Healthcare as a Service: Bringing AI Lung Cancer Diagnosis to Anyone, Anywhere

By Nuruzzaman Faruqui

·

You can stream films and store photos in the cloud. Why not get a lung cancer check the same way? Healthcare as a Service (HAAS) turns an accurate AI diagnosis model into an online service that any clinic connected to the internet could use.

96.07%accuracy of the HAASNet AI model
96.47%average precision of diagnosis reports
Globalaccess via the cloud

The problem

Research labs have built many accurate AI models for spotting lung cancer in CT scans. Very few of them reach patients. The gap is usability: a model sitting on a researcher’s computer helps nobody in a rural hospital.

Early diagnosis saves lives in lung cancer, but early diagnosis depends on access to specialists and equipment that many communities do not have.

What the research did

We borrowed an idea from the software world, Software as a Service, where you use powerful programs through a web browser instead of installing them. We designed Healthcare as a Service: a complete lung cancer diagnosis system that lives in the cloud.

At its heart is HAASNet, a convolutional neural network built to run efficiently on cloud servers. It analyses CT scans, then combines its prediction with physical symptom data from Internet of Medical Things devices to produce a diagnosis report.

What we found

HAASNet classified CT scans with 96.07% accuracy. The full system produced diagnosis reports with an average precision of 96.47%, recall of 95.39% and F1-score of 94.81%.

Because it runs in the cloud, the service is not limited by geography. A clinic only needs an internet connection to send data and receive a report.

Who benefits

  • Communities far from specialist hospitals, who could get earlier screening
  • Health systems in developing countries, which can share one central service instead of buying expensive equipment everywhere
  • Doctors, who receive a structured report that combines imaging and symptoms

Limits and what comes next

HAAS is a research model and has not been deployed with real patients. A real service would need strong data privacy protections, clinical validation, regulatory approval and reliable internet access, which remains a challenge in many regions. It builds on our earlier LungNet research.

The team

I led this work with Mohammad Abu Yousuf (Jahangirnagar University) and Faris Kateb, Md. Abdul Hamid and Muhammad Mostafa Monowar (King Abdulaziz University, Saudi Arabia).

Frequently asked questions

What is Healthcare as a Service?

It is a model where a medical tool, here an AI that diagnoses lung cancer, is offered online through the cloud, so clinics can use it without owning special hardware.

How accurate is HAAS?

Its core AI model, HAASNet, was 96.07% accurate on CT scans, and the full system reached 96.47% average precision.

Is my medical data safe in a cloud service like this?

Any real deployment would need strict privacy and security controls. That is one reason my research also focuses on medical cybersecurity.

About the author

Nuruzzaman Faruqui is an award-winning AI educator and researcher who has taught artificial intelligence and machine learning at university level since 2018. A two-time winner of the Excellence in Teaching Award (2023, 2024) and winner of the Best Researcher Award (2024) at Daffodil International University, he is now a doctoral researcher in AI at Monash University, Australia. He has published 53 papers and six books, and writes these articles to make AI research easy for everyone to understand. His teaching · About Zaman · Get in touch

Keep reading