Treatment Adequacy and Baseline Risk Modify the Association Between Adjuvant Chemotherapy and Survival After Pathological Complete Response in Rectal Cancer

ABSTRACT

The role of adjuvant chemotherapy (ACT) in patients with locally advanced rectal cancer (LARC) who achieve pathological complete response (pCR) after neoadjuvant chemoradiotherapy (nCRT) remains controversial, and it is unclear whether pCR represents a uniformly low‐risk state with respect to long‐term outcomes. We retrospectively analyzed consecutive LARC patients who achieved pCR following nCRT and radical surgery between 2017 and 2022. Survival outcomes were assessed according to postoperative ACT administration, treatment adequacy (≥ 4 cycles vs. < 4 cycles), and baseline risk features. Among 1069 patients treated with nCRT and surgery, 251 (23.5%) achieved pCR. After a median follow‐up of 49 months, no statistically significant differences in overall survival (OS) or disease‐free survival (DFS) were observed between patients who received ACT and those who did not. In contrast, patients who completed an adequate course of ACT (≥ 4 cycles) demonstrated improved 4‐year OS and DFS compared with those receiving fewer cycles or no ACT. This association was largely confined to patients with baseline high‐risk features, while no significant survival differences were observed between different ACT regimens. These findings suggest that pCR does not represent a biologically homogeneous or uniformly low‐risk condition in LARC. Adequate postoperative chemotherapy may confer survival benefit in selected high‐risk patients. A risk‐adapted approach to ACT warrants further investigation and prospective validation.

What's new?

Recent evidence suggests that eradication of the primary tumor does not necessarily exclude the presence of micrometastatic disease in rectal cancer. Consequently, the potential benefit of post‐operative adjuvant chemotherapy in patients who achieved pathological complete response remains unclear. By incorporating treatment adequacy and baseline risk stratification, this real‐world study demonstrates that pathological complete response does not represent a biologically homogeneous state or uniformly low‐risk condition among patients with locally advanced rectal cancer. Survival benefit from post‐operative adjuvant chemotherapy was confined to patients completing an adequate treatment course and those with high‐risk pretreatment features, supporting individualized postoperative decision‐making.

Recent evidence suggests that eradication of the primary tumor does not necessarily exclude the presence of micrometastatic disease in rectal cancer. Consequently, the potential benefit of post‐operative adjuvant chemotherapy in patients who achieved pathological complete response remains unclear. By incorporating treatment adequacy and baseline risk stratification, this real‐world study demonstrates that pathological complete response does not represent a biologically homogeneous state or uniformly low‐risk condition among patients with locally advanced rectal cancer. Survival benefit from post‐operative adjuvant chemotherapy was confined to patients completing an adequate treatment course and those with high‐risk pretreatment features, supporting individualized postoperative decision‐making.

Abbreviations

ACT

adjuvant chemotherapy

AJCC

American Joint Committee on Cancer

ASA

American Society of Anesthesiologists

BMI

body mass index

CAPOX

capecitabine plus oxaliplatin

CI

confidence interval

CRM

circumferential resection margin

CT

computed tomography

ctDNA

circulating tumor DNA

DFS

disease‐free survival

DMFS

distant metastasis‐free survival

EMVI

extramural vascular invasion

HR

hazard ratio

JSCCR

Japanese Society for Cancer of the Colon and Rectum

LARC

locally advanced rectal cancer

MRI

magnetic resonance imaging

NCCN

National Comprehensive Cancer Network

nCRT

neoadjuvant chemoradiotherapy

OS

overall survival

pCR

pathological complete response

PUMCH

Peking Union Medical College Hospital

TME

total mesorectal excision

TNT

total neoadjuvant therapy

1. Introduction

Rectal cancer remains a major global health burden, and neoadjuvant chemoradiotherapy (nCRT) followed by total mesorectal excision (TME) is the standard treatment for patients with locally advanced disease [1]. With the widespread adoption of this multimodal approach, a substantial proportion of patients achieve pathological complete response (pCR), which is consistently associated with favorable long‐term outcomes [2]. Nevertheless, distant metastasis remains the predominant mode of treatment failure even among patients with pCR, suggesting that eradication of the primary tumor does not necessarily equate to elimination of occult systemic disease [3].

