Addressing Existential Distress Through Compassion: Feasibility and Preliminary Effects of Compassion With You (COMwithU), a Compassion‐Based Intervention for Patients With Cancer Authors: Yeonkyu Choi, Jang Sub Kim, Hyejeong Ahn, Soowan Han, Min Cheol Seo, Jae Hyun Ryou, Da Eun Kim, Yoosuk An, Bhumsuk Keam, Tae‐Yong Kim, Eun Seoung Yu, Hye Yoon Park Published: 2026 ABSTRACT ABSTRACTBackgroundCancer‐related distress extends beyond anxiety and depression to existential suffering, including loss of meaning, helplessness, uncertainty, and relational strain. Compassion‐based interventions in psycho‐oncology are promising, but existential outcomes such as demoralization and spiritual well‐being have rarely been examined. Compassion With You (COMwithU) is a brief, five‐session compassion‐based intervention that centers on cultivating self‐compassion, supported by the therapist's compassionate presence and by practices of extending compassion toward others.MethodsThis single‐arm feasibility study evaluated COMwithU in adult patients with cancer using a pre‐post‐1‐month follow‐up design. The intervention comprised five individual 60‐min sessions delivered in person or via videoconferencing. Primary outcomes were feasibility and acceptability. Secondary outcomes included anxiety, depression, spiritual well‐being, demoralization, self‐compassion, and intolerance of uncertainty. Linear mixed models examined changes across baseline, post‐intervention, and follow‐up.ResultsOf 44 participants who started the intervention, 37 attended at least 3 of 5 sessions, yielding a completion rate of 84.1%. Among 34 full completers, 31 (91.2%) completed the post‐intervention assessment, and 30 (88.2%) completed the follow‐up. Helpfulness was rated highly (mean 4.63/5). Anxiety, depression, spiritual well‐being, demoralization, and self‐compassion improved significantly, whereas intolerance of uncertainty did not change significantly. Among participants above the demoralization cutoff at baseline, 68.0% showed reliable improvement, and 56.0% achieved clinically significant change by follow‐up.ConclusionsCOMwithU demonstrated good feasibility, high acceptability, and promising preliminary effects in patients with cancer. The strongest effects were observed for demoralization and spiritual well‐being. Further randomized controlled trials should directly assess self‐compassion, relational compassion processes, and existential outcomes.Trial registrationThis study was registered at the Clinical Research Information Service (CRIS), Republic of Korea (Registration number: KCT0010315) Background The diagnosis and treatment of cancer impose physical burdens and complex psychosocial and spiritual distress [1]. Such multidimensional suffering reduces quality of life and is associated with increased mortality [2, 3]. Psychosocial interventions, including cognitive behavioral therapy, effectively alleviate anxiety and depression [4, 5]. However, existential distress, which encompasses threats to meaning, dignity, and mortality awareness, and persistent uncertainty regarding prognosis, treatment trajectory, and the future, remains a conceptually and empirically distinct clinical domain not fully addressed by improved mood symptoms [6, 7, 8, 9]. Notably, a substantial proportion of patients with cancer experience clinically relevant existential distress even without mental disorders [6], underscoring the need for therapeutic approaches that go beyond symptom reduction to engage fundamental dimensions of meaning, human connection, and shared vulnerability [7, 10]. To address these unmet needs, compassion‐based interventions, including Compassion‐Focused Therapy (CFT) [11] and Mindful Self‐Compassion (MSC) [12], have gained increasing attention in psycho‐oncology [13]. Compassion is commonly defined as a sensitivity to suffering in oneself and others, coupled with a commitment to alleviate and prevent it [14]. By reframing suffering as part of the shared human condition, these approaches may lessen shame and self‐criticism and support psychological adjustment in illness‐imposed limitations on one's life, roles, and future [15]. Lower self‐compassion has been associated with demoralization and psychological distress in patients with cancer [16, 17]. Recent systematic reviews and meta‐analyses indicate that compassion‐based interventions reduce anxiety, depression, and distress and enhance self‐compassion in patients with cancer [13, 18], with a significant reduction in distress reported in a randomized trial of compassion‐focused therapy techniques in patients with cancer recurrence [19]. However, existential outcomes such as demoralization and spiritual well‐being have rarely been examined as primary signals in these trials [10, 13]. Moreover, cultivating self‐compassion may be particularly challenging for patients with cancer: individuals experiencing