Predictive value of image cytometry for diagnosis of lung cancer in heavy smokers

Abstract

The Research Institute for Diagnosis and Treatment of Early Lung Cancer (RIDTELC) Lung Study was initiated to determine whether lung cancer screening by automated sputum cytometry combined with conventional sputum cytology and auto-fluorescence in addition to white light bronchoscopy could enhance the detection rate of early lung cancer. The present study analyses the initial findings to evaluate the efficiency of automated sputum cytology in predicting the diagnosis of lung cancer. In this study, malignancy grade was used as a predictive parameter for lung cancer. In total, 2,480 heavy smokers (>30 pack-yrs), aged 50-74 yrs, with no previous cancer in the last 5 yrs, received chest radiology, conventional sputum cytology and sputum cytometry screening. In total, twenty-seven lung cancers were diagnosed, representing a prevalence of 1.1%, 25 of which provided sputum samples. Positive automated sputum cytology results were seen in 176 smokers (7.2%), 10 (0.4%) of whom had severe dysplasia or higher lesions (positive results) by conventional sputum cytology examination. Out of 25 tumour cases, 20 had suspicious results using automated sputum cytology, representing 80% sensitivity. One patient out of 24 with tumours had positive results on conventional sputum cytology, representing a sensitivity of 4.2%. For all stages of squamous cell lung cancer and later stage adenocarcinoma the sensitivity of automated sputum cytology was 100%. For adenocarcinoma stage I sensitivity was 25%. In conclusion, DNA analysis of sputum slides by automated sputum cytology may be a suitable tool for the detection of early lung cancer and the characterisation of a high-risk group with pre-invasive lesions for follow-up.