The role of adjuvant chemotherapy (ACT) in rectal cancer patients achieving pCR remains highly controversial. Current clinical guidelines offer inconsistent recommendations, reflecting uncertainty regarding whether pCR represents a biologically homogeneous state or encompasses a spectrum of residual metastatic risk [4]. Several large retrospective analyses and meta‐analyses have reported conflicting results, with some suggesting no additional survival benefit from ACT, while others demonstrating improved overall survival [5, 6, 7]. Importantly, most existing studies have treated pCR as a uniform endpoint and have not adequately accounted for pretreatment tumor biology or baseline risk features.

Emerging evidence indicates that baseline clinical and radiological characteristics may retain prognostic relevance despite the achievement of pCR [7, 8]. These observations challenge the conventional assumption that postoperative pathological staging alone sufficiently captures long‐term risk in rectal cancer. From a biological perspective, pCR reflects an excellent local response to nCRT but may not fully exclude the presence of micrometastatic disease, particularly in patients with aggressive pretreatment features.

Against this background, a critical unresolved question is not whether ACT should be routinely administered to all pCR patients but rather whether a risk‐adapted strategy can identify subsets of pCR patients who continue to derive survival benefit from systemic therapy. Addressing this question is of particular clinical importance, as unnecessary chemotherapy may compromise postoperative recovery and long‐term quality of life, whereas omission of ACT in high‐risk patients may forfeit a potentially curative opportunity.

Therefore, we conducted a real‐world retrospective cohort study of consecutively treated patients with locally advanced rectal cancer (LARC) who achieved pCR after nCRT and radical surgery. We hypothesized that the survival benefit from postoperative ACT may be confined to patients with baseline high‐risk features who complete an adequate course of chemotherapy. Accordingly, we sought to evaluate whether treatment adequacy and baseline risk jointly modify the association between ACT and long‐term survival.

2. Methods

2.1. Study Design and Patient Selection

From November 2017 to February 2022, data were collected from LARC patients who underwent nCRT and radical surgery at Peking Union Medical College Hospital (PUMCH). Inclusion criteria: (1) pathologically diagnosed rectal adenocarcinoma; (2) age over 18 years; (3) completion of standard nCRT; (4) postoperative pathological confirmation of pCR. Exclusion criteria: (1) incomplete clinical data; (2) received total neoadjuvant therapy (TNT) prior to surgery; (3) local excision, palliative surgery, and detection of distant metastases preoperatively or intraoperatively; (4) patients who died within 30 days following resection. Patients were categorized into an ACT group and a non‐ACT group based on postoperative management. Baseline high‐risk features were defined according to the Japanese Society for Cancer of the Colon and Rectum (JSCCR) criteria and included patients with clinical stage III disease, as well as those with high‐risk stage II disease characterized by clinical T4 stage, poor histological differentiation, EMVI or MRF involved identified at baseline assessment [9, 10]. Previous studies in colorectal cancer have suggested that shorter durations of ACT may provide comparable efficacy to longer courses [11, 12, 13]. Therefore, patients were stratified according to the number of ACT cycles received. Receipt of ≥ 4 cycles was defined as adequate ACT, whereas receipt of fewer than 4 cycles (including 1–3 cycles or no ACT) was defined as inadequate ACT [14, 15].

2.2. Treatment Protocol

All patients underwent nCRT (long‐course radiotherapy combined with capecitabine‐based chemotherapy). TME was performed 8–10 weeks following the completion of radiotherapy. Patients treated with planned TNT were excluded. Surgical specimens were evaluated in the Department of Pathology at PUMCH through standard procedures. The reports included pathological staging (AJCC 8th edition) [16], the number of involved lymph nodes, vascular and nerve invasion. For patients who achieved pCR, ACT regimens were determined in routine clinical practice at the discretion of the treating surgeons and based on patient preference. Oxaliplatin‐containing regimens were generally considered for younger and fit patients, whereas capecitabine monotherapy was more commonly selected for patients with higher toxicity risk.

2.3. Data Collection and Follow‐Up

Patient demographics, including age, gender, body mass index (BMI), American Society of Anesthesiologist score [17] and concomitant disease, were recorded. All patients had baseline staging data determined via magnetic resonance imaging (MRI). MRI examinations were reviewed by experienced radiologists at PUMCH. Baseline oncological indicators included stage, histological type, tumor distance from the anal verge, mesorectal fascia (MRF), perineural invasion, and the presence of extramural vascular invasion (EMVI). Patients underwent regular postoperative follow‐up, which included colonoscopy and chest‐abdominal‐pelvic computed tomography (CT) to assess tumor recurrence and distant metastasis. OS was defined as the time from radiotherapy to death from any cause. DFS was defined as the time from radiotherapy to the first documented recurrence or death, whichever occurred first. Distant metastasis‐free survival (DMFS) was defined as the time from radiotherapy to the occurrence of distant metastasis.

