Intended for healthcare professionals

Practice Easily Missed?

Lung cancer

BMJ 2014; 349 doi: (Published 06 November 2014) Cite this as: BMJ 2014;349:g6560
  1. Richard D Neal, professor of primary care medicine1,
  2. William Hamilton, professor of primary care diagnostics2,
  3. Trevor K Rogers, consultant respiratory physician3
  1. 1North Wales Centre for Primary Care Research, Bangor University, Wrexham LL13 7YP, UK
  2. 2Department of Primary Care Diagnostics, University of Exeter Medical School, Exeter, UK
  3. 3Doncaster Royal Infirmary, Doncaster and Bassetlaw NHS Foundation Trust, Doncaster, UK
  1. Correspondence to: R D Neal r.neal{at}

The bottom line

  • Suspect lung cancer in all at-risk patients (age, smoking, chronic obstructive pulmonary disease) with a new respiratory symptom, or atypical non-respiratory symptom, even if they appear otherwise well

  • Consider lung cancer in non-smokers with suspicious symptoms, especially haemoptysis and multiple symptoms

  • Chest radiography is cheap, easy, widely available, and relatively harm free, but can be falsely negative. Have a low threshold for repeating or referring for specialist opinion (or considering computed tomography if available) if there are diagnostic suspicions

  • Aim to diagnose patients as quickly as possible to optimise the chance of cure and active anti-cancer treatment

A 65 year old recent ex-smoker with a 20 pack year history and chronic obstructive pulmonary disease (COPD) consults his general practitioner because of worse chest symptoms over the past three weeks, with increasing breathlessness and production of green sputum, and a feeling of being “off colour.” Similar previous spells have been treated with antibiotics and steroids to good effect. On examination his temperature is normal, his heart rate is 88 beats/min, respiratory rate is 16 breaths/min, and oxygen saturation is 94% on air. He has tobacco stained fingers but no clubbing or evidence of weight loss. Scattered crackles and wheezes are found on auscultation of the lungs. After treatment with antibiotics and steroids he feels no better, and after two further consultations he is referred for chest radiography, which is reported as showing a lesion suspicious of lung cancer.

How common is it?

  • In 2011, there were 43 463 new cases of lung cancer in the United Kingdom, making it the second most common cancer and representing 13% of all new cancers1

  • The crude incidence rate is 77 cancer cases per 100 000 men in the UK, and 61 per 100 000 women

  • In most cases there is a history of smoking, and incidence increases with age

  • Patients with chronic obstructive pulmonary disease are about four times more likely to develop lung cancer than those without (1% risk/year)2

  • Lung cancer accounts for 22% of deaths from cancer in the UK. An average GP will see about one new diagnosis of lung cancer each year

Lung cancer

Lung cancer refers to primary tumours of the lung and is initially classified histologically as small cell lung cancer (SCLC) or non-small cell lung cancer (NSCLC) because these two types behave very differently. The main forms of NSCLC are squamous, adenocarcinoma, and neuroendocrine large cell cancers. This subdivision has become more important with the advent of tailored chemotherapy for advanced disease based on the subtype of NSCLC and biological drugs for some adenocarcinomas. These include tyrosine kinase inhibitors in tumours with sensitising mutations in the epidermal growth factor receptor and, even more recently, anaplastic lymphoma kinase inhibitors.

Why is it missed?

Lung cancer is harder to diagnose than most other cancers so the diagnosis is often delayed. There is evidence from recent well conducted cohort studies that:

  • A third of patients consulted their GP about the health problem caused by cancer three or more times before referral (third highest of the 18 cancers reported)3

  • The median diagnostic interval (time from first presentation to diagnosis) is 112 days (interquartile range 45-251)—the second highest of the 15 cancers reported4

  • The median primary care interval (time from first presentation to referral) is 14 days (3-40) (only myeloma is higher) and increases with increasing numbers of pre-referral consultations3

  • Thirty nine per cent of lung cancers present as an emergency (a marker for poor outcomes)—only cancers of the brain and pancreas are higher.5

The diagnosis may be initially missed because of lack of a clear “symptom signature,”6 symptomatic “noise” resulting from COPD and other comorbidities,7 8 chest radiographs being reported as normal or benign findings,9 presentation with non-respiratory or atypical symptoms, or patient mediated factors (such as delay in re-presenting or declining earlier referral).

Why does this matter?

Overall, lung cancer has a poor prognosis. Longer diagnostic intervals are associated with increased mortality,10 with fewer patients being amenable to curative treatments. There may also be morbidity and quality of life benefits from more timely diagnosis, although this has not been proved. Certainly, late diagnosis creates enormous challenges for patients and their families in coming to terms with the diagnosis, planning for their changed circumstances, and resolving their affairs. It is not unusual for patients to present for the first time as a medical emergency and then to die in hospital.5

How is it diagnosed?

