Abstract
Purpose
Patients with human epidermal growth factor receptor 2 (HER2)-positive advanced gastric cancer (AGC) and ascites have poor prognosis and are often excluded from clinical trials. Thus, evidence on the effect of trastuzumab deruxtecan (T-DXd) in this population remains limited.
Materials and Methods
This subgroup analysis of the EN-DEAVOR (UMIN000049032) study assessed outcomes in patients with T-DXd-treated HER2-positive AGC from Japan, stratified by computed tomography-defined ascites severity: none (ascites not detected), low (confined to the upper or lower abdomen), moderate (neither low nor high), and high (diffusely present throughout the abdomen). The outcomes included overall survival (OS), real-world progression-free survival (rwPFS), ascites response rate, and incidence of grade ≥3 adverse events (AEs).
Results
Among 312 patients, 173, 81, 34, and 20 had no, low, moderate, and high ascites severity, respectively. The median OS was 12.35, 6.77, 6.97, and 5.13 months, respectively, and the median rwPFS was 5.32, 3.68, 3.50, and 2.76 months, respectively, for the no, low, moderate, and high severity groups. The overall ascites response rate to ascites severity was 23.0%. The median OS and rwPFS were 11.83 (95% confidence interval [CI], 7.82–18.96) and 6.64 (95% CI, 4.17–9.56) months, respectively, in responders, and 6.28 (95% CI, 4.90–7.03) and 3.15 (95% CI, 2.17–3.94) months, respectively, in non-responders. Grade ≥3 AE incidence was comparable across subgroups.
Conclusions
T-DXd demonstrated efficacy and a manageable safety profile in patients with HER2-positive AGC, including those with severe ascites. The ascites response correlated with improved outcomes.
Keywords: Ascites, Erb-b2 receptor tyrosine kinases, Stomach neoplasms, Trastuzumab deruxtecan, Treatment outcome
INTRODUCTION
Human epidermal growth factor receptor 2 (HER2)-positive advanced gastric cancer (AGC) typically has a poor prognosis, and the presence of ascites further exacerbates this condition [1]. Patients with ascites frequently experience impaired oral intake due to complications such as gastric outlet or intestinal obstruction that worsens their quality of life (QOL) [2,3]. Owing to poor prognosis and QOL, patients with massive ascites are often excluded from clinical trials designed to study new therapies for AGC [4,5]. Consequently, this population remains underrepresented in clinical evidence, creating a significant gap in treatment guidance. Currently, systemic therapy, including cytotoxic chemotherapy, antibody–drug conjugates, and immunotherapy, is weakly recommended in Japanese clinical guidelines for patients with AGC with impaired oral intake or massive ascites because of limited evidence (strength of evidence C) [6]. Therefore, systemic therapy is advised only after a thorough evaluation of the general health and functional status of the patients.
Trastuzumab deruxtecan (T-DXd), an HER2-directed antibody-drug conjugate, was approved in Japan in September 2020 as a third or later treatment option for patients with HER2-positive AGC [7]. In phase 2 of the open-label DESTINY-Gastric01 trial, T-DXd demonstrated efficacy and a manageable safety profile in patients with HER2-positive AGC [5]. However, this clinical trial excluded patients with extensive peritoneal metastasis, severe ascites, or poor performance status. The presence of ascites is usually associated with poor prognosis [1,8]. Ascites can be a sign of advanced disease and peritoneal metastasis; however, many patients without peritoneal metastases may have accompanying ascites at initial diagnosis [9]. Iwai et al. [10] assessed the effect of ascites severity on the overall survival (OS) of patients with AGC. The median OS was only 51.0 days (95% confidence interval [CI], 31.3–70.7) in patients with massive ascites, compared with 216.0 days (95% CI, 0.0–454.3) in those with small ascites and 345.0 days (95% CI, 0.0–808.8) in those with moderate ascites [10]. Thus, ascites severity in gastric cancer can be an important factor influencing treatment decisions and prognosis. EN-DEAVOR, a real-world study conducted in Japan, evaluated the effectiveness and safety of T-DXd in 312 patients with gastric cancer [11]. The prevalence of ascites in the study population was 43.3% (135/312). Previous analyses within this study reported no differences in OS [11] or real-world progression-free survival (rwPFS) [12] based on the presence or absence of ascites. However, a comprehensive evaluation of the differential impact of therapies in patients with AGC and ascites, particularly when stratified by ascites severity, has not yet been performed. Therefore, a subgroup analysis of the EN-DEAVOR study was conducted to evaluate differences in the effectiveness and safety of T-DXd administration in patients with HER2-positive AGC based on ascites severity. Additionally, this analysis explored ascites response rates in patients treated with T-DXd and their subsequent impact on patient prognosis. This study aimed to provide clinically important insights that could inform treatment decisions in patients with ascites.