2.4. Statistical Analysis

For categorical variables, data were presented as frequencies and percentages. Categorical variables were compared using the Chi‐square or Fisher's exact test, while continuous parameters were analyzed using the Student's t‐test or Mann–Whitney U test, depending on the data distribution. Survival was measured from the initiation of nCRT to partially mitigate potential immortal time bias related to postoperative treatment classification. To address potential immortal time bias related to ACT cycle completion, a landmark sensitivity analysis was performed at 8 months after nCRT initiation. Patients with recurrence or death before this time point were excluded. Survival analyses for DFS, DMFS, and OS were performed using the Kaplan–Meier method to estimate survival probabilities. Differences between groups were compared using the log‐rank test. Hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated using univariable Cox proportional hazards models. In Cox models, inadequate ACT was used as the exposure variable, with adequate ACT as the reference category. Given the limited number of outcome events, multivariable adjustment was not performed to avoid overfitting. A two‐tailed p value < 0.05 is considered statistically significant. All statistical analyses were conducted using SPSS version 26.0 software (IBM Corporation, Chicago, IL).

3. Results

3.1. Patients and Clinical Characteristics

From November 2017 to February 2022, a total of 1069 LARC patients at PUMCH underwent neoadjuvant therapy followed by radical surgery. Among them, 251 (23.5%) patients who achieved pCR were selected (Figure 1). There were 161 male patients (64.1%), with an average age of 59 years; 76 (30.3%) patients had positive EMVI, 84 (33.5%) patients had positive MRF, and 231 (92.0%) patients had lymph node metastasis indicated by MRI at diagnosis. There were 51 (20.3%) patients who did not receive ACT postoperatively (Non‐chemo group), while 200 (79.7%) patients did (Chemo group). In the Chemo group, 53 (26.5%) patients were treated with CAPOX, and 147 (73.5%) patients received capecitabine monotherapy. The clinical characteristics of the two groups are listed in Table 1. No significant differences were observed, indicating a well‐balanced cohort suitable for subsequent outcome analyses.

FIGURE 1.

The flow chart of patient selection. Flow diagram illustrating the selection process of patients with locally advanced rectal cancer (LARC) treated at Peking Union Medical College Hospital. Among 1069 patients who received neoadjuvant chemoradiotherapy (nCRT) followed by radical surgery, 251 patients achieved pathological complete response (pCR) and were included in the final analysis. Patients were subsequently categorized according to receipt of postoperative adjuvant chemotherapy (ACT).

TABLE 1.

Clinical characteristics of the non‐chemo and chemo groups.

Variables	Non‐chemo (n = 51)	Chemo (n = 200)	p
Age, M (Q1, Q3)	62.00 (52.50, 68.00)	59.00 (52.00, 67.00)	0.211
Sex, male, n (%)	33 (64.71)	128 (64.00)	0.925
BMI, mean ± SD	24.31 ± 3.27	24.60 ± 6.25	0.742
ASA, n (%)			0.179 a
I	5 (9.80)	36 (18.00)
II	38 (74.51)	146 (73.00)
III	8 (15.69)	18 (9.00)
Distance from anal verge, mean ± SD	5.41 ± 2.48	5.77 ± 2.33	0.339
MRF, positive, n (%)	13 (25.49)	70 (35.00)	0.198
EMVI, positive, n (%)	18 (35.29)	58 (29.00)	0.383
Perineural invasion, positive, n (%)	6 (11.76)	23 (11.50)	0.958
Clinical T stage, T3–T4, n (%)	44 (86.27)	174 (87.00)	0.891
Clinical N stage, N1–N2, n (%)	47 (92.16)	184 (92.00)	1.000

Abbreviations: ASA: American Society of Anesthesiologist score; BMI: body mass index; EMVI: extramural vascular invasion; M: median; MRF: mesorectal fascia; Q₁: 1st quartile; Q₃: 3rd quartile; SD: standard deviation.

^a Fisher exact test.

