Adaptive radiotherapy for virally mediated head and neck cancer

Introduction: Clinical investigation has revealed a subgroup of head and neck cancers that are virally mediated. The relationship between nasopharyngeal cancer and Epstein Barr Virus (EBV) has long been established and more recently, the association between oropharyngeal cancer and Human Papillomavirus (HPV) has been revealed1,2 These cancers often present with nodal involvement and generally respond well to radiation treatment, evidenced by tumour regression1. This results in the need for treatment plan adaptation or re-planning in a subset of patients. Adaptive techniques allow the target region of the radiotherapy treatment plan to be altered in accordance with treatment-induced changes to ensure that under or over dosing does not occur3. It also assists in limiting potential overdosing of surrounding critical normal tissues4. We sought to identify a high-risk group based on nodal size to be evaluated in a future prospective adaptive radiotherapy trial. Method: Between 2005-2010, 121 patients with virally mediated, node positive nasopharyngeal (EBV positive) or oropharyngeal (HPV positive) cancers, receiving curative intent radiotherapy treatment were reviewed. Patients were analysed based on maximum size of the dominant node at diagnosis with a view to grouping them in varying risk categories to determine the need of re-planning. The frequency and timing of the re-planning scans were also evaluated. Results: Sixteen nasopharyngeal and 105 oropharyngeal tumours were reviewed. Twenty-five (21%) patients underwent a re-planning CT at a median of 22 (range, 0-29) fractions with 1 patient requiring re-planning prior to the commencement of treatment. Based on the analysis, patients were subsequently placed into risk categories; ≤35mm (Group 1), 36-45mm (Group 2), ≥46mm (Group 3). Re-planning CT’s were performed in Group 1- 8/68 (11.8%), Group 2- 4/28 (14.3%), Group 3- 13/25 (52%). Conclusion: In this series, patients with virally mediated head and neck cancer and nodal size > 46mm appear to be a high-risk group for the need of re-planning during a course of curative radiotherapy. This finding will now be tested in a prospective adaptive radiotherapy study. ‘Real World’ Implications: This research identifies predictive factors for those patients with virally mediated head and neck cancer that will benefit most from treatment adaptation. This will assist in minimising the side effects experienced by these patients thereby improving their quality of life after treatment.

Growth of confined cancer spheroids : a combined experimental marker, critical modelling approach

A critical step in the dissemination of ovarian cancer is the formation of multicellular spheroids from cells shed from the primary tumour. The objectives of this study were to apply bioengineered three-dimensional (3D) microenvironments for culturing ovarian cancer spheroids in vitro and simultaneously to build on a mathematical model describing the growth of multicellular spheroids in these biomimetic matrices. Cancer cells derived from human epithelial ovarian carcinoma were embedded within biomimetic hydrogels of varying stiffness and grown for up to 4 weeks. Immunohistochemistry, imaging and growth analyses were used to quantify the dependence of cell proliferation and apoptosis on matrix stiffness, long-term culture and treatment with the anti-cancer drug paclitaxel. The mathematical model was formulated as a free boundary problem in which each spheroid was treated as an incompressible porous medium. The functional forms used to describe the rates of cell proliferation and apoptosis were motivated by the experimental work and predictions of the mathematical model compared with the experimental output. This work aimed to establish whether it is possible to simulate solid tumour growth on the basis of data on spheroid size, cell proliferation and cell death within these spheroids. The mathematical model predictions were in agreement with the experimental data set and simulated how the growth of cancer spheroids was influenced by mechanical and biochemical stimuli including matrix stiffness, culture duration and administration of a chemotherapeutic drug. Our computational model provides new perspectives on experimental results and has informed the design of new 3D studies of chemoresistance of multicellular cancer spheroids.