MATERIALS AND METHODS
Study design and patients
A post hoc subgroup analysis was performed using data from EN-DEAVOR, a retrospective, real-world, observational cohort study conducted across 63 sites in Japan between September 25, 2020, and September 30, 2021 (UMIN000049032) [11]. All study procedures were conducted in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1964 and its later versions. The study protocol was approved by the Institutional Ethical Review Board of the Nagoya University Graduate School of Medicine (approval No. 2022-0170). Informed consent was obtained from all participants before their inclusion in the study. However, opt-out registration was allowed for patients when obtaining written informed consent was challenging due to various reasons. The primary analysis included patients aged ≥20 years who had a histopathologically confirmed diagnosis of HER2-positive unresectable advanced or recurrent gastric or gastroesophageal junction adenocarcinoma (immunohistochemistry [IHC] 3+ or IHC 2+ with in situ hybridization-positive) that had progressed following systemic therapy and had initiated treatment with T-DXd [11].
Patients were stratified into subgroups based on the severity of ascites, which was determined using computed tomography (CT) performed by local investigators in routine clinical practice at each participating site [4]. The criteria for the ascites severity classification are presented in Supplementary Table 1.
Outcomes
The primary endpoints included OS, rwPFS, objective response rate (ORR), and ascites response rate across subgroups defined by ascites severity. The exploratory endpoints included OS and rwPFS based on the ascites response status (ascites response vs. ascites non-response). Patients were categorized into 3 groups: those with ascites response, those without ascites response, and those without ascites at baseline [4]. Ascites response was evaluated based on changes in severity observed on follow-up CT scans assessed by local investigators in routine clinical practice at each participating site.
The proportion of ascites response was determined by dividing the number of patients with the best ascites response (complete or partial response) by the total number of patients who initially presented with ascites [4]. The detailed criteria for the ascites response classification are presented in Supplementary Table 1.
The safety endpoint included the incidence of grade ≥3 adverse events (AEs) across subgroups defined by ascites severity.
Statistical analysis
The sample size for the primary analysis was 300 patients, considering the feasibility of data collection at the participating study sites [11]. The analysis population included all patients enrolled in the main study, excluding those with major protocol deviations or those judged ineligible during a case-review meeting. P-values were calculated using the Cox proportional hazards model. A 2-sided significance level of 5% was used for the hypothesis testing. Multiplicity adjustments were not made. Cases that were not evaluable (Supplementary Table 1) were excluded from the ascites response-stratified survival analyses. Hazard ratios (HRs) with 95% CIs were used to present data for OS and rwPFS. The 95% CIs were estimated using the Greenwood’s formula. Statistical analyses were conducted using SAS software version 9.4 (SAS Institute, Cary, NC, USA).
Ethics approval
All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1964 and later versions. The study protocol was approved by Nagoya University Bioethics Review Board (approval No. 2022–0170).
RESULTS
Patient disposition
Among the 312 patients included in the primary analysis, 81, 34, and 20 had low, moderate, and high ascites severity, respectively (Fig. 1). A total of 173 patients had no detectable ascites at baseline, whereas ascites severity was unknown in 4 patients.
Fig. 1. Patient disposition.
Demographics and baseline clinical characteristics
The demographic and baseline clinical characteristics of the patients according to ascites severity are shown in Table 1. An increasing trend in the proportion of patients with an Eastern Cooperative Oncology Group Performance Status of ≥2 (13.6% low, 32.4% moderate, and 35.0% high ascites severity) and ≥2 metastatic sites (60.5% low, 67.6% moderate, and 80.0% high ascites severity) was observed with higher ascites severity. The proportion of patients who had received previous lines of therapy was comparable across all groups; however, a greater number of patients with no ascites had undergone ≥4 lines of therapy (32.4%) than patients with ascites (14.8% low, 17.6% moderate, and 15.0% high ascites severity). Although the differences were modest, the stomach was more frequently the primary site in patients with ascites (91.4% low, 85.3% moderate, and 90.0% high ascites severity) than in those without ascites (80.3%), whereas the gastroesophageal junction primaries were less frequent in patients with ascites (8.6% low, 14.7% moderate, and 10.0% high ascites severity) than in those without ascites (19.7%). Diffuse-type histology was more prevalent in patients with ascites (39.5% low, 32.4% moderate, and 25.0% high ascites severity) than in those without ascites (17.3%), whereas intestinal-type histology was less common in patients with ascites (45.7% low, 52.9% moderate, and 55.0% high ascites severity) than in those without ascites (59.5%).
Table 1. Demographics and baseline clinical characteristics (n=308).
