Periodic Reporting for period 1 - DECIDER (Improved clinical decisions via integrating multiple data levels to overcome chemotherapy resistance in high-grade serous ovarian cancer)
Okres sprawozdawczy: 2021-02-01 do 2022-07-31
Epithelial ovarian cancer kills more than 40,000 women in Europe each year. High-grade serous ovarian cancer (HGSOC) is the most common and most difficult-to-treat subtype of this disease. Despite initially good response to platinum and taxane, the HGSOC tumours typically develop chemotherapy resistance leading to relapses and 5-year survival of only 38%. A reason for the chemotherapy resistance is that HGSOC is often diagnosed when it has already metastasised to the abdomen and the tumours consist of billions of cancer cells with numerous genomic aberrations in their genome, some of which causing the ability to overcome the chemotherapy effect. Accordingly, it is extremely difficult to find effective drugs or combinations of drugs that provide a lasting effect for the patients. Current understanding on many important questions, such as chemotherapy’s effect on the genomes and to what degree tumour genomes differ between poor and good responders, is insufficient to tailor treatment based on an HGSOC patient’s genome.
In the DECIDER project, our overall goal is to deliver tools that enable more effective and cost-efficient treatment of HGSOC patients. Firstly, our goal is to comprehensively characterise chemotherapy resistance mechanisms in HGSOC. To accomplish this goal, we prospectively collect samples from HGSOC patients who have given consent to the DECIDER study. These samples are then analysed with state-of-the-art measurement technologies to reveal aberrations at DNA, RNA or epigenetic levels. These data are combined with data from histopathological and functional imaging, as well as clinical records. Secondly, we will utilise the acquired knowledge and suggest effective treatments to HGSOC patient groups characterised by genomic aberrations.
Thus, an important part of the DECIDER activities is to suggest treatment options, if such exist, for relapsed patients without standard-of-care treatment options and belonging to the DECIDER study based on genomic information or ex vivo cell model systems.
In the DECIDER project, our overall goal is to deliver tools that enable more effective and cost-efficient treatment of HGSOC patients. Firstly, our goal is to comprehensively characterise chemotherapy resistance mechanisms in HGSOC. To accomplish this goal, we prospectively collect samples from HGSOC patients who have given consent to the DECIDER study. These samples are then analysed with state-of-the-art measurement technologies to reveal aberrations at DNA, RNA or epigenetic levels. These data are combined with data from histopathological and functional imaging, as well as clinical records. Secondly, we will utilise the acquired knowledge and suggest effective treatments to HGSOC patient groups characterised by genomic aberrations.
Thus, an important part of the DECIDER activities is to suggest treatment options, if such exist, for relapsed patients without standard-of-care treatment options and belonging to the DECIDER study based on genomic information or ex vivo cell model systems.
The DECIDER project started in February 2021 and the main results are detailed herein.
The basis for all research in the DECIDER project is the prospective collection of tissue samples of HGSOC patients. To that end, we have optimised an efficient system at the Hospital District of Southwest Finland, that ensures the collection of high-quality longitudinal tissue samples, body fluids and relevant clinical data from each patient that is included into the study. Within the first 18 months of the DECIDER project, 63 HGSOC patients have been recruited. Furthermore, patients that were recruited before February 2021 in the HERCULES study belong to the DECIDER project as well. Thus, currently the DECIDER cohort consists of 312 patients. The collected tissues are prepared by experienced personnel at the Hospital District of Southwest Finland for the purposes of 1) tissue dissociation and further processing as organoids and tumoroids to be used in DECIDER, 2) analysis of digitalised haematoxylin and eosin (H&E) stained histopathological samples, and 3) DNA or RNA extraction and subsequent sequencing.
State-of-the-art sequencing at our current service provider includes DNA (whole genome) and RNA (expression of the genes) from individual cells that stem from different parts of the tumour. Furthermore, we analyse the methylation of DNA, and the so-called circulating tumour DNA (ctDNA) from plasma samples. From the 312 prospective patients, 252 have currently whole genome sequencing (WGS) data available. The sequenced data are preprocessed and analysed to reveal mutations, copy number variations, gene fusions, and other changes in in the patients’ genomes. Based on the longitudinal data, and by integrating DNA, transcriptomics and DNA methylation data, we are able to model cancer evolution, to give insides into how the genome of the tumour changes in response to therapy. After analysis of the sequencing results, critical information (like germ line variants related to hereditary ovarian, breast and colorectal cancer risk) is regularly reported back to the treating clinicians of the patients. In the first 18 months alone, we reported back to the clinic relevant information for nine of the HGSOC patients included in DECIDER.
Apart from sequencing results, a pilot version of a deep learning-based AI model was developed that can predict HGSOC patient’s chemotherapy response based on patterns on histopathological slide images from HGSOC tissues. This tool is a critical component of Pillar I in the DECIDER project, which aims to identify patients unlikely to benefit from standard therapy, at the time of diagnosis.
We have also established 28 long-term HGSOC organoids from patient samples collected during the DECIDER project. These organoids are going to be the major ex vivo tool to evaluate the repurposing of drugs that were identified to be potentially usable as targets to treat resistant HGSOC.
