Preclinical experience with cisplatin, gemcitabine, and doxorubicin in pulmonary suffusion.

Background: Because suffusion amplifies lung chemotherapy while limiting systemic toxicity, we tested candidate drugs for treating human lung cancers and pulmonary metastases. Methods: Immature beagle dogs underwent thoracotomy for unilateral lung suffusion of cisplatin (0.125-2 mg/kg; n = 19), doxorubicin (3.75-7.5 mg/kg; n = 7), gemcitabine (168.75 mg/kg; n = 5), or saline (n = 3). After ipsilateral lung circulation isolation and drainage, pulmonary artery chemotherapy was injected, dwelled for 30 minutes, and then aspirated. Bilateral lung biopsies and serum samples assessed delivery and leak. After lung reperfusion, animals recovered for 30 days with scheduled monitoring of vital signs, weights, and behaviors. At experiment termination, necropsy histopathologic tissue analyses assessed tolerability. Results: All 32 animals recovered, except 1 with lung torsion and 2 with pulmonary toxicity that required early euthanasia. Serum concentrations during suffusion for cisplatin (135 ng/mL), doxorubicin (undetectable), and gemcitabine (1452 ng/mL) indicated minimal systemic leakage. Cisplatin escalations showed uniform suffusion deliveries (100% fibrosis at a 100% systemic chemotherapy dose), which was then reduced to a nondamaging 25% threshold. When the equivalent dose of doxorubicin was used, toxicity occurred, but 12.5% (2.5-fold amplification of local delivery) was well tolerated. Gemcitabine, like cisplatin, caused minimal toxicity at 25% of the systemic dose (5-fold amplification). Optimized doses caused no hematologic or metabolic derangements and necropsies showed no gross organ injury other than adhesions. Histopathology demonstrated multifocal ipsilateral lung fibrotic changes without contralateral or extrapulmonary pathology. Conclusions: While suffusion delivery of the vesicant doxorubicin was tolerated less well than cisplatin and gemcitabine, all appear to be safe and feasible for human trials.