| Characteristics | Ascites severity | ||||
|---|---|---|---|---|---|
| None (n=173) | Low (n=81) | Moderate (n=34) | High (n=20) | ||
| Sex | |||||
| Male | 138 (79.8) | 58 (71.6) | 24 (70.6) | 12 (60.0) | |
| Female | 35 (20.2) | 23 (28.4) | 10 (29.4) | 8 (40.0) | |
| Age (yr) | |||||
| Mean ± SD | 68.0±10.5 | 66.9±12.2 | 66.1±9.7 | 70.1±8.0 | |
| Median (range) | 71.0 (28–89) | 70.0 (29–83) | 68.5 (27–79) | 73.0 (50–81) | |
| ECOG PS | |||||
| 0 and 1 | 163 (94.2) | 69 (85.2) | 23 (67.6) | 13 (65.0) | |
| ≥2 | 9 (5.2) | 11 (13.6) | 11 (32.4) | 7 (35.0) | |
| Unknown | 1 (0.6) | 1 (1.2) | 0 (0.0) | 0 (0.0) | |
| HER2 status (IHC and ISH): before T-DXd treatment | |||||
| IHC3+ | 114 (65.9) | 61 (75.3) | 25 (73.5) | 15 (75.0) | |
| IHC2+ and ISH+ | 49 (28.3) | 20 (24.7) | 9 (26.5) | 5 (25.0) | |
| Others | 10 (5.8) | 0 (0.0) | 0 (0.0) | 0 (0.0) | |
| Site of primary lesions | |||||
| Stomach | 139 (80.3) | 74 (91.4) | 29 (85.3) | 18 (90.0) | |
| Gastroesophageal junction | 34 (19.7) | 7 (8.6) | 5 (14.7) | 2 (10.0) | |
| Any surgery for primary lesions | |||||
| Yes | 69 (39.9) | 25 (30.9) | 6 (17.6) | 5 (25.0) | |
| No | 104 (60.1) | 56 (69.1) | 28 (82.4) | 15 (75.0) | |
| Histological type of primary lesions | |||||
| Diffuse | 30 (17.3) | 32 (39.5) | 11 (32.4) | 5 (25.0) | |
| Intestinal | 103 (59.5) | 37 (45.7) | 18 (52.9) | 11 (55.0) | |
| Others | 14 (8.1) | 3 (3.7) | 2 (5.9) | 2 (10.0) | |
| Unknown | 26 (15.0) | 9 (11.1) | 3 (8.8) | 2 (10.0) | |
| Number of metastatic organs | |||||
| 0 | 1 (0.6) | 0 (0.0) | 0 (0.0) | 0 (0.0) | |
| 1 | 71 (41.0) | 32 (39.5) | 11 (32.4) | 4 (20.0) | |
| ≥2 | 101 (58.4) | 49 (60.5) | 23 (67.6) | 16 (80.0) | |
| Metastatic site* | |||||
| Lymph nodes | 124 (71.7) | 49 (60.5) | 15 (44.1) | 15 (75.0) | |
| Liver | 90 (52.0) | 42 (51.9) | 15 (44.1) | 10 (50.0) | |
| Peritoneum | 31 (17.9) | 45 (55.6) | 23 (67.6) | 15 (75.0) | |
| Lung | 37 (21.4) | 17 (21.0) | 8 (23.5) | 5 (25.0) | |
| Bones | 17 (9.8) | 6 (7.4) | 3 (8.8) | 2 (10.0) | |
| Brain | 6 (3.5) | 2 (2.5) | 0 (0.0) | 0 (0.0) | |
| Others | 14 (8.1) | 3 (3.7) | 3 (8.8) | 4 (20.0) | |
| Previous lines of therapy | |||||
| 1 and 2 | 77 (44.5) | 47 (58.0) | 23 (67.6) | 12 (60.0) | |
| 3 | 40 (23.1) | 22 (27.2) | 5 (14.7) | 5 (25.0) | |
| ≥4 | 56 (32.4) | 12 (14.8) | 6 (17.6) | 3 (15.0) | |
Data are presented as number (%) unless otherwise specified.
SD = standard deviation; ECOG PS = Eastern Cooperative Oncology Group Performance Status; HER2 = human epidermal growth factor receptor 2; IHC = immunohistochemistry; ISH = in situ hybridization; T-DXd = trastuzumab deruxtecan.
*Patients may have more than 1 metastatic site.
Primary endpoints
The median (95% CI) OS was 12.35 (9.89–13.96), 6.77 (5.85–8.15), 6.97 (4.44–7.82), and 5.13 (2.14–7.23) months in patients with no, low, moderate, and high ascites severity, respectively (Fig. 2). OS was significantly lower in patients with low (HR, 1.88; 95% CI, 1.38–2.57; P<0.0001), moderate (HR, 1.93; 95% CI, 1.28–2.91; P=0.0017), and high (HR, 2.59; 95% CI, 1.48–4.53; P=0.0009) ascites severity than in patients with no ascites (Fig. 2). The median (95% CI) of rwPFS was 5.32 (4.57–6.24), 3.68 (2.53–4.67), 3.50 (1.84–6.24), and 2.76 (0.89–4.99) months in patients with no, low, moderate, and high ascites severity, respectively (Fig. 3). The risk of progression was significantly higher in patients with low (HR, 1.53; 95% CI, 1.16–2.03; P=0.0029) and moderate (HR, 1.64; 95% CI, 1.11–2.42; P=0.0126) ascites severity than in patients with no ascites (Fig. 3). Among the patients with ascites who achieved a target lesion response, the ORR was 38.1% (32/84), with rates of 39.3% (22/56), 33.3% (6/18), and 40.0% (4/10) in the low, moderate, and high ascites severity groups, respectively (Table 2). The ascites response rate across all groups was 23.0% (31/135), with rates of 16.0%, 32.4%, and 35.0% in patients with low, moderate, and high ascites severity, respectively (data not shown).
