A MORPHING TECHNIQUE TO ESTIMATE LUNG CANCER

DEFORMATION DUE TO BREATHING IN RADIOTHERAPIC

TREATMENT

Grazia Maria Pia Masselli, Luigi Battista, Sergio Silvestri

Faculty of Biomedical Engineering, University Campus Bio-Medico, via Alvaro del Portillo 200, 00128 Rome, Italy

Sara Ramella, Lucio Trodella

Radiotherapy and Oncology Unit, University Hospital, University Campus Bio-Medico

via Alvaro del Portillo 200, 00128 Rome, Italy

Keywords: Medical Image Detection, Acquisition, Analysis and Processing.

Abstract: A morphing technique aimed to correlate lung cancer patient’s chest cross circumference variations with

tumor morphology during quiet respiration is here described. Two CT slices corresponding to the same

tumor section are acquired at forced inspiration and forced expiration and correlated with chest

circumference values. An image sequence has been obtained by applying a linear morphing transformation.

Each image of the sequence has been associated with a chest circumferential value and a sequence subset

images corresponding to subject’s tidal volume has then been selected and compared with a CT slice

acquired at tidal volume. Images showing the minimum pixel differences with slice at tidal volume were

identified and associated with chest circumference values, allowing to estimate in which phase of the

breathing period the CT scan was carried out. CT exams in free-breathing and breath-hold conditions have

been conducted on a lung cancer patient in order to correlate the acquired slices with the variations of

patient’s chest circumference measured with a pneumatic strain gauge. The here described methodology

could allow to define the area to be irradiated during a particular phase of the breathing period, considering

the cancer area in the morphing simulation frame corresponding to this phase as target.

1 INTRODUCTION

Tumour motion due to respiration is an important

key issue for the development of accurate radiation

treatment of neoplasms located in lungs and

abdominal sites since, as it is well known, the

movement and deformation of tumors during the

breathing cycle affect not only the accuracy of CT

imaging but also the possibility of a successful

focused radiation treatment (Webb, 2006). Thus,

organ deformation during radiation delivery is a

geometric uncertainty that must be taken into

account in order to improve the quality and the

accuracy of radiotherapic treatment. The traditional

approach, according to ICRU (International

Commission on Radiation Units and Measurements)

Report 50 (ICRU Report 50, 1993), considers safety

margins around Gross Tumour Volume (GTV)

defined from a free-breathing CT scan: this method

estimates the extent of setup uncertainty and organ

motion and adds margins around a Clinical Tumour

Volume (CTV) to form a Planning Target Volume

(PTV). Several studies of the internal motion of the

tumour have been conducted based on the

hypothesis of rigid motion (Lujian et al, 1999, Wu et

al, 2004, Report 91 del AAPM Task Group 76,

2006) and different models of organ motion due to

respiration interpolating experimental data have

been proposed; moreover, it was observed that

abdominal organs motion due to respiration is well

correlated with diaphragm motion and it is

predominant in the craniocaudal direction. Some

authors (Lujian et al, 1999) have evaluated dose

delivered to the moving organ undergoing radiation

treatment as a function of the dose value predicted in

static case, considering an unidirectional movement

of the organ.The aim of focusing the dose within the

363

Maria Pia Masselli G., Battista L., Silvestri S., Ramella S. and Trodella L. (2010).

A MORPHING TECHNIQUE TO ESTIMATE LUNG CANCER DEFORMATION DUE TO BREATHING IN RADIOTHERAPIC TREATMENT.

In Proceedings of the Third International Conference on Bio-inspired Systems and Signal Processing, pages 363-366

DOI: 10.5220/0002591203630366

 SciTePress

region of interest minimizing the irradiation of

surrounding tissues, can be accomplished with

various methods such as, among others: 1)

continuous tracking systems, that allow the

irradiation of tumors while patient is breathing

normally –still in a research phase– (Murphy, 2007)

; 2) breath-hold techniques (Mageras and Yorke,

2004); 3) respiratory gated radiotherapy (Keall et al,

2006). Visual biofeedback techniques have also been

developed in order to reduce respiratory amplitude

(Masselli et al, 2009) and enhance respiration

reproducibility (Masselli et al, 2009, George et al,

2006)

and, in a preliminary way, for predicting

organ motion due to respiration during radiation

treatment (Briere et al, 2006, Venkat et al, 2008).

