Cancer specialists in the UAE are using artificial intelligence to understand more about why treatment works for some but fails to help others.
Two patients can have the same type and stage of cancer, and receive the same treatment, but experience markedly different outcomes.
Dubai-based Nexomic, which has been established in part by one of the UAE’s leading oncologists, aims to harness next-generation AI to cancer treatment.
Prof Humaid Al Shamsi, who leads cancer care at Burjeel Medical City and is chief executive of the Burjeel Cancer Institute, is exploring new methods that could transform patient care and boost survival rates.
Among women, breast cancer is the most common form accounting for 94 cases per 100,000 people, while colorectal, prostate and lung cancers are the most commonly diagnosed in men.
Some patients can respond well for years, while others deteriorate within months, despite undergoing the same treatment.
Challenges ahead
Understanding why remains one of the biggest challenges in cancer medicine, Prof Al Shamsi said.
“Cancer is much more complicated than a single mutation,” he said. “A patient may have the same diagnosis, the same stage and even the same recognised biomarkers as another patient, but their cancers can behave very differently. The question we are trying to answer is: what is happening underneath that explains the difference?”
Nexomic is developing its own AI that combines numerous layers of tumour biology to identify patterns linked to treatment response, resistance and progression of disease.
Rather than looking only for individual mutations such as HER2, KRAS or BRCA, the company’s DeepOmic platform analyses several types of molecular information together, including gene activity and DNA alterations.
Deeper knowledge
The approach, known as multi-omics, is designed to provide a more complete picture of how a tumour is behaving.
By understanding where a cancer originates and its biological state, doctors hope that those deeper molecular patterns can be translated into clinically useful information for researchers, pharmaceutical companies and oncologists.
Also working on the programme is Dr Ghadeer Ghosheh, who is Nexomic’s chief AI officer.
She brings a background in computer engineering and computational health informatics, rooted in academic training at New York University Abu Dhabi and the University of Oxford.
“Cancer is not a single data point,” Dr Ghosheh said. “It is a complex biological system and each layer of information tells us something different. Our goal with DeepOmic is to bring those layers together so that we can understand the biology more completely.
“The real opportunity is to move beyond identifying individual abnormalities and begin recognising the biological patterns that explain how a cancer behaves and responds to treatment.”
AI is already being used around the world to help doctors diagnose cancer earlier and deliver life-saving treatment before tumours develop further.
It can be used to analyse medical images, digital tissue slides and patient health records to identify tiny patterns that otherwise could be overlooked.
Algorithms can spot subtle signs of breast or lung cancer years before symptoms appear, while pairing a radiologist with AI reduces false positives and false negatives, increasing overall detection accuracy.
Precision cancer medicine has traditionally relied heavily on identifying particular mutations or proteins that can be targeted by medicine.
That approach has transformed treatment for many patients but a large proportion of cancers either lack an obvious target or eventually develop resistance.
Nexomic is attempting to move beyond this one-marker-at-a-time model to learn larger patterns from huge volumes of data.
Data access
Access to that health data is critical to the programme’s success, as a major challenge facing any medical AI system is proving it works outside of the data sets it has been designed for.
To resolve that, Nexomic has entered a formal data-access agreement in the UK with Genomics England and Cancer Research Horizons.
It will allow the company to evaluate and improve computational tools and research using highly governed genomic and clinical information.
Dr Mohammad Ghosheh is Nexomic’s chief executive and leads the development of DeepOmic for pharmaceutical research.
“Medicine has made extraordinary progress, but many treatment decisions are still based on what works for the average patient,” he said. “We are building Nexomic to help close the gap between what we know about cancer biology and what we can use in practice.
“That means developing technology that can support better treatment decisions, help pharmaceutical companies understand why therapies succeed or fail and ultimately bring more personalised medicine closer to patients.”