Fig. 2. Kaplan–Meier analysis of OS based on ascites severity.
The numbers alongside the curves indicate the median OS and 95% CI.
OS = overall survival; CI = confidence interval; HR = hazard ratio.
*P-values were calculated using the Cox proportional hazards model.
Fig. 3. Kaplan–Meier analysis of rwPFS based on ascites severity.
The numbers alongside the curves indicate the median rwPFS and 95% CI.
rwPFS = real-world progression-free survival; CI = confidence interval; HR = hazard ratio.
*P-values were calculated using the Cox proportional hazards model.
Table 2. ORR, DCR, and BOR in patients with a target lesion (n=222).
| Variables | Ascites severity | ||||
|---|---|---|---|---|---|
| None (n=138) | Low (n=56) | Moderate (n=18) | High (n=10) | ||
| ORR | 64 (46.4) | 22 (39.3) | 6 (33.3) | 4 (40.0) | |
| 95% CI* | 37.9–55.1 | 26.5–53.2 | 13.3–59.0 | 12.2–73.8 | |
| DCR | 118 (85.5) | 42 (75.0) | 15 (83.3) | 8 (80.0) | |
| 95% CI* | 78.5–90.9 | 61.6–85.6 | 58.6–96.4 | 44.4–97.5 | |
| BOR | |||||
| CR | 3 (2.2) | 1 (1.8) | 0 (0.0) | 1 (10.0) | |
| PR | 61 (44.2) | 21 (37.5) | 6 (33.3) | 3 (30.0) | |
| SD | 54 (39.1) | 20 (35.7) | 9 (50.0) | 4 (40.0) | |
| PD | 17 (12.3) | 11 (19.6) | 2 (11.1) | 1 (10.0) | |
| NE† | 3 (2.2) | 3 (5.4) | 1 (5.6) | 1 (10.0) | |
Data are presented as number (%) unless otherwise specified.
ORR = objective response rate; DCR = disease control rate; BOR = best overall response; CI = confidence interval; CR = complete response; PR = partial response; SD = stable disease; PD = progressive disease; NE = not evaluable.
*Determined using the Clopper–Pearson method.
†Including non-implemented.
Exploratory endpoints
The demographic and baseline clinical characteristics of the patients stratified by ascites response are shown in Supplementary Table 2. The median OS was 11.83 months (95% CI, 7.82–18.96) in ascites responders, compared with 6.28 months (95% CI, 4.90–7.03) in non-responders (Fig. 4). Ascites responders demonstrated significantly improved survival, with an HR of 0.39 (95% CI, 0.24–0.65; P=0.0003), compared with non-responders (Fig. 4). The median rwPFS was 6.64 months (95% CI, 4.17–9.56) in ascites responders, compared with 3.15 months (95% CI, 2.17–3.94) in non-responders (Fig. 5). Ascites responders had a significantly lower risk of disease progression, with an HR of 0.45 (95% CI, 0.28–0.70; P=0.0005), compared with non-responders (Fig. 5). Among patients with target lesions, the ORR was 64.0% (16/25) in those with an ascites response and 25.5% (12/47) in those without an ascites response (Supplementary Table 3).
Fig. 4. Kaplan–Meier analysis of OS based on ascites response.
The numbers alongside the curves indicate the median OS and 95% CI.
OS = overall survival; CI = confidence interval; HR = hazard ratio.
*P-values were calculated using the Cox proportional hazards model.
Fig. 5. Kaplan–Meier analysis of rwPFS based on ascites response.
The numbers alongside the curves indicate the median rwPFS and 95% CI.
rwPFS = real-world progression-free survival; CI = confidence interval; HR = hazard ratio.
*P-values were calculated using the Cox proportional hazards model.
Safety
The incidence of grade ≥3 AEs stratified by ascites severity is shown in Table 3. A higher proportion of patients with high ascites severity (55.0%) experienced grade ≥3 AEs, followed by those with moderate (47.1%) and low (40.7%) ascites severity. Non-hematotoxicity was more frequent in the high ascites severity group (40.0%) than in the low (28.4%) and moderate (26.5%) ascites severity groups. The most common AEs in patients with high ascites severity were anorexia and decreased white blood cell counts compared to those in patients with no, low, and moderate ascites severity.
