Abstract
-
Purpose
- Young-onset rectal cancer (YORC) is increasingly common and often presents with aggressive histological features; however, its effect on recurrence remains uncertain. This study aimed to estimate the causal effect of age at diagnosis on recurrence in patients with rectal cancer using advanced causal inference methods.
-
Methods
- This multicenter retrospective cohort study was conducted at 2 university hospitals in Thailand. Patients diagnosed with rectal adenocarcinoma between 2013 and 2022 who underwent curative-intent treatment were included. The primary exposure was age at diagnosis, categorized as YORC (<50 years) or late-onset rectal cancer (LORC; ≥50 years). Confounders were prespecified using a directed acyclic graph and addressed using a confounder composite score. Age was also modeled as a continuous variable using natural splines. The primary outcome was cancer recurrence.
-
Results
- Among 585 patients, 116 (19.8%) had YORC and 469 (80.2%) had LORC. Patients with YORC had fewer comorbidities but more unfavorable histological features than those with LORC. Overall, recurrence occurred in 40.2% of patients and most commonly involved distant metastasis. The estimated average causal effect of age on recurrence was −6.5% (95% confidence interval, −16.6% to 3.5%; P=0.202), indicating that younger age was not associated with an increased risk of recurrence, although a protective effect was not confirmed. Natural spline modeling indicated a nonlinear increase in recurrence hazard with advancing age.
-
Conclusion
- Younger age at diagnosis was not associated with worse recurrence outcomes despite adverse pathological features, and a potential protective association could not be excluded.
-
Keywords: Young adult; Age factors; Rectal neoplasms; Recurrence; Causality
INTRODUCTION
Colorectal cancer (CRC) is a major global health concern and ranks among the 3 leading causes of cancer-related mortality worldwide in both male and female patients [1]. Although overall CRC incidence has declined in older populations because of improved screening, CRC incidence has steadily increased among individuals younger than 50 years [2–4]. Approximately 25% of newly diagnosed rectal cancer cases now occur in patients younger than 50 years, underscoring the increasing burden of young-onset colorectal cancer [5].
Although they are often grouped together as CRC, colon cancer and rectal cancer are biologically and clinically distinct entities [6]. Previous studies have reported differences in carcinogenesis, tumor biology, treatment response, and prognosis between colon and rectal cancers [7–10]. These distinctions support separate analyses of colon and rectal cancer in research. Accordingly, young-onset rectal cancer (YORC) warrants specific investigation independent of young-onset colon cancer.
Despite increasing attention, survival outcomes in YORC compared with late-onset rectal cancer (LORC) remain inconclusive, with conflicting findings across studies [11–13]. Age has emerged as a potentially important prognostic factor [14, 15]; however, its effect on recurrence and survival in rectal cancer remains unclear because of methodological limitations in previous studies. Many traditional statistical analyses risk conflating association with causation, which limits the validity of causal interpretations drawn from those findings [16].
To address this gap, this study used a causal modeling framework, specifically, directed acyclic graphs (DAGs), to identify and control for confounding variables [17]. This approach was intended to support a more accurate estimate of the causal effect of age on rectal cancer recurrence. Using causal inference methods, we aimed to clarify whether younger age at diagnosis independently influences recurrence risk in patients with rectal cancer.
METHODS
Ethics statement
This study was approved by the Human Investigation Committee of the Faculty of Medicine Vajira Hospital, Navamindradhiraj University (No. COA 143/67E), and the Faculty of Medicine, Chiang Mai University (No. SUR-2567-0440). Informed consent was waived because of the retrospective design and the use of deidentified patient data. This retrospective chart review involving human participants followed the ethical standards of the institutional and national research committees and the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards. This retrospective study followed the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines [18].
Study design and patient population
Electronic medical records were retrieved from Vajira Hospital and Maharaj Nakorn Chiang Mai Hospital, 2 tertiary medical centers in Thailand. Data were collected for patients diagnosed with rectal cancer between January 2013 and December 2022 who were aged 20 years or older at diagnosis. Patients were identified using International Classification of Diseases, 10th Revision (ICD-10), code C20, which indicates malignant neoplasm of the rectum. The exclusion criteria were absence of a Thai national identification number; a history of another malignancy, including synchronous colon cancer; concomitant inflammatory bowel disease or hereditary colorectal cancer; nonadenocarcinoma histology, such as gastrointestinal stromal tumor, neuroendocrine carcinoma, or squamous cell carcinoma; and no curative-intent treatment. YORC was defined as rectal cancer diagnosed in patients younger than 50 years, according to the American Gastroenterological Association definition [19]. The study flowchart is shown in Fig. 1.
