Intended for healthcare professionals

Practice Rational Testing

Investigating suspected lung cancer

BMJ 2022; 378 doi: (Published 13 July 2022) Cite this as: BMJ 2022;378:e068384
  1. Stephen H Bradley, GP and NIHR academic clinical lecturer1,
  2. Bobby SK Bhartia, consultant thoracic radiologist2,
  3. Martyn PT Kennedy, consultant respiratory physician2,
  4. Lesleigh Kowalski Frank, research scientist3,
  5. Jessica Watson, GP and NIHR academic clinical lecturer4
  1. 1University of Leeds, Leeds, UK
  2. 2Leeds Teaching Hospitals NHS Trust, Leeds
  3. 3University of Washington, Seattle, USA
  4. 4University of Bristol, Bristol, UK
  1. Correspondence to S Bradley medsbra{at}

What you need to know

  • Up to a quarter of lung cancers are diagnosed in people who have never smoked

  • Consider chest radiography and urgent specialist referral or advice for unusual presentations that are not explicitly listed in local referral pathways

  • Unexplained haemoptysis warrants urgent referral (within 14 days) for computed tomography imaging

  • Inform patients when you anticipate that negative radiography and/or blood tests might provide insufficient reassurance to exclude lung cancer/other serious illness and advise that further investigations may be needed

A 49 year old woman who has never smoked has a telephone follow-up appointment with a GP. She originally consulted one month ago with tiredness and non-exertional chest pain. Her phone call today is to receive the results of an electrocardiogram (ECG) and blood tests, which were organised following her initial presentation. She reports that her symptoms have not resolved. Her ECG and blood tests are unremarkable aside from an elevated platelet count of 482 × 109/L. (reference range 150-400 × 109/L).

Globally, an estimated 10-25% of lung cancer cases occur in people who have never smoked; additionally, if lung cancer in non-smokers were considered a separate disease, estimates suggest that it would be the seventh most prevalent cancer worldwide.1 Nonetheless, increasing age and tobacco exposure are still the most important risk factors,2 and lung cancer (irrespective of smoking status) is the leading cause of cancer death worldwide.3

Outcome is closely correlated to stage at diagnosis,4 but the high prevalence of common lung cancer symptoms (such as cough) presenting in primary care can make timely diagnosis difficult.5

How is risk assigned?

Symptoms that indicate risk of lung cancer are typically presented as positive predictive values (PPVs), which represent the risk that a patient with symptoms, or a collection of symptoms, has lung cancer (fig 1). In this article …

View Full Text

Log in

Log in through your institution


* For online subscription