Periodic Reporting for period 4 - PRYSM (Early recognition of intracranial aneurysms to PRevent aneurYSMal subarachnoid hemorrhage)
Período documentado: 2024-08-01 hasta 2025-01-31
HEALTHCARE PROBLEM
Intracranial aneurysms (IAs)—balloon-like bulges in the walls of brain arteries—are common, affecting approximately 3% of the population, or 10 million adults in the EU. While most aneurysms remain asymptomatic, rupture can cause aneurysmal subarachnoid hemorrhage (ASAH), a severe form of stroke. One-third of ASAH patients die, while another third suffer severe disability, requiring lifelong care. ASAH typically occurs around age 50—significantly younger than most other cardiovascular diseases—impacting patients during their working years and placing a heavy burden on families and healthcare systems.
Early detection of aneurysms is possible through imaging, which can enable preventive treatment via endovascular intervention or surgery. However, current screening strategies are inadequate due to a limited understanding of who is at risk. The pathogenesis of IAs and ASAH remains poorly understood, as only a few genetic, environmental, and imaging risk factors have been identified. Furthermore, the interplay between these factors remains unclear, partly due to the lack of large, well-characterized cohorts for adequately powered studies. Notably, the disease exhibits significant heterogeneity, with sex-related differences being particularly striking—two-thirds of ASAH patients are women. However, the reasons behind this increased risk in women remain unknown.
OVERALL OBJECTIVE
The PRYSM (Early recognition of intracranial aneurysms to prevent aneurysmal subarachnoid hemorrhage) project aimed to enhance disease understanding to enable early recognition and prevention of ASAH.
APPROACH
PRYSM introduced a novel, integrated approach combining genetic, clinical, and imaging risk factors, assessed using machine learning. Machine learning—an AI-driven computational method—analyzes data patterns to improve predictive accuracy. Through this approach, imaging markers, genetic variants, including those with roles in arterial wall structure and function, and environmental risk factors were identified. Additionally, interactions between female sex and specific risk factors contributing to the disease were uncovered, and sex-specific prediction models improved predictive accuracy for disease prediction.
CONCLUSION
The combined results of PRYSM advance our understanding of IA development and represent an important step toward individualized risk prediction and precision medicine in early IA detection. By improving early identification of high-risk individuals, these findings contribute to reducing the burden of ASAH and improving patient outcomes.
Intracranial aneurysms (IAs)—balloon-like bulges in the walls of brain arteries—are common, affecting approximately 3% of the population, or 10 million adults in the EU. While most aneurysms remain asymptomatic, rupture can cause aneurysmal subarachnoid hemorrhage (ASAH), a severe form of stroke. One-third of ASAH patients die, while another third suffer severe disability, requiring lifelong care. ASAH typically occurs around age 50—significantly younger than most other cardiovascular diseases—impacting patients during their working years and placing a heavy burden on families and healthcare systems.
Early detection of aneurysms is possible through imaging, which can enable preventive treatment via endovascular intervention or surgery. However, current screening strategies are inadequate due to a limited understanding of who is at risk. The pathogenesis of IAs and ASAH remains poorly understood, as only a few genetic, environmental, and imaging risk factors have been identified. Furthermore, the interplay between these factors remains unclear, partly due to the lack of large, well-characterized cohorts for adequately powered studies. Notably, the disease exhibits significant heterogeneity, with sex-related differences being particularly striking—two-thirds of ASAH patients are women. However, the reasons behind this increased risk in women remain unknown.
OVERALL OBJECTIVE
The PRYSM (Early recognition of intracranial aneurysms to prevent aneurysmal subarachnoid hemorrhage) project aimed to enhance disease understanding to enable early recognition and prevention of ASAH.
APPROACH
PRYSM introduced a novel, integrated approach combining genetic, clinical, and imaging risk factors, assessed using machine learning. Machine learning—an AI-driven computational method—analyzes data patterns to improve predictive accuracy. Through this approach, imaging markers, genetic variants, including those with roles in arterial wall structure and function, and environmental risk factors were identified. Additionally, interactions between female sex and specific risk factors contributing to the disease were uncovered, and sex-specific prediction models improved predictive accuracy for disease prediction.