Data collection and definitions
Data collection included demographic and clinical characteristics at diagnosis. Demographic variables included age, sex, body weight, height, comorbidities, smoking status, and family history of CRC. Clinical characteristics included tumor location, completeness of preoperative colonoscopy, preoperative imaging findings, presenting symptoms, and preoperative carcinoembryonic antigen (CEA) level. Treatment details were also documented, including the need for preoperative diversion ostomy, neoadjuvant treatment and its components, surgical intervention, pathological findings, adjuvant treatment, and completion of neoadjuvant and adjuvant therapies. Curative treatment intent was also assessed. Treatment outcomes included follow-up status (regular follow-up or loss to follow-up), survival status, date of last follow-up for surviving patients, recurrence status, and recurrence type, if present.
Tumor location was defined as the distance from the anal verge in centimeters and classified as upper, middle, or lower rectum. The primary measurement was based on pelvic magnetic resonance imaging (MRI). When imaging findings differed from those of other modalities, such as colonoscopy or digital rectal examination, tumor location was determined by multidisciplinary team (MDT) consensus. Colonoscopy completeness was defined as successful cecal intubation before any treatment. Preoperative imaging data were extracted from radiology reports. Local tumor staging and tumor characteristics were assessed primarily using pelvic MRI, whereas metastatic disease was evaluated using computed tomography (CT) of the chest, abdomen, and pelvis. Preoperative CEA level (ng/mL) was documented before initial treatment.
Treatment and follow-up
Neoadjuvant treatment was categorized as chemotherapy, chemoradiotherapy (CRT), or radiotherapy (RT). Total neoadjuvant therapy (TNT) was also identified. The CRT protocols at our institutions included long-course radiotherapy with concurrent oral capecitabine or infused 5-fluorouracil/leucovorin, as well as short-course RT (SCRT; 5 Gy, 5 fractions). The TNT protocols in this study consisted of either induction chemotherapy followed by CRT or SCRT, or CRT or SCRT followed by consolidation chemotherapy, using fluoropyrimidine-based regimens, such as FOLFOX (fluorouracil, leucovorin, and oxaliplatin) or CAPOX (capecitabine and oxaliplatin), according to institutional protocols. The choice between long-course CRT and SCRT and the sequencing of chemotherapy were determined at the discretion of the treating MDT based on patient and tumor characteristics. The MDT, consisting of surgeons, medical oncologists, radiation oncologists, pathologists, and radiologists, determined the specific treatment protocol for each patient. Details of chemotherapy regimens, including the total number of cycles administered, were collected.
Completion of neoadjuvant or adjuvant treatment was defined as receipt of the full standard-dose chemotherapy regimen and completion of all planned cycles without postponement. In this study, curative-intent treatment was defined as an MDT-determined treatment strategy established before treatment initiation, with the goal of cure for both primary and metastatic disease.
Follow-up was conducted according to the National Comprehensive Cancer Network (NCCN) guidelines and included CT imaging every 3 to 6 months for the first 2 years and annually thereafter, together with CEA monitoring and scheduled colonoscopy. Loss to follow-up was defined as failure to attend scheduled hospital visits. Recurrence was classified as local recurrence or metastatic recurrence and was identified using surveillance imaging, endoscopy, or pathological confirmation. Vital status and dates of death were obtained from the government civil registration database, and follow-up data were updated as of June 2025.
Statistical analysis
For statistical analysis, continuous variables were expressed as mean±standard deviation or median (interquartile range, IQR), and categorical variables were reported as frequencies and percentages. Inferential tests included the independent-samples t-test or Wilcoxon rank sum test for continuous variables and the chi-square test or Fisher exact test for categorical variables, as appropriate. Missing data were not imputed and were handled using complete-case analysis.