Genetic association of the KLK4 locus with risk of prostate cancer

The Kallikrein-related peptidase, KLK4, has been shown to be significantly overexpressed in prostate tumours in numerous studies and is suggested to be a potential biomarker for prostate cancer. KLK4 may also play a role in prostate cancer progression through its involvement in epithelial-mesenchymal transition, a more aggressive phenotype, and metastases to bone. It is well known that genetic variation has the potential to affect gene expression and/or various protein characteristics and hence we sought to investigate the possible role of single nucleotide polymorphisms (SNPs) in the KLK4 gene in prostate cancer. Assessment of 61 SNPs in the KLK4 locus (±10 kb) in approximately 1300 prostate cancer cases and 1300 male controls for associations with prostate cancer risk and/or prostate tumour aggressiveness (Gleason score <7 versus ≥7) revealed 7 SNPs to be associated with a decreased risk of prostate cancer at the Ptrend<0.05 significance level. Three of these SNPs, rs268923, rs56112930 and the HapMap tagSNP rs7248321, are located several kb upstream of KLK4; rs1654551 encodes a non-synonymous serine to alanine substitution at position 22 of the long isoform of the KLK4 protein, and the remaining 3 risk-associated SNPs, rs1701927, rs1090649 and rs806019, are located downstream of KLK4 and are in high linkage disequilibrium with each other (r2≥0.98). Our findings provide suggestive evidence of a role for genetic variation in the KLK4 locus in prostate cancer predisposition.

Combined expression of KLK4, KLK5, KLK6, and KLK7 by ovarian cancer cells leads to decreased adhesion and paclitaxel-induced chemoresistance.

OBJECTIVE: Chemoresistance is a critical feature of advanced ovarian cancer with only 30% of patients surviving longer than 5 years. We have previously shown that four kallikrein-related (KLK) peptidases, KLK4, KLK5, KLK6 and KLK7 (KLK4-7), are implicated in peritoneal invasion and tumour growth, but underlying mechanisms were not identified. We also reported that KLK7 overexpression confers chemoresistance to paclitaxel, and cell survival via integrins. In this study, we further explored the functional consequenses of overexpression of all four KLKs (KLK4-7) simultaneously in the ovarian cancer cell line, OV-MZ-6, and its impact on integrin expression and signalling, cell adhesion and survival as contributors to chemoresistance and metastatic progression. METHODS: Quantitative gene and protein expression analyses, confocal microscopy, cell adhesion and chemosensitivity assays were performed. RESULTS: Expression of α5β1/αvβ3 integrins was downregulated upon combined stable KLK4-7 overexpression in OV-MZ-6 cells. Accordingly, the adhesion of these cells to vitronectin and fibronectin, the extracellular matrix binding proteins of α5β1/αvβ3 integrins and two predominant proteins of the peritoneal matrix, was decreased. KLK4-7-transfected cells were more resistant to paclitaxel (10-100 nmol/L: 38-54%), but not to carboplatin, which was associated with decreased apoptotic stimuli. However, the KLK4-7-induced paclitaxel resistance was not blocked by the MEK1/2 inhibitor, U0126. CONCLUSIONS: This study demonstrates that combined KLK4-7 expression by ovarian cancer cells promotes reduced integrin expression with consequently less cell-matrix attachment, and insensitivity to paclitaxel mediated by complex integrin and MAPK independent interactions, indicative of a malignant phenotype and disease progression suggesting a role for these KLKs in this process.

A snapshot of radiation therapy techniques and technology in Queensland : an aid to mapping undergraduate curriculum

Introduction: Undergraduate students studying the Bachelor of Radiation Therapy at Queensland University of Technology (QUT) attend clinical placements in a number of department sites across Queensland. To ensure that the curriculum prepares students for the most common treatments and current techniques in use in these departments, a curriculum matching exercise was performed. Methods: A cross-sectional census was performed on a pre-determined “Snapshot” date in 2012. This was undertaken by the clinical education staff in each department who used a standardized proforma to count the number of patients as well as prescription, equipment, and technique data for a list of tumour site categories. This information was combined into aggregate anonymized data. Results: All 12 Queensland radiation therapy clinical sites participated in the Snapshot data collection exercise to produce a comprehensive overview of clinical practice on the chosen day. A total of 59 different tumour sites were treated on the chosen day and as expected the most common treatment sites were prostate and breast, comprising 46% of patients treated. Data analysis also indicated that intensity-modulated radiotherapy (IMRT) use is relatively high with 19.6% of patients receiving IMRT treatment on the chosen day. Both IMRT and image-guided radiotherapy (IGRT) indications matched recommendations from the evidence. Conclusion: The Snapshot method proved to be a feasible and efficient method of gathering useful