Recently, “morphing” techniques have been used in

a radiotherapy scenario (Deurloo et al, 2005), in

order to develop a method for the quantification of

tumor form variations in complex organs with

respect to the mean GTV obtained from elaboration

of CT slices acquired during free-breathing.

“Morphing” stands for “metamorphosing” and

indicates one of the first special digital effects used

in motion pictures and animations that allows to

transform a source image into a target image through

a seamless, fluid and gradual transition (Gomes et al,

1999). A morphing technique aimed to estimate lung

tumor deformation due to a lung cancer patient’s

breathing, in order to correlate patient’s chest

circumference variations with tumor morphology is

here proposed. The methodology allows to define

the area to be irradiated during a particular phase of

the breathing period, considering the cancer area in

the morphing simulation frame corresponding to this

phase as target.

2 METHODOLOGY

DESCRIPTION

A morphologic transformation is here used for

creating transitions between two morphological

configurations. Beyond the acquisition of a 2D CT

slice during patient freely breathing at tidal volume

(CT

), in which lung tumour border line has

motion shadings because of breathing, CT images

have been acquired in breath-hold condition, i.e.

forced inspiration and forced expiration in order to

obtain static images of the lung tumour. These two

slices, approximately corresponding to the same

tumour section and referred to the two different

tumour configurations, called CT

MaxExp

and CT

MaxInsp

in the following, have been considered as the source

image and the target image for morphing sequence.

A pneumatic strain gauge (PSG) (Masselli et al,

2009) has been used in order to measure the

variations of patient’s chest cross section

circumference ∆C during the above reported CT

exams carried out in breath-hold and free-breathing

conditions. Values ΔC

MaxExp

and ΔC

MaxInsp

have been

measured in breath-hold conditions. During quiet

respiration the ∆C(t) has been measured, according

to patient’s respiratory pattern, obtaining an interval

of ∆C values comprised between ∆C

TVmin

and

∆C

TVmax

, which are the tidal volume chest

circumference at quiet expiration and inspiration end

respectively. The images CT

MaxInsp

and CT

MaxExp

have been loaded on a 2D morphing program

(http://www.stoik.com/) in order to generate a

simulation of the lung tumour deformation from

forced expiration to forced inspiration. After control

point allocation, the number of frames of the

morphing simulation has been set. Thus, the

program allowed to transform the markers on source

image in markers on target image. Marker

movement was regulated by a distortion curve: we

have considered a linear transformation. The quality

of simulation depends on the number and position of

chosen markers. In order to correlate ∆C values with

lung tumour morphology, the correspondence

between the number of each frame of morphing

simulation and ∆C values has been found

considering a n+2 frames sequence, where n is the

number of simulation frames generated by the

program and 2 are the source image CT

MaxInsp

(frame

0) and the target image CT

MaxExp

(frame n+1) of the

morphing transformation, that in turn refer to

ΔC

MaxExp

and ΔC

MaxInsp

measured values.The interval

between ΔC

MaxExp

and ΔC

MaxInsp

has been divided in

n+1 intervals having the same amplitude, in order to

associate a set of ∆C values with the corresponding

frame of morphing sequence. Thus, it has been

possible to calculate ΔC

MaxInsp

with the following

equation:

u)1n(CCC

MaxExp1nMaxInsp

⋅++==

ΔΔΔ

(1)

and, similarly, a generic value of chest

circumference ∆C

corresponding to the frame m of

morphing sequence:

umCC

MaxExpm

⋅+

(2)

It has been possible to associate the number of

each frame of morphing sequence with ∆C values

measured during CT scans. The sequence was

between the minimal and the maximal tumour

extension, so there were some frames of morphing

BIOSIGNALS 2010 - International Conference on Bio-inspired Systems and Signal Processing

364

sequence describing lung tumour deformation during

free-breathing: by substituting the corresponding set

of ∆C values measured during CT exam in equation

(2), it has been possible to individuate the interval of

consecutive frames which described tumor motion

and deformation during tidal volume respiration.