CONCLUSION
The combined results of PRYSM advance our understanding of IA development and represent an important step toward individualized risk prediction and precision medicine in early IA detection. By improving early identification of high-risk individuals, these findings contribute to reducing the burden of ASAH and improving patient outcomes.
The project was structured into four work packages (WPs):
WP1: IMAGING
We developed a semi-automated tool to measure the diameters, configurations, and angles of intracranial arteries. Using this tool, we identified several vessel characteristics associated with IA development (also see image imaging markers attached).
WP2: IMAGING GENETICS
We analyzed genetic factors underlying the imaging markers identified in WP1. Associations were found between genetic variants near the PKD1L2, TRPC6, and DCSTAMP genes—known for their roles in arterial wall structure and function. Additionally, the TRPC6 gene variant was linked to IA development.
WP3: ENVIRONMENT
Currently, only hypertension and smoking are established environmental risk factors for IA and ASAH. We developed a personalized prediction model using environmental factors but found no new risk factors, as the algorithm relied on known predictors. To overcome this, we employed an alternative approach independent of prediction models, revealing several potential risk factors for ASAH, including blood biomarkers and a lung function marker. However, information on these risk factors are not routinely available in large population cohorts. In this WP we also identified commonly prescribed drugs associated with a reduced risk of ASAH: lisinopril, amlodipine, simvastatin, metformin, and tamsulosin. Last, we found an interaction between sex and smoking with women who smoke having a higher risk of both unruptured IAs and ASAH than men.
WP4: ANEURYSM PREDICTION
We conducted the largest genome-wide association study (GWAS) on IAs to date and developed a polygenic risk score (PRS) for ASAH incidence. While PRS provided modest predictive value beyond known environmental factors, it did not significantly improve IA presence prediction (also see image PRS attached). Similarly, our IA presence and ASAH prediction models underperformed, likely due to interactions between female sex and specific risk factors. We developed sex-specific models, which improved predictive accuracy. Adding imaging markers from WP1 to the IA presence model enhanced performance, though statistical significance was not achieved. Validation in a larger cohort is needed, but such a dataset is currently unavailable.
Professionals have been informed through publications in international, peer reviewed, scientific journals and presentations at international conferences, such as the European Stroke Organization Conference. The results of the genetic analyses specifically have been shared at the international stroke genetics consortium conferences.
WP1: IMAGING
We developed a semi-automated tool to measure the diameters, configurations, and angles of intracranial arteries. Using this tool, we identified several vessel characteristics associated with IA development (also see image imaging markers attached).
WP2: IMAGING GENETICS
We analyzed genetic factors underlying the imaging markers identified in WP1. Associations were found between genetic variants near the PKD1L2, TRPC6, and DCSTAMP genes—known for their roles in arterial wall structure and function. Additionally, the TRPC6 gene variant was linked to IA development.
WP3: ENVIRONMENT
Currently, only hypertension and smoking are established environmental risk factors for IA and ASAH. We developed a personalized prediction model using environmental factors but found no new risk factors, as the algorithm relied on known predictors. To overcome this, we employed an alternative approach independent of prediction models, revealing several potential risk factors for ASAH, including blood biomarkers and a lung function marker. However, information on these risk factors are not routinely available in large population cohorts. In this WP we also identified commonly prescribed drugs associated with a reduced risk of ASAH: lisinopril, amlodipine, simvastatin, metformin, and tamsulosin. Last, we found an interaction between sex and smoking with women who smoke having a higher risk of both unruptured IAs and ASAH than men.
WP4: ANEURYSM PREDICTION
We conducted the largest genome-wide association study (GWAS) on IAs to date and developed a polygenic risk score (PRS) for ASAH incidence. While PRS provided modest predictive value beyond known environmental factors, it did not significantly improve IA presence prediction (also see image PRS attached). Similarly, our IA presence and ASAH prediction models underperformed, likely due to interactions between female sex and specific risk factors. We developed sex-specific models, which improved predictive accuracy. Adding imaging markers from WP1 to the IA presence model enhanced performance, though statistical significance was not achieved. Validation in a larger cohort is needed, but such a dataset is currently unavailable.