3.2. Long‐Term Survival Outcomes in the Overall Cohort

The median follow‐up time for the entire cohort was 49 months (interquartile range, 36–65.5 months). A total of 12 patients (4.8%) experienced tumor progression, with 4 cases in the non‐chemo group and 8 cases in the chemo group (Table S1). 7 (2.8%) patients died, including two deaths unrelated to tumor progression.

Kaplan–Meier survival analysis demonstrated no statistically significant differences in OS (p = 0.054) or DFS (p = 0.088) between patients who received ACT and those who did not. Nevertheless, a consistent trend toward improved OS (Figure 2A) and DFS (Figure 2B) was observed in the ACT group, suggesting a potential survival benefit associated with ACT in this pCR population.

FIGURE 2.

Survival outcomes according to receipt of adjuvant chemotherapy. Kaplan–Meier survival curves comparing (A) overall survival (OS) and (B) disease‐free survival (DFS) between patients who received postoperative adjuvant chemotherapy (ACT) and those who did not. Survival was calculated from the initiation of neoadjuvant chemoradiotherapy. p values were determined using the log‐rank test. [Color figure can be viewed at wileyonlinelibrary.com]

3.3. Association Between ACT Cycles and Survival Outcomes

Given the potential association between the number of ACT cycles and survival outcomes, patients were further stratified according to the number of ACT cycles received (0, 1–3, and ≥ 4 cycles). Kaplan–Meier analyses demonstrated that patients who received 4–6 cycles of ACT achieved significantly improved overall survival compared with those who received no ACT or only 1–3 cycles. The difference in OS (Figure 3A) among the three groups was statistically significant (p = 0.029), whereas no significant difference in DFS was observed between the two groups (Figure 3B). Based on these results, patients were subsequently categorized into an adequate ACT group (defined as receipt of ≥ 4 cycles of ACT) (n = 165) and an inadequate ACT group (defined as 0–3 cycles) (n = 86) for further analyses. No significant differences were observed between the adequate ACT and inadequate ACT groups with respect to baseline clinicopathological features (Table S2), indicating a well‐balanced cohort suitable for subsequent outcome analyses.

FIGURE 3.

Survival outcomes according to number of ACT cycles. Kaplan–Meier curves showing (A) overall survival (OS) and (B) disease‐free survival (DFS) stratified by number of postoperative ACT cycles (0 cycles, 1–3 cycles, and ≥ 4 cycles). Survival differences among groups were assessed using the log‐rank test. [Color figure can be viewed at wileyonlinelibrary.com]

Survival outcomes were compared between patients who received adequate ACT and those who received inadequate ACT. The adequate ACT group demonstrated a significantly higher OS rate than the inadequate ACT group (Figure 4A). During follow‐up, 7 deaths in the inadequate ACT group and 3 occurred in the adequate ACT group. The 4‐year OS rate was 97.5% versus 90.5% (p = 0.008; HR, 5.21; 95% CI, 1.34–20.23).

FIGURE 4.

Survival outcomes according to treatment adequacy. Kaplan–Meier curves comparing patients who received adequate ACT (≥ 4 cycles) and inadequate ACT (0 or 1–3 cycles): (A) overall survival (OS), (B) disease‐free survival (DFS), and (C) distant metastasis‐free survival (DMFS). Hazard ratios (HRs) and 95% confidence intervals (CIs) were derived from Cox proportional hazards models. p values were calculated using the log‐rank test. [Color figure can be viewed at wileyonlinelibrary.com]

Regarding DFS, patients in the adequate ACT group demonstrated more favorable outcomes compared with those in the inadequate ACT group (Figure 4B). A total of 8 DFS events occurred in the inadequate ACT group and 6 in the adequate ACT group. In univariable Cox analysis, inadequate ACT was associated with a higher risk of DFS events (HR 2.85, 95% CI 0.98–8.29; p = 0.044), although the CI was wide, reflecting the limited number of events. Interaction analyses between ACT adequacy and individual pretreatment risk factors (EMVI, MRF involvement, cT4 and cN2) did not demonstrate statistically significant effect modification (all P for interaction > 0.05).

In the adequate ACT group, 3 patients experienced local recurrence and 3 developed distant metastases. In contrast, no local recurrence was observed in the inadequate ACT group, while 6 patients experienced distant metastases. Although no statistically significant difference in DMFS was observed between the two groups (p = 0.077), a favorable trend toward improved DMFS was noted in patients who received adequate ACT (Figure 4C).

