Patient-specific pharmacokinetics of intratumoral cisplatin track pathologic necrosis in early-stage lung cancer.

Intratumoral injection of cisplatin into lung cancer is feasible and can increase local drug exposure while reducing systemic toxicity. However, its success relies on achieving sufficient drug exposure to the tumor, which has not been quantitatively evaluated in a clinical setting. In the context of a Phase I clinical trial of intratumoral cisplatin injections into resectable lung tumors via robotic bronchoscopy (N = 6), we developed a quantitative framework to assess delivery effectiveness and long-term cisplatin exposure in tumors. Initial injectate retention was estimated using a novel image analytics pipeline based on intraoperative cone-beam computed tomography scans, by assuming that emerging high-attenuation regions are due to injectate deposition. Long-term drug exposure was estimated using a novel pharmacokinetic model tailored for intratumoral drug delivery and informed by sequential venous blood draws following the injections. The area under the platinum concentration curve in the tumor compartment at t = 3-21 h and platinum concentration in the tumor compartment (t = 3-6 h), estimated from the fitted pharmacokinetic model, showed a strong monotonic association with percentage pathological necrosis assessed post-resection by three independent, blinded pathologists (Spearman's ρ = 0.93, 95% bootstrap CI 0.67-1.00, unadjusted exact p = 0.033), which was not statistically significant when corrected for multiple comparisons (adjusted p = 0.17). No other pharmacokinetic variable correlated was significantly correlated with necrosis or other pathological response metrics prior to multiple comparison correction. This result aligns with the known time-concentration dependency of cisplatin action in tumor. Further studies will be needed to evaluate this observation for repeatability and clinical relevance. Our approach provides a framework for estimating intratumoral drug exposure in clinical settings, which may help with the interpretation of clinical trial outcomes and future evaluation of injection strategies.