severe distress often fear or resist directing compassion toward themselves [20]. In addition, cancer‐related suffering is inherently relational: patients frequently contend with guilt toward burdened family members, unexpressed gratitude, strained relational dynamics, and the painful renegotiation of caregiving roles, concerns that self‐compassion alone may not fully address [7, 21, 22]. Cultivating compassion toward others may foster shared humanity and is associated with greater psychological well‐being [23]. Furthermore, the therapist's compassionate presence may serve as a particularly powerful relational resource for patients who are physically depleted and socially isolated by their illness [24]. This provides a safe space in which patients can begin to accept their situation, soothe difficult emotions, and extend kindness toward themselves [14]. These relational dimensions may be particularly salient in the Korean context, where family centrality and relational obligations are prominent features of the cancer experience [21, 25, 26]. Accordingly, we developed Compassion With You (COMwithU), a brief, five‐session individual intervention centered on cultivating self‐compassion in patients with cancer, informed by MSC [12] and incorporating the therapist's compassionate presence and practices of compassion toward others as relational supports. As a preliminary step before a fully powered randomized controlled trial, we evaluated its feasibility, acceptability, and preliminary effects using a single‐arm pre–post–follow‐up design [27]. Methods Study Design This single‐arm feasibility study evaluated the feasibility, acceptability, and preliminary effects of COMwithU using a pre–post–1‐month follow‐up design, with supplementary open‐ended participant feedback. The study was conducted in two sequential phases. The first phase was developmental, comprising 30 participants, during which the workbook, case report form, and assessment battery were refined based on clinical observations and participant feedback. The outcome assessment battery was substantially revised between the two phases, precluding direct comparison of data across phases. The subsequent feasibility testing phase comprised 50 participants enrolled after finalization of the protocol and study procedures. Based on a one‐sample proportion test for the primary feasibility endpoint (program completion rate ≥ 80%; α = 0.05), a target enrollment of 50 was set, anticipated to yield approximately 35 evaluable participants with 83% power [27]. Accordingly, all feasibility, acceptability, and preliminary outcome analyses reported here are based on participants who received the finalized intervention using the finalized assessment procedures. This study was approved by the Institutional Review Board of Seoul National University Hospital (IRB No. 2212–059–1384) and was registered at the Clinical Research Information Service (CRIS; KCT0010315). All participants provided written informed consent. Reporting follows the CONSORT extension for pilot and feasibility trials [28]. Participants and Recruitment Patients with cancer were recruited from outpatient clinics at Seoul National University Hospital or via hospital posters between May 2024 and December 2025. Eligible participants were adults (≥ 19 years) who understood the study procedures and provided written informed consent. We excluded patients with clinically significant agitation, irritability, suicide risk, or aggression; delirium; significant cognitive impairment; or severe physical or mental frailty hindering participation. No restrictions were placed on cancer type or stage to maximize generalizability in this feasibility stage. Intervention COMwithU is a manualized five‐session individual psychological intervention designed to help patients with cancer embrace their suffering with compassion and reframe illness narratives. The program draws primarily on MSC [12], adapting its core meditation practices and approach of turning toward suffering with acceptance, openness, and curiosity. Building on this foundation, we incorporated two relational components—the therapist's compassionate presence and structured practices of compassion toward others—designed to support the cultivation of self‐compassion. Given the physical and psychological vulnerability of patients with cancer, we proposed that the therapist's compassionate presence may provide a relational foundation that helps patients begin to cultivate self‐compassion throughout the intervention. Accordingly, it comprises four core therapeutic components: (1) self‐compassion cultivation, drawn from MSC; (2) compassion directed toward others; (3) narrative techniques for reframing illness experiences; and (4) the therapist's compassionate presence, a therapeutic stance characterized by mindfulness, respect, and acceptance, rooted in the spiritual care