New tools to spot people at risk of lung and stomach cancers

By Lilian Anekwe

BMJ Group News

What do we know already?

Lung and stomach cancers can be difficult to diagnose because they both have symptoms that can also occur in people who don’t have these diseases. Often doctors can’t be sure without sending people for tests, which can be distressing.

But treating lung and stomach cancer is more successful when the tumours are picked up early. So it would be useful to have ways to weigh up the most common symptoms and calculate people’s risk of having these diseases.

So researchers looked at records of patients aged between 30 and 84 from 564 GP practices in England and Wales. They looked at some of the most common symptoms and factors that affect people’s risk of lung and stomach cancer (risk factors), then compared how often those symptoms occurred in people who had been diagnosed with either stomach or lung cancer with people who were not diagnosed with cancer.

They calculated what the risk was for each symptom, and then by combining the symptoms people had, worked out what their risk was of either having or getting lung or stomach cancer in the following two years.

What does the new study say?

Overall there were 5,981 lung cancers and 3,870 stomach cancers (either gastric or oesophageal cancers) diagnosed during the two year study.

The symptoms and risk factors most strongly linked to stomach cancer were:

• Difficulty swallowing
• Vomiting blood
• Loss of appetite
• Stomach aches or pain
• Weight loss
• Being former or current smokers
• Lack of iron in the blood (anaemia)

For lung cancer, the most common symptoms and risk factors were:

• Coughing up blood
• Being former or current smokers
• Weight loss
• Loss of appetite
• A cough within the previous 12 months
• Lack of iron in the blood (anaemia)

The researchers calculated that people who had more than a 0.2 percent risk of getting a stomach cancer within two years were considered at highest risk. There were 1,028 people diagnosed with stomach cancer in this group. Those who had a score of more than 0.4 percent risk of getting lung cancer within two years were at the highest risk. There were 1,697 people diagnosed with lung cancer in this group.

That means that, for every 100 people in the high risk groups identified by these calculators, around one of them would have either stomach or lung cancer within two years.

How reliable is the research?

The estimates of risk were calculated using a large, reliable set of data, and the researchers also checked against another big group of people to make sure it was accurate.

But we don’t know how much information was missing from the patients’ records, and also not all patients with symptoms of lung or stomach cancer will go and see their GP. So it’s possible that, if some of the symptoms were left out, these estimates of risk have been underestimated.

Also, this study doesn’t tell us whether using symptom-based tools is likely to lead to earlier identification of lung or stomach cancer, at a stage when treatment or surgery is more likely to be possible. We need a trial comparing these tools in general practices against usual practice to help answer this question.