Diagnosis of Lung Cancer – Improving Survival Rates

Lung cancer is a major global health burden: it was responsible for 1.3 million deaths in 2004, equating to 2.3% of all deaths. Death rates from lung cancer are predicted to continue to rise, with the disease being responsible for 2.8% of all deaths (1.67 million) by 2015.¹

Despite advances in treatment, survival rates from lung cancer in the UK have improved by only a few per cent in the last 40 years. The fiveyear survival rate for patients diagnosed between 1991 and 1993 was 5%.² The EUROpean CAncer REgistry-based study on survival and CARE of cancer patients 4 (EUROCARE-4)³ has highlighted the difference in survival between England and other European countries. The fiveyear survival rate in England for patients diagnosed between 1995 and 1999 was 8.4% compared with the average European rate of 12%. These figures are in even greater contrast to reported five-year survival rates in the US of 15.7% for patients diagnosed between 1995 and 2001.⁴ Analysis of EUROCARE-4 also showed that one-year survival rates in England were lower than the European average, probably reflecting poorer access to care. This would suggest a particular need to promote earlier diagnosis in the UK, to try to improve survival.

Survival is dependent on the disease stage at diagnosis, with a marked variation between earlier- and later-stage disease. Five-year survival for localised disease is around 49% compared with 2% for disease with distant metastases at presentation.⁴ Unfortunately, the majority of lung cancers have already been disseminated at the time of presentation.^4,5

Screening
Much interest has focused on diagnosing lung cancer earlier in order to try to improve radical treatment rates and reduce mortality. Initially, this interest focused on screening. The first randomised controlled trial took place in London in the 1960s.⁶ This looked at a chest X-ray every six months for three years versus a chest X-ray at the beginning and end of the three-year period. Diagnosis and resection rates were higher in the group receiving more frequent chest X-rays, but lung cancer mortality was similar in both groups. Three US studies⁷ in the 1970s and 1980s looked at the use of either chest X-ray alone or in combination with sputum cytology.

The Mayo Lung Project⁸ compared chest X-ray and sputum cytology every four months with standard care. Participants randomised to the standard care arm were advised to have a yearly chest X-ray and sputum cytology. This showed that resection rates increased by 14% (32–46%) in the group undergoing screening compared with the group receiving standard care alone, but no stage shift was evident. Fiveyear survival in the screened group reached 33% compared with 15% in the non-screened group.

To avoid confusion due to the biases inherent in screening, the ultimate proof of benefit is disease-specific mortality. Unfortunately, lung cancer mortality was not different in the two groups (3.2/1,000 person-years versus 3.0/1,000 person-years). This lack of improvement in mortality was also evident in the other two studies: the Johns Hopkins⁸ and Memorial Sloan-Kettering studies.⁹ Both looked at the addition of sputum cytology every four months to annual chest X-ray. There was also a contemporaneous study¹⁰ in Czechoslovakia comparing chest X-ray plus sputum cytology every six months for three years versus chest X-ray and sputum cytology at the beginning and end of the three years. This study essentially replicated the findings of the Mayo Lung Project, with an increased number of cancers detected in the more frequently screened group but with no difference in lung-cancer-associated mortality.¹¹None of these studies had a ‘no-screening’ control group (see Table 1). Therefore, based on these studies, it would not be possible to recommend either chest X-ray or sputum cytology as a screening test for lung cancer. Indeed, a Cochrane systematic review¹² has suggested that more frequent screening with chest X-rays is associated with an 11% relative increase in lung cancer mortality compared with less frequent screening.

