research report

University of Hull 

2007/08 


Use of high-field strength MR derived biomarkers in prostate cancer: potential for improved cancer staging and prognostication  
Dr P Gibbs, Dr MD Pickles, Dr GP Liney, Dr B Zelhof, Prof. LW Turnbull  

This study aims to assess the staging and prognostic potential of MR based biomarkers in prostate cancer. Commencing in August 2006, approximately 70 men, potentially scheduled for radical prostatectomy, with biopsy proven cancer have been recruited to date. MR imaging is being performed at 3.0 Tesla using an endorectal coil in addition to a multi channel pelvic phased array coil to improve image quality. Dynamic contrast enhanced images of Gadolinium based contrast agent uptake are acquired to enable subsequent pharmacokinetic modelling. Diffusion weighted images and T2 weighted images are obtained from which maps of the apparent diffusion coefficient (ADC) and T2 of water are calculated respectively. To date, 20 patients have proceeded to radical prostatectomy. Whole mounted slices are stained with hematoxylin and eosin (H & E) for determination of tumour grade, location and volume. Comparison of ADC and T2 maps with H & E staining have revealed  significant correlations between these parameters and cell density for both cancer and normal peripheral zone tissue. Further sections have been stained for endoglin and CD31 (which reflect microvessel density), hypoxia inducible factor 1α  and MIB-1 (a cell density marker) and are awaiting analysis.


Centre for Magnetic Resonance Investigations
Scientific Director: Professor LW Turnbull  

Improved detection and staging of invasive and in situ breast cancer  
Prof L W Turnbull, Dr P Gibbs, Dr GP Liney, Dr M Lowry, Dr D Manton, Dr M Pickles  

We have already shown that MRI provides valuable information for the detection, loco-regional staging and clinical management of breast cancer. A multi-centre study developed in Hull seeks to determine if the addition of DCE-MRI to conventional triple assessment reduces the rate of re-operation or mastectomy following wide local excision for primary breast cancer. This study includes extensive quality assurance and comparative histopathology, and psychological assessment of the sequelae of investigation.  

The National Institute for Clinical Excellence has advised on the use of MRI for screening women with a high risk of developing breast cancer from their family history. The reported sensitivity and specificity of 77% (CI 80-90%) and 81% (CI 80-83%) for MRI at 1.5T requires the development of imaging at 3T. Together with the detection of early invasive disease, MR must detect pre-invasive, ductal carcinoma in situ (DCIS). DCIS typically results in the development of microcalcifications detectable by X-ray mammography. However their absence in about 35% of cases results in failure to detect all DCIS at an early stage and inability to accurately delineate the extent of DCIS present, resulting in inadequate resection margins. We are developing very high spatial resolution imaging for textural analysis, together with functional techniques including T2 mapping, dynamic contrast-enhanced and diffusion weighted imaging to further increase specificity.  

Development of surrogate biomarkers of tumour response to chemo- or hormonal therapy  
Prof L W Turnbull. Dr M Pickles, Dr G Liney, Dr P Gibbs, Dr M Lowry, Dr D Manton  

There is considerable interest in developing reliable surrogate biomarkers of response of tumours to neoadjuvant chemotherapy (NAC) to minimise ineffectual treatment. We have used breast cancer to examine the utility of MRI as a surrogate biomarker as surgical evaluation of response is performed in most cases, but these techniques are transferable to other solid tumours.  

Measurement of response of breast cancer to NAC currently uses X-ray mammography and ultrasound scanning, but both have established disadvantages. It is thought that vascular and metabolic parameters may provide a more sensitive indicator of early tumour response than changes in volume. Treatment induced changes in vascularity examined using DCE-MRI and quantified using pharmacokinetic modelling have provided prognostic information in our hands. We have reported early changes in the apparent diffusion coefficient of water after commencing NAC and have shown that spectroscopically determined water T2 values and changes in percentage water in tissue provide prognostic information. Our pilot study demonstrated a strong association between pre-treatment tumour morphology and treatment outcome and quantification of textural parameters aided lesion discrimination.  

A multi-centre study is now required to establish whether MR imaging can reliably predict the effectiveness of NAC in locally advanced breast cancer at an early stage of treatment. To ensure feasibility we have initiated a phase 2 multi-centre, prospective, longitudinal, observational cohort study to establish quality assured foundations for a  multi-centre study by testing the technical achievability of acquiring multi-parameter, multi-MR system standardised protocols, ensuring compatibility of DICOM information between manufacturers and using semi-automated tumour segmentation and analysis tools for data extraction.  