Among these frames, there was the “best frame”,

which was the most similar to the tumour

configuration described by free-breathing CT slice

(CT

) related to the same tumour section as CT

slices used for morphing simulation. This frame has

been found through the calculation of mean grey

levels of the difference image between each

simulation frame and the slice CT

related to the

tumour configuration during free-breathing

acquisition: the frame whose difference image had

the smaller mean grey levels has been the most

similar to CT

. By Eq. (2), ∆C value corresponding

to this best frame was calculated and compared with

the interval of ∆C

values: in this way it was

possible to know in which phase of the breathing

period CT exam at tidal volume was carried out and

to define the target volume to be irradiated

corresponding to this phase.

3 RESULTS AND DISCUSSION

CT scans during free-breathing and in breath-hold

conditions have been conducted on a lung cancer

patient undergoing radiotherapic treatment. During

CT exams, patient’s chest circumference variations

have been measured with the PSG, obtaining

∆C

TVmin

=0 mm at the end of quiet expiration and

∆C

TVmax

=10 mm at the end of quiet inspiration on

average. ∆C

MaxInsp

=14 mm and ∆C

MaxExp

=-5 mm

during maximal inspiration and expiration,

respectively. We have considered ∆C

TVmin

as zero-

reference for ∆C measurements. Slices CT

MaxExp

and

MaxInsp

acquired during CT carried out in forced

expiration and in forced inspiration respectively,

were associated with the measured ∆C

MaxExp

and

∆C

MaxInsp

values and have been loaded on the

morphing program for creating the interpolation

sequence. 56 markers have been placed on the

border line of the tumour (Fig. 1), a frame number

equal to 100 and a linear transformation have been

chosen for morphing sequence generation. By

substituting the above reported measured ∆C

TVmin

and ∆C

TVmax

in equation (2), the numbers of the

simulation frames corresponding to tumor

configuration during tidal volume breathing have

been calculated. Thus, the frames between 25 and 78

of the morphing simulation represented the minimal

and the maximal tumor extension during free-

breathing (Fig. 2). In order to verify the accuracy of

the method, the differences between simulation

frames and the image CT

acquired during free-

breathing, after converting this slice from DICOM

format in 8 bit bitmap format with 256 grey levels

were calculated. In table 1 the numbers of simulation

frames, along with the corresponding mean grey

levels of the difference image are reported. From an

exam of table 1, it emerges that the frames more

similar to CT slice acquired during free breathing

are frames 25-60, because the corresponding

difference images have the smaller means of grey

levels. ∆C

values corresponding to these frames and

calculated by equation (2) were equal to 0-7 mm. By

comparing these values with the ∆C

values

measured in the CT exam carried out during quiet

respiration, it emerges that the slice CT

was

acquired approximately at the end of

expiration/beginning of inspiration, excluding the

inspiratory peak. Thus, in order to define the area to

be irradiated during each phase of tidal volume

respiration, the frames corresponding to the ∆C

interval between ∆C

TVmin

and ∆C

TVmax

have to be

considered. Target tumor can be delimited on these

frames, adding only a set up margin as safety

margin, since it allows for the uncertainties on

treatment reproducibility.

Figure 1: Markers placement on source image (forced

expiration) and on target image (forced inspiration).

Figure 2: Correspondence between simulation frames and

∆C values.

A MORPHING TECHNIQUE TO ESTIMATE LUNG CANCER DEFORMATION DUE TO BREATHING IN

RADIOTHERAPIC TREATMENT

365

Table 1: Results of comparison between simulation frames

and CT slice acquired during free breathing.

Frame number Mean of grey levels

0-5 17

10-20 16

25-60 15

65-70 16

75-80 17

85-90 18

95-101 19

4 CONCLUSIONS

The here proposed technique allows to estimate the

lung tumor morphology during patient’s free-

breathing by acquiring CT slice at forced expiration

and forced inspiration. The technique could give an

important contribution for the improvement of

radiation treatment planning, always considering the

set up margin, which allows for the uncertainties on

treatment reproducibility. In order to define the area

to be irradiated during a particular phase of the

breathing period, the cancer area in the simulation

frame corresponding to this phase has to be

considered as target: this allows the absence of

motion shadings. The present work represents only a

first stage study which could allow to deliver a high

dose to the tumour while minimizing the dose

delivered to the surrounding healthy tissue, though

further researches with more subjects are still needed

in order to test the accuracy of the presented

methodology.

