Digital Diagnostics at the Point-of-Care

Medical diagnostics is undergoing a transition where an increasing number of processes can be supported by digitalization and artificial intelligence (AI) at the point-of-care (POC). Our research group has developed a diagnostic system which includes obtaining a sample, digitizing the sample at the POC with a mobile microscope scanner, image transfer over mobile networks, AI-analysis, and verification by a remote expert and feed-back of the results back to the POC for decision support. We have conducted proof-of-concept studies regarding the novel method that combines AI and mobile digital microscopy for cell-based cervical cancer screening in resource-limited settings in East Africa. Next, we will assess the usefulness of the diagnostic method in the form of three validation studies in Tanzania and Kenya: A) for screening of gynecological cell samples with the aim of detecting precancers for the prevention of cervical cancer, B) as a monitoring method for drug efficacy within control programs for soil-transmitted helminth infections in school children, and C) for malaria diagnostics.

Project descriptions

Ongoing projects

In the following years clinical studies regarding cancer and infectious diseases will be carried out and within these the diagnostic accuracy is evaluated in comparison with conventional diagnostics, as well as time consumption, technical feasibility, cost per test, and acceptance of the method in a point-of-care setting.

1. Validation study on screening for cervical cancer.

We currently scale up the use of the new diagnostic method in the form of a validation study in Kenya and Tanzania with the aim of detecting cervical atypia for the purpose of cervical cancer screening/prevention. During 2022-2023 a total of 1,500 patient´s cervical smears will be collected and digitized for the detection of premalignant cells using a deep learning-based AI algorithm. In 2023-24 the method's diagnostic accuracy, technical feasibility, cost and time per test, and acceptance of the AI method is evaluated and compared to conventional diagnostics. This study is initiated in 2022 and completed 2023.

2. A field trial on fever diagnostics at the point-of-care.

In this trial, the diagnostic performance of the AI-based method developed to detect malaria (P. falciparum) is compared to diagnostics using conventional microscopy and rapid diagnostic tests. Patients (n=800) attending a primary health care facility in Yombo village in Bagamoyo District, Tanzania and at Kinondo Hospital, Kenya will be recruited. This study is initiated in 2023 and completed 2024.

3. A surveillance trial on neglected tropical diseases in school children within a control program.

We will evaluate the diagnostic accuracy and cost effectiveness of the method as compared with conventional microscopy for Neglected Tropical Diseases control (soil-transmitted helminths and Schistosoma mansoni) after wide scale use of mass drug administration (Albendazole) in a school based anti-helminthic control program in Kwale County, Kenya and in the Mwanza Region, Tanzania. Stool samples will be collected from children attending primary schools in the region (n =1,500 children). This study will be conducted in 2022-2024.

Collaborators

Nina Linder, UU and University of Helsinki
Andreas Mårtensson, UU
Johan Lundin, KI, and University of Helsinki
Billy Ngasala, MUHAS (Muhimbili University of Health and Allied Sciences) and UU
Lwidiko E. Mhamilawa, MUHAS and UU
Mary Inziani Matilu, KEMRI (Kenya Medical Research Institute)

Funding
Swedish Research Council, The Erling-Persson Foundation, InDevelop

Responsible researcher(s)/contact persons
Nina Linder, nina.linder@kbh.uu.se
Andreas Mårtensson, andreas.martensson@kbh.uu.se