Structure Creation Using Limited Point Topology-Evidence in Radiotherapy

Shape Variability Analysis and Image Segmentation in Radiotherapy

Introduction to SCULPTER

The SCULPTER project addresses an urgent problem: to cut unacceptably high intra and inter-observer variability of tumour and OAR delineation in radiotherapy, of the kind recently reported for the breast, prostate, lung, rectum and cervix.

These sites encompass high incidence cancers, which could otherwise benefit from improved effectiveness, comparison and adoption of new cancer treatment techniques. SCULPTER will create and use a radically new deformable clinical target volume paradigm supported by real-time 3D computation and visualisation. The proposal focuses on a delineation approach that combines a range of expert sources of information and presents these to the radiotherapy planning process in a dynamic and interactive way. Results will emerge sequentially to build a unique body of tumour and organ knowledge and an evolving delineation tool for parallel clinical investigations and potentially wide ranging exploitation.

With the above points in mind the SCULPTER proposal has the following specific aims.

  • To develop novel 3Dt structure creation consistent with the demands of international protocols.
  • To provide evidence based point definition with local and global control in 3D.
  • To develop self-mapping, expert-topology primitives for diverse anatomical sites using RBFs.
  • To capture and exploit expert range and confidence indices for evidence based points.
  • To develop topology transfer between volumes visible in dynamic data sets.
  • To analyse the evidence basis and suggest objective criteria for standardised clinical practice.
  • Reduce variability of delineation when using large scale static and dynamic imaging studies.

How does SCULPTER work

Sculpter works by interacting with 2D slices of the image dataset (CT/MR). The first and last slice on which the bio-structure (organ to delineate) is visible. The operator then defines a number of initial (evidence based) points that define the surface of the bio-structure. Normally three or four slices are picked and often no more than 8 points are needed on each slice. A mapping process is then used to "shrink wrap" a primitive object over the defined points.

Results are extremely good for the simple GTV/CTV/PTV structures in the pelvis e.g. prostate, because the organ's principal axes are appropriately aligned. Investigation of bladder and rectum suggests itself. SCULPTER will be extended to complex structures with very angular topology, such as the pelvis bones. The femoral heads are often irradiated in prostate treatments and so form potentially mobile but non-deformable OAR. The pelvis also forms the cage-like constraints essential for modelling the motion and deformation patterns of the prostate-bladder-rectum.

In figure 2, (below) the left hand side shows two views of a GTV defined by an expert observer (data kindly provided by William Beaumont Hospital, Detroit). The right hand side shows the same two views of a surface defined using SCULPTER with < 40 evidence based points. The intra observer variability shown by the crinkle cut nature of the surface defined by the expert user is not present in the surface defined using the SCULPTER tool.

In figures 3a (below) and 3b (below) the white boxes indicate the position of the evidence based points.

These have been defined on only five slices. The green contour shows the results of SCULPTER's mapping process. The blue contour shows the expert observers definition of the bio-structure. In figure 3a the fifth slice shows the expert user has decided the posterior border is concave. This is inconsistent with the adjacent slices as suggested by the contour calculated by the SCULPTER tool.

This is just one example of how definition of a minimal set of points for which the observer has a high degree of confidence can produce a more consistent bio-structure definition than if the observer is forced to make a decision for hundreds of points or describe the structure using a free-hand graphics tool.

Research Systems International (Kodak)

The SCULPTER tool is being written in IDL (Interactive Data Language, RSI). RSI have taken great interest in this work and are supporting the research of the developing technologies section for its development.


[1] Crowne E C, Wallace W H B, Moore C J, Mitchell R, Robertson W R, and Shalet S M,"Degree of activation of the pituitary-testicular axis in early pubertal boys with constitutional delay of growth and puberty (CDGP) detrmines the growth response to treatment with testosterone or oxandralone", Jnl Clin Endocrinol Metab, 80, No 6, 1869-1875, 1995.

[2] Crowne E C, Wallace W H B, Moore C J, Mitchell R, Robertson W H, Holly J and Shalet S M, "Effect of low dose oxandralone and testosterone treatment on the pituitary-testicular and GH axes in boys with constitutional delay of growth and puberty (CDGP)", Clinical Endocrinol, 46, 209-216, 1997.

