How AI Can Help Address The Global Shortage of Radiologists

Share on linkedin
Share on facebook
Share on reddit
Share on pinterest

There’s a global shortage of radiologists. Over 2/3 of people on earth do not have access to radiologists. There are big disparities between countries and within countries. Some countries like the US have tens of thousands of radiologists whereas 14 African countries have no radiologists at all. In India there is approximately one radiologist for every 100,000 people whereas in the US there is one radiologist for every 10,000 people.

There are also disparities within countries. For example in São Paulo, Brazil there are 10 radiologists per every 100,000 inhabitants but in northern Brazil the ratio is 3:100,000. Although there are about 37,000 radiologists in the US, they are not evenly distributed. California, Texas, Florida, Massachusetts, and New York have thousands of radiologists, some states have just 40. One hospital in Boston, Massachusetts General Hospital, has 126 radiologists.

Access to trained radiologists is also limited for millions of people who live on 11,000 islands around the world. Most of these islands are located in remote places without hospitals, imaging equipment, or doctors. Some islands with larger populations have radiographers or medical radiation technologists who are trained to operate scanning devices but they are not trained to interpret results or make a diagnosis.

This map shows the substantial differences in the number of trained radiologists between countries. Source: Medical imaging and nuclear medicine: a Lancet Oncology Commission

This map shows how radiologists are dispersed across the US. Source: Harvey L. Neiman Health Policy Institute

How AI Can Help Bridge the Gap

AI can help bridge the gap between remote communities and state of the art medical centers. Automated screening with AI can help radiologists to triage patients by flagging abnormal medical images. AI embedded in mobile devices can help screen people who live in rural areas without hospitals. These devices are designed to work in places without electricity and without Internet access.

Using AI can help radiologists improve workflow and efficiency especially in publicly-funded healthcare systems. In the UK and Canada hundreds of thousands of people wait over 30 days to have medical images read by a radiologist. Using AI can help reduce delays in identifying and acting on abnormal medical images. This is especially important in chest and brain imaging where time is critical. 40% of all diagnostic imaging performed worldwide are images of the chest.

Examples of AI applications for medical imaging

  • In May 2022 the FDA cleared the world’s first AI-driven portable and automated 3D breast ultrasound scanner. In just 2 minutes, the system automatically scans the entire breast volume and offers 3D visualization of the breast tissue. The ATUSA developed by iSono Health system does not require a radiologist or ultrasound technologist for image acquisition, however, interpretation of images requires a physician with training in breast ultrasound.
  • In a recent study, Gleamer BoneView AI software helped hospitals reduce missed fractures by 29% and increased sensitivity by 30% for exams with more than one fracture.
  • Studies at the AI Precision Health Institute at the University of Hawaiʻi Cancer Center have demonstrated that AI can predict which women will develop breast cancer and those who will not.
  • Audace Nakeshimana, a Rwandan scientist who graduated from MIT founded Insightiv to make radiology services remotely accessible to Rwandans using AI. Nkeshimana and his engineering team built a system that is compatible with CT, MRI, and X-Ray, and has capabilities such as workflow management, reporting, and is able to interoperate with third-party medical imaging solutions including advanced medical imaging viewers. The system is undergoing the regulatory process in Rwanda and has not been used commercially yet, however, it has been demo in private hospitals and public healthcare institutions and there has been good interest in it.
  • Niramai, founded by Geetha Manjunath, PhD, has developed a low cost portable device to detect early stage breast cancer using machine learning. This portable device can be used in places without electricity and without internet access and can even screen women for breast cancer in the privacy of their own homes.
  • Dr CADx is a startup in Zimbabwe that has developed an AI powered diagnostics system to help doctors diagnose medical images more accurately, at low cost, and remotely. The startup has developed a prototype that can analyze chest X-rays for TB and 14 other chest pathologies, including pneumonia lung cancer nodules and pneumothorax. A pilot validation study for TB showed a standalone accuracy of 96% the gold standard. A test on the other 14 chest findings yielded an average accuracy of 84%, which is comparable to an expert radiologist.
  • AI scientists at the AI Precision Health Institute at the University of Hawai‘i Cancer Center in collaboration with the University of Guam Cancer Research Center are using AI to bridge the gap between developing regions and state of the art medical centers in the Pacific.
  • AI scientists at the AI Precision Health Institute at the University of Hawai‘i Cancer Center are collaborating with doctors in underserved communities and enrolling underrepresented populations in clinical studies.
  • The Digital Health Innovation Research Group (mDHIRG) is a multidisciplinary research team based at Muhimbili University of Health and Allied Sciences in Dar es Salaam, Tanzania. The mDHIRG is developing innovative digital health technology to provide equitable access to healthcare services in line with the Tanzania Development Vision 2025.
  • Aidoc’s AI based decision support software analyzes medical imaging data, flags findings, and alerts radiologists, surgeons, and neurologists of suspected positive cases.
  • In a recent study by scientists at Université Lorraine in France and Tecnológico de Monterrey in Mexico in collaboration with the Hub de IA del Tec de Monterrey in Mexico, scientists demonstrated that AI helped identify hard to detect endoscopic kidney stones with high accuracy.
  • Jameel Clinic at MIT in collaboration with Mass General Hospital is developing new AI models that can predict the risk of lung cancer up to 6 years in advance from CT imaging.
  • Researchers in Germany demonstrated that AI can detect breast cancer signs that radiologists can’t detect.
  • ChestLink AI reviewed 500,000 X-rays across multiple locations with 99% sensitivity and zero clinically relevant errors.
  • Scientists at University of Copenhagen helped radiologists reduce workload by 3/5 using AI to help screen for breast cancer.
  • In India Synapsica’s AI software helps radiologists automate numerous tasks. The software can analyze MRIs in less than a minute and serves over 1 million people.
  • Scientists at University of Michigan Medical School used AI to help doctors assess the effectiveness of bladder cancer treatment.
  • Denmark provides free healthcare services to all citizens. The Danish Council of Radiographers is studying how AI can be used to improve patient care in Denmark.
  • RAD-AID International has more than 12,000 volunteers who provide medical imaging services to more than 80 hospitals in 38 countries where people do not have access to radiologists. RAD-AID manages a rapidly growing portfolio of AI tools (thanks to generous partners contributing software) for helping low-resource hospitals to learn and potentially use AI for increasing and improving clinical patient care.

Researching the Global Supply of Radiologists

There is no global registry of radiologists and it’s difficult to determine the exact number of radiologists in each country because there is limited up-to-date information that is publicly available. For the past 5 years I have been researching the radiologist supply worldwide to get a sense of the disparities to see where AI might be the most useful in improving access to care.

The have compiled a list showing the number of registered radiologists in every country around the world. The list is arranged in descending order starting with China, the country with the highest population, to Tokelau, a group of atolls in the South Pacific Ocean, halfway between Hawaii and New Zealand, with 1,500 people. The list shows the estimated number of radiologists in each country next to the population of each country.

0 replies on “How AI Can Help Address The Global Shortage of Radiologists”

Related Post