Non-small-cell lung cancer

Why the numbers are moving in opposite directions. In the United States lung cancer incidence is falling 1.9 percent a year and mortality 4.1 percent a year, the mortality falling faster than the incidence because more cancers are found early and because targeted drugs and immunotherapy work. In the UK incidence is down 11 percent since the early 1990s and mortality down 40 percent since the early 1970s, but the projection to 2038 to 2040 is for around 66,200 cases a year: rates fall while the population ages, so the count rises. What does not move is the share found late. In England 34 percent of staged cases in 2022 were stage I or II; in the United States 24 percent are localised at diagnosis, 21 percent regional and 51 percent already distant (Cancer Research UK; SEER).

Smoking, and how much of the disease it explains. This is an account of why the UK has as many lung cancers as it does, not of why any one person has one, and it changes nothing about the treatment on this page. Cancer Research UK attributes 72 percent of UK lung cancer cases to tobacco, 71 percent to active smoking and 1 percent to second-hand smoke, and about 86 percent of UK lung cancer deaths to smoking. Seventy-nine percent of UK cases are judged preventable in total. Stopping works and works better the earlier it happens: of British men who smoke all their lives 15.9 percent die of lung cancer by age 75, against 9.9 percent of those who stop by 60, 6 percent by 50, 3 percent by 40 and 1.7 percent by 30; in women the figures are 9.5 percent for lifelong smokers, 5.3 percent for stopping by 60 and 2.2 percent by 50. Risk in people who stopped about 7 years ago is 43 percent lower than in current smokers, and 72 percent lower about 12 years after stopping. Starting younger is worse: lung cancer death risk is 37 percent higher per five years younger at starting (Cancer Research UK, quoting the studies behind each figure).

The other causes, with the share of UK cases each accounts for. Workplace exposures account for 13 percent of UK cases, and asbestos alone is linked to 6 to 8 percent of UK lung cancer deaths each year, with lung cancer mortality 77 percent higher in asbestos-exposed workers than in the general population. Outdoor air pollution accounts for 8 percent of UK cases; adenocarcinoma risk is 40 percent higher per 10 micrograms per cubic metre of fine particulate matter, and overall lung cancer risk 9 percent higher, with no significant association for the coarser PM10 fraction. Ionising radiation accounts for 5 percent, of which indoor radon is the largest part: an estimated 9 percent of European lung cancer deaths are linked to indoor radon, and risk rises 16 percent per 100 becquerels per cubic metre of usual home radon. Second-hand smoke accounts for 1 percent of all cases but an estimated 15 percent of cases in people who have never smoked, raising their risk by 31 percent. Family history matters independently of smoking: risk is 82 percent higher with an affected sibling and 25 to 37 percent higher with an affected parent. Previous lung disease matters too, with risk 43 to 57 percent higher after pneumonia and 104 to 144 percent higher after emphysema. The smaller occupational exposures each account for a fraction of a percent of British cases: silica 0.02 percent, diesel engine exhaust 0.02 percent, painting 0.01 percent, welding 0.001 percent (Cancer Research UK).

Lung cancer in people who have never smoked. Fifteen percent of lung cancers diagnosed in the UK arise in never-smokers, which would make it the 8th commonest cancer in the country if it were counted as a disease of its own (BJC Reports 2023); worldwide the figure is put at 15 to 25 percent depending on the region, and if counted separately it would be the 7th commonest cause of cancer death (OLIVE protocol, BMJ Open Respiratory Research 2026; Health Science Reports 2026). It is a different disease: overwhelmingly adenocarcinoma, commoner in women and in east Asian populations, and far more often driven by a targetable mutation. The UK's largest prospective look at it, the Million Women Study, followed 634,039 never-smoking women for 14 years and found 1,469 lung cancers, 0.2 percent of the cohort; of 34 candidate risk factors only three reached significance, non-white ethnicity (relative risk 2.34), asthma requiring treatment (1.32) and height of 165 cm or more (1.16), and restricting the analysis to adenocarcinoma did not change the result. The practical consequence is that the risk factors that select people for screening do not select these patients, and the assumption that a non-smoker's cough is not cancer delays their diagnosis.

