9.1 ? 3.7
9.7 ? 7.9
11.1 ? 4.1
12.0 ? 10.1
8.4 ? 3.1
9.0 ? 7.1
15.0 ? 3.8
13.8 ? 7.8
18.1 ? 3.3
12.1 ? 6.3
11.9 ? 6.2
4.3 ? 3.8
14.6 ? 6.4
5.1 ? 5.1
15.7 ? 3.015.2 ? 4.8 NS—
Abbreviations: OAR ? organs at risk; PTV ? planning target volume; ST 3D-CRT ? standard 3D conformal radiotherapy treatment (technique);
CONKISS ? conformal kidneys sparing (method); Vx (%) ? percentage of total volume receiving ? Gy; NS ? not significant (p ? 0.05).
Fig. 4. Balancing the load among the OARs. ST ? standard;
CONKISS ? conformal kidneys sparing (method).
Medical DosimetryVolume 36, Number 1, 2011 38
for a few patients similar to other reported methods.18
The previous experiences were used to develop the
CONKISS method, which had better results for all of our
patients. Similarly to Higgins et al.,22we found that the
6-MV plans were superior to the 18-MV plans. Osborne
et al.23reported a comparison of noncoplanar and copla-
nar techniques to treat pancreatic cancer based on normal
tissue complication probability (NTCP) and on total
weighted equivalent uniform dose (EUD) calculations.
They found that noncoplanar techniques have an overall
benefit compared with coplanar techniques. Our experi-
ences similarly showed that coplanar beam arrangements
were worse than the noncoplanar CONKISS method.
The lower SD values of the CONKISS method
show that the reproduction of its result is easier than that
of the ST technique.
Advantages of lateral beam directions
Based on various reports,24,25the respiration-
induced movement of the pancreas and the OARs in the
AP direction is the least compared with the movements
in other directions: the movements of the pancreas in the
craniocaudal direction an average 21.6 mm, in the LR
direction an average 12.0 mm, and in the AP direction an
average 6.0 mm.24The use of mostly lateral fields al-
lowed a higher probability in delivering the planned dose
to the PTV and to the OARs. Another advantage was that
the kidneys received the least dose when the lateral fields
went through the least kidney area seen from BEV.
CONKISS vs. IMRT comparison
Brown et al.9compared 2 IMRT and 1 conformal
technique for 15 patients retrospectively. The average
volume of their PTV was similar: 678.2 cm3(PTV1).
Their prescription dose was different: 45 Gy to the PTV
(PTV1); 59.4 Gy to the PTV-0.5 cm (PTV2); and 64.8
Gy to the PTV-1 cm (PTV3). To compare the results, we
increased the number of fractions to 64.8 Gy without
reducing the original PTV, thus the doses to the OARs
were considerably overestimated. Table 3 shows the
comparison of the OAR percentage volumes.
In our plans, the V20 for the total kidney was still
smaller than for the IMRTi and IMRTs techniques
(19.9% for the CONKISS plans and 27.7 and 22.3% for
the IMRTs and IMRTi plans, respectively).
Balancing the dose to the OARs
According to Wilkowski et al., concurrent chemo-
therapy can significantly reduce the tolerance level of the
kidneys; therefore they aimed not to expose 30% of a
kidney to more than 20 Gy.12If one kidney is not
functioning well than it can be sacrificed to spare the
other well-functioning kidney as much as possible. With
the 1/2¡2/3 rule, the CONKISS method takes into con-
sideration what could be more important: lower dose to
both or only one kidney or better PTV coverage.