An MDT developed DAGs to identify confounding factors affecting the association between age and recurrence and to support confounder control. Variables included in the DAG were selected a priori based on clinical knowledge and prior literature, and intervariable correlations were assessed as a supplementary step to avoid redundancy (Supplementary Fig. 1). DAG analysis was performed using Dagitty.net (Supplementary Fig. 2). Based on the DAG, baseline variables, including sex, body mass index (BMI), presenting symptoms, and tumor location, were identified as potential confounders because they may influence both age and recurrence risk. In contrast, treatment-related variables, such as neoadjuvant therapy, clinical staging, and adjuvant treatment, were conceptualized as post-exposure variables that may lie along the causal pathway between age and recurrence; therefore, they were considered potential mediators rather than confounders. The final adjustment set was based on the minimal sufficient adjustment set derived from the DAG, with the aim of blocking noncausal backdoor pathways while preserving the causal pathway of interest. Because the primary objective was to estimate the total causal effect of age on recurrence, post-exposure variables were not included in the adjustment model to avoid overadjustment bias. Kaplan-Meier survival analysis was performed to evaluate recurrence-free survival (RFS) and overall survival (OS). RFS was defined as the time from surgery to first documented tumor recurrence; death without prior recurrence was treated as a censoring event. Because not all patients underwent surgery, RFS analyses were restricted to the surgical cohort. OS was defined as the time from diagnosis to death from any cause, with patients who were alive censored at the date of last follow-up.
The average causal effect (ACE) of age on recurrence was estimated as a risk difference between YORC and LORC using a DAG-informed confounder composite score (CCS). The CCS, conceptually similar to a disease risk score, was incorporated as a covariate in regression models for confounding adjustment. The robustness of the findings was evaluated in sensitivity analyses using alternative effect measures, including risk ratios, odds ratios, and hazard ratios (HRs) with 95% confidence intervals (CIs). Statistical significance was defined as P<0.05. Statistical analyses were performed using RStudio ver. 2023.06.0+421 (Posit) with packages dplyr, ggplot2, epiDisplay, tidyr, coxme, logistf, lubridate, survival, survminer, forcats, psych, polychor, corrplot, fmsb, and splines. The ACE was further estimated using Stata ver. 18.0 (StataCorp) with multiple sensitivity analysis methods.
RESULTS
A total of 585 patients with rectal cancer were included. Of these, 116 (19.8%) were classified as having YORC and 469 (80.2%) as having LORC. As shown in Table 1, the YORC group had a significantly lower prevalence of comorbidities than the LORC group (18.1% vs. 58.6%, P<0.001), and tumors were located closer to the anal verge in the YORC group (median, 5 cm vs. 6 cm; P=0.018). Patients with YORC were also more likely than those with LORC to have poorly differentiated, mucinous, or signet ring cell histology (25.0% vs. 6.8%, P<0.001) and had lower pretreatment CEA levels (median, 4.3 [IQR, 2.2–23.7] vs. 7.3 [IQR, 3.1–27.4]; P=0.025). No statistically significant differences were observed between groups in sex distribution, BMI, presenting symptoms, clinical stage, or colonoscopy completeness. Supplementary Fig. 3 shows a radar plot of standardized differences in selected baseline characteristics between groups.
Table 2 summarizes treatment modalities and short-term oncological outcomes. The distribution of neoadjuvant treatment strategies, including TNT use, was comparable between the YORC and LORC groups, with no statistically significant differences. Among patients who received TNT, both induction and consolidation chemotherapy strategies were used. Induction chemotherapy was more common in the YORC group, whereas consolidation chemotherapy was more common in the LORC group. Most patients in both cohorts completed neoadjuvant treatment (91.2% vs. 82.9%, P=0.364), indicating high treatment adherence across age groups. Detailed distributions of treatment strategies are provided in Supplementary Table 1. Sphincter-preserving procedures were performed at similar rates in both groups (75.9% vs. 77.2%, P=0.762). Pathological complete response of the primary tumor (ypT0) was observed in 11 patients (10.2%) in the YORC group and 36 patients (8.0%) in the LORC group (P=0.891). The proportion of patients with a positive circumferential resection margin was low and did not differ significantly between groups (7.4% vs. 6.0%, P=0.655). Pathological nodal involvement also did not differ significantly between the groups (33.3% vs. 40.8%, P=0.152).
