Quantitative detection of lung cancer cells by fluorescence in situ hybridization: comparison with conventional...

The aim of this study was to clarify whether fluorescence in situ hybridization (FISH) can diagnose lung cancer in various clinical specimens in comparison with conventional cytology.

Prospective study.

University hospital in a metropolitan area.

Fifty consecutive patients with abnormal chest radiography or CT scan findings were enrolled. The patients included 32 men and 18 women, with an average age of 64 years. The final definitive diagnosis was made by histologic examination, as follows: 38 primary lung cancers (24 adenocarcinomas, 8 squamous cell carcinomas, 2 large cell carcinomas, and 4 small cell carcinomas); 1 metastatic renal cell carcinoma; and 11 benign lesions.

Four types of clinical specimens were analyzed. Cells obtained by transbronchial brushing and transbronchial fine-needle aspiration using a fiberoptic bronchoscope under fluoroscopy, CT scan-guided percutaneous needle biopsy, and bronchial washings. On every examination, duplicate slides were made for analyses of conventional cytology and FISH.

Classifications according to conventional cytology were as follows: class I, 4 patients; class II, 15 patients; class IIIa, 3 patients; class IIIb, 5 patients; and class V, 23 patients. A classification higher than class IIIb was considered to be positive for cancer. For cytology, we found no false-positive cases and 11 false-negative cases. The specificity was 100%, and the sensitivity was 71.8%. By FISH, 34 cases showed aberrant copy numbers in either chromosome 3 or 17. We found no false-positive cases and five false-negative cases. The specificity was 100%, and the sensitivity was 87.1%.

The ability of FISH to detect aneusomic lung cancer cells is superior to conventional cytology for the diagnosis of lung cancer.