Chest X-rays are less sensitive for the detection of lung cancer than computed tomography (CT). Evaluation of the chest X-rays taken as part of the Mayo Lung Project identified that 90% of peripheral lung cancers and 65–70% of central tumours that were detected in the chest X-ray every four months had in retrospect been visible on previous X-rays.¹⁴ Interest has therefore moved to the use of CT for screening. This was made possible by the advent of low-dose spiral CT, which reduced both the radiation dose and scan time.¹⁵ Early reports showed increased rates of detection over chest X-ray, with the vast majority of detected tumours being stage I.^16-18 The largest observational report of CT screening is the International Early Lung Cancer Action Project (I ELCAP).¹⁹ A total of 31,567 participants over 40 years of age and deemed to be at lung cancer risk due to either cigarette smoking or occupational exposure had a baseline CT. All patients had to be fit to undergo thoracic surgery if required. Twentyseven thousand, four hundred and fifty-six repeat scans were performed between seven and 18 months after the previous screening. Four hundred and seventy-nine cancers were identified, 405 on the initial scan and 74 on annual screening. There were five interim diagnoses. Overall, 85% of tumours were clinical stage I with 72% confirmed pathological stage I. For all participants, the 10-year lung cancer survival rate was 80%, increasing to 88% in those with clinical stage I disease. However, a major criticism of CT is that it identifies nodules that will ultimately turn out to be non-malignant. During the prevalence screen in the I-ELCAP study,19 13% of CTs identified non-calcified nodules requiring further investigation,including serial CTs, positron-emission tomography–CT (PET–CT) and percutaneous biopsy. Only 9.6% of these participants were subsequently proved to have lung cancer. Even higher rates of benign nodule identification have been quoted in other studies^20–23 and up to 20% of invasive procedures following CT screening are for benign disease.^20,21,23An additional concern is overdiagnosis, such that patients receive treatment for slowly growing tumours that may never have caused them any problems in their lifetime, a phenomenon that is already recognised in other screening programmes.^24,25 Several studies have calculated tumour volume doubling times for screen-detected cancers. These have shown that many of the screen-identified tumours are slow-growing, with doubling times well in excess of the 40–70 days calculated from epidemiological data of non-screen-identified cancers.²⁶

Currently, there is insufficient evidence to recommend low-dose CT as a screening test for lung cancer, although there are several randomised controlled trials currently under way seeking to definitively answer this question.^27–29

There are several other methods currently under investigation for use as screening tests for lung cancer, including narrow-band and autofluorescence bronchoscopy. One observational study looked at the use of bronchoscopy along with CT as a primary tool in screening.^20,30 Volunteer current and former smokers underwent sputum induction for quantitative cytometry and CT before being offered autofluorescence bronchoscopy. Five hundred and sixty-one subjects were enrolled in the study, with 378 undergoing bronchoscopy. Fourteen primary lung cancers were identified, of which four (29%) were CT occult and only detected by autofluorescence bronchoscopy. All of these CT occult cancers were squamous cell carcinomas. Because of the observational nature of the study, the significance of the use of this approach on mortality is unknown.

Biological tools, such as testing serum for tumour-associated antibodies, detection of gene-promoter hypermethylation in sputum samples, exhaled breath volatile organic compounds and detection of novel proteins in serum or sputum, are also in development.³¹ Unfortunately, none of these is currently ready for use in clinical practice and no form of screening for lung cancer can be recommended.

Symptom Recognition and Reporting
Interest has now switched to looking at whether lung cancer can be diagnosed earlier in its natural history by focusing on promoting symptom recognition and reporting. Ninety per cent of patients are symptomatic at the time of diagnosis,³² often experiencing multiple symptoms.^33–35 Many of those presenting will have been symptomatic for many months, with reported delays to healthcare of up to two years.³³ Much work has focused on investigating this, with reported median delays from onset of symptoms to presenting to healthcare ranging from seven to 31 days.^35–39 Public knowledge of lung cancer symptoms generally appears to be poor.^35,37,40,41 Patients often develop symptoms but are unaware that they could be related to a sinister cause: it appears that between 50 and 75% of lung cancer patients may not be aware of the significance of their symptoms.^35,37 Only when further symptoms develop, or their general health deteriorates will they seek advice.^33,35 In particular, systemic symptoms such as lethargy and weight loss seem to be associated with longer delays, whereas haemoptysis tends to prompt a more rapid response.^33,39