Breast cancer detection from thermal images using bispectral invariant features

Highly sensitive infrared (IR) cameras provide high-resolution diagnostic images of the temperature and vascular changes of breasts. These images can be processed to emphasize hot spots that exhibit early and subtle changes owing to pathology. The resulting images show clusters that appear random in shape and spatial distribution but carry class dependent information in shape and texture. Automated pattern recognition techniques are challenged because of changes in location, size and orientation of these clusters. Higher order spectral invariant features provide robustness to such transformations and are suited for texture and shape dependent information extraction from noisy images. In this work, the effectiveness of bispectral invariant features in diagnostic classification of breast thermal images into malignant, benign and normal classes is evaluated and a phase-only variant of these features is proposed. High resolution IR images of breasts, captured with measuring accuracy of ±0.4% (full scale) and temperature resolution of 0.1 °C black body, depicting malignant, benign and normal pathologies are used in this study. Breast images are registered using their lower boundaries, automatically extracted using landmark points whose locations are learned during training. Boundaries are extracted using Canny edge detection and elimination of inner edges. Breast images are then segmented using fuzzy c-means clustering and the hottest regions are selected for feature extraction. Bispectral invariant features are extracted from Radon projections of these images. An Adaboost classifier is used to select and fuse the best features during training and then classify unseen test images into malignant, benign and normal classes. A data set comprising 9 malignant, 12 benign and 11 normal cases is used for evaluation of performance. Malignant cases are detected with 95% accuracy. A variant of the features using the normalized bispectrum, which discards all magnitude information, is shown to perform better for classification between benign and normal cases, with 83% accuracy compared to 66% for the original.

Body temperature and its effect on leukocyte mobilization, cytokines and markers of neutrophil activation during and after exercise

We investigated the influence of rectal temperature on the immune system during and after exercise. Ten well-trained male cyclists completed exercise trials (90 min cycling at 60% VO(2max) + 16.1 - km time trial) on three separate occasions: once in 18 degrees C and twice in 32 degrees C. Twenty minutes after the trials in 32 degrees C, the cyclists sat for approximately 20 min in cold water (14 degrees C) on one occasion, whereas on another occasion they sat at room temperature. Rectal temperature increased significantly during cycling in both conditions, and was significantly higher after cycling in 32 degrees C than in 18 degrees C (P < 0.05). Leukocyte counts increased significantly during cycling but did not differ between the conditions. The concentrations of serum interleukin (IL)-6, IL-8 and IL-10, plasma catecholamines, granulocyte-colony stimulating factor, myeloperoxidase and calprotectin increased significantly following cycling in both conditions. The concentrations of serum IL-8 (25%), IL-10 (120%), IL-1 receptor antagonist (70%), tumour necrosis factor-alpha (17%), plasma myeloperoxidase (26%) and norepinephrine (130%) were significantly higher after cycling in 32 degrees C than in 18 degrees C. During recovery from exercise in 32 degrees C, rectal temperature was significantly lower in response to sitting in cold water than at room temperature. However, immune changes during 90 min of recovery did not differ significantly between sitting in cold water and at room temperature. The greater rise in rectal temperature during exercise in 32 degrees C increased the concentrations of serum IL-8, IL-10, IL-1ra, TNF-alpha and plasma myeloperoxidase, whereas the greater decline in rectal temperature during cold water immersion after exercise did not affect immune responses.

Changes in markers of muscle damage, inflammation and HSP70 after an Ironman triathlon race

We investigated the effects of an Ironman triathlon race on markers of muscle damage, inflammation and heat shock protein 70 (HSP70). Nine well-trained male triathletes (mean +/- SD age 34 +/- 5 years; VO(2peak) 66.4 ml kg(-1) min(-1)) participated in the 2004 Western Australia Ironman triathlon race (3.8 km swim, 180 km cycle, 42.2 km run). We assessed jump height, muscle strength and soreness, and collected venous blood samples 2 days before the race, within 30 min and 14-20 h after the race. Plasma samples were analysed for muscle proteins, acute phase proteins, cytokines, heat shock protein 70 (HSP70), and clinical biochemical variables related to dehydration, haemolysis, liver and renal functions. Muscular strength and jump height decreased significantly (P < 0.05) after the race, whereas muscle soreness and the plasma concentrations of muscle proteins increased. The cytokines interleukin (IL)-1 receptor antagonist, IL-6 and IL-10, and HSP70 increased markedly after the race, while IL-12p40 and granulocyte colony-stimulating factor (G-CSF) were also elevated. IL-4, IL-1beta and tumour necrosis factor-alpha did not change significantly, despite elevated C-reactive protein and serum amyloid protein A on the day after the race. Plasma creatinine, uric acid and total bilirubin concentrations and gamma-glutamyl transferase activity also changed after the race. In conclusion, despite evidence of muscle damage and an acute phase response after the race, the pro-inflammatory cytokine response was minimal and anti-inflammatory cytokines were induced. HSP70 is released into the circulation as a function of exercise duration.