Table 3. Incidence rate of grade ≥3 adverse events.
| Variables | Ascites severity | |||
|---|---|---|---|---|
| None (n=173) | Low (n=81) | Moderate (n=34) | High (n=20) | |
| Adverse events | 90 (52.0) | 33 (40.7) | 16 (47.1) | 11 (55.0) |
| Hematotoxicity | 53 (30.6) | 20 (24.7) | 10 (29.4) | 6 (30.0) |
| Non-hematotoxicity | 47 (27.2) | 23 (28.4) | 9 (26.5) | 8 (40.0) |
| Nausea | 8 (4.6) | 2 (2.5) | 1 (2.9) | 1 (5.0) |
| Vomiting | 5 (2.9) | 0 (0.0) | 1 (2.9) | 0 (0.0) |
| Anorexia | 14 (8.1) | 8 (9.9) | 2 (5.9) | 4 (20.0) |
| Malaise | 8 (4.6) | 2 (2.5) | 4 (11.8) | 0 (0.0) |
| Anemia | 17 (9.8) | 7 (8.6) | 3 (8.8) | 2 (10.0) |
| Diarrhea | 2 (1.2) | 3 (3.7) | 0 (0.0) | 0 (0.0) |
| Constipation | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) |
| Fatigue | 2 (1.2) | 1 (1.2) | 0 (0.0) | 0 (0.0) |
| Neutrophil count decreased | 34 (19.7) | 16 (19.8) | 7 (20.6) | 4 (20.0) |
| Platelet count decreased | 8 (4.6) | 3 (3.7) | 1 (2.9) | 1 (5.0) |
| White blood cell count decreased | 4 (2.3) | 4 (4.9) | 0 (0.0) | 2 (10.0) |
| Lymphocyte count decreased | 0 (0.0) | 1 (1.2) | 0 (0.0) | 0 (0.0) |
| Febrile neutropenia | 3 (1.7) | 1 (1.2) | 0 (0.0) | 0 (0.0) |
| Interstitial pneumonia | 10 (5.8) | 2 (2.5) | 1 (2.9) | 1 (5.0) |
| Pneumonitis | 3 (1.7) | 1 (1.2) | 0 (0.0) | 1 (5.0) |
| Infusion-related reaction | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) |
| Others | 13 (7.5) | 7 (8.6) | 5 (14.7) | 2 (10.0) |
Data are presented as number (%).
DISCUSSION
In this study, T-DXd demonstrated a clinically meaningful response in patients with HER2-positive AGC, including those with severe ascites, with comparable safety profiles. Clinically meaningful survival and disease progression rates were found in all ascites severity groups, particularly among the ascites responders. Importantly, the OS and rwPFS outcomes of patients who responded to ascites were comparable to those of patients without ascites, suggesting that ascites resolution may have mitigated the associated poor prognosis.
A previous analysis of the EN-DEAVOR study indicated that the overall ascites status was not a prognostic factor for OS [11] or rwPFS [12]. However, stratification by ascites severity indicated that, irrespective of severity, the presence of ascites was associated with lower survival and a higher risk of progression than the absence of ascites. Before T-DXd treatment, the median (95% CI) of OS (5.13 [2.14–7.23] months) and rwPFS (2.76 [0.89–4.99] months) were shorter in patients with high ascites severity than in those with low and moderate ascites severity. The lower survival and progression rates observed in this group may be attributed to the higher proportion of patients presenting with more severe disease and poorer conditions at baseline [13]. The survival and progression rates observed in this study among patients with high ascites severity were comparable to those reported in previous studies. A real-world study has evaluated the efficacy of T-DXd in patients with AGC, including those with ascites [14]. Of the total number of patients with AGC and ascites, 87.5% had high ascites severity. The study reported a median OS and progression-free survival (PFS) of 3.9 months and 2.75 months, respectively, in patients with ascites. In addition to T-DXd, previous studies have assessed the effectiveness of immune checkpoint inhibitors such as nivolumab in patients with AGC and ascites [15,16]. Takahashi et al. [15] reported a median OS of 3.66 months and median PFS of 1.53 months in patients with ascites. Suzuki et al. [16] classified patients with moderate or massive ascites as having a high ascites burden (HAB); in this group, the study reported a median OS of 2.5 months and a median PFS of 1.0 months following nivolumab treatment. Previous studies have reported a median OS ranging from 4.6 to 6.0 months in patients with AGC and high ascites severity treated with systemic therapy [2,17]. Although direct comparisons are limited owing to differences in study design, the present findings suggest that the effectiveness of T-DXd may be comparable to that of other therapies, given the relatively higher survival and progression rates observed in patients with AGC and high ascites severity.
The comparable survival and disease progression rates observed in this study may be partly attributed to the ascites response rates of 16.0%, 32.4%, and 35.0% observed in the patients with low, moderate, and high ascites severity, respectively. Furthermore, patients who demonstrated an ascites response experienced substantially prolonged OS and rwPFS compared to non-responders. Given that patients with ascites typically have poor prognoses and limited treatment options, the observed benefits of T-DXd in improving both survival and disease progression indicate a potential therapeutic benefit for symptom control.
Matsumoto et al. [14] also reported an ascites response rate of 25% in patients with AGC following T-DXd treatment. However, the study did not explore differences in OS or rwPFS based on the ascites response status. Suzuki et al. [16] reported an ascites response rate of 27% in patients with HAB after nivolumab therapy. In subgroup analyses, patients who showed improvement in ascites also showed a favorable trend in OS [16]. Although the underlying mechanisms contributing to the higher response rate in patients with high ascites severity remain unclear, careful monitoring is warranted when treating these patients with T-DXd.
