Investigating the risk of drug-induced interstitial lung disease in patients with metastatic breast cancer treated with antiHER2 regimens through a network meta-analysis approach

The advent of antibody-drug conjugates such as trastuzumab deruxtecan (T-Dxd), has significantly increased the awareness towards drug-induced Interstitial Lung Disease (ILD), a pulmonary disorder caused by several conditions including drugs, which may cause a wide range of pathological processes, from inflammation to interstitial fibrosis. However, the differential magnitude of risk among the available treatments is still unknown. This network meta-analysis (NMA) aims to provide first preliminary data on the risk of ILD across antiHER2 regimens. After a systematic literature review, a Bayesian NMA using a Markov-Chain Monte Carlo simulation was performed using STATA software (Stata Corp, version 17). Treatments were classified as: Trastuzumab (T) + chemotherapy (CT), T/Pertuzumab (P)+CT, Everolimus (Eve)/T+CT, T-DM1, T-DM1/P, T-DM1+Atezolizumab (Atezo) and T-Dxd. Outcomes were reported with corresponding 95% credible intervals (CrIs). Treatments were ranked using the surface under the cumulative ranking curve (SUCRA). BOLERO-1, BOLERO-3, CLEOPATRA, KATE2, MARIANNE and DESTINY-Breast03 were the main trials included in the NMA. In the experimental arm, ILD of any grade was detected in 10.5% of patients for DESTINY-Breast03, 4.5% and 3.6% respectively for BOLERO-1 and BOLERO-3, 0.7% in KATE2, 0.5% in CLEOPATRA and 0.3% in MARIANNE. In terms of tolerability, T-DM1 ranked first (SUCRA 80.1%), followed by TDM1/P/T + CT (SUCRA 77.5%). Instead, T-Dxd and Eve/T + CT were the most prone to ILD development (respectively SUCRA 18.7% and 11.1%). No significant differences were shown between T+CT and T-DM1 or T-DM1+Atezo in NMA, while a significantly lower risk ratio was observed for T+CT when compared to T/P+ CT (RR 0.04, 95%CrI 0.01-0.34), T-Dxd (RR 0.03, 95%CrI 0.00 - 0.56) and Eve/T+CT (RR 0.02, 95%CrI 0.00 - 0.08). Our NMA showed how drug induced ILD could differentially affect antiHER2 treatments for MBC patients. Further investigations are needed to understand the underlying pathological mechanisms causing drug induced ILD and to identify predictive biomarkers focused on the early identification of patients at higher risk.