The median follow-up duration for survival outcomes was 6.34 years (Supplementary Table 2). At the time of analysis, 44.0% of patients in the YORC group and 39.0% in the LORC group were alive. The 5-year RFS rate was 68.4% for YORC and 56.3% for LORC, and the 5-year OS rates were 55.6% and 51.3%, respectively. However, RFS and OS did not differ significantly between groups. After adjustment for confounders identified using DAGs, the adjusted HR (aHR) for RFS comparing YORC with LORC was 1.34 (95% CI, 0.90–1.99; P=0.155), and the aHR for OS was 1.12 (95% CI, 0.83–1.53; P=0.455), indicating no statistically significant difference in survival. Fig. 2 shows the Kaplan-Meier curves for RFS and OS, which demonstrate comparable survival distributions. The median RFS was 6.18 years for YORC and 6.36 years for LORC (P=0.440), whereas the median OS was 5.38 and 5.04 years, respectively (P=0.556).
Recurrence occurred in 37.9% of patients with YORC and 40.7% of patients with LORC (P=0.583). Recurrence patterns were broadly similar between groups, with a slightly higher proportion of distant metastasis in the LORC group (Fig. 3). Among patients with recurrent YORC, 29.5% had locoregional recurrence, 61.4% had metastatic recurrence, and 9.1% had both locoregional and metastatic recurrence. Among patients with recurrent LORC, 23.0% had locoregional recurrence, 61.3% had metastatic recurrence, and 15.7% had both. These differences were not statistically significant (P=0.426) (Supplementary Table 2). The distribution of locoregional recurrence subtypes, including central/anterior, lateral, posterior, and composite recurrence, also did not differ significantly between groups (P=0.269).
To investigate the causal relationship between age at diagnosis and rectal cancer recurrence, we performed a DAG-informed analysis using a CCS for confounding adjustment. This approach estimated that patients younger than 50 years had a 6.5% lower absolute risk of recurrence than those aged 50 years or older (ACE, −6.5%; 95% CI, −16.6% to 3.5%), although the difference was not statistically significant (P=0.202). A natural spline-based Cox regression model demonstrated a continuous, nonlinear increase in the hazard of recurrence with advancing age, particularly after 50 years of age (Fig. 4). Covariate balance before and after CCS-based confounding adjustment was assessed using a Love plot, which showed substantial improvement in standardized mean differences (SMDs) across covariates (Supplementary Fig. 4). Sensitivity analyses using alternative effect measures were performed to assess the robustness of the model. Across these analyses, effect estimates consistently favored the younger group, indicating a stable negative association between younger age and recurrence risk (Supplementary Fig. 5).
DISCUSSION
Several clinically relevant clinicopathological differences were observed between patients with YORC and those with LORC in this cohort, many of which are consistent with previous reports. As expected, younger patients had fewer concomitant comorbidities, reflecting generally better overall health status. Tumors in patients with YORC were more likely to show unfavorable histopathological features, including poor differentiation and mucinous or signet ring cell histology, consistent with previous literature suggesting a more aggressive tumor phenotype in younger patients [11]. Tumors in the YORC group also tended to be located lower in the rectum, with a median distance from the anal verge of 5 cm compared with 6 cm in the LORC group. However, this difference may have limited clinical relevance because all low- and mid-rectal tumors typically require total mesorectal excision. Treatment adherence, defined as completion of neoadjuvant and adjuvant therapies, did not differ significantly between groups; this contrasts with some previous studies that reported higher adherence among younger patients [20]. The overall recurrence rate in this study was within the range reported previously (25%–65%) [21–23], but distant recurrence predominated over locoregional recurrence. This pattern differs from earlier reports in which pelvic recurrence was more common [24, 25]. Among patients with locoregional failure, the central or anterior compartment was the most frequently involved site.
Although OS was comparable between the YORC and LORC groups, this finding should be interpreted cautiously. OS reflects both cancer-related and non–cancer-related mortality, and age is a well-established determinant of natural mortality risk. The median follow-up duration of 6.34 years may not have been sufficient to fully capture age-related differences in long-term OS, particularly among older patients. Therefore, the absence of a difference in OS does not necessarily indicate equivalent overall prognosis between groups. In contrast, recurrence-based outcomes, such as RFS, may more directly reflect cancer-related risk and are therefore more appropriate for evaluating the effect of age in this context.
To further contextualize the OS findings, stage-stratified survival analyses were performed and are presented in Supplementary Fig. 6 (OS) and Supplementary Fig. 7 (RFS). These analyses demonstrated substantial heterogeneity in survival outcomes across disease stages, with progressively worse survival in more advanced stages. The relatively lower OS rates observed in this cohort are therefore likely attributable, at least in part, to the underlying stage distribution, including patients with locally advanced disease and stage IV disease treated with curative intent. This stage-specific variability highlights the importance of considering disease stage when interpreting OS outcomes and reinforces that unstratified OS estimates may not fully reflect prognosis across heterogeneous patient populations.