tradition [29]. The five sessions were sequenced to progressively expand the focus of compassion, beginning with self‐narrative and self‐compassion, then engagement with suffering, compassion for others, and finally an integrative reflective exercise (Supporting Information S2). Each 60‐min session was delivered individually, either in‐person or via video conferencing (Zoom, with camera on) to accommodate patients' physical limitations and treatment schedules. All therapists completed a training program specialized in psycho‐oncology and compassion‐based interventions, consisting of three phases: (1) didactic instruction covering psychological and existential issues in patients with cancer, the theoretical foundations of MSC, and the spiritual care tradition underlying compassionate presence; (2) experiential training through observed co‐facilitation of pilot sessions with patients with cancer; and (3) feedback provided by the principal investigator through case‐based discussion of session verbatim transcripts. Four therapists completed 30 pilot cases before independent delivery; their readiness to proceed was determined by the principal investigator based on their performance. Ongoing clinical support was provided throughout the study period through weekly group supervision addressing protocol adherence and maintaining compassionate presence as a therapeutic stance. Measures Primary Outcomes: Feasibility and Acceptability Feasibility was assessed in terms of recruitment, retention, and session completion [27]. Retention was defined as the proportion of enrolled participants who completed all five sessions and provided the post‐intervention assessment (T2). Session completion was defined as the proportion of participants who attended at least three of the five sessions among those who started the intervention, as the first three sessions covered the core therapeutic components of the program, including self‐compassion and compassionate engagement with suffering. A priori feasibility benchmarks were set at ≥ 80% for program completion and ≥ 70% for assessment retention, informed by published retention rates in cancer psychosocial trials [30]. Acceptability was assessed at post‐intervention using a program evaluation questionnaire consisting of nine five‐point Likert‐scale items. These included four items assessing overall program evaluation (helpfulness, necessity, benefit, and willingness to recommend) and five items assessing the helpfulness of each session component, along with three open‐ended questions on perceived benefits and suggestions for improvement. Secondary Outcomes: Preliminary Effects Anxiety and Depression The Korean version of the Hospital Anxiety and Depression Scale (K‐HADS) was used to assess anxiety (7 items) and depression (7 items) separately, each rated on a four‐point scale (0–3). A score of ≥ 8 on either subscale indicates clinically significant symptoms [31]. In the present study, Cronbach's alpha was 0.90 for anxiety and 0.81 for depression. Spiritual Well‐Being The Korean version of the Functional Assessment of Chronic Illness Therapy–Spiritual Well‐Being Scale (FACIT‐Sp) assessed meaning/peace (8 items) and faith (4 items) on a five‐point scale (0–4) [32]. In the present study, Cronbach's alpha was 0.90. Existential Distress Demoralization was assessed using the Korean version of the Demoralization Scale‐II (DS‐II‐Kr), a 16‐item measure rated on a 3‐point scale (0–2; total range 0–32) comprising Meaning and Purpose and Distress and Coping Ability subscales [33]. A cut‐off score of ≥ 10 indicates clinically significant demoralization. In the present study, Cronbach's alpha was 0.94. Self‐Compassion The Korean version of the Self‐Compassion Scale (K‐SCS) was used to assess self‐compassion [34]. The 26‐item scale comprises six subscales on a 5‐point Likert scale (1–5), with higher scores indicating greater self‐compassion. In the present study, Cronbach's alpha was 0.95. Intolerance of Uncertainty The Intolerance of Uncertainty Scale‐12 (IUS‐12) assessed intolerance of uncertainty using 12 items on a 5‐point scale (1–5) [35]. In the present study, Cronbach's alpha was 0.90. Data Analysis Statistical analyses were performed using R (version 4.5.2). Descriptive statistics (means and standard deviations [SD] or median and interquartile range [IQR]) were calculated for demographic characteristics and feasibility metrics. To examine changes in outcome measures across three time points (T1, T2, T3), linear mixed models (LMMs) were employed using the lme4 and lmerTest packages. Time was entered as a fixed effect, and random intercepts were included for each participant to account for individual variability at baseline. Age and gender were included as covariates. LMMs use all available data under the missing‐at‐random assumption, accommodating incomplete observations without listwise