It has also been noted that even those deemed to be at risk of lung cancer, predominantly current and ex-smokers, do not always perceive themselves to be at risk.^35,41 Even when patients recognise a change in their health, there are many barriers to presentation. Themes that have been identified include fear of wasting the doctor’s time, feeling unworthy of treatment (particularly in relation to being a smoker), being unsure as to whether the symptom/change experienced is ‘normal’, putting the symptom down to being part of the ageing process, minimising symptoms, stoicism and the and associating the symptom with a known, pre-existing, condition.^40,41
Delays have also been identified once patients present to their primary care team, with many patients having to present on more than one occasion before onward referral/further investigation. This is despite clear advice in the British National Institute for Clinical Excellence (NICE) guidelines in terms of chest X-ray referral.⁴² The delay from first presentation to referral to a respiratory specialist has been reported to range from a mean of 3438 to 73 days (range 0 to> 175).^{37,36,39,43,44} Bowen and Raynor’s study³⁷ also showed that of the 76% of patients who first consulted their own family doctor, only one-third were referred following their initial consultation, with a further third referred by another doctor in the practice, suggesting a second consultation. A Danish study looked at potential reasons for the delay in onward referral of symptomatic patients and identified several contributing factors. In patients with co-morbid diseases, symptoms were often ascribed to this rather than potential lung cancer.⁴⁵

Chest X-rays reported as normal were associated with a longer delay, with primary care teams being falsely reassured. Twenty-five per cent of lung cancer diagnoses in the UK are made during an acute admission, despite the patient having presented previously to their primary healthcare team with a symptom that could be indicative of lung cancer.⁴⁶ Improving the early diagnosis of lung cancer in Britain has become a government priority, with the formation of the National Awareness and Early Detection Initiative (NAEDI), an important component of the 2007 cancer-reform strategy.⁴⁷ This hypothesises that delays lead to more advanced disease at diagnosis with associated poor one- and five-year survival rates and potentially avoidable deaths. Abdel- Rehman et al. calculated that if UK survival rates were similar to those in Europe, nearly 1,000 deaths per year within five years of the diagnosis of lung cancer, could be avoided.⁴⁸ The NAEDI pathway highlights many areas that could be targeted in order to try to promote earlier diagnosis (see Figure 1).⁴⁹

One such strategy is to use social marketing techniques to raise awareness of lung cancer symptoms and to encourage a more timely presentation to healthcare services. Social marketing uses commercial marketing techniques to change individual and organisational behaviours and policies.⁵⁰

Similar approaches have already been used in other cancers, an example of which is the West of Scotland Cancer Awareness Project (WoSCAP).⁵¹ This project used a mass media campaign combined with general practice education to raise awareness of the symptoms of oral and colorectal cancer. Awareness of symptoms was improved and, in those presenting who were aware of the campaign, presentation was more timely in 60%.

An initial social marketing pilot has been carried out in lung cancer in Doncaster, the largest metropolitan borough in the UK, which has a high rate of lung cancer^52,53 and a high rate of social deprivation. The social marketing campaign and primary care education programme were initially designed as a way of addressing a recognised health inequality. Six areas, covered by 11 general practice surgeries, with the highest lung cancer risk were identified. In these areas, brief intervention training was undertaken with the general practitioners, practice nurses and local pharmacists. Following this, there was a public awareness campaign launched comprising leaflets, advertising on bus banners and billboards (see Figure 2), local media events and bus stops that were fitted with sound chips to make them ‘cough’ and in so doing drew attention to the campaign posters.

This project increased awareness of the importance of seeking medical advice for a prolonged cough and resulted in a statistically significant increase in chest X-ray referrals. Lung cancer diagnosis rates were also increased, although this was not statistically significant. No stage shift was evident, but the numbers at different lung cancer stages were too small for subgroup analysis.^53,54

If the responses to this campaign could be replicated on a larger scale, and people could be encouraged to present earlier with their symptoms, their disease should be identifiable in a more timely fashion. In turn, this will hopefully increase the number of patients suitable for curative treatment as well as influencing the numbers receiving active treatment (chemotherapy and/or radiotherapy). Both actions should lead to improved survival of patients with lung cancer. ■