Building on these observed differences, a major strength of this study is its use of a causal inference framework to move beyond traditional associations and evaluate whether age at diagnosis affects recurrence risk in rectal cancer. Using a DAG-informed causal framework, we identified a minimally sufficient adjustment set and estimated the causal effect of age on recurrence using a CCS for confounding adjustment. Unlike previous studies that relied primarily on conventional multivariable regression [26, 27], this approach used the causal structure specified in the DAG to guide covariate adjustment. Although the estimated ACE was not statistically significant, the consistent direction of effect across models suggests that younger age is not associated with increased recurrence risk and may be associated with a modest protective effect. Contrary to some previous reports suggesting poorer outcomes in younger patients [12, 28], these findings may reflect better treatment tolerance, fewer comorbidities, and more intensive multimodal therapy among younger individuals. Differences in inclusion criteria and analytic methods across studies, particularly the absence of causal adjustment in earlier analyses, may partly explain these discrepancies. This finding is consistent with several previous studies reporting similar trends among younger patients [26, 27].
These findings suggest that younger patients with rectal cancer may present with more aggressive tumor histology, but this does not necessarily translate into worse recurrence outcomes. The estimated causal effect and the consistent direction of effect across models indicated that younger age may be associated with a modest reduction in recurrence risk, possibly reflecting better treatment tolerance, fewer competing comorbidities, and more intensive multidisciplinary management.
Because some patients did not undergo surgical resection despite curative-intent management, including patients with complete clinical response managed nonoperatively, those who declined surgery, and those deemed medically inoperable, we conducted a sensitivity analysis restricted to patients who underwent surgery to assess potential heterogeneity. The results remained consistent with the primary analysis, with a slightly larger effect size in the surgery-only cohort (ACE, −8.7%; 95% CI, −18.9% to 1.5%; P=0.095). The consistency in both direction and magnitude suggests that inclusion of nonsurgical patients did not meaningfully influence the estimated causal effect, supporting the robustness of the findings.
In addition, although patients with stage IV disease were included only when treated with curative intent, metastatic tumor burden may not have been fully captured by conventional staging and could have contributed to residual confounding. To address this concern, we performed a sensitivity analysis excluding patients with stage IV disease. The estimated effect remained consistent in both direction and magnitude (ACE, −6.6%; 95% CI, −18.0% to 4.9%; P=0.262), suggesting that the observed association between age and recurrence was not driven by the inclusion of patients with metastatic disease.
The spline-based analysis demonstrated a nonlinear increase in recurrence hazard with advancing age, particularly after 50 years of age. This finding challenges the common practice of dichotomizing age in prognostic modeling. These results may indicate that age, although clinically relevant, should not be interpreted in isolation. Instead, prognostic estimates and treatment decisions should consider both biological and contextual patient factors. From a methodological perspective, this study supports the broader integration of causal inference frameworks into cancer outcomes research to generate evidence that is more informative for clinical decision-making.
Limitations
Several limitations must be acknowledged. First, as in all retrospective studies, residual confounding and selection bias remain possible despite robust adjustment. In particular, although disease stage was included and sensitivity analyses excluding patients with stage IV disease were performed, tumor burden within each stage, such as the number and extent of metastases, was not captured. This limitation may have influenced treatment decisions and recurrence risk, representing a potential source of residual confounding. Second, important molecular data, such as mismatch repair status or consensus molecular subtype, were not available; these data may have provided additional insight into age-related tumor biology. Third, loss to follow-up was more frequent among older patients, potentially leading to underestimation of recurrence or survival outcomes in the LORC group. Fourth, although recurrence patterns were categorized as locoregional and metastatic and no significant differences were observed between age groups, separate causal analyses were not performed for recurrence subtypes because the primary objective was to estimate the overall causal effect of age on recurrence. Therefore, potential differential effects of age on specific recurrence patterns cannot be excluded. Finally, although the study was conducted at 2 tertiary centers, the findings may not be generalizable to broader populations or different health care settings. A post hoc power calculation was not performed because such analyses are considered uninformative once the data have been observed. Instead, the width of the CIs reflects the precision of the estimates and indicates the need for larger, prospectively designed studies to validate these findings. Future research should examine the biological mechanisms underlying age-related differences in tumor behavior, including genetic alterations, microsatellite instability, and the tumor immune microenvironment.
