ABSTRACT
Exfoliated cancer cells have been described in the sputum of lung cancer patients since the 1930s.1 Papanicolaou and Saccomanno, pioneers of exfoliative cytology, demonstrated that premalignant cytological changes could be detected several years before a clinical diagnosis of lung cancer in high-risk people.2,3 From the time of the initial report, it took almost 40 years to begin a systematic randomized evaluation of exfoliated sputum morphology as a screening test for lung cancer. Surprisingly, these NCI Collaborative Lung Cancer Screening Trials found no mortality benefit from morphologic screening of sputum specimens, and they were underpowered to detect a small benefit of chest X-ray screening.4 The NCI trials established that frankly malignant cytology specimens are often associated with roentgenographically occult central lesions, most of which are squamous cell carcinomas.5-7 Roentgenographically occult squamous cell lung cancer detected by cytology often (over 80%) is early cancer (stage 0 or stage 1), where the actual 5-year cure by surgery is excellent.8 However, even in a screened population, only 14% of lung cancers are roentgenographically occult squamous cell tumors detectable only by cytology.9 A recent evaluation of sputum morphology confirmed a relative risk of 3.3 for carcinoma in individuals with moderate sputum atypia or worse when compared to matched controls with normal sputum cytology.10 Although insensitive to the majority of lung cancers, sputum atypia remains a standard of practice against which molecular methods should be compared. Subsequently, microscopic imaging and molecular biology have been applied to improve the sensitivity of this easily collected material for lung cancer screening.11,12
Today, there is a growing interest in the early identification of asymptomatic lung cancer by chest computerized tomography (CT). This interest may have resulted from the initial report of the Early Lung Cancer Action Project (ELCAP), which detected a lung cancer prevalence of 2.7% in asymptomatic heavy smokers over the age of 60; most of these lesions were early stage.13 This represents a clear improvement over chest X-ray screening where the low density of up to 77% of surgically curable lesions detectable by CT would have caused them to be missed by chest X-ray.14 Nevertheless, the improved sensitivity of helical CT comes at the cost of a lowered specificity (higher false-positive rate). New, non-calcified nodules are seen in up to 50% of screening helical CT examinations, increasing in frequency with thinner CT slices and decreasing with followup examinations.15-19 Up to 90% of the abnormalities detected on the initial CT examination are not malignant.13 Further, the CT scan may be insensitive to superficial, preinvasive/microinvasive cancers in the central airways that may be manifest in
sputum.11 Within 5 years, the NCI has initiated a new randomized trial of lung cancer screening to determine whether low-radiation-dose spiral CT (LDCT) scanning can lead to reduced lung cancer mortality.20 This multicenter NCI Lung Screening Study of helical CT screening will require another 6 years to complete. Although no new sputum tests were sufficiently validated to perform contemporaneously during this trial, NCI-LSS sputum specimens have been collected and stored for future evaluation of a sputum screening test. This discussion will consider candidate sputum tests that might be considered for evaluation as complementary to helical CT screening for lung cancer.
