Cancer Research · Active
AI-Assisted Cancer Triage in Low-Resource Clinics
Late-stage presentation dominates cancer outcomes across the region because pathology capacity is concentrated in a handful of urban centres.
Principal investigator: Dr. Tendai Mahaso
Programme
40+
Active programmes across clinical, agri and environmental science
60+
Partner institutions, hospitals and grower networks
180
Peer-reviewed publications and field reports to date
Objectives
- Develop an image and biomarker triage model for district clinics
- Validate against 12,000 histopathology-confirmed cases
- Reduce median time-to-diagnosis by 40%
Methodology
- 1Retrospective cohort assembly across 9 partner hospitals
- 2Multimodal model training on de-identified imaging and laboratory data
- 3Prospective clinical validation with pathologist adjudication
Partners
- University Teaching Hospital
- National Cancer Registry
- Ministry of Health
Publications
Multimodal triage of suspected malignancy in district clinics
Lancet Digital Health · 2025
Connected
Laboratories and services behind this study
Genomics Laboratory
Sequencing, genotyping and bioinformatics
ExploreMolecular Biology Laboratory
Real-time PCR, amplification and nucleic acid workflows
ExploreCancer Genomic Profiling
Targeted tumour sequencing with therapy-matching interpretation.
ExploreResearch lifecycle
From a question worth asking to a result people can use
01
Scoping
Define the decision, the endpoint and the population or field site with our partners.
02
Protocol & ethics
Pre-registered protocol, ethics or biosafety approval and consent frameworks.
03
Execution
Accredited laboratory and field workflows with monitored data capture.
04
Translation
Peer-reviewed publication, regulatory dossiers and hand-off to service teams.
Partner with us
Collaborate on AI-Assisted Cancer Triage in Low-Resource Clinics
Bring us a study, a sample set or a field problem. We scope feasibility, ethics and timelines within one working week.
Publications & data access
Request pre-prints, anonymised datasets or replication protocols from any completed programme.