Forty-one patients with known lung carcinoma or suspicion of bearing this disease were studied with 111In-antiCEA F023C5-F(ab')2. Five patients with positive results were further studied to assess in vivo specificity using 111In-4C4-F(ab')2 (MoAb antihepatitis). Immunoscintigraphic results have been compared to immunohistochemistry in 16 patients. Tumour visualization with non-specific MoAb was present in all five patients although the maximum tumour/background (T/B) ratio obtained was 1.51. Therefore a 'specificity criterion' has been applied to examinations performed with MoAb antiCEA. There was considered to be a positive result when the T/B ratio was greater than 1.60. Results obtained in this way for thoracic lesions were 25/34 true positive (TP) (73%) and 6/7 true negative (TN) (86%). When studies were analysed visually only, the results were 32/34 TP (94%) and 3/7 TN (43%). In addition, immunoscintigraphic results are more concordant with immunohistochemical results when the specificity criterion is applied. In metastatic lesions results were 9/13 TP (69%) with only 1/3 TP (33%) for liver metastasis (anatomic site of lowest detection ability). SPECT imaging did not improve results over planar images, although it clearly contributed to a better anatomic location of lesions. The constant presence of non-specific uptake in thoracic lesions makes it necessary to establish a T/B ratio limit in order to obtain adequate specificity with the immunoscintigraphic technique.