Use and outcomes of trastuzumab deruxtecan in HER2-positive and HER2-low metastatic breast cancer in a real-world setting: a nationwide cohort study Authors: H Jourdain, A Di Meglio, I Mansouri, D Desplas, M Zureik, N Haddy Published: 2024 ABSTRACT AbstractBackgroundSince 2020, trastuzumab deruxtecan (T-DXd) has been used in France for patients with previously treated human epidermal growth factor receptor 2 (HER2)-positive or HER2-low metastatic breast cancer (mBC). We aimed to describe the clinical characteristics, outcomes, and potential toxicities among patients receiving T-DXd for HER2-positive and HER2-low mBC.Patients and methodsUsing the French National Health Data System (SNDS), we identified patients who initiated T-DXd for mBC from 30 September 2020 to 30 September 2023. Follow-up data were available through 31 December 2023. Patients were categorized into three groups according to HER2 expression and line of treatment: HER2-positive mBC receiving T-DXd in the third (HER2+ 3L) or second line (HER2+ 2L) and HER2-low mBC receiving T-DXd in the second line (HER2-low2L). We describe their characteristics and report the Kaplan–Meier estimates of overall survival (OS) and incidence of hospitalization.ResultsThe cohort comprised 5890 patients, including 2010 (34.1%) HER2+ 3L, 1260 (21.4%) HER2+ 2L, and 2620 (44.5%) HER2-low2L. For the three respective groups, the median age at inclusion was 59 years [interquartile range (IQR) 51-69 years], 59 years (50-68 years), and 61 years (52-70 years); 34.8%, 30.2%, and 16.0% had brain metastases; 14.2%, 13.7%, and 13.4% had a current or history of cardiovascular disease. Median OS was 30.2 months [95% confidence interval (CI) 28.1-33.5 months] for HER2+ 3L patients, was not reached for HER2+ 2L patients, and was 16.8 months (95% CI 14.5 months-not reached) for HER2-low2L patients. The incidence of hospitalization for cardiac, respiratory, digestive, and hematological disorders was similar for HER2-positive patients treated in the second or third line, whereas HER2-low patients had higher incidence rates for these events.ConclusionIn this large French observational study, T-DXd users were older, had more comorbidities, and had more brain metastases than patients included in registration trials. The rapid expansion of clinical indications of T-DXd calls for proactive surveillance and timely management of potentially life-threatening T-DXd-related toxicity.Key words: trastuzumab deruxtecan (T-DXd), metastatic breast cancer, outcomes, early access program, French National Health Data System (SNDS) Introduction Metastatic breast cancer (mBC) remains a leading cause of cancer-related mortality among women worldwide, necessitating ongoing advances in therapeutic strategies.1 The persistent challenge of managing this disease, especially in its metastatic forms, has driven the need for more targeted and effective treatments. For patients with human epidermal growth factor receptor 2 (HER2)-overexpressing cancers, representing 15%-20% of breast cancers, the therapeutic care has been substantially improved due to HER2-targeting monoclonal antibodies.2 The advent of antibody–drug conjugates3 (ADCs) has significantly modified the treatment landscape of HER2-positive breast cancers, offering targeted delivery of cytotoxic agents and, thus, improving clinical outcomes. Trastuzumab deruxtecan (T-DXd), a novel ADC, has emerged as a promising therapeutic agent in this domain. The DESTINY clinical development program investigated the efficacy and safety of T-DXd in HER2-positive and HER2-low mBC. Initially, the DESTINY-Breast014 and Breast025 trials established T-DXd efficacy in third-line treatment for HER2-positive mBC, leading to its accelerated approval. The DESTINY-Breast036 trial subsequently demonstrated the superiority of T-DXd over trastuzumab emtansine (T-DM1) in second-line therapy, notably improving progression-free and overall survival (OS) rates. The DESTINY-Breast047 trial further expanded the use of T-DXd to HER2-low mBC. In terms of safety, T-DXd has shown a generally manageable profile, with low-grade hematological and gastrointestinal adverse events, although interstitial lung disease observed in patients receiving T-DXd can be severe.8 Patients included in the DESTINY-Breast clinical trials had a median age of 54-58 years, presented with asymptomatic brain metastases in 5%-18% of