Family doctors need more help in recognising patients at a high risk of lung cancer, new research has found.

One in three lung cancer sufferers die within 90 days of diagnosis, suggesting GPs are struggling to spot the early signs of the disease.

Respiratory physician Dr Emma O'Dowd who carried out the research at Britain's Nottingham University, said family doctors needed more help in recognising patients at a high risk of lung cancer.

The research found that of those who die within three months of discovering the cancer, one in 10 die within a month while one in 20 are not diagnosed until they have died.

Dr O'Dowd said "surprisingly" her research found that those who are given a late diagnosis on average visited their their GP five times in the few months beforehand, countering a common preconception that they would not seek medical attention.

It was therefore "key" to find out what symptoms they were displaying in those consultations and how they could be better identified by GPs as being linked to lung cancer.

"We're losing a lot of patients early on. I wanted to find out more about these patients who died early and if there are features that can help us to diagnose them earlier," Dr O'Dowd said.

"I started off with the preconception that people who died early didn't ever see their GP. Actually, they saw their GPs more before diagnosis compared to those patients which lived longer.

"Lung cancer can be difficult for doctors to distinguish from other lung diseases so we need to give them some tools that will help identify a patient as high risk," she said.

"If we can diagnose patients at an earlier stage hopefully they can get curable treatment rather than palliative treatment which is what most patients are getting at the moment."

Dr O'Dowd, whose research has been published in the British Medical Journal, said most GPs are only likely to see one new case of lung cancer a year which is why it is important to promote risk assessment tools.

New software is being developed for doctors that would flag up the risk of lung cancer in patients by tracking their symptoms and lifestyle.

Copyright AAP 2014