T-DXd demonstrated a manageable safety profile in different patient groups. Despite some variations, the overall incidence of grade ≥3 AEs was relatively consistent among patients with low (40.7%), moderate (47.1%), and high (55.0%) ascites severity. A modest increase in the incidence of anorexia and decreased white blood cell count were observed in patients with high ascites severity compared to those with moderate or low ascites severity. This may reflect a more fragile clinical status or altered pharmacokinetics in patients with advanced disease, although causality cannot be definitively established from these data [4].
This study provides important real-world evidence regarding the safety and effectiveness of T-DXd in routine clinical practice for patients with HER2-positive AGC and ascites and offers insights that extend beyond the controlled settings of clinical trials. To the best of our knowledge, this is the largest real-world study evaluating the clinical outcomes of T-DXd in this specific patient population. This is also the first study to evaluate the effectiveness and safety of T-DXd in patients with AGC stratified by ascites severity while incorporating ascites response as a clinical outcome.
This study has few limitations. Although the sample size of patients with ascites in this study was relatively large, patients with moderate or severe ascites were enrolled in small numbers, highlighting the need for further investigations with larger sample sizes. Because this was an observational study, the comparative efficacy of T-DXd versus other treatments for patients with ascites could not be definitively established. Brain metastases were captured retrospectively from the medical records, and the protocol for the EN-DEAVOR study did not mandate systematic brain imaging or specific evaluation criteria. Consequently, the reported prevalence rate should be viewed as descriptive and interpreted with caution. In addition, the protocol did not predefine uniform criteria for “peritoneal metastasis”; consequently, peritoneal metastasis status was based on site-level investigator reporting and clinical judgment, which may have introduced inter-site variability. Ascites severity and ascites response were also assessed by site investigators, and inter-observer variability across centers is possible. Furthermore, evaluation of the ascites response requires follow-up CT imaging. This analysis may underrepresent patients with poor prognoses who do not survive long enough for follow-up CT imaging. Additionally, because the dataset captured only the ascites response status without the date of the first response confirmation, a time-dependent analysis could not be performed. Thus, survival comparisons based on ascites response should be interpreted as exploratory data and not as causal data. Detailed data on the use and timing of paracentesis/cell-free and concentrated ascites reinfusion therapies were not systematically collected; however, such cases were classified as not evaluable and excluded from the ascites response-stratified analyses. Future studies should prospectively evaluate the outcomes in patients with ascites and incorporate routine measurements of ascites severity and response throughout the study period.
In conclusion, in this subgroup analysis, T-DXd demonstrated a clinically meaningful and durable antitumor response across all ascites severity groups with a manageable safety profile. Patients who showed an ascites response following T-DXd treatment demonstrated improved survival outcomes, highlighting the ascites response as a potential surrogate marker of clinical benefit. The incidence of grade ≥3 AEs was comparable in patients with different levels of ascites severity. These findings suggest that T-DXd may offer therapeutic value not only for tumor control but also for symptom relief in patients with ascites. However, further validation through prospective studies and randomized controlled trials is essential to confirm these observations, refine patient selection, and optimize treatment strategies for patients with ascites.
ACKNOWLEDGMENTS
