The aim of this study was to investigate the possible additional diagnostic information provided by imprint cytology when performing ultrasound-guided transthoracic core biopsy and to evaluate whether it could optimise the biopsy procedure.A total of 155 transthoracic core biopsies with touch imprint smears were performed under ultrasound guidance, with 127 malignant and 28 benign lesions. The imprint smears were stained using Riu9s method and interpreted by a cytopathologist. These were compared with the histopathology of core biopsy specimens and the final diagnosis of malignant versus benign disease.The overall diagnostic accuracy of imprint cytology was 94% (146 out of 155). Histopathological analysis showed an overall accuracy of 94% (146 out of 155), with a sensitivity of 94% (119 out of 127) and negative predictive value of 79% (27 out of 34). The combination of these two methodologies had an increased overall accuracy and negative predictive value of 98% (152 out of 155) and 90% (28 out of 31), respectively. The results of imprint cytology and histopathology were in agreement in 143 patients (92%).In conclusion, imprint cytology of ultrasound-guided transthoracic core biopsy is a sensitive procedure for diagnosing peripheral thoracic lesions, and it may increase the diagnostic accuracy and cancer negative prediction of biopsy alone. With an on-site approach, imprint cytology may help to assess the adequacy of biopsy specimens and optimise the biopsy procedure. Eur Respir J 2004; 24: 905-909.