Pathological Validation of Functional Imaging in Head and Neck Squamous Cell Carcinoma

Overview

Chemo-radiotherapy (CRT) is currently the cornerstone in the management of locoregional advanced head and neck cancer (HNC). Optimization of the quality of RT is therefore an important issue, if the investigators want to improve the therapeutic index in HNC. This could be achieved by a more accurate definition of the tumor volume and by identification of radioresistant volumes within the tumor. Recent literature puts in this regard the incorporation of functional imaging (FI) in the RT treatment planning forward as a promising tool. FI modalities provide an outstanding contrast between tumor and surrounding tissues. This is in contrast to anatomical imaging. Using anatomical imaging in RT treatment planning, sufficient margins need to be placed around the tumor volume in order to compensate for geometric uncertainties. Consequently many surrounding functional structures receive high doses of irradiation, resulting in side effects. It is expected that, using FI in RT treatment planning will make these margins smaller or even unnecessary, which will result in less irradiation of the surrounding tissues. So far only one study has reported a comparison between tumor volume on anatomical (CT and MRI) and FI (PET-CT) modalities with pathological tumor volume. This study showed indeed that the tumor volumes delineated on PET-CT correlated more to tumor volumes defined by pathology and were significantly smaller. Furthermore, FI provides us with a deeper insight in the tumor's underlying biological activity and microstructure. These techniques can thus help to identify radioresistant subvolumes which might benefit from treatment intensification. A validation of these FI modalities with pathology is necessary to investigate their true power in tumor delineation and in the identification of radioresistant subvolumes.