Chest radiography is the main investigation that leads to diagnosis. In about 10% of patients subsequently diagnosed with lung cancer, the initial radiograph is reported as normal, and indeterminate abnormalities are found in a further 13%.9 Computed tomography is used largely in specialist practice, either when the radiograph is negative but cancer is still suspected or when the radiograph shows an abnormality. Further invasive tests are then undertaken to provide a tissue diagnosis. Thus, identification in primary care, where most patients initially present,11 generally hinges on identifying features that prompt a chest radiograph or that require specialist investigation even if the radiograph is normal, such as persistent haemoptysis in an older smoker. Thus, we concentrate on features that may prompt a request for chest radiography.

A recent systematic review collated five primary care reports of lung cancer symptoms.6 Positive predictive values (PPVs, which express the risk of cancer numerically) of symptoms were: cough 0.4% in two studies, weight loss 1.1% and 6.1% in two studies, and appetite loss 0.9% and 4.7% in two studies. The classic feature of lung cancer—haemoptysis—had a PPV of 2.4-7.5%. However, only a minority of primary care patients with lung cancer report haemoptysis.11 PPVs increase with age, current smoking, and multiple or persistent symptoms.11 12 13

Three algorithms for the diagnosis of lung cancer have been created: the simplest, a risk assessment tool, offers PPVs for pairs of symptoms or for repeated symptoms, stratified into smokers and non-smokers. In a before and after study, use of the tool was accompanied by increased requests for chest radiography and lung cancer diagnosis, including early stage cancers.14 Q-cancer and a second algorithm are multivariable equations that incorporate data on risk factors as well as symptoms.12 13 Their theoretical performance is good, but no reports of actual performance are available. Risk assessment tools and Q-cancer have been incorporated into all UK based primary care clinical software systems and can be programmed to prompt the GP once a lung cancer risk above an agreed threshold has been estimated.

Patients who present with symptoms recommended by the National Institute for Health and Care Excellence (NICE) for investigation have shorter times to diagnosis compared with those without these symptoms.4 15 This reflects possible alternative diagnoses for the “low risk but not no risk” symptoms, such as cough, plus the atypical presentations, especially with metastases.

How is it managed?

Treatment in the United Kingdom is based on NICE guidance and depends on histological type, disease stage, fitness, performance status, and patient preference.16 Tissue diagnosis is commonly achieved at bronchoscopy, sometimes using endobronchial ultrasound, or imaging guided needle biopsy. In NSCLC, if the tumour appears localised on the diagnostic computed tomogram, staging positron emission tomography-computed tomography is needed to establish eligibility for curative treatment. When the disease is localised and the patient has adequate physiological reserve, surgery is the treatment of choice. If the tumour is unresectable, radiotherapy is an alternative radical option, provided the disease can be encompassed in a treatment volume. Stereotactic ablative radiotherapy is increasingly available as a curative option in patients with peripheral tumours and borderline fitness. Multimodality treatment may increase cure rates and should be considered in fitter patients. When radical treatment is not possible in NSCLC and functional status is good, chemotherapy can improve life expectancy and symptom control. Highly effective biological agents are now available for the minority of adenocarcinomas that carry specific sensitising mutations.

For localised SCLC, chemoradiotherapy can be given with curative intent, and chemotherapy has prognostic and palliative value in the remaining patients, provided they are fit enough. The substantial majority of patients who present with advanced disease of whatever histological type can benefit from palliative treatments, including radiotherapy and specialist palliative care. However, resection rates are improving,17 and there are encouraging signs that early diagnostic initiatives, such as the use of risk assessment tools, may have benefit14—particularly in reducing the number of emergency admissions.


Cite this as: BMJ 2014;349:g6560


  • This is one of a series of occasional articles highlighting conditions that may be more common than many doctors realise or may be missed at first presentation. The series advisers are Anthony Harnden, professor of primary care, Department of Primary Care Health Sciences, University of Oxford, and Richard Lehman, general practitioner, Banbury. To suggest a topic for this series, please email us at practice{at}

  • Contributors: RDN wrote the first draft of most of the paper. WH wrote the first draft of the “How is it diagnosed?” section. TKR wrote the first draft of the “How is it managed?” section. All three authors then revised the paper critically. All authors agree to be accountable for the work. RDN is guarantor.

  • Competing interests: We have read and understood BMJ policy on declaration of interests and declare the following interests: WH is clinical adviser for the current revision of the NICE 2005 guidance, CG27, on the topic of selection of patients for cancer investigation. His contribution to this article is in a personal capacity, and does not represent the view of the Guideline Development Group, or of NICE itself. RDN sits on two funding boards (Roy Castle Lung Cancer Foundation and the RCGP scientific foundation board) that may fund research into the early diagnosis of lung cancer. RDN and WH hold grants from several bodies (including the National Institute for Health Research, Department of Health, Cancer Research UK, Tenovus, Cancer Research Wales, Betsi Cadwaladr University Health Board, Public Health Wales, and National Institute for Social Care and Health Research) to undertake research into the early diagnosis of cancer. WH has received funding from charities including Macmillan to give lectures and attend conferences.

  • Provenance and peer review: Commissioned; externally peer reviewed.

  • Patient consent not required (patient anonymised, dead, or hypothetical).


View Abstract