Data Collection, included patients

We conducted a retrospective cohort study of consecutive patients with breast cancer who received at least one dose of T-DXd at The Ohio State University (OSU) from January 1, 2019, to February 20, 2024. Study data were compiled using REDCap electronic data capture tools at OSU. The study protocol was reviewed and approved by The OSU Institutional Review Board (2023C0104), and a waiver of informed consent was granted due to the retrospective nature of the study.

Demographic variables

The following potential risk factors for T-DXd pneumonitis were collected: age, sex assigned at birth, race, body mass index, smoking status, cancer stage, pulmonary involvement of cancer (excluding non-specific pulmonary nodules, chest wall and rib involvement), hormone status (estrogen receptor [ER], progesterone receptor [PR], and HER2), line of therapy, prior systemic therapy, prior chest or breast radiation, total doses of T-DXd, T-DXd concentration, pre-existing lung disease (chronic obstructive pulmonary disease [COPD], asthma, ILD), and baseline SpO2 and oxygen requirements.

Pneumonitis diagnosis and treatment

T-DXd pneumonitis was determined by the treating clinician at the time of toxicity and retrospectively reviewed for agreement by a member of the study team. Attribution of pneumonitis to T-DXd was based on association with T-DXd use, corroborative chest computed topography (CT) findings, and reasonable exclusion of alternate etiologies of lung inflammation including infection and cancer progression. Pneumonitis severity was assigned based on the Common Terminology Criteria for Adverse Events (CTCAE v4), with mild disease defined as grade 1–2 and severe disease as grade 3–5. Pneumonitis treatment course, resolution of pneumonitis, and re-challenge with T-DXd following pneumonitis (if applicable) were recorded.

Radiation dosimetry variables

Prior history of chest and breast radiation preceding T-DXd initiation was recorded for all patients. For patients who had available radiation treatment data, the mean lung dose (MLD), the lung volume receiving ≥ 20 Gy (V20), and regional nodal irradiation (RNI) were recorded. If patients received multiple courses of chest radiation, only the dosimetry variables for definitive treatment of breast cancer were included.

Evaluation of pre-existing chest imaging abnormalities

As previously described [10], we defined the group of patients with chest imaging abnormalities as those who exhibited at least one of the following pre-treatment CT findings in the lung: ground glass appearance, reticular opacity, consolidation (excluding infectious pneumonia), traction bronchiectasis, centrilobular nodularity, and honeycombing. Interstitial changes that were thought to be related to underlying malignancy were not included. The extent of imaging abnormality was measured on a five-point scale for upper, middle, and lower zones of the lung (0, none, 1 < 5%, 2, 5–25%, 3, 25–50%, 4 > 50%). Determination of imaging abnormality and severity scoring was done by board-certified pulmonologists on the study team who were blinded to all other aspects of the patient history, and interobserver disagreements were resolved by consensus.

Evaluation of Pre-treatment Pulmonary function testing

Spirometry (forced expiratory volume in 1 s [FEV1], forced vital capacity [FVC], FEV1/FVC ratio, bronchodilator responsiveness testing), total lung volume (TLC), residual volume (RV), diffusion capacity (DLCO), diffusion capacity corrected for hemoglobin (DLCO corrected for hemoglobin), 6-min walk test, and oxygen requirement within 12 and 18 months of ICI initiation were recorded by the study team and served as the baseline lung function. Testing that did not meet ATS criteria for acceptability and repeatability were not included [11]. If there were multiple PFTs before ICI treatment, only the PFTs closest to the start of ICI treatment were considered. All values were recorded as raw values (liters for FEV1, FVC, TLC, RV, mmol/min/mmHg for DLCO and DLCO corrected for hemoglobin) or as a percent predicted (pp) adjusted for patient’s sex, height, age, and race. When available, post-bronchodilator FEV1/FVC was utilized to determine the presence of obstructive disease [12].

Statistical analysis

Patient characteristics and risk factors were compared between patients with and without T-DXd pneumonitis. Categorical variables were summarized as frequencies (n) and percentages (%) and compared using chi-squared test or fisher exact test as appropriate. Continuous variables were summarized as mean and standard deviation (SD) and compared using two sample t-test if they are normally distributed, otherwise median with interquartile range (IQR, 25th and 75th percentile) is used to describe variable distributions and compared using Wilcoxon rank sum test. Multivariable logistic regressions were used to study the association between potential risk factors and T-DXd pneumonitis, and model selection was based on clinical variables of interest and significant associations found on univariate analysis. The Kaplan-Meier survival curves with log-rank tests were used to compare overall survival between patients with T-DXd pneumonitis and patients without T-DXd pneumonitis, as well as between patients with mild pneumonitis (grade 1–2) and severe pneumonitis (grade 3–5). The statistical analysis was performed using R software (version 4.3.0; R Core Team, R Foundation for Statistical Computing, Vienna, Austria).