[Reliability of bronchial tumor typing base on fibroscopically obtained biopsies].
This study is based on the bronchial tumors biopsied and later examined between 1971 and 1986 at the Institute of Pathology, Lausanne. In each case the diagnosis based on the initial biopsy material is compared with the final surgical or autopsy diagnosis (Reference: Histological typing of lung tumours, second edition, WHO, 1981, 2). The series studied is constituted by 163 cases: 144 men (88.3%) and 19 women (11.7%). In 136 cases (83.4%), the biopsy diagnoses are identical to the diagnoses based upon surgical or autopsy material; in 27 cases (16.6%) these diagnoses differ. The positive predictive value of bronchial biopsy appears to be excellent (100%) for small cell and epidermoid carcinoma (CA). It is satisfactory for adenocarcinoma (85.7%) but insufficient for large cell CA (42.3%) and the tumors grouped under "others" (50%). The diagnosis of large cell CA, in situ CA, papillary tumor, undifferentiated CA, or carcinomatous lymphangitis should be considered with caution and in some cases it is advisable to repeat the biopsy. The discrepancies between the initial and final diagnoses can, in all cases, be attributed either to the biopsy specimen being too small and therefore nonrepresentative or, less frequently, to crushing and necrosis of the tissues.