MD, PhD Candidate Henning Langen Stokmo gave a lecture at CarciNor Annual Conference in March, about image analysis and neuroendocrine carcinomas.
Stokmo is a current research associate at Oslo University Hospital, and was awarded research funds from CarciNor’s Research Fund in 2015. He is currently working as a junior consultant at Oslo University Hospital, and as a radiology resident at Vestre Viken Hospital Trust.
Text and photo: Mari Sandvold, editor
Clinical usefulness of advanced image analysis of 18F-FDG PET/CT for in vivo characterisation of neuroendocrine carcinomas
In our project, we will perform advanced image analyses of 18F-fluorodeoxyglucose positron emission tomography (18F-FDG PET) and computer tomography (CT) images and compare these new imaging parameters with blood- and immunohistochemical biomarkers as well as clinical outcome. The goal is to improve tumour classification, diagnosis, prognosis and treatment response in neuroendocrine carcinomas (NECs).
Today, the World Health Organization (WHO) classification of NECs is based on the Ki-67 cell proliferation index using immunohistochemistry analysis of tumour tissue biopsies. This WHO classification is the basis for treatment selection. NECs are characterised by a biological heterogeneity and there may be large variations in the Ki-67 index within, as well as between tumours and metastases within the same patient. Inevitably, classification based on a limited number of biopsies is prone to sampling errors as the biopsies may not be representative for the disease.
Cancer cells have altered metabolism to promote growth, survival and proliferation, and a common feature of cancer cells is increased glucose uptake and metabolism, a phenomenon known for many years as the Warburg effect. Although described as early as in the 1920’s by Warburg and colleagues, a renewed interest in this phenomenon has risen due to increased research and knowledge in cancer cell biology and metabolism. In this study, we will use advanced image analysis to study the glucose uptake in neuroendocrine neoplasias (NENs). Our hypothesis is that tumour aggressiveness is positively correlated to both level of the Ki-67 index and glucose demand. The glucose uptake in vivo can be depicted and quantified by intravenous administration of a low dose of the radiolabelled glucose analogue, 18F-FDG and imaged by PET.
Texture analysis is mathematical analysis of the inter-relationships between pixels in images. NECs are spatially and temporarily heterogeneous and texture analysis of the PET and CT images will increase the information that can be extracted and used characterise tumour heterogeneity.
Whole body 18F-FDG PET/CT depicts the entire disease burden and its biology (glucose metabolism and heterogeneity) within different lesions. Comparing the results from imaging with biomarkers and genetic profiling from blood and immunohistochemistry from selected lesions holds the promise to improve the classification of NECs and consequently, the correct selection of treatment. This also allows the extracted image characteristics (the radiomics features) to be directly validated using the genetic profile attained from liquid biopsies. Combining these methods, our multi-disciplinary research team can establish new ways to more accurately predict and improve outcome for NEC patients.
Text: MD, PhD Candidate Henning Langen Stokmo