Medical writing support was provided by Syed Obaidur Rahman, PhD, from Cactus Life Sciences (part of Cactus Communications), which was funded by Daiichi Sankyo Co., Ltd.
Footnotes
Funding: This study was funded by Daiichi Sankyo Co., Ltd. In March 2019, AstraZeneca entered a global development and commercialization collaboration agreement with Daiichi Sankyo for T-DXd.
Conflict of Interest: Keiko Minashi has received research funding from Ono Pharmaceutical, MSD, Astellas Pharma, BeiGene, Amgen, and Daiichi Sankyo/UCB Japan. Hisato Kawakami is a consultant for Astellas Pharma and Daiichi Sankyo; has received honoraria from Bristol Myers Squibb, Eli Lilly Japan K.K., Ono Pharmaceutical, Daiichi Sankyo, Takeda Pharmaceutical, Teijin Pharma, Otsuka Pharmaceutical, GlaxoSmithKline K.K., Bayer Yakuhin, MSD K.K., Chugai Pharmaceutical, Merck Biopharma, Yakult Pharmaceutical Industry, Taiho Pharmaceutical, and Nippon Kayaku; and has received research funding from Bristol Myers Squibb, Taiho Pharmaceutical, Daiichi Sankyo, Eisai, and Kobayashi Pharmaceuticals. Koki Nakanishi has received honoraria from Taiho Pharmaceutical, Ono Pharmaceutical, Daiichi Sankyo, and Bristol Myers Squibb. Akitaka Makiyama is a consultant for Lilly Japan and has received honoraria from Lilly Japan, Chugai Pharma, Takeda, Daiichi Sankyo, Taiho Pharmaceutical, Ono Pharmaceutical, and Bristol Myers Squibb Japan. Naotoshi Sugimoto has received research funding from MSD, Ono Pharmaceutical, Taiho Pharmaceutical, Lilly Japan, Daiichi Sankyo, Sumitomo Dainippon Pharma, Chugai Pharma, BeiGene, Solasia Pharma, Astellas Pharma, and Eisai. Satoshi Morita has received honoraria from AstraZeneca Japan, Bristol Myers Squibb, Chugai Pharma, Lilly Japan, Merck Sharp & Dohme, Nippon Boehringer Ingelheim, Ono Pharmaceutical, Pfizer, and Eisai; has received research funding from Eisai; and is one of the statistical editor for Gastric Cancer. Yukiya Narita is a consultant for Daiichi Sankyo/AstraZeneca; was at the speakers’ bureaus of Ono Pharmaceutical, Lilly, Taiho Pharmaceutical, Bristol Myers Squibb, and Daiichi Sankyo/AstraZeneca; and has received research funding from Ono Pharmaceutical, Daiichi Sankyo/AstraZeneca, and Astellas Pharma. Takeshi Kawakami has received honoraria from Bristol Myers Squibb and Ono Pharmaceutical. Yasuhiro Kodera is a consultant for Daiichi Sankyo; has received honoraria from Chugai Pharma, Yakult Honsha, Taiho Pharmaceutical, Ono Pharmaceutical, MSD K.K., Tsumura & Co., Daiichi Sankyo, Nippon Kayaku, Lilly Japan, Miyarisan Pharmaceutical, Olympus, AbbVie, Amgen, Bristol Myers Squibb, Edwards Lifesciences, and Covidien; has received research funding from Chugai Pharma, Daiichi Sankyo, Taiho Pharmaceutical, Abbott Japan, Lilly Japan, Kaken Pharmaceutical, Tsumura & Co., Covidien, Nippon Kayaku, Shionogi, Johnson & Johnson, Bayer, TOA Pharmaceuticals, AbbVie, EA Pharma, Pfizer, and Otsuka; and is the Editor-in-Chief of Gastric Cancer. Hiroki Kume, Keita Yamaguchi, and Wataru Hashimoto are employees of Daiichi Sankyo Japan. Kei Muro is a consultant or served as an advisor for Chugai Pharma, AstraZeneca, Ono Pharmaceutical, Amgen, and Solasia Pharma; and has received honoraria from Takeda, Chugai Pharma, Taiho Pharmaceutical, Lilly, Ono Pharmaceutical, Bayer, Sanofi, Bristol Myers Squibb, Daiichi Sankyo/UCB Japan, MSD, Daiichi Sankyo, Pfizer, Sanofi, Taiho Pharmaceutical, Astellas Pharma, Eisai, Chugai Pharma, Amgen, Novartis, and Merck KGaA.
Rin Inamoto, Hirotaka Konishi, Motohiro Imano, Tomohiro Nishina, and Motohisa Hagiwara have nothing to disclose.
- Conceptualization: 1M.K., I.R., 1K.H., N.K., M.A., S.N., 2K.H., M.S., K.Y., 3K.H., Y.K., 2M.K.
- Data curation: 1M.K., I.R., 1K.H., N.K., M.A., S.N., 2K.H., M.S., N.Y., I.M., N.T., K.T., H.M., K.Y., 3K.H., Y.K., H.W., 2M.K.
- Formal analysis: H.W.
- Investigation: 1M.K., I.R., 1K.H., N.K., M.A., S.N., 2K.H., M.S., N.Y., I.M., N.T., K.T., H.M., K.Y., 2M.K.
- Resources: 3K.H., Y.K.
- Writing - original draft: 1M.K., I.R., 1K.H., N.K., M.A., S.N., 2K.H., M.S., N.Y., I.M., N.T., K.T., H.M., K.Y., 3K.H., Y.K., H.W., 2M.K.
- Writing - review & editing: 1M.K., I.R., 1K.H., N.K., M.A., S.N., 2K.H., M.S., N.Y., I.M., N.T., K.T., H.M., K.Y., 3K.H., Y.K., H.W., 2M.K.
