The purpose of this study was to evaluate the feasibility, efficacy, safety, and pharmacokinetics of trastuzumab plus cisplatin and gemcitabine in patients with Her2-overexpressing stages IIIB or IV non-small cell lung cancer (NSCLC) and to study the relationship between results from the two methods for determining levels of Her2 overexpression. Chemonaive patients were eligible if they had stages IIIB or IV NSCLC with either a Her2 score of at least 1+ by immunohistochemical (IHC) analysis or a serum Her2 shed antigen level of at least 15 ng/ml by enzyme-linked immunosorbent assay (ELISA). Treatment consisted of cisplatin 75 mg/m(2) day one plus gemcitabine 1250 mg/m(2) days one and eight plus trastuzumab 4 mg/kg day one and 2 mg/kg weekly thereafter on a 21-day cycle for six cycles followed by weekly maintenance trastuzumab therapy. Of the 21 patients enrolled, 8 (38%) patients had a partial response. The 1-year survival rate was 62% (13/21). Median time to progression was 36 weeks. Pharmacokinetic studies revealed no interaction between trastuzumab and gemcitabine plus cisplatin. In patients screened for this study, Her2 expression was zero in 283/360 (79%); 1+ in 32/360 (9%); 2+ in 27/360 (8%); and 3+ in 18/360 patients (5%). Serum Her2 shed antigen was >15 ng/ml in 27/ 288 (9%) patients. Of patients who had both Her2 assays, 24% (4/17) with ELISA scores >15 ng/ml had IHC scores of 3+, compared with only 2% (3/145) of the patients <15 ng/ml and 4% (7/162) of all patients. The addition of trastuzumab to cisplatin and gemcitabine was well tolerated, but further study will be required to determine whether this combination is superior to chemotherapy alone. This may be demonstrated if only those patients with Her2, having a score of IHC 3+ were eligible. Since IHC 3+ is rare in NSCLC, performing IHC in only those patients with serum Her2 shed antigen >15 ng/ml would greatly increase the efficiency of IHC screening though at the cost of excluding nearly half the patients with Her2 scores of 3+ on IHC analysis. Thus, if sequential screening consisting of serum ELISA followed by IHC analysis is implemented, it may make a trastuzumab trial feasible but should ultimately be supplanted by another screening system if trastuzumab is shown to be beneficial to some patients with IHC Her2 scores of 3+.