cases, and were heavily pre-treated. Although T-DXd demonstrated efficacy in this setting, real-world data on patient characteristics, treatment duration and subsequent treatment to manage progression,9 and OS and toxicity are limited.10, 11, 12 The accelerated integration of T-DXd into clinical practice highlights the urgent need for real-world evidence to clarify its safety and efficacy in comprehensive patient populations. We aimed to evaluate the real-world use of T-DXd from its market introduction in September 2020 to September 2023 in treating HER2-positive and HER2-low mBC. We specifically analyzed patient demographics, treatment patterns, treatment duration, OS rates, and the incidence of serious adverse events requiring hospitalization. Patients and Methods Study design and participants This was a new-user, descriptive cohort study. The results are reported following the European Society for Medical Oncology Guidance for Reporting Oncology Real-World Evidence (ESMO-GROW).13 Since September 2020, T-DXd has been delivered in France as part of the early access program. We used data from National Early Access and Expensive Drug Cohort (AP-MO ‘Accès précoce—Médicaments Onéreux’, see Supplementary Methods, available at https://doi.org/10.1016/j.esmoop.2024.104083), created by our team, EPI-PHARE, using data from the French National Health Data System (SNDS) covering >99% of the French population (68 million residents).14, 15, 16, 17, 18, 19 The AP-MO cohort also contains medical indications for each treatment delivery. We included patients to whom T-DXd was delivered at least once (ATC L01FD04, Supplementary Table S1, available at https://doi.org/10.1016/j.esmoop.2024.104083) between 30 September 2020 (starting date of the first early access program) and 30 September 2023. The first delivery of T-DXd defined the index date for the study participants, whom we followed until the end of the study (31 December 2023, last day for which data were available at the time of the study) or death. Procedures T-DXd has been approved for four indications within the early access program: (i) HER2-positive mBC at third line or beyond (HER2+ 3L+), following a history of T-DM1 treatment; (ii) HER2-positive mBC at second line or beyond (HER2+ 2L+), without prior T-DM1 treatment; (iii) HER2-low mBC in the second line (HER2-low 2L+); and (iv) HER2-positive metastatic gastric cancer. Medical indications in the AP-MO database were occasionally incomplete or inaccurately recorded, which we addressed using statistical methods supplemented by clinical validation. We restricted our analyses to patients treated for mBC only (Supplementary Methods, Supplementary Figure S1, available at https://doi.org/10.1016/j.esmoop.2024.104083). We analyzed sociodemographic characteristics and baseline comorbidities. Details on breast cancer diagnosis, treatment histories over the past 5 years, and other relevant medical data are detailed in the Supplementary Methods, available at https://doi.org/10.1016/j.esmoop.2024.104083. Discontinuation of T-DXd treatment was defined as a period of 21 days (T-DXd theoretical coverage) plus 60 days (discontinuation washout) with no delivery of T-DXd. The date of treatment discontinuation was defined as the last date of administration plus 21 days (or, if the patient died, the date of death). The time to treatment discontinuation (TTD) was defined as the duration between the index date and the discontinuation date. The time to first subsequent treatment (TFST) was assessed from the initiation of treatment to the initiation of a new treatment or death, whichever occurred first. OS was estimated from the initiation of the T-DXd treatment to the date of death. Patients still on treatment at the end of the study period were censored on 31 December 2023. Putative toxicity-related events were defined as inpatient or outpatient hospitalizations, for cardiac, pulmonary, digestive, hematological or brain-related disorders, sepsis, and pain or fatigue. These events were defined based on the first hospital discharge, with the ICD-10 code as the principal or related diagnosis during the follow-up period (Supplementary Table S2, available at https://doi.org/10.1016/j.esmoop.2024.104083). Statistical analysis Patients’ characteristics at inclusion were described by medical indication. The Kaplan–Meier method was used to estimate TTD, TFST, and