Functional imaging for individualized cancer treatment

Kreft i livmor og i livmorhals er vanlige kreftformer hos kvinner. Viktige utfordringer i behandlingen av disse kreftformene er: (1) overbehandling av pasienter med lav risiko for tilbakefall, (2) underbehandling av pasienter med høy risiko for tilbakefall og (3) dårlig overlevelse hos pasienter med utbredt sykdom. Mangelfulle verktøy for å skille mellom høy- og lavrisikopasienter og til å skreddersy behandlingen, er noe av årsaken til disse utfordringene. Ved bruk av bl.a. kunstig intelligens og maskinlæring på radiologiske bilder, vil vi prøve å lære maskinen å finne svulsten i bildene. Videre skal svulstens egenskaper basert på bildedata sammenliknes med resultater fra vevsprøver av svulsten. Ved bruk av maskinlæring for å analysere alle dataene fra bildene og vevsprøvene samlet, ønsker vi å avdekke nye sammenhenger av betydning for pasientens prognose og hvilken behandling som forventes å være mest effektiv. Dette prosjektet har som målsetting å gi bedre verktøy for å tilby mer skreddersydd behandlingen ved gynekologisk kreft i fremtiden.

In gynecologic cancer there is an immediate need for accurate preoperative tumor assessments to safely individualize therapy. Overtreatment of early stage disease and poor survival in metastatic disease partially relates to lack of predictive preoperative imaging markers. Gynecologic cancer patients prospectively subjected to preoperative imaging by magnetic resonance imaging (MRI) and positron emission tomography (PET) / computed tomography (CT), are studied to identify novel imaging biomarkers relevant for prognosis and treatment. In parallel tumor tissue of the same patients are analyzed for genetic and molecular markers in relation to clinical findings and outcome. Furthermore, the research group has recently established orthotopic mouse endometrial cancer models derived from cell lines and from patient derived biopsies. These models will be studied by optical imaging, MRI and PET-CT during disease progression. Functional imaging characteristics will be explored in these animal models and studied in relation to corresponding molecular characteristics of the tumor. The research group has extensive experience with preoperative imaging in the context of clinical trials and tissue markers in gynecologic cancer. This research platform allows for novel comparison between advanced functional imaging findings and genetic, molecular, metabolic and angiogenic characteristics of the corresponding tumor tissue. Building on previous experience, the current project will explore functional imaging findings in gynecologic cancer in patients and in animal models in relation to stage, outcome, and genetic and molecular tumor markers. The overall aim of the project is to provide improved imaging tools with which to identify high-risk patients and disease progression, and thus facilitate better individualized treatment, reduced morbidity and implementation of targeted therapy in gynecologic cancer patients.