In the landmark analysis, no deaths occurred within the landmark window. Two patients experienced metastasis at 7 months postoperatively and were excluded from the DFS analysis. After landmark restriction, the association between adequate ACT and improved DFS remained directionally consistent with the primary analysis, although statistical significance was attenuated. The HR estimates were similar in magnitude to the primary analysis (Figure S2).

3.4. Subgroup Analysis

Among patients who received ACT, further subgroup analyses were performed according to the ACT regimen, including capecitabine monotherapy (n = 147), CAPOX (capecitabine plus oxaliplatin) therapy (n = 53), and no ACT (n = 51). No significant difference in OS (Figure S1A) and DFS (Figure S1B) was observed between the two chemotherapy regimens. However, both ACT regimens were associated with superior OS compared with no ACT.

According to the JSCCR guidelines, baseline high‐risk disease was defined as clinical stage III or high‐risk stage II, with the latter including clinical T4 tumors, poor differentiation, or EMVI/MRF positive [9]. Two hundred and thirty‐four patients with baseline high‐risk stage II disease (n = 3) and stage III (n = 231) rectal cancer were classified into a high‐risk group and further stratified according to receipt of adequate ACT. Within this high‐risk population, patients who received adequate ACT demonstrated a significantly improved 4‐year OS compared with those who received inadequate ACT (p = 0.008; HR, 5.19; 95% CI, 1.34–20.16) (Figure 5A). A similar association was observed for 4‐year DFS (p = 0.049; HR, 2.79; 95% CI, 0.96–8.11), although the CI was wide and crossed unity, warranting cautious interpretation (Figure 5B).

FIGURE 5.

Survival outcomes in patients with baseline high‐risk features. Kaplan–Meier curves of (A) overall survival (OS) and (B) disease‐free survival (DFS) among patients with baseline high‐risk stage II or stage III disease, stratified by receipt of adequate (≥ 4 cycles) versus inadequate (0–3 cycles) ACT. Survival differences were assessed using the log‐rank test, and hazard ratios (HRs) with 95% confidence intervals (CIs) were calculated using Cox proportional hazards models. [Color figure can be viewed at wileyonlinelibrary.com]

4. Discussion

In this real‐world cohort of patients with LARC who achieved pCR following nCRT, we observed that postoperative ACT was not uniformly associated with improved survival when analyzed as a binary intervention. Instead, our findings indicate that the potential benefit of ACT was highly dependent on treatment adequacy and baseline oncologic risk. Patients who completed an adequate course of ACT (≥ 4 cycles) experienced significantly improved overall and disease‐free survival, whereas those receiving fewer cycles had outcomes comparable to patients who did not receive adjuvant therapy. These results suggest that pCR represents a favorable but biologically heterogeneous state, within which selected patients may continue to benefit from postoperative systemic treatment.

The clinical significance of pCR has been well established, and patients achieving pCR generally experience excellent local control and favorable prognosis. However, distant metastasis remains the predominant mode of failure even among patients achieving pCR, indicating that eradication of the primary tumor does not necessarily equate to elimination of systemic disease [4]. Increasing evidence suggests that pretreatment tumor characteristics, including advanced clinical stage and high‐risk MRI features, may retain prognostic relevance despite complete pathological regression [9]. Our findings support this concept by demonstrating that survival outcomes among pCR patients continued to differ according to baseline risk profiles and receipt of adequate systemic therapy.

A particularly notable observation of this study is that ACT, when considered merely as administered or not administered, was not associated with statistically significant survival differences. In contrast, a clear survival advantage emerged when treatment adequacy was taken into account. Additional interaction analyses incorporating MRI‐defined risk factors did not demonstrate statistically significant effect modification. However, these analyses were limited by small event numbers and should be interpreted cautiously. Patients receiving fewer than four cycles of ACT demonstrated survival outcomes similar to those of patients who did not receive chemotherapy at all [18]. This finding suggests that subtherapeutic exposure to postoperative systemic therapy may be insufficient to achieve meaningful control of residual micro‐metastatic disease, especially in patients with persistent systemic risk. These results highlight the importance of treatment completion rather than treatment initiation alone when evaluating the role of ACT in pCR patients. Such methodological differences may partially account for the neutral survival results reported in some previous large‐scale studies [19, 20].

Importantly, the survival benefit associated with adequate ACT was numerically more evident in high‐risk patients. Even after achieving pCR, these patients continued to exhibit differential outcomes according to postoperative treatment, underscoring that pretreatment oncologic risk is not fully neutralized by pathological downstaging [21]. This observation aligns with emerging data suggesting that baseline tumor biology and systemic disease potential remain critical determinants of long‐term prognosis, and it supports the concept that adjuvant treatment decisions should incorporate pretreatment risk stratification rather than rely solely on postoperative pathological findings [22, 23].