OS. The main risk factors for death, treatment discontinuation, treatment modification, and cause-specific hospitalization were estimated using a multivariable Cox proportional hazards regression (see Supplementary Methods, available at https://doi.org/10.1016/j.esmoop.2024.104083). The incidence of hospitalization for each event was reported independently, without prioritization or censoring of the first event to occur. The number of cases before and after T-DXd initiation was compared using the McNemar test,20 with a statistically significant threshold of 5%. As a sensitivity analysis, we computed the incidence of hospitalization within 21 days after the last T-DXd delivery to estimate the hospitalization incidence rate while the patients were being treated. We also computed the incidence of hospitalization and restricted to overnight stays to specifically assess serious adverse events. Analyses were carried out with R21 version 3.5.2 and SAS Enterprise Guide software, version 7.15. Results Between September 2020 and September 2023, our cohort included 5890 patients treated with T-DXd for mBC. Among them, 2010 HER2-positive patients (34.1%) received T-DXd as third-line therapy and 1260 HER2-positive patients (21.4%), as second-line therapy. In addition, 2620 patients (44.5%) were treated with T-DXd for HER2-low mBC (see flow chart in Supplementary Figure S1 and initiation dynamics in Supplementary Figure S2, available at https://doi.org/10.1016/j.esmoop.2024.104083). Detailed demographic and baseline characteristics are summarized in Table 1. The median age at treatment initiation was 60 years [interquartile range, (IQR) 51-69 years], with HER2-low patients slightly older at 61 years (IQR 52-70 years) than HER2-positive patients at 59 years (IQR 51-69 years). Table 1Baseline characteristics of the study population by medical indicationHER2+HER2-lowOverall populationOverallHER2+ 3L+HER2+ 2L+2L+n = 5890n = 3270n = 2010n = 1260n = 2620Year of inclusion 2020189 (3.2)189 (5.8)174 (8.7)15 (1.2)0 (0.0) 2021872 (14.8)872 (26.7)771 (38.4)101 (8.0)0 (0.0) 20221691 (28.7)1263 (38.6)726 (36.1)537 (42.6)428 (16.3) 20233138 (53.3)946 (28.9)339 (16.9)607 (48.2)2192 (83.7)Median age, years (IQR)60.0 (51.0-69.0)59.0 (51.0-69.0)59.0 (51.0-69.0)59.0 (50.0-68.0)61.0 (52.0-70.0)Age categories (years) 18-491193 (20.3)709 (21.7)421 (20.9)288 (22.9)484 (18.5) 50-591671 (28.4)960 (29.4)598 (29.8)362 (28.7)711 (27.1) 60-691587 (26.9)845 (25.8)517 (25.7)328 (26.0)742 (28.3) 70+1439 (24.4)756 (23.1)474 (23.6)282 (22.4)683 (26.1)Sex: female5843 (99.2)3243 (99.2)1989 (99.0)1254 (99.5)2600 (99.2)Duration of breast cancer (years) ≤31544 (26.2)1010 (30.9)504 (25.1)506 (40.2)534 (20.4) 3-82753 (46.7)1469 (44.9)984 (49.0)485 (38.5)1284 (49.0) >81593 (27.0)791 (24.2)522 (26.0)269 (21.3)802 (30.6)Pre-metastatic treatments Breast surgery3120 (53.0)1805 (55.2)1147 (57.1)658 (52.2)1315 (50.2) Radiotherapy3909 (66.4)2229 (68.2)1444 (71.8)785 (62.3)1680 (64.1) Hormonotherapy4309 (73.2)1993 (60.9)1246 (62.0)747 (59.3)2316 (88.4)Number of previous treatments in the past 5 years, median (IQR)2.0 (2.0-4.0)3.0 (2.0-4.0)3.0 (2.0-5.0)1.0 (1.0-2.0)2.0 (2.0-3.0)Number of previous treatments in the past 5 years 095 (1.6)12 (0.4)0 (0.0)12 (1.0)83 (3.2) 11109 (18.8)653 (20.0)29 (1.4)624 (49.5)456 (17.4) 21796 (30.5)946 (28.9)608 (30.2)338 (26.8)850 (32.4) 31356 (23.0)607 (18.6)487 (24.2)120 (9.5)749 (28.6) 4+1534 (26.0)1052 (32.2)886 (44.1)166 (13.2)482 (18.4)HER2-targeting therapy3289 (55.8)3245 (99.2)2010 (100.0)1235 (98.0)44 (1.7) Trastuzumab emtansine2010 (34.1)2010 (61.5)2010 (100.0)0 (0.0)0 (0.0) Trastuzumab and pertuzumab2689 (45.7)2689 (82.2)1593 (79.3)1096 (87.0)0 (0.0) Trastuzumab alone1917 (32.5)1876 (57.4)1354 (67.4)522 (41.4)41 (1.6) Anti-HER2 TKIs700 (11.9)697 (21.3)567 (28.2)130 (10.3)3 (0.1)Other Per os chemotherapy3268 (55.5)1216 (37.2)924 (46.0)292 (23.2)2052 (78.3) Anti-CDK4/62501 (42.5)378 (11.6)207 (10.3)171 (13.6)2123 (81.0) Anti-PD-1 or anti-PD-L1121 (2.1)35 (1.1)19 (0.9)16 (1.3)86 (3.3) PARPi143 (2.4)19 (0.6)9 (0.4)10 (0.8)124 (4.7) Sacituzumab govitecan292 (5.0)8 (0.2)4 (0.2)4 (0.3)284 (10.8)Duration of metastatic disease (years) ≤11067 (18.1)572 (17.5)215 (10.7)357 (28.3)495 (18.9) 1-42509 (42.6)1391 (42.5)907 (45.1)484 (38.4)1118 (42.7) >41661 (28.2)892 (27.3)651 (32.4)241 (19.1)769 (29.4) No metastasis record653 (11.1)415 (12.7)237 (11.8)178 (14.1)238 (9.1)Digestive metastases2543 (43.2)1228 (37.6)740 (36.8)488 (38.7)1315 (50.2)Brain metastases1498 (25.4)1079 (33.0)699 (34.8)380 (30.2)419 (16.0)Other comorbid conditions2304 (39.1)1310 (40.1)792 (39.4)518 (41.1)994 (37.9) Cardiovascular disease811 (13.8)459 (14.0)286 (14.2)173 (13.7)352 (13.4) Diabetes563 (9.6)307 (9.4)179 (8.9)128 (10.2)256 (9.8) Respiratory conditions448 (7.6)234 (7.2)133 (6.6)101 (8.0)214 (8.2) Psychiatric conditions256 (4.3)151 (4.6)101 (5.0)50 (4.0)105 (4.0) Obesity530 (9.0)322 (9.8)194 (9.7)128 (10.2)208 (7.9) Tobacco564 (9.6)335 (10.2)199 (9.9)136 (10.8)229 (8.7)Data are reported as n (%) unless otherwise stated.CDK4/6, cyclin-dependent kinase 4/6; CI, confidence interval; HER2, human epidermal growth factor receptor 2; HER2+ 3L+, HER2-positive patients receiving T-DXd as third-line therapy and beyond, with a history of trastuzumab emtansine; HER2+ 2L+, HER2-positive patients receiving T-DXd as second-line therapy and beyond, with no history of trastuzumab emtansine; HER2-low 2L+, HER2-low patients receiving T-DXd as second-line therapy and beyond; IQR, interquartile range; PARPi, poly (ADP-ribose) polymerase inhibitors; PD-1, Programmed cell death protein 1; PD-L1, Programmed death-ligand 1; SD, standard deviation; T-DXd, trastuzumab deruxtecan; TKIs, tyrosine kinase inhibitors. Baseline characteristics of the study population by medical indication Data are reported as n (%) unless otherwise stated. CDK4/6, cyclin-dependent kinase 4/6; CI, confidence interval; HER2, human epidermal growth factor receptor 2; HER2+ 3L+, HER2-positive patients receiving T-DXd as third-line therapy and beyond, with a history of trastuzumab emtansine; HER2+ 2L+, HER2-positive patients receiving T-DXd as second-line therapy and beyond, with no history of trastuzumab emtansine; HER2-low 2L+, HER2-low patients receiving T-DXd as second-line therapy and beyond; IQR, interquartile range; PARPi, poly (ADP-ribose) polymerase inhibitors; PD-1, Programmed cell death protein 1; PD-L1, Programmed death-ligand 1; SD, standard deviation; T-DXd, trastuzumab deruxtecan; TKIs, tyrosine kinase inhibitors. In the HER2-positive 3L+ group, 79.3% received a pertuzumab-trastuzumab combination as a previous line of treatment, 67.4% trastuzumab, and 28.2% a HER2-targeting tyrosine kinase inhibitor (TKI), while 46.0% underwent oral chemotherapy. For the HER2-positive 2L+ group, 87.0% had been treated with the pertuzumab-trastuzumab combination, 41.4% with trastuzumab, and 10.3% with a HER2-targeting TKI. Among the HER2-low group, 81.0% received a cyclin-dependent kinase 4/6 inhibitor, 78.3% underwent oral chemotherapy, and 10.8% had previously received sacituzumab govitecan ADC. Brain and digestive metastases were identified in a third of patients with HER2 cancer (n = 1079, 33.0% and n = 1228, 37.6%, respectively). Half of the HER2-low patients (50.2%, n = 1315) presented with digestive metastases at treatment initiation, while brain metastases were found in 16.0% of patients (n = 419). The main comorbid conditions at baseline included cardiovascular diseases (n = 811, 13.8%), diabetes (n = 563, 9.6%), and respiratory conditions (n = 448, 7.6%) (Table 1). The median duration of follow-up was 17.0 months (IQR 9.6-24.7 months) for HER2+ 3L+, 11.1 months (IQR 6.6-15.3 months) for HER2+ 2L+, and 7.3 months (IQR 4.9-10.2 months) for HER2-low 2L+ mBC patients. Detailed follow-up characteristics are summarized in Table 2. Table 2Follow-up characteristics of the study population by medical indicationHER2+ 3L+HER2+ 2L+HER2-low 2L+n = 2010n = 1260n = 2620Follow-up duration Mean follow-up, months (SD)17.5 (9.7)11.8 (6.7)7.5 (3.5) Median follow-up, months (Q1-Q3)17.0 (9.6-24.7)11.1 (6.6-15.3)7.3 (4.9-10.2)Overall survival Median overall survival, months (95% CI)30.2 (28.1-33.5)NR16.4 (14.5-NR) Survival at 12 months, probability (95% CI)80.5 (78.7-82.3)85.6 (83.4-87.9)62.3 (59.7-65) Survival at 24 months, probability (95% CI)58.6 (56.1-61.2)66.7 (61.4-72.5)— Mortality at the end of the study, n (%)749 (37.3)209 (16.6)714 (27.3)Treatment duration Median time to treatment discontinuation, months (95% CI)10.8 (10.4-11.5)11.7 (11.0-12.9)5.6 (5.5-5.9) Median number of treatment deliveries (IQR)12.0 (6.0-19.0)10.0 (6.0-17.0)7.0 (4.0-10.0)Median time to first subsequent treatment, months (95% CI)12.4 (11.9-13.1)14.6 (13.8-16.4)8.4 (8.1-8.7)CI, confidence interval; HER2, human