New players for advanced prostate cancer and the rationalisation of insulin-sensitising medication

Obesity and type 2 diabetes are recognised risk factors for the development of some cancers and, increasingly, predict more aggressive disease, treatment failure, and cancer-specific mortality. Many factors may contribute to this clinical observation. Hyperinsulinaemia, dyslipidaemia, hypoxia, ER stress, and inflammation associated with expanded adipose tissue are thought to be among the main culprits driving malignant growth and cancer advancement. This observation has led to the proposal of the potential utility of “old players” for the treatment of type 2 diabetes and metabolic syndrome as new cancer adjuvant therapeutics. Androgen-regulated pathways drive proliferation, differentiation, and survival of benign and malignant prostate tissue. Androgen deprivation therapy (ADT) exploits this dependence to systemically treat advanced prostate cancer resulting in anticancer response and improvement of cancer symptoms. However, the initial therapeutic response from ADT eventually progresses to castrate resistant prostate cancer (CRPC) which is currently incurable. ADT rapidly induces hyperinsulinaemia which is associated with more rapid treatment failure. We discuss current observations of cancer in the context of obesity, diabetes, and insulin-lowering medication. We provide an update on current treatments for advanced prostate cancer and discuss whether metabolic dysfunction, developed during ADT, provides a unique therapeutic window for rapid translation of insulin-sensitising medication as combination therapy with antiandrogen targeting agents for the management of advanced prostate cancer.

Predicting fitness for chemotherapy in older cancer patients

Background: The Vulnerable Elders Survey-13 (VES-13) is increasingly used to screen for older patients who can proceed to intensive chemotherapy without further comprehensive assessment. This study compared the VES-13 determination of fitness for treatment with the oncologist's assessments of fitness. Method: Sample: Consecutive series of solid tumour patients ≥65 years (n=175; M=72; range=65-86) from an Australian cancer centre. Patients were screened with the VES-13 before proceeding to usual treatment. Blinded to screening, oncologists concurrently predicted patient fitness for chemotherapy. A sample of 175 can detect, with 90% power, kappa coefficients of agreement between VES-13 and oncologists’ assessments >0.90 ("almost perfect agreement"). Separate backward stepwise logistic regression analyses assessed potential predictors of VES-13 and oncologists’ ratings of fitness. Results: Kappa coefficient for agreement between VES-13 and oncologists’ ratings of fitness was 0.41 (p<0.001). VES-13 and oncologists’ assessments agreed in 71% of ratings. VES-13 sensitivity = 83.3%; specificity = 57%; positive predictive value = 69%; negative predictive value = 75%. Logistic regression modelling indicated that the odds of being vulnerable to chemotherapy (VES-13) increased with increasing depression (OR=1.42; 95% CI: 1.18, 1.71) and decreased with increased functional independence assessed on the Bartel Index (OR=0.82; CI: 0.74, 0.92) and Lawton instrumental activities of daily living (OR=0.44; CI: 0.30, 0.65); RSquare=.65. Similarly, the odds of a patient being vulnerable to chemotherapy, when assessed by physicians, increased with increasing age (OR=1.15; CI: 1.07, 1.23) and depression (OR=1.23; CI: 1.06, 1.43), and decreased with increasing functional independence (OR=0.91; CI: 0.85, 0.98); RSquare=.32. Conclusions: Our data indicate moderate agreement between VES-13 and clinician assessments of patients’ fitness for chemotherapy. Current ‘one-step’ screening processes to determine fitness have limits. Nonetheless, screening tools do have the potential for modification and enhanced predictive properties in cancer care by adding relevant items, thus enabling fit patients to be immediately referred for chemotherapy.