1. BACKGROUND AND SETTING 1.1. INTRODUCTION Concurrent (chemo-) radiotherapy (CRT) is the current standard of care for patients with locally advanced head and neck squamous cell carcinoma (HNSCC). The proximity of important functional structures with the tumour makes treatment however highly complex. Treatment related toxicity can be severe with xerostomia and dysphagia gravely complicating the patient's quality of life. The use of altered fractionation schedules and/or concurrent chemotherapy has resulted in substantial gains in loco-regional control, leading to significant improvements in overall survival. Although the prognosis of patients treated with RT for locally advanced HNSCC is continually improving, there are still a high number of locoregional failures. Using highly conformal radiotherapy the investigators can increase the therapeutic index by giving higher doses to the more radio-resistant parts of the tumour while maintaining or even reducing dose to the surrounding tissues. Functional imaging can help us to define these radio-resistant subvolumes and help us to delineate the gross tumour volume (GTV) better in the future. Positron emission tomography (PET), Diffusion weighted magnetic resonance imaging (DWI) and Dynamic contrast enhanced MRI (DCE-MRI) can give us further insight in the tumour's underlaying biological and microstructural characteristics. However, at the moment the investigators are not sure how to interpret the different imaging parameters obtained from these functional imaging modalities. Therefore, the investigators want to correlate the imaging parameters with the pathological characteristics of the tumour. On the other hand the investigators want to correlate the GTV defined on the functional imaging modalities with the pathological GTV, and see if these modalities can help us to delineate the GTV more accurately in the future. 1.2. IMAGING MODALITIES 1.2.1. FDG-Positron Emission Tomography (FDG-PET) Increased glucose metabolism is a fundamental characteristic of many malignant tumours, including HNSCC. Positron emission tomography (PET) after injection of radioactively labeled 2' fluorodeoxyglucose (18F-FDG) can help us measure and quantify this metabolism in tumours, using the standardized uptake value (SUV). Several studies have correlated high SUV prior to treatment with significantly worse outcome in HNSCC. Furthermore it appears that most local recurrences occur within the FDG-PET defined GTV. The molecular base for this is not completely understood. Proliferating tumour cells consume glucose at a high rate and release lactate and not CO2. This way they omit the mitochondria driven oxidative phosphorylation of glucose from the production of ATP (=Warburg-effect). Several molecular parameters have been associated with this glycolytic switch, such as activation of the hypoxia induced HIF1α and the increased presence of GLUT glucose transport proteins. So far no clear correlation was found between the presence of hypoxia and FDG-uptake. 1.2.2. Diffusion-weighted Magnetic resonance imaging (DWI) Diffusion weighted magnetic resonance imaging (DWI) can characterize tissues based on the random displacement of water molecules. In biological tissues this displacement is limited by underlying tissue-specific barriers and this difference can be quantified and visualized using apparent diffusion coefficient (ADC) values. A study on 165 patients in our centre with HNSCC demonstrated that the pre-treatment ADC value is a strong and independent prognostic factor for outcome. Patients with a high ADC value respond worse to ionizing radiation than patients with a lower ADC value. There is currently no clear explanation why tumours with a higher cellular density are more radiosensitive than tumours with a low density. A number of microscopic features effect water diffusivity in tissue (presence of necrosis, cellular density, inflammation, integrity of cellular membranes). Many of these also affect the radio- and chemosensitivity of the tumours. 1.2.3. Dynamic contrast enhanced Magnetic resonance imaging (DCE-MRI) Dynamic contrast enhanced MRI (DCE-MRI) is a non-invasive imaging modality which has been developed and evaluated for the characterization of vascular properties of tissues. Adequate oxygen supply is essential for the effectivity if ionizing radiation. However tumor angiogenesis is far from perfect and newly formed vessels display permeability, tortuosity and a generally poor functionality. These vascular properties could provide us insight in the aggressiveness and radiosensitivity of the tumor. Dynamic computed tomography (CT) has demonstrated that perfusion measurements can predict outcome in head and neck cancer after radiotherapy. Similarly quantitative assessment of DCE-MRI has also been correlated with response to ionizing radiation in patients with HNSCC. A recent trial on a limited number of patients correlated certain DCE-MRI parameters with intratumoural hypoxia. However, so far no spatial correlation has been done. 2. STUDY OBJECTIVES 2.1. PRIMARY OBJECTIVES The main objective of this study is to correlate DWI, DCE-MRI and FDG-PET with the spatial distribution of hypoxia in patients with head and neck cancer. 2.2. SECONDARY OBJECTIVES A. To correlate findings on functional imaging with intrinsic molecular parameters depicting proliferation, hypoxia and glucose metabolism. B. To validate the use of the hypoxia markers and the 15-gene hypoxia gene expression classifier. C. To compare tumour volume as derived from pathology with the volumes delineated on the several anatomical and functional imaging modalities. 3. STUDY DESIGN The target group for this trial is patients with a histologically proven squamous cell carcinoma of the larynx, eligible for surgery, staged T3-4. Prior to treatment patients will undergo an FDG-PET/CT and MRI with DW and DCE, dedicated for optimal visualization of the primary tumor and analysis. The imaging modalities will be performed as close to the surgery as possible (PET/CT one week before surgery; MRI one day before surgery). Prior to treatment, three slices on the imaging modalities will be selected by a radiologist based on their visual quality and heterogeneity of functional parameters. These regions of interest will be compared to different histopathological parameters on the resection specimen. The radiologist will also select regions, based on DWI and DCE parameters, who are suspect to be hypoxic. From these regions a biopsy will be taken. After total laryngectomy, the resection specimen will be oriented, and biopsy material will be obtained from the regions in the tumour that have been selected on the imaging modalities The rest of the specimen will be fixed in formaldehyde. To account for the shrinkage of the tumour due to fixation, the resection specimen will be placed in a cardboard box with an agarose solution and scanned with an MRI with T1w/T2 images. From this a shrinkage factor will be calculated comparing the delineated T1 and T2 weighted images. After fixation the tumour will be cut up into macroslices with a thickness of about 5mm. On each macroslice tumour volume will be delineated. Using this, the pathological tumour volume will be determined taking into account the previously determined shrinkage factor. This pathological volume will be compared to the tumour volume delineated on the imaging modalities. On the different pre-surgery functional imaging modalities, 4 separate observers will delineate the GTV. The different volumes will be compared and overlap will be calculated after 3D specimen reconstruction and registration. The GTV obtained from the functional imaging will be correlated with the pathological tumor volume of the resection specimen. This information will give us more insight into the true power of the investigated functional imaging techniques in determining tumour volume. This is important to assess the role of functional imaging modalities in treatment planning in the future. At the three levels, chosen by the radiologist on the imaging modalities, 4µm thick slices will be taken. On each level an immunohistochemical staining will be carried out (GLUT-1, CA-IX, HIF-1α, VEGF and KI 67 staining will be performed ). IHC will be scored according to a semiquantitative scoring system using a field analysis where each field will be assigend a score of 1-4 according to the approximate area of immunostaining (0:0%, 1:1-5%, 2: 5-15%, 3: 15-30% and 4: \>30%. 4. SAMPLE SIZE Twenty patients with squamous cell carcinoma of the larynx, planned to undergo total laryngectomy, will be included.