epidermal growth factor receptor 2; HER2+ 3L+, HER2-positive patients receiving T-DXd as third-line therapy and beyond, with a history of trastuzumab emtansine; HER2+ 2L+, HER2-positive patients receiving T-DXd as second-line therapy and beyond, with no history of trastuzumab emtansine; HER2-low 2L+, HER2-low patients receiving T-DXd as second-line therapy and beyond; IQR, interquartile range; NR, not reached; SD, standard deviation; T-DXd, trastuzumab deruxtecan. Follow-up characteristics of the study population by medical indication CI, confidence interval; HER2, human epidermal growth factor receptor 2; HER2+ 3L+, HER2-positive patients receiving T-DXd as third-line therapy and beyond, with a history of trastuzumab emtansine; HER2+ 2L+, HER2-positive patients receiving T-DXd as second-line therapy and beyond, with no history of trastuzumab emtansine; HER2-low 2L+, HER2-low patients receiving T-DXd as second-line therapy and beyond; IQR, interquartile range; NR, not reached; SD, standard deviation; T-DXd, trastuzumab deruxtecan. In the group of patients treated for HER2-positive mBC in the third line, 749 (37.3%) died during the follow-up period. For these patients, median OS was 30.2 months [95% confidence interval (CI) 28.1-33.5 months], with a 1-year survival rate of 80.5% (95% CI 78.7% to 82.3%). In the second-line therapy HER2-positive group, 209 (16.6%) died during the follow-up period, the median OS was not assessable, and the 1-year survival rate was 85.6% (95% CI 83.4% to 87.9%). In the group of patients treated for HER2-low mBC, 714 (27.3%) died during the follow-up. Median OS was 16.8 months (95% CI 14.5 months-not reached) and the 1-year survival rate was 62.3% (95% CI 59.7% to 65.0%) (Figure 1). Overall survival by medical indication. (A) HER2+ 3L+: HER2-positive patients receiving T-DXd as third-line therapy and beyond, with a history of trastuzumab emtansine. (B) HER2+ 2L+: HER2-positive patients receiving T-DXd as second-line therapy and beyond, with no history of trastuzumab emtansine. (C) HER2-low 2L+: HER2-low patients receiving T-DXd as second-line therapy and beyond. 2L, second-line; 3L, third-line; HER2, human epidermal growth factor receptor 2; mOS, median overall survival; NR, not reached; T-DXd, trastuzumab deruxtecan. The median TTD was 10.8 months (95% CI 10.4-11.5 months) for HER2+ 3L+ patients, 11.7 months (95% CI 11.0-12.9 months) for HER2+ 2L+ patients, and 5.6 months (95% CI 5.5-5.9 months) for HER2-low 2L+ patients (Table 2, Supplementary Figure S3, available at https://doi.org/10.1016/j.esmoop.2024.104083). For HER2-positive patients receiving third-line therapy, the TFST was 12.4 months (95% CI 11.9-13.1 months), and among treatment switchers, 58.8% of patients transitioned to trastuzumab alone (n = 594/1011) and 18.9% to per os chemotherapy (n = 191/1011). In the second-line HER2-positive group, the median TFST extended to 14.6 months (95% CI 13.8-16.4 months), with trastuzumab alone as the main subsequent therapy (45.4%, n = 189/416). For HER2-low patients, the TFST was 8.4 months (95% CI 8.1-8.7 months), with the most common subsequent therapies including per os chemotherapy (48.7%, n = 399/819) and sacituzumab govitecan (38.7%, n = 317/819) (Supplementary Table S3 and Figure S4, available at https://doi.org/10.1016/j.esmoop.2024.104083). Multivariate Cox regression showed that increasing age, the number of previous treatment lines, digestive or brain metastases, a recent diagnosis for breast cancer and metastatic disease, and the presence of other comorbid conditions were associated with an increased risk of death. Socioeconomic disadvantage was not associated with the effectiveness and safety outcomes. On the contrary, the presence of comorbid conditions, in particular diabetes, cardiovascular diseases and respiratory conditions, and the presence of brain metastases at baseline were the main risk factors of hospitalization for cardiovascular, respiratory, gastric, hematological, and brain disorders and for pain or fatigue (Supplementary Tables S4-S12, available at https://doi.org/10.1016/j.esmoop.2024.104083). The number of hospitalizations, incidence rates of hospitalization, and McNemar’s P values are presented in Tables 3 and 4. Table 3Incidence of hospitalization per 100 person-years for the main toxicity events for HER2+ mBC T-DXd usersToxicityObserved events n (%)PYIncidence