deletion. Pairwise comparisons between time points were conducted using estimated marginal means with Holm correction for multiple comparisons via the emmeans package. Paired Cohen's d (dz = M_diff/SD_diff) was calculated from within‐subject difference scores among participants with complete data at both time points. Because LMMs and effect sizes use different analytic samples, minor discrepancies may arise. All tests were two‐tailed with a significance level of p < 0.05. Open‐ended responses were summarized descriptively. Session‐specific helpfulness was compared using a Friedman test with post‐hoc Wilcoxon signed‐rank tests (Holm correction). Reliable change index (RCI) analyses were conducted for HADS‐Anxiety, HADS‐Depression, and DS‐II. Reliable improvement was defined as an RCI exceeding the threshold in the direction of symptom improvement; clinically significant change required reliable improvement and a score below the clinical cutoff at the corresponding follow‐up assessment (computational details in Table 1 note). As a secondary sensitivity analysis, LMMs were repeated within the subgroup of participants scoring at least the clinical cutoffs at baseline. TABLE 1Reliable change index and clinical significance analyses.HADS‐anxietyHADS‐depressionDS‐IICutoff≥ 8≥ 8≥ 10Cronbach's α 0.8950.8110.943Baseline clinical, n (%)27/36 (75.0)27/36 (75.0)29/37 (78.4)S_diff2.212.702.76Min. Reliable change (pts)4.335.305.41T1 → T2 n (Available pairs)222225Reliably improved5 (22.7)2 (9.1)13 (52.0)Clinically significant a 3 (13.6)2 (9.1)10 (40.0)No reliable change17 (77.3)19 (86.4)12 (48.0)Reliably deteriorated0 (0.0)1 (4.5)0 (0.0)T1 → T3 n (Available pairs)192225Reliably improved8 (42.1)3 (13.6)17 (68.0)Clinically significant a 3 (15.8)3 (13.6)14 (56.0)No reliable change11 (57.9)19 (86.4)8 (32.0)Reliably deteriorated0 (0.0)0 (0.0)0 (0.0) Note: Based on Jacobson & Truax (1991). RCI = (Xpost − Xpre)/S_diff; S_diff = √2 × SD_pre × √(1 − α). Reliably Improved: |RCI| > 1.96 in the direction of improvement. a Subset of reliably improved; requires both reliable improvement and post‐score below clinical cutoff. Analyses are restricted to participants scoring at or above the clinical cutoff at baseline. Values are n (%). One participant did not complete the HADS at baseline (denominator = 36). Item‐level missing data on HADS‐Anxiety at T3 resulted in fewer available pairs (n = 19) than for depression (n = 22). Cutoff crossing analyses are reported in Supporting Information S2. Reliable change index and clinical significance analyses. Note: Based on Jacobson & Truax (1991). RCI = (Xpost − Xpre)/S_diff; S_diff = √2 × SD_pre × √(1 − α). Reliably Improved: |RCI| > 1.96 in the direction of improvement. Subset of reliably improved; requires both reliable improvement and post‐score below clinical cutoff. Analyses are restricted to participants scoring at or above the clinical cutoff at baseline. Values are n (%). One participant did not complete the HADS at baseline (denominator = 36). Item‐level missing data on HADS‐Anxiety at T3 resulted in fewer available pairs (n = 19) than for depression (n = 22). Cutoff crossing analyses are reported in Supporting Information S2. Results Participant Characteristics Table 2 presents the baseline demographic and clinical characteristics of the 37 participants. Most participants were women (n = 30, 81.1%) with a mean age of 48.2 ± 11.4 years. Most participants were highly educated, with 33 of 37 (89.2%) having a college education or higher. The most common cancer type was breast cancer (n = 20, 54.1%); 56.8% had early‐stage cancer (stages 0–II) and 43.2% advanced‐stage (stages III–IV). The median time from diagnosis was 14.0 months (IQR: 6.0–28.0). Seventeen participants (45.9%) were taking psychotropic medication. Most (75.7%) received in‐person intervention entirely, while the remaining 24.3% received mixed interventions (in‐person and video conferencing). TABLE 2Baseline demographic and clinical characteristics (N = 37).Variable n (%) or M ± SDSociodemographic characteristicsGender, female30 (81.1)Age, years, M ± SD48.2 ± 11.4Married21 (56.8)Education≤ high school4 (10.8)College/University19 (51.4)Graduate school14 (37.8)Currently employed or in school11 (29.7)Annual household income c < $27,00011 (29.7)$27,000–55,00012 (32.4)≥ $55,00014 (37.8)ReligionNone15 (40.5)Catholic8 (21.6)Protestant7 (18.9)Buddhist5 (13.5)Other2 (5.4)Clinical characteristicsCancer typeBreast20 (54.1)Pancreaticobiliary3 (8.1)Lung2 (5.4)Gynecological2 (5.4)Head and neck2 (5.4)Gastric1 (2.7)Other7 (18.9)Cancer stage02 (5.4)I11 (29.7)II8 (21.6)III6 (16.2)IV10 (27.0)Recurrence a 6 (16.2)Metastasis b 9 (24.3)Current treatment (multiple response)Hormone therapy9 (24.3)Cytotoxic chemotherapy9 (24.3)Surgery7 (18.9)Targeted therapy5 (13.5)Immunotherapy3 (8.1)Radiation therapy2 (5.4)Time since diagnosis, months, median (IQR)14.0 (6.0–28.0)Comorbidity13 (35.1)Psychiatric medication use17 (45.9)Abbreviations: IQR = interquartile range; M = mean; SD = standard deviation. a Unknown for 1 participant. b Unknown for 2 participants. c Assessed as monthly household income in Korean Won (KRW) and converted to approximate annual USD equivalents (1 USD ≈ 1300 KRW). Baseline demographic and clinical characteristics (N = 37). Abbreviations: IQR = interquartile range; M = mean; SD = standard deviation. Unknown for 1 participant. Unknown for 2 participants. Assessed as monthly household income in Korean Won (KRW) and converted to approximate annual USD equivalents (1 USD ≈ 1300 KRW). Feasibility and Acceptability Participant Flow and Feasibility Participant flow and feasibility outcomes are summarized in Figure 1. Of the 50 patients assessed for eligibility, all consented to participate; 44 initiated the intervention, whereas 6 withdrew before the first session. Of these 44 participants, 37 attended at least three of the five sessions and were classified as session completers, yielding a session completion rate of 84.1%, which exceeded the prespecified threshold of 80%. Among these 37 participants, 34 completed all five sessions and were classified as full completers, whereas three participants completed at least three sessions but did not complete the full program. Of the 34 full completers, 31 (91.2%) provided post‐intervention data (T2), exceeding the prespecified retention threshold of 70%, and 30 (88.2%) also provided follow‐up data (T3). Considering all 44 participants who initiated the intervention, the overall post‐intervention assessment rate was 70.5% (31/44). There were no significant differences in age, sex, or any baseline outcome variables between participants with missing data and those with complete data (all p > 0.10). Participant flow diagram. Flow of participants through eligibility assessment, intervention initiation, program completion, and post‐intervention (T2) and 1‐month follow‐up (T3) assessments in the feasibility study of COMwithU. Of the 44 participants who initiated the intervention, 37 attended at least three of 5 sessions and were classified as session completers; among them, 34 completed all 5 sessions. Missingness at T2 and T3 partially overlapped: 29 participants provided data at both time points, 2 at T2 only, 1 at T3 only, and 2 at neither. COMwithU, Compassion With You; T2, post‐intervention; T3, 1‐month follow‐up. Acceptability Of the 31 participants who provided post‐intervention data, 30 completed the Likert‐scale evaluation (one participant with missing data was excluded). Overall perceived helpfulness of the program was rated favorably (4.63 ± 0.56 on a five‐point scale). Perceived necessity of the program (4.60 ± 0.67), perceived benefit (4.60 ± 0.72), and willingness to recommend (4.47 ± 0.97) were also high. Session‐specific helpfulness ratings ranged from 4.20 to 4.73, with all five sessions rated > 4.0. On the Friedman test, perceived helpfulness across sessions differed significantly (χ2(4) = 26.82, p < 0.001); post‐hoc comparisons showed that sessions addressing compassion toward others and engagement with suffering were rated somewhat lower than self‐compassion–focused sessions. In open‐ended feedback (Supporting Information S2), the most frequently reported benefits were increased self‐kindness, reduced self‐criticism, and practical use of meditation exercises. Approximately one fourth of participants (7/30) further described relational shifts toward family members. Participants primarily suggested additional sessions and longer‐term follow‐up. No adverse events or unintended effects were reported during the intervention or follow‐up period. Changes in Psychological Outcomes Table 3 presents the results of LMMs examining changes in psychological outcomes across the three time points. Five of six psychological outcomes demonstrated statistically significant time effects. TABLE 3Changes in psychological outcomes across time points: Linear mixed model results.OutcomeT1T2T3F P d12 d13 p T1–T2 p T1–T3HADS‐anxiety10.3 (4.8)8.0 (4.7)8.0 (4.5)9.67< 0.001*** 0.840.610.001** 0.001** HADS‐depression9.9 (4.4)8.0 (4.1)8.1 (4.5)6.130.004** 0.510.490.007** 0.013* FACIT‐Sp total21.2 (10.5)28.8 (10.8)28.6 (10.1)12.96< 0.001*** 0.730.72< 0.001*** < 0.001*** Meaning/Peace 15.2 (7.1)20.6 (7.0)20.3 (6.6)14.61< 0.001*** 0.830.88< 0.001*** < 0.001*** Faith 6.0 (4.9)8.3 (5.1)8.4 (5.0)4.820.011* 0.380.400.024* 0.024* DS‐II total17.3 (8.2)12.6 (8.3)11.3 (9.3)14.33< 0.001*** 0.731.02< 0.001*** < 0.001*** Meaning/Purpose 6.8 (5.0)4.4 (4.7)3.8 (4.5)11.83< 0.001*** 0.710.89< 0.001*** < 0.001*** Distress/Coping 10.5 (3.7)8.3 (4.2)7.5 (5.2)13.02< 0.001*** 0.670.910.001** < 0.001*** SCS Total76.8 (20.9)85.4 (22.6)87.3 (24.4)8.65< 0.001*** 0.520.850.004** 0.001** Self‐Kindness 12.5 (4.5)14.7 (5.7)15.2 (5.8)8.51< 0.001*** 0.530.690.003** 