We did not observe significant differences in survival outcomes between fluoropyrimidine monotherapy and oxaliplatin‐containing regimens among patients receiving ACT. No firm conclusions regarding regimen superiority can be drawn. Given the favorable prognosis of patients achieving pCR, balancing treatment intensity against potential toxicity while ensuring adequate exposure to systemic therapy may represent a reasonable consideration in postoperative management [24].

From a clinical perspective, these findings argue against the routine administration of ACT to all patients achieving pCR. While omission of postoperative therapy may be appropriate for selected low‐risk patients, inadequate or prematurely discontinued chemotherapy appears unlikely to confer meaningful benefit. Conversely, patients with baseline high‐risk features may derive a survival advantage from completing an adequate course of adjuvant therapy. Together, these results support a risk‐adapted approach to postoperative management that balances oncologic benefit against treatment‐related burden.

Several limitations should be acknowledged. First, the retrospective design precludes causal inference. Classification based on the number of ACT cycles may introduce immortal time bias, as patients must survive long enough to complete adequate chemotherapy. Although survival was calculated from the initiation of nCRT and a landmark analysis was performed to partially mitigate this issue, residual bias cannot be excluded. Second, important clinical factors that may have influenced treatment adequacy—such as postoperative performance status, treatment‐related toxicity, comorbidities, or patient preference—were not uniformly captured and may have confounded the association between ACT completion and survival. Third, the number of outcome events, particularly deaths, was limited, resulting in relatively wide CIs and restricted statistical power. Therefore, the observed associations should be interpreted cautiously.

Future studies should aim to refine patient selection for ACT using more precise markers of residual systemic disease. Biomarkers such as circulating tumor DNA may complement baseline clinical risk factors and pathological response, enabling more individualized postoperative strategies [25]. Prospective validation of risk‐adapted approaches will be essential to optimize long‐term outcomes while avoiding unnecessary treatment‐related toxicity [10].

5. Conclusion

In patients with LARC achieving pCR after nCRT, pCR does not necessarily represent a uniformly low‐risk state. In this real‐world cohort, receipt of an adequate course of postoperative ACT was associated with improved survival, particularly among patients with baseline high‐risk features. These findings suggest that a risk‐adapted approach to ACT may merit further investigation, but prospective validation is required before such a strategy can be incorporated into routine clinical practice.

Author Contributions

Chentong Wang: writing – original draft, methodology, investigation. Xiao Zhang: visualization, data curation, formal analysis. Yang An: writing – review and editing, investigation, formal analysis. Chunkang Liu: visualization, writing – review and editing. Bin Wu: project administration, data curation. Yi Xiao: data curation, project administration. Junyang Lu: methodology, data curation. Guole Lin: data curation, writing – review and editing, conceptualization, funding acquisition, methodology.

Funding

This work was supported by the National High Level Hospital Clinical Research Funding (2022‐PUMCH‐C‐005) and Peking Union Medical College 2023 Education and Teaching Reform of Central Universities (2023zlgl002). The funders had no role in the design and conduct of the study, including collection, management, analysis, and interpretation of the data, and preparation, review, or approval of the manuscript.

Ethics Statement

The study was approved by the Ethics Committee of PUMCH (ID JS‐1296).

Consent

The requirement for informed consent was waived owing to the retrospective design and minimal risk to participants.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: Survival outcomes according to ACT regimen. Kaplan–Meier curves comparing (A) overall survival (OS) and (B) disease‐free survival (DFS) among patients receiving capecitabine monotherapy, capecitabine plus oxaliplatin (CAPOX), or no adjuvant chemotherapy. Differences were evaluated using the log‐rank test.

Figure S2: Disease‐free survival according to treatment adequacy in the landmark analysis. Kaplan–Meier curves comparing patients who received adequate ACT and inadequate ACT. Hazard ratios (HRs) and 95% confidence intervals (CIs) were derived from Cox proportional hazards models. P values were calculated using the log rank test.

Table S1: Patterns of recurrence and death.

Table S2: Clinical characteristics of the adequate and inadequate groups.

Data Availability Statement

Anonymized data may be made available from the corresponding author upon reasonable request and subject to institutional approval.