Medprime Technologies

Problem: India was home to more than 1.6 lakh malaria cases in 2021. While malaria-related deaths have been dropping recently, the case numbers continue to be high. Evidence shows that conventional microscopy-based detection misses about 25% malaria cases in India; this problem is further exacerbated in hard-to-reach remote areas. Analog microscopy as a diagnostic tool for malaria has evolved very little since its first use. Even though the basic principles of optics remain constant, the subjectivity that comes with dependence on skills and training of officials who use the microscopes can result in wide variation in disease interpretation. These limitations in traditional microscopy-based detection result in underdiagnosis and poor surveillance of malaria and other diseases.

Solution: IHF engaged with Medprime Technologies for faster, more cost-effective and more accurate microscopy for disease diagnosis. Medprime’s artificial intelligence algorithm, locally trained by machine learning, automatically detects, identifies and differentiates between malarial parasites (P. falciparum and P. vivax). This tool will be particularly useful in identifying infections with low-parasite loads, which are commonly missed by conventional microscopy.

Impact: The algorithm is microscope, disease (communicable and non-communicable) and sample (blood, urine, stool, pus) agnostic and will slash- by third- the time taken for slide viewing, image processing and result reporting. The algorithm will also overcome the need for trained manpower and will prove important for training and research for medical, paramedical and allied health staff that need remote collaboration and interoperability. IHF’s support to Medprime will enable the development and pilot testing of this AI/Machine Learning (ML)-led multiplex diagnosis software. The platform algorithm can be used beyond malaria in WBC differentiation, histopathology and cytopathology for cancer, sickle cell disease etc. and can run samples like blood, urine, stool, pus.