AI in Radiology in Africa and the Middle East: The Evidence

Few medical fields have attracted as much artificial intelligence research as radiology, and few regions need extra reading capacity as badly as parts of Africa. The idea is appealing: software that reads images where radiologists are scarce. The reality is more specific. AI is well established for a few narrow tasks, especially tuberculosis screening, and much less proven elsewhere. Deployment also depends on infrastructure, regulation and data governance as much as on algorithms.

The workforce gap

Radiologist numbers vary enormously across the continent. At a 2019 meeting in Cairo reported by the European Society of Radiology, African radiologists noted that the radiologist ratio ranges from about 1 to 80 per million population, depending on the country. In Ghana, a 2025 article citing 2019 data reported fewer than two radiologists per million people, concentrated in major cities.

In such settings, imaging equipment may exist without anyone available to report the studies quickly. Delayed or missing reports are a real clinical problem. This is the gap AI promoters point to, and in some tasks the case is strong.

Where the evidence is strongest: tuberculosis

The clearest example of AI entering routine practice is computer-aided detection (CAD) for tuberculosis on chest X-rays. In 2021, the World Health Organization (WHO) issued a conditional recommendation that CAD software may be used in place of human readers to interpret digital chest X-rays for TB screening and triage, in people aged 15 and older. WHO rated the certainty of the evidence as low, but the recommendation was a milestone. For the first time, a global health body accepted AI as a substitute for human reading in a defined use.

The reasoning was practical. Many high-burden countries do not have enough qualified readers for large screening programs. CAD tools are now used in TB screening programs in high-burden countries, including mobile and community screening. Programs still need to:

• Choose a threshold score suited to the local population and the program's goals.
• Confirm positive screens with bacteriological testing.
• Re-evaluate each new software version, since products are updated frequently.

Beyond TB: a more cautious picture

For other uses, such as fracture detection, stroke CT, mammography and general chest X-ray triage, most published validation studies come from Europe, North America and East Asia. Performance can drop when an algorithm meets different equipment, image quality, patient populations and disease patterns. Before relying on any tool, ask:

• Was it validated on data from a population similar to yours?
• Was it tested on the X-ray or CT machines you actually use, including older or portable units?
• Does it hold the regulatory clearance your country requires?
• What happens to the images: processed locally or in the cloud?

The Middle East: infrastructure ready, governance tightening

Many hospitals in the Gulf already run digital PACS and RIS systems that can integrate AI tools. For them, the main constraints are often governance and data location rather than equipment:

• In the UAE, Federal Law No. 2 of 2019 on the use of ICT in health fields restricts transferring health data related to services provided in the UAE outside the country, except in cases defined by the health authorities. Cloud-based AI tools that send images abroad must be assessed against this rule.
• In Saudi Arabia, the Personal Data Protection Law has been fully enforceable since September 2024, with additional controls for health data.

For hospitals, this favors AI tools that run on local servers or in national cloud regions, and contracts that specify exactly where images are processed.

Practical conditions for success

Experience from TB programs and early hospital deployments points to a few conditions:

1. A clear clinical question. "Triage chest X-rays for TB" works better than "AI for radiology".
2. Local validation. Run the tool silently on a few hundred local studies and compare it with expert reads before using it clinically.
3. Defined responsibility. Decide who reviews AI flags, how quickly and who signs the report.
4. Reliable infrastructure. Power, network, PACS integration and maintenance often matter more than the algorithm.
5. Monitoring. Track performance over time and after every software update.

AI can extend radiology capacity, but it does not replace the need to train and retain radiologists, radiographers and reporting staff. In most settings, the realistic goal is faster triage and more consistent reports, not autonomous reading.

Key takeaways

• Radiologist density in Africa ranges widely, with some countries reporting fewer than two per million people.
• WHO has endorsed computer-aided detection for TB screening on chest X-rays since 2021, the best-established AI use in the region.
• For other tasks, check whether the tool was validated on populations and equipment similar to yours.
• In the Gulf, data localization and personal data laws shape which AI deployments are allowed.
• Start with a narrow question, validate locally, and define who is responsible for the final report.

Frequently asked questions

Does WHO recommend AI for reading chest X-rays?

For tuberculosis screening and triage, yes. Since 2021, WHO conditionally recommends that computer-aided detection may replace human readers for digital chest X-rays in people aged 15 and older.

Can AI replace radiologists in Africa?

Not in general. AI is well established for narrow tasks such as TB screening, but most other uses still need radiologist oversight and local validation.

Can hospitals in the UAE use cloud AI for imaging?

Only after legal review. Federal Law No. 2 of 2019 restricts transferring health data outside the UAE except in cases defined by the health authorities.

Sources

1. European Society of Radiology blog — African radiologists call for more cooperation with the ESR in radiation protection
2. Citi Newsroom — Solving Ghana's and Africa's radiology crisis: AI as an engine for equitable care (2025)
3. PATH — How computers can help screen for TB
4. Al Tamimi & Co — The Federal Law regulating the use of ICT in the UAE healthcare sector
5. Clyde & Co — Saudi Arabia's Personal Data Protection Law becomes enforceable (2024)

Dictate your reports, nothing leaves your computer

Nabady Whisper transcribes your voice offline in English, French or Arabic, with report templates for every specialty.

Try it free → Watch the demo

General information, checked at the publication date; it is neither medical nor legal advice.