The issue concerning the liver seems to be contro-
versial. On one hand, Dawson et al.—based on NTCP
estimation—indicated higher tolerance of the liver tis-
sue: just 5% risk of radiogenic liver damage at 47 Gy, or
31 Gy for 66% or 100% of the liver volume,13respec-
tively. On the other hand, according to Wilkowski et al.,
the dose tolerance limit of the liver should be further
reduced as a result of concurrent chemotherapy to a
maximum 25 Gy, or 37.5 Gy for 50% or 25% of the liver
volume, respectively.12Based on a liver function test,
a patient-specific liver dose tolerance limit should be
Using the CONKISS method the dose to the kid-
neys and the liver will be almost the same in percentage
of their tolerance limits: left kidney 64%, right kidney
76%, total kidney 70%, liver 72%; thus the CONKISS
method makes a balance between the kidneys and the
The fact that the mean dose to the liver increased
while its V35 decreased shows that the increase in the
overall biological effect, because of increased mean liver
Table 3. Comparison of the OAR percentage volumes for the 3D-CRT, IMRTi, IMRTs, ST 3D-CRT, and
CONKISS plans for a total 64.8 Gy dose
Tolerance Limit 3D-CRT
% of ST 3D-CRT% of 3D-CRT% of 3D-CRT
PTV mean dose
45 ? 9 ? 5.4 ? 9 Gy
Abbreviations: PTV ? planning target volume; ST 3D-CRT ? standard 3D conformal radiotherapy treatment (technique); CONKISS ? conformal
kidneys sparing (method); IMRT ? intensity-modulated radiotherapy; IMRTi ? simultaneous integrated IMRT boost; IMRTs ? sequential IMRT
boosting;9Vx (%) ? percentage of total volume receiving ? Gy.
CONKISS 3D-CRT pancreas treatment method ● Z. SEBESTYÉN et al.
dose, would presumably not be so severe because simul- Download full-text
taneously the liver V35 decreased.
The CONKISS method is an effective and individu-
alizable treatment planning method to significantly re-
duce the dose to kidneys, without any significant change
in the conformity and homogeneity. This OAR sparing
could potentially allow either dose escalation, thus fur-
ther enhancing the loco regional control or further de-
creasing the possibility of OAR related side effects, thus
ensuring the possibility to apply any further chemother-
apy regimens. The WEDDE algorithm gives possibility
to develop other new conformal planning techniques to
improve OAR sparing, similarly to the CONKISS
method. Using 3D-CRT, the CONKISS method can be a
simple, smart alternative to IMRT.
Acknowledgment—The authors thank Markus Alber, Judit Boda-
Heggemann, Pierre Pilette, and Katalin Hideghety for the critical
reading of the manuscript.
1. Brade, A.; Brierley, J.; Oza, A.; et al. Concurrent gemcitabine and
radiotherapy with and without neoadjuvant gemcitabine for locally
advanced unresectable or resected pancreatic cancer: a phase I-II
study. Int. J. Radiat. Oncol. Biol. Phys. 67:1027–36; 2007.
2. Jemal, A.; Siegel, R.; Ward, E.; et al. Cancer statistics, 2008. CA.
Cancer. J. Clin. 58:71–96; 2008.
3. Ko, A.H.; Quivey, J.M.; Venook, A.P.; et al. A phase II study of
fixed-dose rate gemcitabine plus low-dose cisplatin followed by
consolidative chemoradiation for locally advanced pancreatic can-
cer. Int. J. Radiat. Oncol. Biol. Phys. 68:809–16; 2007.
4. Keene, K.S.; Rich, T.A.; Penberthy, D.R.; et al. Clinical experi-
ence with chronomodulated infusional 5-fluorouracil chemoradio-
therapy for pancreatic adenocarcinoma. Int. J. Radiat. Oncol. Biol.
Phys. 62:97–103; 2005.
5. Murphy, J.D.; Adusumilli, S.; Griffith, K.A.; et al. Full-dose gem-
citabine and concurrent radiotherapy for unresectable pancreatic
cancer. Int. J. Radiat. Oncol. Biol. Phys. 68:801–8; 2007.
6. Ben-Josef, E.; Shields, A.F.; Vaishampayan, U.; et al. Intensity
modulated radiotherapy (IMRT) and concurrent capecitabine for
pancreatic cancer. Int. J. Radiat. Oncol. Biol. Phys. 59:454–9;
7. Spry, N.; Harvey, J.; MacLeod, C.; et al. 3D Radiotherapy can be
safely combined with sandwich systemic gemcitabine chemother-
apy in the management of pancreatic cancer: Actors influencing
outcome. Int. J. Radiat. Oncol. Biol. Phys. 70:1438–46; 2008.