Humanised xenograft models of bone metastasis revisited : novel insights into species-specific mechanisms of cancer cell osteotropism

The determinants and key mechanisms of cancer cell osteotropism have not been identified, mainly due to the lack of reproducible animal models representing the biological, genetic and clinical features seen in humans. An ideal model should be capable of recapitulating as many steps of the metastatic cascade as possible, thus facilitating the development of prognostic markers and novel therapeutic strategies. Most animal models of bone metastasis still have to be derived experimentally as most syngeneic and transgeneic approaches do not provide a robust skeletal phenotype and do not recapitulate the biological processes seen in humans. The xenotransplantation of human cancer cells or tumour tissue into immunocompromised murine hosts provides the possibility to simulate early and late stages of the human disease. Human bone or tissue-engineered human bone constructs can be implanted into the animal to recapitulate more subtle, species-specific aspects of the mutual interaction between human cancer cells and the human bone microenvironment. Moreover, the replication of the entire "organ" bone makes it possible to analyse the interaction between cancer cells and the haematopoietic niche and to confer at least a partial human immunity to the murine host. This process of humanisation is facilitated by novel immunocompromised mouse strains that allow a high engraftment rate of human cells or tissue. These humanised xenograft models provide an important research tool to study human biological processes of bone metastasis.

The development of Monte Carlo techniques for the verification of radiotherapy treatments

Introduction: Recent advances in the planning and delivery of radiotherapy treatments have resulted in improvements in the accuracy and precision with which therapeutic radiation can be administered. As the complexity of the treatments increases it becomes more difficult to predict the dose distribution in the patient accurately. Monte Carlo (MC) methods have the potential to improve the accuracy of the dose calculations and are increasingly being recognised as the ‘gold standard’ for predicting dose deposition in the patient [1]. This project has three main aims: 1. To develop tools that enable the transfer of treatment plan information from the treatment planning system (TPS) to a MC dose calculation engine. 2. To develop tools for comparing the 3D dose distributions calculated by the TPS and the MC dose engine. 3. To investigate the radiobiological significance of any errors between the TPS patient dose distribution and the MC dose distribution in terms of Tumour Control Probability (TCP) and Normal Tissue Complication Probabilities (NTCP). The work presented here addresses the first two aims. Methods: (1a) Plan Importing: A database of commissioned accelerator models (Elekta Precise and Varian 2100CD) has been developed for treatment simulations in the MC system (EGSnrc/BEAMnrc). Beam descriptions can be exported from the TPS using the widespread DICOM framework, and the resultant files are parsed with the assistance of a software library (PixelMed Java DICOM Toolkit). The information in these files (such as the monitor units, the jaw positions and gantry orientation) is used to construct a plan-specific accelerator model which allows an accurate simulation of the patient treatment field. (1b) Dose Simulation: The calculation of a dose distribution requires patient CT images which are prepared for the MC simulation using a tool (CTCREATE) packaged with the system. Beam simulation results are converted to absolute dose per- MU using calibration factors recorded during the commissioning process and treatment simulation. These distributions are combined according to the MU meter settings stored in the exported plan to produce an accurate description of the prescribed dose to the patient. (2) Dose Comparison: TPS dose calculations can be obtained using either a DICOM export or by direct retrieval of binary dose files from the file system. Dose difference, gamma evaluation and normalised dose difference algorithms [2] were employed for the comparison of the TPS dose distribution and the MC dose distribution. These implementations are spatial resolution independent and able to interpolate for comparisons. Results and Discussion: The tools successfully produced Monte Carlo input files for a variety of plans exported from the Eclipse (Varian Medical Systems) and Pinnacle (Philips Medical Systems) planning systems: ranging in complexity from a single uniform square field to a five-field step and shoot IMRT treatment. The simulation of collimated beams has been verified geometrically, and validation of dose distributions in a simple body phantom (QUASAR) will follow. The developed dose comparison algorithms have also been tested with controlled dose distribution changes. Conclusion: The capability of the developed code to independently process treatment plans has been demonstrated. A number of limitations exist: only static fields are currently supported (dynamic wedges and dynamic IMRT will require further development), and the process has not been tested for planning systems other than Eclipse and Pinnacle. The tools will be used to independently assess the accuracy of the current treatment planning system dose calculation algorithms for complex treatment deliveries such as IMRT in treatment sites where patient inhomogeneities are expected to be significant. Acknowledgements: Computational resources and services used in this work were provided by the HPC and Research Support Group, Queensland University of Technology, Brisbane, Australia. Pinnacle dose parsing made possible with the help of Paul Reich, North Coast Cancer Institute, North Coast, New South Wales.