in the T-DXd population/100 PY, Poisson (95% CI)Number of past observed events n (%)McNemarP valueCardiac disease188 (5.7)40574.6 (4.0-5.3)151 (4.6)0.04 Heart failure28 (0.9)41630.7 (0.4-1.0)23 (0.7)0.57 Pulmonary embolism34 (1.0)41580.8 (0.6-1.1)23 (0.7)0.17Diseases of the respiratory system308 (9.4)39847.7 (6.9-8.6)141 (4.3)<0.01 Interstitial lung disease36 (1.1)41440.9 (0.6-1.2)4 (0.1)<0.01 Pleural effusion91 (2.8)41312.2 (1.8-2.7)55 (1.7)<0.01 Pneumonitis102 (3.1)41132.5 (2.0-3.0)39 (1.2)<0.01Diseases of the digestive system418 (12.8)387210.8 (9.8-11.9)261 (8.0)<0.01 Acute liver injury15 (0.5)41660.4 (0.2-0.6)14 (0.4)1 Nausea and vomiting45 (1.4)41301.1 (0.8-1.5)25 (0.8)0.02 Ascites54 (1.7)41581.3 (1.0-1.7)9 (0.3)<0.01Hematological disease234 (7.2)40185.8 (5.1-6.6)133 (4.1)<0.01 Thrombocytopenia15 (0.5)41670.4 (0.2-0.6)5 (0.2)0.04Sepsis30 (0.9)41540.7 (0.5-1.0)14 (0.4)0.02Headache, pain, malaise, and fatigue225 (6.9)40395.6 (4.9-6.3)171 (5.2)<0.01Brain-related disorders100 (3.1)41092.4 (2.0-3.0)80 (2.4)0.14Number of cases before and after T-DXd initiation were compared using the McNemar test, with a statistically significant threshold of 5%.Bold values highlight a P-value inferior to 0.01.CI, confidence interval; HER2, human epidermal growth factor receptor 2; mBC, metastatic breast cancer; PY, person-years; T-DXd, trastuzumab deruxtecan. Incidence of hospitalization per 100 person-years for the main toxicity events for HER2+ mBC T-DXd users Number of cases before and after T-DXd initiation were compared using the McNemar test, with a statistically significant threshold of 5%. Bold values highlight a P-value inferior to 0.01. CI, confidence interval; HER2, human epidermal growth factor receptor 2; mBC, metastatic breast cancer; PY, person-years; T-DXd, trastuzumab deruxtecan. Table 4Incidence of hospitalization per 100 person-years for the main toxicity events for HER2-low mBC T-DXd usersToxicityObserved events n (%)PYIncidence in the T-DXd population/100 PY, Poisson (95% CI)Number of past observed events n (%)McNemarP valueCardiac disease119 (4.5)16017.4 (6.2-8.9)83 (3.2)0.01 Heart failure26 (1.0)16381.6 (1.0-2.3)8 (0.3)<0.01 Pulmonary embolism30 (1.1)16331.8 (1.2-2.6)16 (0.6)0.06Diseases of the respiratory system232 (8.9)158214.7 (12.8-16.7)105 (4.0)<0.01 Interstitial lung disease23 (0.9)16411.4 (0.9-2.1)3 (0.1)<0.01 Pleural effusion121 (4.6)16127.5 (6.2-9.0)54 (2.1)<0.01 Pneumonitis52 (2.0)16343.2 (2.4-4.2)26 (1.0)<0.01Diseases of the digestive system306 (11.7)155819.6 (17.5-22.0)154 (5.9)<0.01 Acute liver injury36 (1.4)16412.2 (1.5-3.0)10 (0.4)<0.01 Nausea and vomiting32 (1.2)16362.0 (1.3-2.8)13 (0.5)<0.01 Ascites87 (3.3)16305.3 (4.3-6.6)19 (0.7)<0.01Hematological disease211 (8.1)157413.4 (11.7-15.3)129 (4.9)<0.01 Thrombocytopenia22 (0.8)16401.3 (0.8-2.0)9 (0.3)0.02Sepsis36 (1.4)16402.2 (1.5-3.0)4 (0.2)<0.01Headache, pain, malaise, and fatigue187 (7.1)158411.8 (10.2-13.6)112 (4.3)<0.01Brain-related disorders37 (1.4)16352.3 (1.6-3.1)17 (0.6)<0.01Number of cases before and after T-DXd initiation were compared using McNemar test, with a statistically significant threshold of 5%.Bold values highlight a P-value inferior to 0.01.CI, confidence interval; HER2, human epidermal growth factor receptor 2; mBC, metastatic breast cancer; PY, person-years; T-DXd, trastuzumab deruxtecan. Incidence of hospitalization per 100 person-years for the main toxicity events for HER2-low mBC T-DXd users Number of cases before and after T-DXd initiation were compared using McNemar test, with a statistically significant threshold of 5%. Bold values highlight a P-value inferior to 0.01. CI, confidence interval; HER2, human epidermal growth factor receptor 2; mBC, metastatic breast cancer; PY, person-years; T-DXd, trastuzumab deruxtecan. In the HER2-positive group, for ∼4000 person-years (PY) of follow-up, the incidence rates of hospitalization per 100 PY for cardiac, respiratory, digestive, hematological, and brain disorders were 4.6, 7.7, 10.8, 5.8, and 2.4, respectively. Except for cardiac and brain disorders, the incidence of these events was significantly higher after T-DXd initiation than for the same period before T-DXd treatment. In the HER2-low group, with a follow-up of ∼1500 PY, the incidence rates of hospitalization for cardiac, respiratory, digestive, hematological, and brain disorders were 7.4, 14.7, 19.6, 13.4, and 2.3 per 100 