Use of high-field strength MR derived biomarkers in prostate cancer: potential for improved cancer staging and prognostication.  
Dr P Gibbs, Mr B Zelhof, Dr M Lowry, Prof L W Turnbull  

The objectives of this study are to determine the tumour staging and prognostic potential of ultra high resolution MR imaging and MR based biomarkers, using a whole-body 3.0 Tesla scanner employing both a multiple channel phased array and an endo-rectal receiver coil, with comparison of the staging data obtained with whole mounted radical prostatectomy specimens. Various MR sequences are being examined for their relative staging accuracy, and confirmation of the underlying mechanisms of action is underway using appropriate immunohistochemical techniques. Our diagnostic accuracy data at 3T for extra-capsular tumour extension using PPA coil alone is 92%, with sensitivity 70%, specificity 97%, PPV 87% and NPV 93%. Inclusion of endo-rectal coil information results in further improvements in staging accuracy.  
The whole mounted specimens enable determination of tumour grade, location and cross-sectional area, the later allowing total tumour volume determination. Further sections are stained for: endoglin and CD31 which reflect microvessel density enabling direct comparison with the pharmacokinetic modelling parameters; hypoxia inducible factor 1a, which is known to upregulate the transcription of a number of genes connected with glucose uptake and metabolism for comparison with T2 measurements and pharmacokinetic modelling parameters; and MIB-1, a proliferation marker enabling comparison with ADC data. The subsequent efficacy of MR derived biomarkers in determining prognosis will be made, by comparison with Gleason grade, stage at radical prostatectomy and the disease free survival period.  

Prediction of response of locally advanced cervical carcinoma to chemoradiotherapy  
Prof L W Turnbull, Dr S Booth, Dr M Pickles  

Assessment of response to cancer is necessary since patients do not respond equally to first-line treatments, necessitating, for some, a switch to an alternative usually more aggressive treatment. While it is true that for patients with cervical carcinoma no realistic second-line treatment is currently available, novel therapies such as antiangiogenic chemotherapeutic agents, photodynamic therapy and gene therapy are undergoing pre-clinical evaluation. In addition the order in which more traditional treatments are delivered is also under investigation. Without the appropriate means to predict and/or assess treatment response, prior to or early during first-line treatment, patients may undergo unnecessary toxicity, without the benefit of the desired level of tumour cell kill.  

Prediction of treatment response via tissue analysis would provide invaluable information, indeed several studies have demonstrated that markers of apoptosis, cellular proliferation, angiogenesis and hypoxia are likely predictors of eventual treatment response.  

The aims of this pilot study are to: determine the efficacy of functional in vivo MR imaging and spectroscopy techniques and ex vivo high resolution magic angle spinning spectroscopy of punch biopsy specimens in predicting response prior to and early during combined chemo-radiotherapy, in patients with biopsy proven cervical cancer; determine the efficacy of pre-treatment molecular pathology obtained from cone biopsy samples in predicting response; assess the ability of in vivo and ex vivo MR and molecular pathology, in predicting longer term disease free and overall survival; and determine the optimal techniques necessary to facilitate a multi-centre study to provide the necessary statistical power to influence clinical practice.  

Assessment of bone strength and marrow composition after chemotherapy using MRI and MRS  
Dr G Liney, Prof L W Turnbull  

Both chemotherapy and hormonal therapy can cause a loss of bone density and increase the risk of osteoporosis for both men and women, causing at times severe morbidity. Once bone is lost it cannot be replaced using current treatment, but the rate of osteoporosis can be slowed down. Early intervention improves outcome as osteoporosis therapy is most effective in preventing further deterioration in bone mass. Detailed data and long-term experience to assess the risks are urgently needed and constitute an important research topic for the coming years and decades. MR imaging and spectroscopy offer the potential to non-invasively study bone strength and marrow composition, both recognised to alter following cancer treatment.   

We have already shown that MRI demonstrates the proportion of red and yellow bone marrow, hypothesised to predict remodelling capacity associated with treatment induced bone loss. Prototype chemical shift imaging sequences in use in Hull are able to demonstrate marrow distribution in the lumbar spine at high resolution, allowing late effects of cancer treatment to be examined in detail and compared with qualitative single voxel MR spectroscopy, conventional DEXA scanning and qCT for which an imaging and post-processing methodology for extremities has also been optimised. This will allow us to explore bone integrity in terms of both structural and textural parameters.  

Role of MRI in radiotherapy planning  
Dr G Liney, Prof L W Turnbull  

This study has examined the efficacy of incorporating dynamic contrast enhanced imaging into the treatment planning of head & neck tumours. Software has been developed to enable functional and anatomical images to be fused in a DICOM compliant format so that it can be input into the radiotherapy planning system. In addition, this software produces morphologically optimised enhancement data which is more suitable for contouring and planning radiotherapy distributions. On going work is utilising these images to obtain conformal plans which demonstrate the strategy of improving tumour control and reducing side effects by boosting dose to the vascular part of the planning target volume. In parallel with the main part of the work, gel dosimetry (in collaboration with Prof J Goodby at York University) is being used to verify these complex dose distributions for which routine film dosimetry is no longer suitable. Computer generated IMRT planned head & neck treatment data has been compared to MR dose maps and good accuracy has been demonstrated.  

The techniques developed in this study are being translated into other areas (e.g. prostate) and will continue to have wide reaching implications for the efficacy of advanced MRI in radiotherapy of cancer.