[3] Moore C J, Slevin    N and Winstanley    S, 'Characterising Vowel Phonation By Fundamental Spectral Normalisation of Lx-Waveforms', Procs 1st Intl Workshop Models and Analysis of Vocal Emissions for Biomedical Appls, Florence Italy, 1-6, 1999.

[4] Moore C J and Jones S, 'Stimulated production of Vowel-like LX-Waveforms and Spectral Features in the Absence of Phonation', Jnl Med Eng Phys, accepted (expected Sept 2002)

[5] Moore C J, Doctoral Thesis 'mathematical Analysis and Picture Encoding Methods applied to Large Stores of Archived Digital Images', University of Manchester, 1988.

[6] Moore C J, "Transform Processing of CT Images; The Critical Role of Phase in Determining Image Structure and Limiting Data Compression", IPSM Report 62, Chpt 15, 100-106, 1990.

[7] Notley H M, Moore C J and Wilkinson J M, "CT scanning of Selectron insertions" Chapter 3, Sections 3.1 - 3.4, Report 45, Publishers Inst Phys Sci in Med, 1986.

[8] Lukka H R, Moore C J and Hunter R D, "The relationship between the bladder and the cervix in patients undergoing intracavitary therapy". Brit Jnl Rad, Vol 60, No 712, 355-359, Apr 1987.

[9] Hounsell A R, Sharrock P, Moore C J, Shaw A, Wilkinson J M and Williams P, "Computer assisted Generation of Multi-leaf Collimator Settings for Conformal Therapy". BJR, 65, 321-326, 1992.

[10] Moore C J, "Computerised Conformal Radiation Therapy - A Critical Process", IEE Computing and Control Engineering Jnl, Vol 6, No 5, 205-210,October 1995.

[11]Moore C J, Graham P A, mace R Sharrock P J, and I "Multi-modal Surface/Outline Projection and Simulation of Target/Critical tissue Movement", IEEE Computer Society, Porch Intl Con on Information Visualisation, 10-17, London, England, August 1997.

[12]Graham P A, Moore C J, mace R I and Sharrock P J, " Dynamic Surface Matching for Patient Positioning in Radiotherapy", IEEE Computer Society, Porch 17th Intl Con on Information Visualisation, pp 16-24, London, England, July 1998.

[13]Moore C J, Sharrock P J, Mackay R I and Graham P A, " 3D Closure of Incomplete Gross Tumour Volumes for ICRU-50 based Radiotherapy Treatment Planning",Procs EMBEC'99 Vienna, Med Biol Eng and Computing, Vol 37, Suppl 2, 652-653, 1999.

[14]Moore C J, Sharrock P J, MacKay R I and Graham P A, " Configurable, Space Variant 3D Transformation of Gross Tumour Volume into Planning Target Volume",Procs EMBEC'99 Vienna, Med Biol Eng and Computing, Vol 37, Suppl 2, 654-655, 1999.

[15]Graham P A, Moore C J, MacKay R I and Sharrock P J, "Dynamic Surface Matching Using Simulated Annealing for Patient Positioning in Radiotherapy", Procs EMBEC'99 Vienna, Med Biol Eng and Computing, Vol 37, Suppl 2, 38-39, 1999.

[16]Moore C J and Graham P A, '3D Dynamic Body Surface Sensing and CT-Body Matching A Tool for Patient Set-up and Monitoring in Radiotherapy', (revision submitted) Jnl Computer Aided Surgery, Special Edition on Planning and Image-Guidance in Radiation Therapy, 2000.

[17]MacKay R I, Graham P A, Sharrock P J and Moore C J, "Patient Positioning using detailed 3D Surface Data for Conformal Therapy of the Prostate: A feasibility Study", (in press) Intl Jnl Rad Onc Biol, Phys, 1999.

[18]MacKay R I, Graham P A, Moore C J, Logue J P and Sharrock P J, "Concurrent Animation and Radiobiological Analysis of 3D Motion in Conformal Radiotherapy", (in press) Radiotherapy and Oncol, 2000.

[19]Graham P A, Thompson B, Stratford J, MacKay R I, Moore C J and Sharrock P J, "Multi-Modal Verification of Patient Positioning in Radiotherapy", IEEE Computer Society, Proc 18thIntl Conf on Information Visualisation, London, England, pp 97-102, July 1999.