The DECIDER project is already producing a vast amount of data for each patient, consisting of clinical and several levels of sequencing data, predictive models, and histopathological data. An open-source AI-based software is currently being developed to visualise these different data and to give the project’s researchers and clinicians an effective tool to integrate and interpret them. This will facilitate clinical decision making based on the DECIDER findings.
The basis for all research in the DECIDER project is the prospective collection of tissue samples of HGSOC patients. To that end, we have optimised an efficient system at the Hospital District of Southwest Finland, that ensures the collection of high-quality longitudinal tissue samples, body fluids and relevant clinical data from each patient that is included into the study. Within the first 18 months of the DECIDER project, 63 HGSOC patients have been recruited. Furthermore, patients that were recruited before February 2021 in the HERCULES study belong to the DECIDER project as well. Thus, currently the DECIDER cohort consists of 312 patients. The collected tissues are prepared by experienced personnel at the Hospital District of Southwest Finland for the purposes of 1) tissue dissociation and further processing as organoids and tumoroids to be used in DECIDER, 2) analysis of digitalised haematoxylin and eosin (H&E) stained histopathological samples, and 3) DNA or RNA extraction and subsequent sequencing.
State-of-the-art sequencing at our current service provider includes DNA (whole genome) and RNA (expression of the genes) from individual cells that stem from different parts of the tumour. Furthermore, we analyse the methylation of DNA, and the so-called circulating tumour DNA (ctDNA) from plasma samples. From the 312 prospective patients, 252 have currently whole genome sequencing (WGS) data available. The sequenced data are preprocessed and analysed to reveal mutations, copy number variations, gene fusions, and other changes in in the patients’ genomes. Based on the longitudinal data, and by integrating DNA, transcriptomics and DNA methylation data, we are able to model cancer evolution, to give insides into how the genome of the tumour changes in response to therapy. After analysis of the sequencing results, critical information (like germ line variants related to hereditary ovarian, breast and colorectal cancer risk) is regularly reported back to the treating clinicians of the patients. In the first 18 months alone, we reported back to the clinic relevant information for nine of the HGSOC patients included in DECIDER.
Apart from sequencing results, a pilot version of a deep learning-based AI model was developed that can predict HGSOC patient’s chemotherapy response based on patterns on histopathological slide images from HGSOC tissues. This tool is a critical component of Pillar I in the DECIDER project, which aims to identify patients unlikely to benefit from standard therapy, at the time of diagnosis.
We have also established 28 long-term HGSOC organoids from patient samples collected during the DECIDER project. These organoids are going to be the major ex vivo tool to evaluate the repurposing of drugs that were identified to be potentially usable as targets to treat resistant HGSOC.
The DECIDER project is already producing a vast amount of data for each patient, consisting of clinical and several levels of sequencing data, predictive models, and histopathological data. An open-source AI-based software is currently being developed to visualise these different data and to give the project’s researchers and clinicians an effective tool to integrate and interpret them. This will facilitate clinical decision making based on the DECIDER findings.
In the DECIDER project, we have a unique set-up with a highly efficient system of collecting, processing, and sequencing of tissue samples from HGSOC patients. In combination with the state-of-the-art sequencing technologies and the expertise of our project partners in genomics, molecular biology, computer science and AI, we have an optimal setting to take on the challenges of chemoresistant HGSOC.
Currently, cell characteristics that make a tumour resistant to chemotherapy are to a large degree unknown. Therapy can hardly be tailored to the patient, due to the wide heterogeneity in genetic changes, and the current therapy options are insufficient and ineffective.
Until the end of the project, we expect to optimise and accelerate our analyses of genetic changes responsible for chemoresistance in individual patients and optimise the high-throughput drug screening experiments using ex vivo cell model systems. Based on that, we expect to gain a comprehensive understanding of the genomic set-up in patients’ tumour tissues, and the ability to suggest effective personalised treatment options. Already now, results stemming from this project have been used to facilitate clinical decision making, and we expect that by the end of the project this will be done in an even more extensive and efficient way.
More targeted treatments for chemoresistant HGSOC patients will reduce the number of expensive but inefficient treatments and will thus improve patient survival and well-being and save health care costs.
Currently, cell characteristics that make a tumour resistant to chemotherapy are to a large degree unknown. Therapy can hardly be tailored to the patient, due to the wide heterogeneity in genetic changes, and the current therapy options are insufficient and ineffective.
Until the end of the project, we expect to optimise and accelerate our analyses of genetic changes responsible for chemoresistance in individual patients and optimise the high-throughput drug screening experiments using ex vivo cell model systems. Based on that, we expect to gain a comprehensive understanding of the genomic set-up in patients’ tumour tissues, and the ability to suggest effective personalised treatment options. Already now, results stemming from this project have been used to facilitate clinical decision making, and we expect that by the end of the project this will be done in an even more extensive and efficient way.
More targeted treatments for chemoresistant HGSOC patients will reduce the number of expensive but inefficient treatments and will thus improve patient survival and well-being and save health care costs.