0.001** Self‐Judgment 17.4 (5.4)18.8 (4.9)19.2 (4.8)3.660.032* 0.360.520.0660.053 Common Humanity 9.6 (3.8)11.3 (4.5)11.5 (5.1)5.910.005** 0.490.630.013* 0.010* Isolation 14.4 (4.2)14.7 (4.3)15.0 (4.8)0.620.5420.090.261.0000.815 Mindfulness 11.2 (4.1)13.0 (4.1)12.8 (4.2)4.950.010* 0.440.440.021* 0.024* Over‐Identification 11.8 (4.1)12.9 (4.3)13.7 (4.0)5.950.004** 0.340.710.1010.003** IUS‐1238.5 (8.7)37.2 (11.3)36.8 (11.6)1.820.1710.250.400.3020.246 Note: Values are M (SD). Model: outcome ∼ time + age + gender + (1|ID), Satterthwaite degrees of freedom. p values for pairwise comparisons are Holm‐adjusted. d = paired Cohen's d (absolute values). Positive d indicates improvement (score decrease for HADS‐A/D, DS‐II, IUS; score increase for FACIT‐Sp, SCS). All T2–T3 comparisons were non‐significant (p > 0.05). Subscales are indented under their parent scale. * p < 0.05. ** p < 0.01. *** p < 0.001. Changes in psychological outcomes across time points: Linear mixed model results. Note: Values are M (SD). Model: outcome ∼ time + age + gender + (1|ID), Satterthwaite degrees of freedom. p values for pairwise comparisons are Holm‐adjusted. d = paired Cohen's d (absolute values). Positive d indicates improvement (score decrease for HADS‐A/D, DS‐II, IUS; score increase for FACIT‐Sp, SCS). All T2–T3 comparisons were non‐significant (p > 0.05). Subscales are indented under their parent scale. p < 0.05. p < 0.01. p < 0.001. Anxiety HADS‐Anxiety showed a significant time effect, F(2, 57.0) = 9.67, p < 0.001, with a large effect from baseline to post‐intervention (d = 0.84) that was maintained at follow‐up (d = 0.61). Pairwise comparisons confirmed significant reductions from T1 to T2 (p = 0.001) and T1 to T3 (p = 0.001), with no significant change from T2 to T3 (p = 0.846). Depression HADS‐Depression also showed a significant time effect, F(2, 59.3) = 6.13, p = 0.004, with medium effect sizes (d = 0.51 at T2, d = 0.49 at T3). T1–T2 (p = 0.007) and T1–T3 (p = 0.013) comparisons were significant, while T2–T3 was not (p = 0.766). Spiritual Well‐Being FACIT‐Sp Total scores increased significantly, F(2, 60.5) = 12.96, p < 0.001, with medium‐to‐large effects (d = 0.73 and 0.72). At the subscale level, the Meaning/Peace subscale showed the largest improvement (d = 0.83–0.88, both p < 0.001), and the Faith subscale also significantly improved (d = 0.38–0.40, both p = 0.024). Demoralization DS‐II Total scores decreased significantly, F(2, 60.1) = 14.33, p < 0.001, yielding the largest effect size observed across all outcome measures at follow‐up (d = 1.02). Notably, the Meaning/Purpose (d = 0.71–0.89) and Distress/Coping (d = 0.67–0.91) subscales demonstrated significant reductions (both p < 0.001 at T1–T3). The effect size continued to increase from T2 to T3, suggesting sustained or continuing improvement. Self‐Compassion SCS total scores increased significantly, F(2, 59.6) = 8.65, p < 0.001, with effects growing from medium at T2 (d = 0.52) to large at T3 (d = 0.85). At the subscale level, Self‐Kindness, Common Humanity, Mindfulness, and Over‐Identification significantly improved, whereas Self‐Judgment reached only a trend level, and Isolation did not change significantly (Table 3). Notably, Over‐Identification showed a delayed effect pattern, with a significantly large T1–T3 effect (d = 0.71, p = 0.003) despite a non‐significant T1–T2 change. Intolerance of Uncertainty IUS‐12 scores did not significantly change with time, F(2, 58.1) = 1.82, p = 0.171, with small effect sizes (d = 0.25–0.40). Across all significant outcomes, improvements occurred primarily from T1 to T2, with T2–T3 comparisons uniformly non‐significant (all p > 0.05), indicating that treatment gains were maintained at 1‐month follow‐up. Individual‐level RCI analyses indicated that the proportion of participants with reliable improvement in demoralization was numerically higher at follow‐up than at post‐intervention. Clinical Significance Table 1 presents RCI and clinical significance analyses for participants above the clinical cutoff at baseline. Demoralization showed the strongest individual‐level change: 52.0% demonstrated reliable improvement by T2, increasing to 68.0% by T3, with clinically significant change in 40.0% and 56.0% of participants, respectively, and no reliable deterioration. For HADS‐Anxiety and HADS‐Depression, reliable improvement rates were lower (9%–42%), consistent with the smaller group‐level effect sizes observed for these outcomes. Clinical Subgroup Analyses Separate LMMs for participants with clinically elevated symptoms (HADS‐A or HADS‐D ≥ 8; n = 27 each) confirmed significant time effects and pairwise reductions consistent with full‐sample findings (all p < 0.01). Additional exploratory interaction analyses by psychiatric medication use, cancer stage, delivery format, and model diagnostics are reported in the Supporting Information