心血管疾病：一种新的预警系统和稀缺性-Horus ML

DECEMBER2025CASESTUDY Cardiovascular Diseases:A New Early Warning Systemand Scarcity Horus ML has developed a low-cost cardiovascular screening system,which uses AI to rapidly analyse retinal images in primary care settings will also address gender bias in cardiovascular diagnosisthrough automated, unbiased screening. To acceleratedevelopment, the start-up has secured more than over€750,000 in public funding. ExecutiveSummary Approximately 50% of heart attacks occur in individualspreviously classified as low risk. Preventative healthcare isheld back by a lack of effective cardiovascular screeningin primary care centres, which don’t have the necessaryspecialised equipment and trained personnel, accordingto Horus ML, a Madrid-based startup. Now performing clinical validation of the new systemwith Madrid’s Hospital Infanta Leonor, Horus is aiming tohave AItheroscope in use by 20 primary care centres andthree hospitals in Spain by the end of 2025. Following asuccessful Spanish deployment, it is planning to sell thesystem internationally, targeting countries with similarhealthcare systems. Horus has developed a new cardiovascular screeningsystem, called AItheroscope, which combines advanceddeep learning algorithms with RetinoLens, a low-costoptical device, to analyse retinal images and detectsubclinical atherosclerosis - asymptomatic plaque buildupin the arteries.Designed to generate a diagnosis in lessthan 10 seconds, AItheroscope could be used by primarycare clinicians and even individuals in their homes. Horus estimates that AItheroscope can address a globalmarket of one billion people who currently have noaccess to effective cardiovascular screening, as they livein underserved rural and resource-limited areas wherespecialised cardiovascular care is unavailable. As well as lowering diagnostic costs, the system promisesto prevent progression to severe cardiovascular events,and reduce hospitalisations. Horus believes AItheroscope Cardiovascular diseases are the leading cause ofmortality worldwide, according to the World HealthOrganisation.1Studies show that approximately50%2of heart attacks occur in individuals previouslyclassified as low risk. The system processes retinal images throughconvolutional neural networks and transformermodels, providing cardiovascular risk assessmentsin under 10 seconds, explains Jesús Prada Alonso,founder and CEO of Horus. This rapid processingspeed is designed to enable integration into routineprimary care workflows without disrupting clinicaloperations. “This approach democratises access toadvanced cardiovascular screening by eliminatingthe need for expensive ultrasound equipment orspecialised medical personnel,” Jesús Prada Alonsoadds. “The technology integrates seamlessly withexisting healthcare information systems throughAPIs, making implementation straightforward forfacilities of varying sizes.” About 60% of people aged 40-55 have subclinicalatherosclerosis (asymptomatic plaque buildup inthe arteries3), according to a research lead by BorjaIbanez, Scientific Director at CNIC (Spanish NationalCenter for Cardiological Research)4. But 58% ofthese patients are classified as low risk by traditionalrisk calculators and therefore remain undiagnosed.This critical gap in preventive healthcare is due to alack of specialised equipment and trained personnelto perform early cardiovascular screening in primarycare centres, according to Horus ML. Founded in Madrid in October 2022, Horus ML isdeveloping an AI-based cardiovascular screeningsystem, called AItheroscope, which analysesretinal images to detect subclinical atherosclerosis.The solution combines advanced deep learningalgorithms with RetinoLens, a low-cost opticaldevice designed for intuitive use in primary careenvironments. One billion people couldbenefit Moving screening from hospitals to primary care,where the vast majority of people receive theirregular medical attention, would enable the earlydetection in asymptomatic patients classifiedas low risk – that could be a game changer forcardiovascular disease prevention. Horus MLestimates that AItheroscope can address a globalmarket of one billion people who currently have noaccess to effective cardiovascular screening. The system processes retinalimages through convolutionalneural networks and transformermodels, providing cardiovascularrisk assessments in under 10seconds “The technology reduces diagnostic costs, preventsprogression to severe cardiovascular events, anddecreases hospitalisations,” explains Jesús PradaAlonso. “Beyond efficiency gains, the systemaddresses gender bias in cardiovascular diagnosisthrough automated, unbiased screening. Jesús Prada Alonso- Founder and CEO of Horus This democratisation of access to advanceddiagnostic technology is particularly significant forunderserved rural and resource-limited areas wherespecialised cardiovascular care is unavailable.” detect subclinical atherosclerosis. As it engageswith primary care centres and the Sp