REFERENCES

Webb, S. (2006) ‘Motion effects in (intensity modulated)

radiation therapy: a review’, Phys. Med. Biol., vol. 51,

pp. R403-R425.

ICRU Report 50 (1993) Available:

http://www.aitro.it/public/Crs100icru.pdf

[20 Mar 2009].

Lujian, A. E., Larsen, E. W., Balter, J. M., Haken, R. K.

T. (1999) ‘A method for incorporating organ motion

into 3D dose calculations’, Med. Phys., vol. 26, pp.

715-720.

Wu, H., Sharp, G. C., Saltzberg, B., Kaeli, D., Shirato, H.,

Jang, S. B. (2004) ‘A finite state model for respiratory

motion analysis in image guided radiation therapy’,

Phys. Med. Biol., vol. 49, pp. 5357-5372.

Report 91 del AAPM Task Group 76 (2006) ‘The

management of respiratory motion in radiation

oncology’, American Association of Physicists in

Medicine.

Murphy, M. J., Balter, J., Balter, S., BenComo, J. A. Jr,

Das, I. J., Jiang, S. B., Ma, C. M., Olivera, G. H.,

Rodebaugh, R. F., Ruchala, K. J., Shirato, H., Yin, F.

F. (2007) ‘The management of imaging dose during

image-guided radiotherapy: report of the AAPM Task

Group 75’, Med. Phys., vol. 34, pp. 4041-4063.

Mageras, G. S., Yorke, E. (2004) ‘Deep inspiration breath

hold and respiratory gating strategies for reducing

organ motion in radiation treatment’, Semin. Radiat.

Oncol., vol. 14, pp. 65-75.

Keall, P. J., Mageras, G. S., Balter, J. M., Emery, R. S.,

Forster, K. M., Jiang, S. B., Kapatoes, J. M., Low, D.

A., Murphy, M. J., Murray, B. R., Ramsey, C. R., Van

Herk, M. B., Vedam, S. S., Wong, J. W., Yorke, E.

(2006) ‘The management of respiratory motion in

radiation oncology report of AAPM Task Group 76’,

Med. Phys., vol. 33, pp. 3874-3900.

Masselli, G. M. P., Silvestri, S., Ramella, S., Trodella, L.

(2009) ‘Design and evaluation of a methodology to

perform personalized visual biofeedback for reducing

respiratory amplitude in radiation treatment’, Med.

Phys., vol. 35, no. 5, pp. 1467-1472.

George, R., Chung, T. D., Vedam, S. S., Ramakrishnan,

V., Mohan, R., Weiss, E., Keall, P. J. (2006) ‘Audio-

visual biofeedback for respiratory-gated radiotherapy:

impact of audio instruction and audio-visual

biofeedback on respiratory-gated radiotherapy’, Int. J.

Radiat. Oncol. Biol. Phys., vol. 65, pp. 924-933.

Briere, T., Jhaveri, P., Krishnan, S., Crane, C., Balter, P.,

Gillin, M., Mohan, R., Beddar, A. (2006) ‘SU-FF-T-

114: Breath Coaching with Visual Feedback for End-

Expiratory Gated Radiotherapy’, Med. Phys., vol. 33,

p. 2075.

Venkat, R. B., Sawant, A., Suh, Y., George, R., Keall, P.

J. (2008) ‘Development and preliminary evaluation of

a prototype audiovisual biofeedback device

incorporating a patient-specific guiding waveform’,

Phys. Med. Biol., vol. 53, pp. N197-N208.

Deurloo, K. E. I., Steenbakkers, R. H. M., Zijp L. J. et al.

(2005) ‘Quantification of shape variation of prostate

and seminal vesicles during external beam

radiotherapy’, Int. J. Radiation Oncology Biol. Phys.,

vol. 61, pp. 228-238.

Gomes, J., Darsa, L., Costa, B., Velho, L. (1999)

‘Warping and Morphing of Graphical Objects’,

Morgan Kaufmann Editor.

http://www.stoik.com/ [25 Mar 2009].

BIOSIGNALS 2010 - International Conference on Bio-inspired Systems and Signal Processing

366