Conclusions
Our study found that younger age at diagnosis was not independently associated with worse recurrence outcomes in rectal cancer and may be associated with a modest protective effect after causal adjustment. Despite more aggressive histological features, YORC had recurrence rates comparable to those of LORC.
SUPPLEMENTARY MATERIALS
Supplementary Fig. 3.
Radar plot comparing standardized differences in baseline characteristics between young-onset rectal cancer (YORC) and late-onset rectal cancer (LORC).
ac-2026-00171-0024-Supplementary-Fig-3.pdf
Supplementary materials are available from https://doi.org/10.3393/ac.2026.00171.0024.
ARTICLE INFORMATION
-
Conflict of interest
No potential conflict of interest relevant to this article was reported.
-
Funding
This study was partially supported by Chiang Mai University.
-
Author contributions
Conceptualization: TT, PS, SS, ES, JK, SJ, AT; Data curation: TT, AT; Formal analysis: TT, AT; Funding acquisition: SS; Investigation: KB, BN, NS, RT, PC; Methodology: TT, PS, SS, ES, JK, SJ, AT; Project administration: PS, AT; Supervision: PS, SS, ES, JK, SJ, KB, BN, NS, RT, AT; Validation: SS, JK, AT; Visualization: TT; Writing–original draft: TT; Writing–review & editing: all authors. All authors read and approved the final manuscript.
Fig. 1.Study flow diagram. ICD-10, International Classification of Diseases, 10th Revision; DAG, directed acyclic graph; YORC, young-onset rectal cancer; LORC, late-onset rectal cancer.
Fig. 2.Kaplan-Meier survival curves stratified by age group. (A) Recurrence-free survival (RFS) and (B) overall survival (OS) curves are shown for patients with young-onset rectal cancer (YORC) and late-onset rectal cancer (LORC). No statistically significant survival difference was observed between groups. CI, confidence interval.
Fig. 3.Recurrence patterns according to age group. (A) Locoregional disease. (B) Metastatic disease. YORC, young-onset rectal cancer; LORC, late-onset rectal cancer.
Fig. 4.Adjusted hazard ratio (aHR) for recurrence across age. The continuous relationship between age at diagnosis and the aHR of rectal cancer recurrence is illustrated, modeled using natural splines. The curve was adjusted for clinical confounders identified using directed acyclic graphs. The red dashed line marks the 50-year age threshold.
Table 1.Baseline characteristics of patients with rectal cancer stratified by age group (n=585)
|
Characteristic |
YORC (n=116, 19.8%) |
LORC (n=469, 80.2%) |
P-value |
|
Age (yr) |
43.5 (37.8–47) |
63 (58–71) |
<0.001 |
|
Sex |
|
|
0.310 |
|
Male |
64 (55.2) |
283 (60.3) |
|
|
Female |
52 (44.8) |
186 (39.7) |
|
|
BMI (kg/m2) |
22.1 (19.5–25.1) |
22.3 (19.4–25.3) |
0.988 |
|
Missing |
3 (2.6) |
8 (1.7) |
|
|
Presence of comorbidity |
21 (18.1) |
275 (58.6) |
<0.001 |
|
Distance of tumor from AV (cm) |
5 (4–8) |
6 (4–10) |
0.018 |
|
Missing |
2 (1.7) |
3 (0.6) |
|
|
Location of tumor |
|
|
0.100 |
|
Upper |
20 (17.2) |
82 (17.5) |
|
|
Middle |
33 (28.4) |
180 (38.4) |
|
|
Lower |
63 (54.3) |
207 (44.1) |
|
|
Presenting symptom |
|
|
0.695 |
|
Screening |
1 (0.9) |
9 (1.9) |
|
|
Symptomatic |
113 (97.4) |
457 (97.4) |
|
|
Missing |
