17072 Background: Published literature has demonstrated minimal activity for VNR in the salvage treatment of NSCLC, with response rates of only 0.8% (Fossella, et al, JCO 2000; 18(12), 2354–2362). However, in our clinical experience, we have observed a number of patients (pts) with clinical benefit. Methods: Retrospective evaluation was performed of all pts with NSCLC who had received VNR at FCCC from 6/02–6/05 in the salvage setting. Pts were evaluable if full medical records (including radiographic imaging for disease assessment) were available, had received ≥2 cycles of VNR, and had advanced, recurrent or metastatic NSCLC. Primary endpoint was response rate (RR); secondary endpoints included safety, median overall survival (OS), and median time to progression (TTP). Results: 52 pts were included, 28 pts were ultimately evaluable (13 pts received <2 cycles of VNR, 7 pts did not have NSCLC, and 4 pts received VNR with RT or other agents). Median age was 64 yrs, 64% were female; ECOG-PS 0–18%, 1–50%, 2–28.5%, 3–3.5%. 82% exhausted 2 lines of systemic therapy before VNR; 32% exhausted 3; 21.4% received prior EGFR-TKI; 18% received other prior protocol therapies. 96.3% of pts had stage IV, stage IIIB, or recurrent disease. 21.4% had bone mets; 25% had brain mets; and 42.8% had visceral metastases. 85.8% had significant comorbidities. Median number of VNR cycles was 4. RR:PR 10.7%, SD/MR 35.7%, and PD 50%. 25% required dose reductions, predominantly for gr 4 hematologic toxicities. Non-hematologic toxicities were generally mild, and included neuropathy, fatigue and GI distress; there were no Tx-related deaths. 32% were able to receive other systemic therapy after VNR. Median OS was 5 months, and median TTP was 3 months. Conclusions: Vinorelbine is active and relatively well-tolerated in the salvage treatment of NSCLC, including heavily pretreated pts. Response rates exceed that observed in the literature. [Table: see text]