S1. Discussion This study examined the feasibility, acceptability, and preliminary effects of COMwithU, a compassion‐based intervention centered on self‐compassion and supported by the therapist's compassionate presence and compassion‐toward‐others practices, for patients with cancer. The intervention demonstrated good feasibility, with a session completion rate of 84.1% exceeding the pre‐specified threshold of 80%, and high acceptability, with an overall helpfulness rating of 4.63 out of 5. Preliminary outcome analyses showed significant improvements in anxiety, depression, self‐compassion, spiritual well‐being, and demoralization, whereas intolerance of uncertainty did not significantly change. Among these outcomes, demoralization and spiritual well‐being had the most notable effects, suggesting that COMwithU may be particularly relevant to existential dimensions of distress. The feasibility findings are clinically meaningful given the fluctuating physical condition and treatment burden of patients with cancer [30]. COMwithU was designed to maximize practicality through an individual format, a brief five‐session structure, and flexible hybrid delivery. The observed completion and satisfaction rates suggest this design may be suitable in psycho‐oncology, where longer interventions are difficult to sustain [30]. Notably, non‐completion appeared to reflect worsening medical condition or early withdrawal rather than dissatisfaction with the program itself, which further supports the acceptability of the intervention. Although significant improvements were observed in anxiety and depression, the largest and clinically most meaningful changes were found in demoralization and spiritual well‐being. These domains map closely onto existential distress—loss of meaning, helplessness, and disconnection from sources of value [7, 9, 36]. Consistent with evidence that spiritual well‐being is inversely associated with demoralization and may buffer against it in patients with cancer [37], the parallel improvements across these two domains point to a shared existential pathway rather than two independent effects. This interpretation was reinforced by the clinical significance analyses: among participants above the demoralization cutoff at baseline, 68.0% improved and 56.0% showed a clinically significant change on follow‐up, with no deterioration. The observed effects for demoralization and spiritual well‐being were of a magnitude within the range reported in meta‐analyses of compassion‐based [13] and existential [10] interventions in oncology, although uncontrolled pre–post effects are inflated by regression to the mean, expectancy, and natural recovery and preclude direct comparison with controlled between‐group estimates. The prominence of existential outcomes may reflect conceptual alignment between the intervention and existential distress. Whereas cognitive behavioral approaches target symptom reduction through restructuring [4, 5], COMwithU engages the capacity to face suffering with kindness and openness rather than avoidance [12]—an approach particularly suited to existential suffering, which may be eased through a changed relationship with the experience itself [9, 15]. Improvements in spiritual well‐being, particularly the Meaning and Peace subscales, are consistent with this interpretation. This suggests that participants developed a more compassionate way of facing existential concerns. Together, these findings suggest COMwithU may be particularly relevant to existential distress, while also benefiting broader psychological symptoms. Several pathways may hypothetically link compassion processes to the observed existential improvements, although none were formally tested. Demoralization involves helplessness, subjective incompetence, and loss of meaning [36]; self‐kindness and the recognition of suffering as part of the shared human condition may directly counter the self‐critical appraisals that sustain it, restoring a sense of agency and meaning [15]. In parallel, the capacity to turn toward suffering with acceptance rather than avoidance—reflected in improved mindfulness and reduced over‐identification—may support meaning‐focused coping, consistent with the largest gains being observed on the Meaning/Peace subscale of spiritual well‐being. Notably, the SCS Isolation subscale did not change and intolerance of uncertainty remained stable, suggesting that the program's preliminary benefits operated primarily through self‐kindness and decentering rather than through reduced felt isolation. The therapist's compassionate presence and the compassion‐toward‐others practices were designed as relational supports for these processes; although a subset of participants spontaneously described relational shifts in open‐ended feedback, these components were not directly measured, and their specific contribution remains an open