2 (1.7) |
3 (0.6) |
|
|
Completeness of colonoscopy |
42 (36.2) |
189 (40.3) |
0.391 |
|
Clinical staging |
|
|
0.209 |
|
I |
2 (1.7) |
14 (3.0) |
|
|
II |
13 (11.2) |
85 (18.1) |
|
|
III |
70 (60.3) |
262 (55.9) |
|
|
IV |
25 (21.6) |
80 (17.1) |
|
|
Missing |
6 (5.2) |
28 (6.0) |
|
|
Tumor grading |
|
|
<0.001 |
|
Well differentiated |
30 (25.9) |
126 (26.9) |
|
|
Moderately differentiated |
57 (49.1) |
311 (66.3) |
|
|
Poorly differentiated, mucinous, or signet |
29 (25.0) |
32 (6.8) |
|
|
CEA level (ng/mL) |
4.3 (2.2–23.7) |
7.3 (3.1–27.4) |
0.025 |
|
Missing |
18 (15.5) |
48 (10.2) |
|
Table 2.Treatment strategies by age group (n=585)
|
Characteristic |
YORC (n=116) |
LORC (n=469) |
P-value |
|
Neoadjuvant treatment |
|
|
0.890 |
|
No |
47 (40.9) |
172 (37.0) |
|
|
Chemotherapy |
14 (12.2) |
62 (13.3) |
|
|
CRT |
52 (45.2) |
224 (48.2) |
|
|
RT |
2 (1.7) |
7 (1.5) |
|
|
Missing |
1 (0.9) |
4 (0.9) |
|
|
Received TNTa
|
|
|
0.973 |
|
No |
55 (80.9) |
237 (80.9) |
|
|
Yes |
13 (19.1) |
56 (19.1) |
|
|
Completion of neoadjuvant treatmenta
|
|
|
0.364 |
|
No |
6 (8.8) |
50 (17.1) |
|
|
Yes |
62 (91.2) |
243 (82.9) |
|
|
Neoadjuvant responsea
|
|
|
0.518 |
|
Complete response |
2 (3.8) |
15 (7.6) |
|
|
Stable disease |
16 (30.8) |
74 (37.4) |
|
|
Partial response |
32 (61.5) |
98 (49.5) |
|
|
Progressive disease |
2 (3.8) |
11 (5.6) |
|
|
Missing |
16 (23.5) |
95 (32.4) |
|
|
Underwent surgery |
|
|
0.152 |
|
No |
8 (6.9) |
18 (3.8) |
|
|
Yes |
108 (93.1) |
451 (96.2) |
|
|
Surgery approachb
|
|
|
0.944 |
|
Open |
40 (37.0) |
166 (36.8) |
|
|
Laparoscopic |
62 (57.4) |
256 (56.8) |
|
|
Laparoscopic conversion to open |
4 (3.7) |
20 (4.4) |
|
|
Missing |
2 (1.9) |
9 (2.0) |
|
|
Type of surgeryb
|
|
|
0.762 |
|
Sphincter-preserving procedure |
82 (75.9) |
348 (77.2) |
|
|
Abdominoperineal resection |
24 (22.2) |
94 (20.8) |
|
|
Missing |
2 (1.9) |
9 (2.0) |
|
|
Pathological T category (pT/ypT)b
|
|
|
0.891 |
|
T0 |
11 (10.2) |
36 (8.0) |
|
|
T1–T2 |
27 (25.0) |
119 (26.4) |
|
|
T3 |
56 (51.9) |
240 (53.2) |
|
|
T4 |
14 (13.0) |
56 (12.4) |
|
|
Pathological N category (pN/ypN)b
|
|
|
0.152 |
|
N0 |
68 (63.0) |
251 (55.7) |
|
|
N+ |
36 (33.3) |
184 (40.8) |
|
|
Missing |
4 (3.7) |
16 (3.5) |
|
|
Surgical marginb
|
|
|
0.655 |
|
Negative |
99 (91.7) |
420 (93.1) |
|
|
Distal positive |
1 (0.9) |
3 (0.7) |
|
|
CRM positive |
8 (7.4) |
27 (6.0) |
|
|
Missing |
0 (0) |
1 (0.2) |
|
|
Adjuvant treatmentb
|
|
|
0.279 |
|
No |
20 (18.5) |
100 (22.2) |
|
|
Chemotherapy |
80 (74.1) |
293 (65.0) |
|
|
CRT |
8 (7.4) |
56 (12.4) |
|
|
RT |
0 (0) |
2 (0.4) |
|
|
Completion of adjuvant treatmentc
|
|
|
0.621 |
|
No |
40 (45.5) |
148 (42.1) |
|
|
Yes |
48 (54.5) |
200 (57.0) |
|
|
Missing |
0 (0) |
3 (0.9) |
|
REFERENCES
- 1. Ervik M, Laversanne M, Lam F, Colombet M, Ferlay J, Filho AM, et al. Global Cancer Observatory: cancer tomorrow (2024 estimates, version 1.0). International Agency for Research on Cancer (IARC); 2026 [cited 2025 Sep 9]. Available from: https://gco.iarc.who.int/tomorrow
- 2. Chambers AC, Dixon SW, White P, Williams AC, Thomas MG, Messenger DE. Demographic trends in the incidence of young-onset colorectal cancer: a population-based study. Br J Surg 2020;107:595–605. ArticlePubMedPMCPDF