PY, respectively, representing 4.5%, 8.9%, 11.7%, 8.1%, and 1.4% of the HER2-low population. All the studied adverse events showed a higher incidence of hospitalization after T-DXd initiation than for the same period before T-DXd treatment. The results of the sensitivity analyses (Supplementary Tables S13-S18, available at https://doi.org/10.1016/j.esmoop.2024.104083) were consistent with the main findings. Discussion With almost 6000 patients included and 3 years of hindsight, this study provides comprehensive insights into the real-world use of T-DXd by French patients with mBC, highlighting both the evolution of its clinical indications and the associated outcomes. This real-world evidence is particularly valuable, as it underscores the variability in patient profiles, which often differ from the controlled conditions of clinical trials. The patients using T-DXd in our cohort were older than those included in randomized clinical trials (RCTs), with a median age of 60 years for HER2-positive patients and 61 years for HER2-low patients. This contrasts with the median age of 54 years reported for the DESTINY-Breast025 and DESTINY-Breast03 studies22 and 57 years for the DESTINY-Breast04 study.7 Furthermore, 13.8% of patients in our cohort had a cardiovascular comorbid condition, either active or under surveillance, whereas the RCTs excluded patients with active cardiovascular disease. Diabetes and respiratory conditions represented 9.6% and 7.6%, respectively, of the population using T-DXd. In addition, baseline brain metastases were present in one of three of HER2+ 3L/2L+ patients and one of six HER2-low patients, whereas they were less frequent in the RCTs (18%, 16%, and 6.4%, respectively), in which only asymptomatic brain metastases were included. Smaller-scale observational studies have also reported a higher prevalence of brain metastases and older patient ages compared with RCTs.10,11 In our study, we showed that comorbidities and the presence of brain metastases—characteristics for which RCT and real-world patients differ significantly—were independent predictors for either survival or cause-specific hospitalization. Median OS in this real-world cohort was 30.2 months (95% CI 28.1-33.5 months) for HER2+ 3L+ patients and 16.4 months (95% CI 14.5 months-not reached) for HER2-low 2L+ patients. These values are lower than those reported in phase III RCTs, which reported a median OS of 39.2 months (95% CI 32.7 months-not estimated) and 23.4 months (95% CI 20.0-24.8 months), respectively. The median OS for the HER2+ 3L+ group in our study was close to the 29.1 months (95% CI 24.6-36.1 months) observed in the phase II DESTINY-Breast01 study,23 which included heavily pre-treated patients (median of six lines of previous therapy, including T-DM1). In our study, the HER2-positive and HER2-low groups showed a significant increase in hospital admissions for respiratory, digestive, and hematological disorders after T-DXd initiation relative to the prior period. However, the reasons for hospitalization may not be exclusively related to T-DXd, but also to the underlying metastatic disease and associated comorbidities and cancer-related frailty that may lead to fluctuating health and performance status. Hospitalizations due to respiratory disorders extended beyond interstitial lung disease, the most significant toxicity associated with T-DXd in clinical trials22,23 (ILD, 1.1% of hospitalizations for HER2-positive patients in our study) to include pleural effusion (4.6% for HER2-low patients in our study) and infectious pneumonitis, events less noticed in clinical trials. Our real-world data are in accordance with the occurrence of other adverse events reported in the DESTINY-Breast03 trial, including nausea, fatigue, vomiting, and ascites.6 Hospitalization related to hematological disorders, events reported in pharmacovigilance studies,24 doubled among HER2-positive patients after T-DXd administration. None the less, the objective of this study was not to infer causality between T-DXd exposure and hospitalization outcomes but, rather, to assess the overall hospitalization risk to aid in the management of patients treated with T-DXd. These insights call for a deeper investigation into these effects to optimize patient treatment. The main strength of our study was the large size of the population included, made possible by the exhaustiveness