How it presents, and when the NHS refers. The NHS lists a cough that is not going away, shortness of breath, coughing up blood, chest or shoulder pain, repeated chest infections or one that will not clear, losing weight without trying, loss of appetite and constant tiredness as the common symptoms, and swelling of the face or neck, difficulty swallowing, a hoarse voice that does not go away and finger clubbing as the less common ones; early lung cancer often causes nothing at all. NICE NG12 turns that into three rules for general practice. Refer on the suspected cancer pathway if the chest X-ray suggests lung cancer, or if the person is 40 or over with unexplained coughing of blood (1.1.1). Offer an urgent direct access chest X-ray at 40 and over for two or more of cough, fatigue, shortness of breath, chest pain, weight loss and appetite loss, or for one of them in anyone who has ever smoked (1.1.2). Consider an urgent chest X-ray at 40 and over for persistent or recurrent chest infection, finger clubbing, supraclavicular or persistent cervical lymphadenopathy, chest signs consistent with lung cancer, or a raised platelet count (1.1.3). Mesothelioma has its own rules in the same section, with asbestos exposure lowering the threshold. In England in 2019 about 26 percent of lung cancers were diagnosed after an urgent suspected cancer referral and 33 percent after an emergency presentation (Cancer Research UK).

Screening: the evidence, the programme and the harms. The evidence is two randomised trials. The National Lung Screening Trial gave 53,454 Americans aged 55 to 74 with 30 or more pack-years three annual low-dose CT scans or chest radiographs and cut lung cancer mortality by 20.0 percent (95 percent confidence interval 6.8 to 26.7) and all-cause mortality by 6.7 percent. NELSON gave 13,195 men and 2,594 women aged 50 to 74 volume CT at baseline and years 1, 3 and 5.5, and found a lung cancer death rate ratio of 0.76 (0.61 to 0.94) in men at ten years, with only 2.1 percent of participants referred for a suspicious nodule. The number needed to screen to prevent one lung cancer death is 323 over 6.5 years in the National Lung Screening Trial and 130 over ten years in NELSON. The harms are real and are the reason screening is targeted rather than universal: 24.2 percent of low-dose CT screens in the National Lung Screening Trial were positive and 96.4 percent of those positives were false; false positives led to 17 invasive procedures per 1,000 people screened, with fewer than one major complication; incidental findings were reported in 4.4 to 40.7 percent of people screened; and estimates of overdiagnosis, the cancers found that would never have caused harm, ranged from 0 to 67 percent across studies (US Preventive Services Task Force evidence report, 2021). In the United States the Task Force recommends annual low-dose CT from 50 to 80 for people with a 20 pack-year history who smoke or stopped within the past 15 years (grade B, 2021), widened from the 55 to 80 and 30 pack-year criteria of 2013.

Screening in the United Kingdom. The UK National Screening Committee recommended targeted screening in June 2022 for people aged 55 to 74 identified as being at high risk, with smoking cessation built into the service, and named England's Targeted Lung Health Check programme as the starting point. The national programme was announced on 26 June 2023, costs £270 million a year at full size and is expected to deliver almost one million scans and find as many as 9,000 cancers a year. Eligibility is not age alone: a person's GP record identifies ever-smokers aged 55 to 74, and a risk model decides who is scanned, using PLCOm2012 or the Liverpool Lung Project version 2 in England, at a PLCOm2012 threshold of 1.51 percent risk over six years. Scanning is every two years for those at high risk. The first phase invited about 900,000 people, made 375,000 risk assessments, carried out 200,000 scans and found more than 2,000 cancers, of which 76 percent were early stage against 29 percent outside the programme in 2019, and about 70 percent of the scanning was done in mobile units parked in places such as supermarket car parks, deliberately in more deprived areas where people are four times more likely to smoke. By March 2025 the programme had invited over two million people and diagnosed 7,193 lung cancers, 63.1 percent at stage 1 and 12.6 percent at stage 2, and the early-stage share of all lung cancer in England rose over the five years, most in the most deprived regions. Full national coverage is expected in 2030.