[20] Moore C J, "Transform Processing of CT Images; The Critical Role of Phase in Determining Image Structure and Limiting Data Compression", IPSM Report 62, Chpt 15, 100-106, 1990.

[21] Zindy E, PhD Thesis, 'Definition of Gross Tumour Volumes using Minimum Datasets', Liverpool John Moores University, 2000.

[22] M. Nirmalan, T. Willard, A. Khan and P. Nightingale,' Effect of assuming a fixed (CaO2-CvO2) on shunt calculations', Br J Anaesth 1997; 79: 135P.

[23] M. Nirmalan, T. Willard, A. Khan and P. Nightingale,'Evaluation of two non-invasive indices as alternative measures of oxygen in critically ill patients', Intensive Care Medicine 1997; 23 (Supplement 1): S26.

[24] M. Nirmalan, T. Willard, A. Khan and P. Nightingale,'Changes in arterial-mixed venous oxygen content difference (CaO2-CvO2) and the effect on shunt calculations in critically ill patients', Br J Anaesth 1998; 80: 829-831.

[25] T. Willard , A. Bryan, B. J. Pollard,'A computer-controlled varying stimulus train pattern stimulator with digital recorder and display system for use in anaesthetised humans. ', Journal of Physiology 1995: 483P. p4P-5P

[26] M. Nirmalan, T. Willard, C. Pomfrett, P. Nightingale and R. Little,'Loss of heart rate variability in critically ill patients', Br J Anaesth 1998; 81: 283P

[27] T. Willard, 'Entropy as a quantification of complexity in data', Presented at the joint meeting of the Association of North Western Intensive Care units and the North Western Injury Research Centre (MRC Trauma Group), University of Manchester, February 1999, at the Royal Bolton Hospital Postgraduate Centre.

[28] M. Nirmalan, T. Willard, P.Sinnot, J. E. Eddleston and R. A. Little,'Heart rate variability in patients with severe brain injury', Br J Anaesth 2001; 86: 312P

[29] M. Nirmalan, T. Willard and P. Nightingale,'Computer simulation of pulmonary oxygen transfer in Acute Respiratory Distress Syndrome (ARDS)', Br J Anaesth 1999; 83: 170P

[30] M. Nirmalan, T. Willard, M. O. Columb and P. Nightingale,'Effect of arterial-mixed venous oxygen content difference (C[a-v]O2) on indices of pulmonary oxygen transfer in an ARDS lung model', Br J Anaesth 2000; 84: 273P

[31] M. Nirmalan, T. Willard M O Columb and P. Nightingale,' Effect of changes in arterial-mixed venous oxygen content difference (C[a-v]O2) on indices of pulmonary oxygen transfer in a model ARDS lung', Br J Anaesth 2001; 86: 477-85

[32] Thind, G.S., Rubin, A.P. (2001). Editorial I: KISS and indices of pulmonary oxygen transfer. Br J Anaesth 86: 465-466

[33] J R Sykes "Removal of Diffusion in MR Dosimetry Gels", 1994, MSC thesis, University of Manchester, Faculty of Medicine

[34] A P Jones, D G Hughes, L Robinson, J R Sykes, D S Brettle, Q Aziz, S Hamdy, D G Thompson, S W G Derbyshire, A C N Chen and A K P Jones "Experiences with functional magnetic resonance" 1998, The British Journal of Radiology 71, 160-166

[35] Andrew P Jones, Johnathan R Sykes, Bernadette M Carrington, Julie A Loncaster, Susan M Todd, David L Buckley, Catherine M L, "Prediction of Clinical Outcome for Cancer of the Cervix using Dynamic Contrast Enhanced MR Imaging", ISMRM (International Society for Magnetic Resonance in Medicine), Glasgow 2001

[36] Susan M Todd, Andrew P Jones, Rachel A Cooper, Julie A Loncaster, Catherine M L West, Johnathan R Sykes, Bernadette M Carrington, "Dynamic contrast-enhanced MR Imaging and Tumour Oxygenation Measurements in Cervical Cancer", ISMRM (International Society for Magnetic Resonance in Medicine), Glasgow 2001

[37] Juliette A. Loncaster, Bernadette M. Carrington, Jonathan R. Sykes, Andrew P. Jones, Susan M. Todd, Rachel Cooper, David L. Buckley, Susan E. Davidson, John P. Logue, Robin D. Hunter, and Catharine M.L. West, "Prediction Of Radiotherapy Outcome Using Dynamic Contrast Enhanced Magnetic Resonance Imaging Of Carcinoma Of The Cervix",Int.J.Radiation.Phys.Chem.Biol, Accepted for publication June 2002.