Professionals have been informed through publications in international, peer reviewed, scientific journals and presentations at international conferences, such as the European Stroke Organization Conference. The results of the genetic analyses specifically have been shared at the international stroke genetics consortium conferences.
The following analyses we are still working on and are beyond the state of the art:
WP 1 IMAGING
We were unable to develop an automatic CoW segmentation and quantification tool, as it proved too technically challenging. Even after organizing a challenge at the Medical Image Computing and Computer-Assisted Intervention (MICCAI) Conference 2023 to solicit input from other researchers in the imaging field, no viable solution emerged. This outcome further highlights the complexity of the problem. As a result, we have established new collaborations, with whom we co-organized the challenge. These partnerships allow us to continue working together and combining efforts to address this challenge.
WP 2 IMAGING GENETICS
Our sex-stratified follow-up GWAS is currently underway with a cohort four times larger than the one assessed in the current ERC Starting Grant project. The analyses are still in progress, and the results will be reported upon completion. Additionally, we will describe the findings from WP2, where we investigated the association of genetic loci of arterial characteristics with IAs and identified an association between the TRPC6 gene and IAs.
WP 3 ENVIRONMENT
Our findings from WP3, which identified five drugs associated with a reduced risk of ASAH, were unexpected but highly intriguing. Preliminary analyses indicate that all five drugs may inhibit endothelial-to-mesenchymal transition (EndMT), while additional cross-disciplinary mechanistic insights I have gathered, suggest that EndMT plays a central causative role in IA and ASAH. Although identifying therapeutic targets was not a direct objective of this project, these results provide critical insights into the disease’s causal pathway and potential treatment strategies. We are actively investigating the causal role of EndMT and assessing the effectiveness of these drugs in reducing ASAH incidence within my recently awarded ERC Proof of Concept (PoC) project.
WP4: ANEURYSM PREDICTION
As suitable cohorts to test whether imaging markers from WP1 can enhance IA presence prediction in the general population are unavailable, we are instead evaluating the added value of the imaging markers in risk prediction in 1,000 individuals screened for IAs, with IAs detected in 10% (ongoing analyses).
WP 1 IMAGING
We were unable to develop an automatic CoW segmentation and quantification tool, as it proved too technically challenging. Even after organizing a challenge at the Medical Image Computing and Computer-Assisted Intervention (MICCAI) Conference 2023 to solicit input from other researchers in the imaging field, no viable solution emerged. This outcome further highlights the complexity of the problem. As a result, we have established new collaborations, with whom we co-organized the challenge. These partnerships allow us to continue working together and combining efforts to address this challenge.
WP 2 IMAGING GENETICS
Our sex-stratified follow-up GWAS is currently underway with a cohort four times larger than the one assessed in the current ERC Starting Grant project. The analyses are still in progress, and the results will be reported upon completion. Additionally, we will describe the findings from WP2, where we investigated the association of genetic loci of arterial characteristics with IAs and identified an association between the TRPC6 gene and IAs.
WP 3 ENVIRONMENT
Our findings from WP3, which identified five drugs associated with a reduced risk of ASAH, were unexpected but highly intriguing. Preliminary analyses indicate that all five drugs may inhibit endothelial-to-mesenchymal transition (EndMT), while additional cross-disciplinary mechanistic insights I have gathered, suggest that EndMT plays a central causative role in IA and ASAH. Although identifying therapeutic targets was not a direct objective of this project, these results provide critical insights into the disease’s causal pathway and potential treatment strategies. We are actively investigating the causal role of EndMT and assessing the effectiveness of these drugs in reducing ASAH incidence within my recently awarded ERC Proof of Concept (PoC) project.
WP4: ANEURYSM PREDICTION
As suitable cohorts to test whether imaging markers from WP1 can enhance IA presence prediction in the general population are unavailable, we are instead evaluating the added value of the imaging markers in risk prediction in 1,000 individuals screened for IAs, with IAs detected in 10% (ongoing analyses).