1M.K., Keiko Minashi; 2M.K., Kei Muro
1K.H., Hisato Kawakami; 2K.H., Hirotaka Konishi; 3K.H., Hiroki Kume
SUPPLEMENTARY MATERIALS
Criteria for ascites severity and response classification [S1]
Demographic and baseline clinical characteristics stratified by ascites response
ORR in patients with a target lesion stratified by ascites response (n=72)
References
- 1.Zheng LN, Wen F, Xu P, Zhang S. Prognostic significance of malignant ascites in gastric cancer patients with peritoneal metastasis: a systemic review and meta-analysis. World J Clin Cases. 2019;7:3247–3258. doi: 10.12998/wjcc.v7.i20.3247. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Hara H, Kadowaki S, Asayama M, Ooki A, Yamada T, Yoshii T, et al. First-line bolus 5-fluorouracil plus leucovorin for peritoneally disseminated gastric cancer with massive ascites or inadequate oral intake. Int J Clin Oncol. 2018;23:275–280. doi: 10.1007/s10147-017-1198-7. [DOI] [PubMed] [Google Scholar]
- 3.Iwasa S, Nakajima TE, Nakamura K, Takashima A, Kato K, Hamaguchi T, et al. Systemic chemotherapy for peritoneal disseminated gastric cancer with inadequate oral intake: a retrospective study. Int J Clin Oncol. 2011;16:57–62. doi: 10.1007/s10147-010-0135-9. [DOI] [PubMed] [Google Scholar]
- 4.Nakajima TE, Yamaguchi K, Boku N, Hyodo I, Mizusawa J, Hara H, et al. Randomized phase II/III study of 5-fluorouracil/l-leucovorin versus 5-fluorouracil/l-leucovorin plus paclitaxel administered to patients with severe peritoneal metastases of gastric cancer (JCOG1108/WJOG7312G) Gastric Cancer. 2020;23:677–688. doi: 10.1007/s10120-020-01043-x. [DOI] [PubMed] [Google Scholar]
- 5.Shitara K, Bang YJ, Iwasa S, Sugimoto N, Ryu MH, Sakai D, et al. Trastuzumab deruxtecan in previously treated HER2-positive gastric cancer. N Engl J Med. 2020;382:2419–2430. doi: 10.1056/NEJMoa2004413. [DOI] [PubMed] [Google Scholar]
- 6.Japanese Gastric Cancer Association. Japanese Gastric Cancer Treatment Guidelines 2021 (6th edition) Gastric Cancer. 2023;26:1–25. doi: 10.1007/s10120-022-01331-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Kotani D, Shitara K. Trastuzumab deruxtecan for the treatment of patients with HER2-positive gastric cancer. Ther Adv Med Oncol. 2021;13:1758835920986518. doi: 10.1177/1758835920986518. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Maeda H, Kobayashi M, Sakamoto J. Evaluation and treatment of malignant ascites secondary to gastric cancer. World J Gastroenterol. 2015;21:10936–10947. doi: 10.3748/wjg.v21.i39.10936. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Xu M, Liu D, Wang L, Sun S, Liu S, Zhou Z. Clinical implications of CT-detected ascites in gastric cancer: association with peritoneal metastasis and systemic inflammatory response. Insights Imaging. 2024;15:237. doi: 10.1186/s13244-024-01818-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Iwai N, Ohara T, Okuda T, Oka K, Sakai H, Kajiwara-Kubota M, et al. Prognostic value of moderate or massive ascites in patients with advanced gastric cancer. Oncol Lett. 2024;27:116. doi: 10.3892/ol.2024.14249. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Kawakami H, Nakanishi K, Makiyama A, Konishi H, Morita S, Narita Y, et al. Real-world effectiveness and safety of trastuzumab-deruxtecan in Japanese patients with HER2-positive advanced gastric cancer (EN-DEAVOR study) Gastric Cancer. 2025;28:51–61. doi: 10.1007/s10120-024-01555-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Nakanishi K, Sugimoto N, Kodera Y, Kawakami H, Makiyama A, Konishi H, et al. Prognostic factors and treatment response in HER2-positive gastric cancer patients receiving trastuzumab deruxtecan: secondary analysis of the EN-DEAVOR study. ESMO Gastrointest Oncol. 2025;8:100184. doi: 10.1016/j.esmogo.2025.100184. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Matsumoto H, Kawazoe A, Shimada K, Fukuoka S, Kuboki Y, Bando H, et al. A retrospective study of the safety and efficacy of paclitaxel plus ramucirumab in patients with advanced or recurrent gastric cancer with ascites. BMC Cancer. 2018;18:120. doi: 10.1186/s12885-018-4057-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Matsumoto T, Yamamura S, Ikoma T, Kurioka Y, Doi K, Boku S, et al. Real-world data of trastuzumab deruxtecan for advanced gastric cancer: a multi-institutional retrospective study. J Clin Med. 2022;11:2247. doi: 10.3390/jcm11082247. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Takahashi Y, Sunakawa Y, Inoue E, Kawabata R, Ishiguro A, Kito Y, et al. Real-world effectiveness of nivolumab in advanced gastric cancer: the DELIVER trial (JACCRO GC-08) Gastric Cancer. 2022;25:235–244. doi: 10.1007/s10120-021-01237-x. [DOI] [PubMed] [Google Scholar]
- 16.Suzuki H, Yamada T, Sugaya A, Ueyama S, Yamamoto Y, Moriwaki T, et al. Retrospective analysis for the efficacy and safety of nivolumab in advanced gastric cancer patients according to ascites burden. Int J Clin Oncol. 2021;26:370–377. doi: 10.1007/s10147-020-01810-x. [DOI] [PubMed] [Google Scholar]
- 17.Iwasa S, Nakajima TE, Nakamura K, Takashima A, Kato K, Hamaguchi T, et al. First-line fluorouracil-based chemotherapy for patients with severe peritoneal disseminated gastric cancer. Gastric Cancer. 2012;15:21–26. doi: 10.1007/s10120-011-0056-y. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Criteria for ascites severity and response classification [S1]
Demographic and baseline clinical characteristics stratified by ascites response
ORR in patients with a target lesion stratified by ascites response (n=72)