- 3. Howren A, Sayre EC, Loree JM, Gill S, Brown CJ, Raval MJ, et al. Trends in the incidence of young-onset colorectal cancer with a focus on years approaching screening age: a population-based longitudinal study. J Natl Cancer Inst 2021;113:863–8. ArticlePMCPDF
- 4. Kim BJ, Hanna MH. Colorectal cancer in young adults. J Surg Oncol 2023;127:1247–51. ArticlePubMed
- 5. White MG, Damania A, Alshenaifi J, Sahasrabhojane P, Peacock O, Losh J, et al. Young-onset rectal cancer: unique tumoral microbiome and correlation with response to neoadjuvant therapy. Ann Surg 2023;278:538–48. ArticlePubMedPMC
- 6. Barreto SG. Young-onset rectal cancer patients: in need of answers. Future Oncol 2019;15:1053–5. ArticlePubMed
- 7. Shamsuddin AM, Phelps PC, Trump BF. Human large intestinal epithelium: light microscopy, histochemistry, and ultrastructure. Hum Pathol 1982;13:790–803. ArticlePubMed
- 8. Li FY, Lai MD. Colorectal cancer, one entity or three. J Zhejiang Univ Sci B 2009;10:219–29. ArticlePubMedPMCPDF
- 9. Arai T, Kino I. Morphometrical and cell kinetic studies of normal human colorectal mucosa. Comparison between the proximal and the distal large intestine. Acta Pathol Jpn 1989;39:725–30. ArticlePubMed
- 10. Lee YC, Lee YL, Chuang JP, Lee JC. Differences in survival between colon and rectal cancer from SEER data. PLoS One 2013;8:e78709. ArticlePubMedPMC
- 11. You YN, Dozois EJ, Boardman LA, Aakre J, Huebner M, Larson DW. Young-onset rectal cancer: presentation, pattern of care and long-term oncologic outcomes compared to a matched older-onset cohort. Ann Surg Oncol 2011;18:2469–76. ArticlePubMedPDF
- 12. Barreto SG, Chaubal GN, Talole S, DeSouza A, Suradkar K, Gaikwad V, et al. Rectal cancer in young Indians--are these cancers different compared to their older counterparts? Indian J Gastroenterol 2014;33:146–50. ArticlePubMedPDF
- 13. Kolarich A, George TJ, Hughes SJ, Delitto D, Allegra CJ, Hall WA, et al. Rectal cancer patients younger than 50 years lack a survival benefit from NCCN guideline-directed treatment for stage II and III disease. Cancer 2018;124:3510–9. ArticlePubMedPMCPDF
- 14. O'Connell JB, Maggard MA, Liu JH, Etzioni DA, Livingston EH, Ko CY. Do young colon cancer patients have worse outcomes? World J Surg 2004;28:558–62. ArticlePubMedPDF
- 15. Sanford NN, Dee EC, Ahn C, Kazmi SA, Beg MS, Folkert MR, et al. Recent trends and overall survival of young versus older adults with stage II to III rectal cancer treated with and without surgery in the United States, 2010-2015. Am J Clin Oncol 2020;43:694–700. ArticlePubMed
- 16. Ramspek CL, Steyerberg EW, Riley RD, Rosendaal FR, Dekkers OM, Dekker FW, et al. Prediction or causality? A scoping review of their conflation within current observational research. Eur J Epidemiol 2021;36:889–98. ArticlePubMedPMCPDF
- 17. Tennant PW, Murray EJ, Arnold KF, Berrie L, Fox MP, Gadd SC, et al. Use of directed acyclic graphs (DAGs) to identify confounders in applied health research: review and recommendations. Int J Epidemiol 2021;50:620–32. ArticlePubMedPMCPDF
- 18. Cuschieri S. The STROBE guidelines. