Introduction

Every year, there are about 3,000 cases of lung cancer deaths in Hong Kong. Lung cancer is the commonest cancer killer in Hong Kong and worldwide. Twice as many women die annually of lung cancer as from breast cancer. Early detection of lung cancer is uncommon in clinical practice, with only 16% of all new cases being localized (ie, stages I or II) at diagnosis. Clinicians routinely screen for breast, colon, nasopharyngeal, cervical, stomach and other cancers, but not for lung cancer. However, the majority of lung cancer cases are diagnosed at stages III or IV, and the 5-year survival rates for patients with advanced disease are in single digits even if they are treated with modern chemotherapy in combination with radiotherapy.

Trials on Radiologic and Sputum Cytology Detection of Lung Cancer

In the 1960's to 1980's, investigators used regular chest X-rays (CXRs) and sputum cytology in high-risk patients with the hope of detecting lung cancer at an early and resectable stage. Three large randomized trials were conducted in the United States and Europe between 1960 and 1980. Subjects underwent CXRs every 4 to 6 months for 3 to 6 years. These trials found that screening resulted in the detection of more resectable lung cancers and higher 5-year survival rates. However, for unclear reasons, lung cancer mortality, overall mortality, and the number of unresectable lung cancers were not reduced.

Low-dose computed tomography (LDCT) of the thorax is currently being evaluated as a screening modality for lung cancer. LDCT is twice as sensitive as CXR in detecting nodules and masses. More recently, several uncontrolled trials on CT of the thorax have reported detection of lung cancers at early and resectable stages. In the Early Lung Cancer Action Project (ELCAP), CT detected almost 6 times as many stage I lung cancers as CXRs. (Figure 1)

The bias of overdiagnosis is an unresolved issue in lung cancer screening, as the cancers detected might be slow growing and the patients might die of other competing causes (eg, cerebrovascular accidents). Moreover, the increased positivity rate and sensitivity for lung cancer detection with LDCT could result in an increased burden of follow-up. LDCT may also identify a greater number of noncancer abnormalities (eg, chronic obstructive pulmonary disease, scarring, cardiac abnormalities) than CXR. To the extent that these abnormalities are followed up, this could result in a further increased burden of follow-up for LDCT, possibly unnecessary interventions, and unjustified patient anxiety.

Bronchoscopy in the Detection of Lung Cancer

The development of fibreoptic bronchoscopy represents a cornerstone of the diagnosis and treatment of lung cancer. In experienced hands, bronchoscopy provides diagnostic sampling in the majority of patients with lung cancer. Sampling techniques such as biopsy, brushing and bronchoalveolar lavage provide a diagnostic yield in up to 60% of patients. While frank tumour-like lesions are usually readily recognized by trained eyes, dysplastic and preneoplastic lesions, which are amenable to curative bronchoscopic treatment such as laser or cryotherapy, could be missed.

Autofluorescence bronchoscopy (AFB) is now an established mode of novel bronchoscopy. It allows the detection of dysplastic lesions or carcinoma in situ, which are macroscopically normal when examined under white light bronchoscopy. The principle of AFB is based on the spectral differences between normal and abnormal bronchial mucosa. It utilizes differences in the biochemical, metabolic and structural composition of normal, preneoplastic and neoplastic tissues. When the bronchial surface is illuminated by violet or blue light, normal tissues fluoresce strongly in green. Preneoplastic and neoplastic lesions appear to be brown, brownish red or red, depending on the severity of the lesions, the presence of endogenous porphyrins, and vascularity. These differences allow the bronchoscopist to utilize AFB for the detection of early tumours or preneoplastic lesions in the airways. (Figure 2)

Conclusions

Early diagnosis of lung cancer is a tough challenge for all clinicians. The use of sputum cytology, regular CXR and, more recently, LDCT thorax and AFB could undoubtedly lead to earlier cancer detection. However, there is so far no evidence of improved survival by such undertaking. Novel techniques in imaging, proteomics, bronchoscopy, testing of cancer-related chemicals, blood tests and sputum examination are being developed, and will hopefully help diagnose lung cancer earlier in the near future.

Figure 1. Low-dose CT scan showing a tumour in the right lung (arrow) of a male smoker who had no symptoms and a normal chest X-ray

Figure 2. Normal bronchoscopy showing no definite tumour (left panel), but this showed up as a purple patch under autofluorescence bronchoscopy (right panel), which can detect early tumours not visible on routine white light bronchoscopy