8. Dobbs, J.; Barrett, A.; Ash, D. Practical radiotherapy planning. 3rd
ed. Bristol: Arnold and Oxford University Press; 1999:247–52.
9. Brown, M.W.; Ning, H.; Arora, B.; et al. A dosimetric analysis of
dose escalation using two intensity-modulated radiation therapy
techniques in locally advanced pancreatic carcinoma. Int. J. Ra-
diat. Oncol. Biol. Phys. 65:274–83; 2006.
10. ICRU. Prescribing, recording, and reporting photon beam therapy.
Report 50. Bethesda, MD: International Commission on Radiation
Units and Measurements; 1993.
11. ICRU. Prescribing and reporting photon beam therapy. Report 62.
Washington D.C.: International Commission on Radiation Units
and Measurements; 1999.
12. Wilkowski, R.; Thoma, M.; Weingandt, H.; et al. Chemoradiation
for ductal pancreatic carcinoma: Principles of combining chemo-
therapy with radiation, definition of target volume and radiation
dose [review]. JOP. 6:216–30; 2005.
13. Dawson, L.A.; Ten Haken R.K.; Lawrence, T.S. Partial irradiation
of the liver [abstract]. Semin. Radiat. Oncol. 11:240–6; 2001.
14. Gulyban, Á.; Kovács, P.; Sebestyén, Z.; et al. Multisegmented
tangential breast fields: A rational way to treat breast cancer.
Strahlenther. Oncol. 184:262–9; 2008.
15. Van’t Riet, A.; Mak, A.C.; Moerland, M.A.; et al. A conformation
number to quantify the degree of conformity in brachytherapy and
external irradiation: Application to the prostate. Int. J. Radiat.
Oncol. Biol. Phys. 37:731–6; 1997.
16. Feuvret, L.; Noël, G.; Mazeron, J.J.; et al. Conformity index: A
review. Int. J. Radiat. Oncol. Biol. Phys. 64:333–42; 2006.
17. Van Asselen, B.; Raaijmakers, C.P.J.; Hofman, P.; et al. An improved
breast irradiation technique using three-dimensional geometrical in-
formation and intensity modulation. Radiother. Oncol. 58:341–7;
18. Hsiung-Stripp, D.C.; McDonough, J.; Masters, H.M.; et al. Com-
parative treatment planning between proton and X-ray therapy in
pancreatic cancer. Med. Dosim. 26:255–9; 2001.
19. Menhel, J.; Levin, D.; Alezra, D.; et al. Assessing the quality of
conformal treatment planning: A new tool for quantitative com-
parison. Phys. Med. Biol. 51:5363–75; 2006.
20. Lomax, N.J.; Scheib, S.G. Quantifying the degree of conformity in
radiosurgery treatment planning. Int. J. Radiat. Oncol. Biol. Phys.
21. Weber, D.C.; Trofimov, A.V.; Delaney, T.F.; et al. A treatment
planning comparison of intensity modulated photon and proton
therapy for paraspinal sarcomas. Int. J. Radiat. Oncol. Biol. Phys.
22. Higgins, P.D.; Sohn, J.W.; Fine, R.M.; et al. Three-dimensional con-
formal pancreas treatment: Comparison of four- to six-field tech-
niques [abstract]. Int. J. Radiat. Oncol. Biol. Phys. 31:605–9; 1995.
23. Osborne, C.; Bydder, S.A.; Ebert, M.A.; et al. Comparison of
non-coplanar and coplanar techniques to treat cancer of the pan-
creas. Australas. Radiol. 50:463–7; 2006.
24. Bussels, B.; Goethals, L.; Feron, M.; et al. Respiration-induced
movement of the upper abdominal organs: A pitfall for the three-
dimensional conformal radiation treatment of pancreatic cancer.
Radiother. Oncol. 68:69–74; 2003.
25. Gierga, D.P.; Chen, G.T.Y.; Kung, J.H.; et al. Quantification of
respiration-induced abdominal tumor motion and its impact on
IMRT dose distributions. Int. J. Radiat. Oncol. Biol. Phys. 58:
Medical DosimetryVolume 36, Number 1, 2011 40