Pixel value to electron density conversion for a Megavoltage Cone Beam CT System

Introduction: The motivation for developing megavoltage (and kilovoltage) cone beam CT (MV CBCT) capabilities in the radiotherapy treatment room was primarily based on the need to improve patient set-up accuracy. There has recently been an interest in using the cone beam CT data for treatment planning. Accurate treatment planning, however, requires knowledge of the electron density of the tissues receiving radiation in order to calculate dose distributions. This is obtained from CT, utilising a conversion between CT number and electron density of various tissues. The use of MV CBCT has particular advantages compared to treatment planning with kilovoltage CT in the presence of high atomic number materials and requires the conversion of pixel values from the image sets to electron density. Therefore, a study was undertaken to characterise the pixel value to electron density relationship for the Siemens MV CBCT system, MVision, and determine the effect, if any, of differing the number of monitor units used for acquisition. If a significant difference with number of monitor units was seen then pixel value to ED conversions may be required for each of the clinical settings. The calibration of the MV CT images for electron density offers the possibility for a daily recalculation of the dose distribution and the introduction of new adaptive radiotherapy treatment strategies. Methods: A Gammex Electron Density CT Phantom was imaged with the MVCB CT system. The pixel value for each of the sixteen inserts, which ranged from 0.292 to 1.707 relative electron density to the background solid water, was determined by taking the mean value from within a region of interest centred on the insert, over 5 slices within the centre of the phantom. These results were averaged and plotted against the relative electron densities of each insert with a linear least squares fit was preformed. This procedure was performed for images acquired with 5, 8, 15 and 60 monitor units. Results: The linear relationship between MVCT pixel value and ED was demonstrated for all monitor unit settings and over a range of electron densities. The number of monitor units utilised was found to have no significant impact on this relationship. Discussion: It was found that the number of MU utilised does not significantly alter the pixel value obtained for different ED materials. However, to ensure the most accurate and reproducible MV to ED calibration, one MU setting should be chosen and used routinely. To ensure accuracy for the clinical situation this MU setting should correspond to that which is used clinically. If more than one MU setting is used clinically then an average of the CT values acquired with different numbers of MU could be utilized without loss in accuracy. Conclusions: No significant differences have been shown between the pixel value to ED conversion for the Siemens MV CT cone beam unit with change in monitor units. Thus as single conversion curve could be utilised for MV CT treatment planning. To fully utilise MV CT imaging for radiotherapy treatment planning further work will be undertaken to ensure all corrections have been made and dose calculations verified. These dose calculations may be either for treatment planning purposes or for reconstructing the delivered dose distribution from transit dosimetry measurements made using electronic portal imaging devices. This will potentially allow the cumulative dose distribution to be determined through the patient’s multi-fraction treatment and adaptive treatment strategies developed to optimize the tumour response.

Detecting distributed Denial-of-Service attacks and Flash Events

This thesis investigates and develops techniques for accurately detecting Internet-based Distributed Denial-of-Service (DDoS) Attacks where an adversary harnesses the power of thousands of compromised machines to disrupt the normal operations of a Web-service provider, resulting in significant down-time and financial losses. This thesis also develops methods to differentiate these attacks from similar-looking benign surges in web-traffic known as Flash Events (FEs). This thesis also addresses an intrinsic challenge in research associated with DDoS attacks, namely, the extreme scarcity of public domain datasets (due to legal and privacy issues) by developing techniques to realistically emulate DDoS attack and FE traffic.

Segmentation of cone-beam CT using a hidden Markov random field with informative priors

Purpose: Flat-detector, cone-beam computed tomography (CBCT) has enormous potential to improve the accuracy of treatment delivery in image-guided radiotherapy (IGRT). To assist radiotherapists in interpreting these images, we use a Bayesian statistical model to label each voxel according to its tissue type. Methods: The rich sources of prior information in IGRT are incorporated into a hidden Markov random field (MRF) model of the 3D image lattice. Tissue densities in the reference CT scan are estimated using inverse regression and then rescaled to approximate the corresponding CBCT intensity values. The treatment planning contours are combined with published studies of physiological variability to produce a spatial prior distribution for changes in the size, shape and position of the tumour volume and organs at risk (OAR). The voxel labels are estimated using the iterated conditional modes (ICM) algorithm. Results: The accuracy of the method has been evaluated using 27 CBCT scans of an electron density phantom (CIRS, Inc. model 062). The mean voxel-wise misclassification rate was 6.2%, with Dice similarity coefficient of 0.73 for liver, muscle, breast and adipose tissue. Conclusions: By incorporating prior information, we are able to successfully segment CBCT images. This could be a viable approach for automated, online image analysis in radiotherapy.