of the SNDS database for the French population. We were able to systematically assess the incidence of hospitalization for various events of interest and had access to OS TTD and first subsequent treatment for all patients. In addition, the National Early Access and Expensive Drugs Cohort is an augmented medico-administrative database, as medical indications are recorded for each treatment delivery. We thus were able to analyze T-DXd users according to their RCT indication. Our study nevertheless had limitations. Firstly, while the study extended beyond 3 years, the median follow-up was <1 year due to continuous enrollment and expanding treatment indications. Our findings would thus benefit from an update of T-DXd outcomes in the upcoming months. Secondly, safety events were defined by serious adverse events leading to hospital admission, which could be considered as a proxy for grade 3 or 4 adverse events. Thus, milder safety events were not characterized (specific but not sensitive approach), which explains the lower incidence of adverse events compared with clinical trials or observational studies based on registries. Thirdly, defining treatment indications can be challenging. Our algorithm based on manually recorded data is susceptible to errors. However, by using the most common code and re-classifying based on patient characteristics and medical validation, we should have minimized classification bias. Finally, due to the nature of our database, we lack access to contextual data on treatment delivery, toxicity, treatment discontinuation, and disease progression. Consequently, the progression-free survival could not be directly estimated. Nevertheless, by assessing the OS, TTD, and TFST, we were able to provide valuable insights into treatment outcomes. Conclusions In this large French study evaluating the use of T-DXd in the real world, patients treated with T-DXd were generally older and had more comorbid conditions and more often brain metastases than patients enrolled in registration trials. The rapid expansion of clinical indications of T-DXd, including both HER2-positive and HER2-low mBC in earlier lines, necessitates further validation of its efficacy and safety over a longer follow-up period. This is particularly important for the relatively newer indication of HER2-low mBC. Funding None declared. Disclosure The authors have declared no conflicts of interest. Data Sharing The French National Health Data System (SNDS) is a strictly anonymous database set, including all mandatory health insurance reimbursement data of the French population, in particular data from the processing of health care reimbursements (electronic or paper treatment sheet) and data from health establishments (PMSI). EPI-PHARE has permanent regulatory access to the data from the French National Health Data System (SNDS) via its constitutive bodies ANSM and CNAM, in application of the provisions of the French Decree No. 2016-1871 of 26 December 2016 relating to the processing of personal data called the ‘National Health Data System’ 15, the French law articles Art. R. 1461-1316 and R. 1461-1417 from the French Public Health Code and the French Data Protection Authority (CNIL) decision CNIL-2016-316.18. In accordance with the permanent regulatory access granted to EPI-PHARE via ANSM and CNAM, this work did not require any specific opinion from the French Ethical and Scientific Committee for Research, Studies and Evaluations health (CESREES) nor approval from the CNIL. All requests in the database were made by duly authorized people. The study was registered on the study register of EPI-PHARE concerning studies from SNDS data under the reference T-2024-05-516. In accordance with data protection legislation and the French regulation, the authors are not allowed to release or make public the data from the SNDS. However, any person or structure, public or private, for-profit or nonprofit, is able to access SNDS data in order to carry out a study, research, or an evaluation in the public interest, upon authorization from the French Data Protection Office (CNIL), via the French Health Data Hub (https://www.health-data-hub.fr/). All information for data subject is available on the EPI-PHARE website at https://www.epi-phare.fr/en/regulation-snds/. Supplementary data Supplementary Data References Associated Data Supplementary Materials Supplementary Data