Diagnosis and staging. NICE NG122 sets the order. Contrast-enhanced CT of the chest, including the liver, adrenals and lower neck, comes before any biopsy. Everyone who could have treatment with curative intent is offered PET-CT. Where the nodal stage would change treatment, nodes are sampled through the airway or oesophageal wall: endobronchial ultrasound-guided transbronchial needle aspiration, endoscopic ultrasound-guided fine-needle aspiration, or both, with surgical mediastinal staging kept for a negative result when suspicion remains high. Enlarged nodes (10 mm or more in short axis on CT) or other lesions are biopsied in preference to the primary when the nodal stage decides treatment. Brain imaging is staged to the stage: none for clinical stage 1 without neurological symptoms, where the prevalence of brain metastases is about 4 percent; contrast-enhanced brain CT for clinical stage 2; contrast-enhanced brain MRI for stage 3. Samples must be big enough for subtyping and molecular markers, and the panel to run is set by the National Genomics Test Directory rather than by the guideline. One caution a reader should know: NG122's recommendations were written against the 7th edition of the American Joint Committee on Cancer staging system, which NICE says on the page, while staging in the clinic has moved two editions on.

The ninth edition of TNM, in force since 1 January 2025. The International Association for the Study of Lung Cancer built a new database of 124,581 patients diagnosed between 2011 and 2019 and analysed 76,518 of them. The T descriptors were tested and left alone: they are identical to the eighth edition. The N descriptors changed, splitting N2 into N2a, a single ipsilateral mediastinal or subcarinal station, and N2b, multiple ipsilateral mediastinal stations with or without the subcarinal station. The M descriptors changed at the far end, splitting M1c into M1c1, multiple metastases in one extrathoracic organ system, and M1c2, multiple metastases in several organ systems; both stay in stage group IVB. The stage groups moved to follow: T1N1 to IIA, T1N2a to IIB, T3N2a to IIIA, and T2aN2b and T2bN2b to IIIB. The practical effect is finer prognostic separation at the borders of resectability rather than a change in who is operated on, and a discontinuity in every survival series that spans the change.

Pathology: how the histology is decided. The 2021 WHO classification keeps the same order of evidence: morphology first, immunohistochemistry to support it, molecular tests last. That matters because most lung cancers are diagnosed on a small biopsy or a cytology specimen rather than a resection, and the fifth edition added a section dedicated to classifying small samples for exactly that reason. In practice a poorly differentiated tumour is assigned by a two-stain panel: TTF-1 marks adenocarcinoma, p40 marks squamous cell carcinoma, and a tumour positive for one and negative for the other is called that type rather than left as non-small-cell carcinoma not otherwise specified, because the name decides whether the tumour is sequenced and whether pemetrexed and bevacizumab are safe. Other changes in the fifth edition bear on the corpus's pages: invasive non-mucinous adenocarcinomas are graded by the percentage of each growth pattern, only the invasive part counts towards the T size in part-lepidic tumours, spread through air spaces is recognised as a prognostic feature, lymphoepithelial carcinoma moved into the squamous cell carcinomas, and thoracic SMARCA4-deficient undifferentiated tumour was recognised as an entity.

Prevention. Because tobacco causes about 72 percent of UK cases, prevention is tobacco control and stopping smoking, and nothing else comes close. NICE NG209 covers the whole of it, from preventing uptake in people aged 24 and under to treating dependence in everyone aged 12 and over, and was last updated on 4 February 2025 to add recommendations on cytisinicline. Lung cancer screening is also a prevention service and not only a detection service: the UK National Screening Committee's recommendation was explicitly for screening with integrated smoking cessation, and the screening appointment reaches people who smoke at a moment when they will act. The other preventable exposures are radon, which can be measured in a home and reduced by sealing and ventilation, and workplace exposures to asbestos, silica, diesel exhaust and the rest, which are regulated rather than chosen.