question for the planned randomized controlled trial. Regarding session‐level acceptability, sessions addressing compassion toward others and engagement with suffering received somewhat lower ratings, likely reflecting higher emotional demands rather than reduced therapeutic value. The non‐significant finding for intolerance of uncertainty may help clarify COMwithU's therapeutic focus. Intolerance of uncertainty is generally conceptualized as a stable cognitive vulnerability [38], whereas COMwithU is oriented toward transforming one's relationship to suffering through compassion and meaning‐making. The combination of unchanged IUS‐12 scores and robust demoralization improvements suggests that the program may operate by helping patients remain connected to meaning in the presence of uncertainty. This program complements rather than directly restructures uncertainty‐related cognition [39]. Patients primarily improved from baseline to post‐intervention, and this improvement was maintained at follow‐up. At the individual level, demoralization showed the strongest pattern of continued improvement, with reliable improvement rates increasing from 52.0% at T2 to 68.0% at T3 and no deterioration, whereas depression improved modestly (9.1%–13.6%), raising the possibility that existential gains may be more responsive to brief compassion‐based intervention than mood‐related symptoms [40]. Implications COMwithU may be particularly relevant for patients facing demoralization, loss of meaning, and existential suffering rather than syndromic psychiatric symptoms alone. Moreover, its brief, flexible format may make it feasible to integrate into outpatient psycho‐oncology services. The present feasibility study demonstrated that the program can be delivered and is acceptable to patients with cancer, but it was not designed to isolate the contribution of each component. A critical next step is a controlled trial that directly assesses compassion toward and from others, to examine whether the relational components add value beyond self‐compassion alone and through which mechanisms change occurs. Study Limitations Several limitations should be considered when interpreting these findings. First, the single‐arm design precludes causal inference. Without a control group, improvements cannot be definitively attributed to the intervention rather than to natural recovery, regression to the mean, or the passage of time. This limitation is inherent to the feasibility design and should temper causal interpretation. At the same time, the consistency of improvement across multiple outcomes, along with the clinically meaningful individual‐level change observed for demoralization, supports the view that the intervention warrants more rigorous evaluation in a controlled trial. Second, the sample was drawn from a single tertiary academic hospital and was predominantly female, highly educated, and enriched for breast cancer, which limits generalizability to more diverse cancer populations and clinical settings. Third, although COMwithU incorporated relational components—the therapist's compassionate presence and compassion‐toward‐others practices—as supports for cultivating self‐compassion, this feasibility study assessed only self‐compassion; compassion toward others and received compassion were not directly measured. The assessment battery was designed around the program's self‐compassion‐centered rationale, in which the relational components served as means of supporting change rather than as target outcomes. Consequently, the study cannot determine whether the observed benefits depended on these relational components. This limitation is being addressed in an ongoing randomized controlled trial that includes dedicated measures of compassion toward and from others. Conclusions COMwithU demonstrated feasibility, acceptability, and promising preliminary effects in patients with cancer. The most notable improvements were observed in demoralization and spiritual well‐being. This supports the potential value of compassion‐based approaches for existential distress in psycho‐oncology. These findings provide a rationale for a future randomized controlled trial with extended follow‐up and direct assessment of self‐compassion, relational compassion processes, and existential outcomes. Author Contributions Conceptualization: H.Y.P., J.S.K., M.C.S., E.S.Y. Methodology: H.Y.P., J.S.K., M.C.S., S.H., E.S.Y. Formal analysis: Y.C., J.S.K. Investigation: Y.C., J.S.K., H.A., S.H., D.E.K., M.C.S., J.H.R., Y.A., B.K., T.K., H.Y.P. Data curation: Y.C., J.S.K., H.A., S.H. Writing – original draft: Y.C., H.Y.P. Writing – review and editing: All authors. Project administration: J.S.K., H.A., S.H., H.Y.P. Funding acquisition: H.Y.P. Conflicts of Interest The authors declare no conflicts of interest. Supporting information Supporting Information S1 Supporting Information S2