[38] J R Sykes and P C Williams, "An Experimental Investigation of the Tongue and Groove Effect for the Philips Multileaf Collimator", 1998, Phys.Med.Biol 43 3157-3165

[39] G J Budgell, J R Sykes and J M Wilkinson, "Rectangular Edge Synchronisation for Intensity Modulated Radiation Therapy with Dynamic Multileaf Collimation", 1998, Phys.Med.Biol 43 2769-2784

[40] J R Sykes, G J Budgell and J M Wilkinson "Leaf Collision avoidance in dynamic intensity modulation with a Philips multileaf collimator and the consequences for dose outside the field", proceedings of the World Congress in Medical Physics and Biomedical Engineering, Nice, France 1997

[41] J R Sykes, H V James and P C Williams, "How much does film sensitivity increase at depth for larger field sizes?" 1999, Med.Phys. 26 329-330

[42] H Hricak and B Carrington, 'MRI of the Pelvis: A Text Atlas', Martin Dunitz, 1991

[43] International Commission on Radiation Units and Measurements, ICRU Report 50, 'Prescribing, Recording and Reporting Photon Beam Therapy', September 1993.

[44] International Commission on Radiation Units and Measurements, ICRU Report 62, 'Prescribing, Recording and Reporting Photon Beam Therapy (Supplement to ICRU Report 50)', 1999.

[45] Mazonakis M, Damilakis J, Varveris H and Prassopoulos P, 'Image segmentation in treatment planning for prostate cancer using the region growing technique', Brit Jnl Radiol, 74 (879), 243-8, 2001.

[46] Pitkanen M A, Holli K A, Ojala A T and Laippala P, 'Quality assurance in radiotherapy of breast cancer: variability in planning target volume delineation', Acata Oncol, 40(1), 50-55, 2001.

[47] Fiorino C, Reni M, Bolognesi A, Cattaneo G M and Calandrino R,'Intra- and Inter-observer variability in contouring prostate and seminal vesicles: implications for conformal treatment planning', Radiother Oncol, 47(3), 285-292, 1998.

[48] Van de Steene J, Linthout N, de Mey J, Vinh-Hung V, classens C, Noppen M, Bel A and Storme G, 'Definition of gross tumour volume in lung cancer: inter-observer variability', Radiother Oncol, 62(1), 37-49, 2002.

[49] Saarnak A E, Boersma M, van Brunningen B N, Wolterlink R and Steggerd M J, ' Interobserver variation in delineation of bladder and rectum contours for brachytherapy of cervical cancer', Radiother Oncol, 56(1), 37-42, 2000.

[50] Senan S, van Sornsen de Koste J, Samson M, Tankink H, Jansen P, Nowak P J, Krol A D, Schmitz P and Lagerward F J, 'Evaluation of a target contouring protocol for 3D conformal radiotherapy in non-small ling cell cancer', Radiother Oncol, 53(3), 247-255, 1999. 

[51] Bardinett E, Cohen L D and Ayache N, 'Analyzing the deformation of the left ventricle of the heart with a parametric deformable model', Technical Report RR-2797, Institut National de Recherche en Informatique et en Automatique (INRIA), 1996.

[52] Carr J C, Fright W R and Beatson W K, 'Surface Interpolation with Radial Basis Functions for Medical Imaging', IEEE Trans Medical Imaging, 16 (1), 96-107, 1997.

[53] EC Biomedicine and Health Research Programme, Project SOLUS-3D, Coordinator J Clements Addenbrookes Hospital, Cambridge UK, EC contract BMH4-CT-95-0476, 1996-1999. 

[54] Fornefett M, Rohr K and Stiehl S, 'Elastic Registration of Medical Images using Radial Basis Functions with Compact Support', Proc Computer Vision and Pattern Recognition, IEEE Computer Society PR00149, 402-407, 1999.

[55] Zindy E and Moore C J, 'Rapid 3D Deformation of Tumour Volumes and Organs at Risk from Minimum Datasets using Deformable Modelling', Revue Internationale de CFAO et d'Infographie, Vol 16, 243-265, 2001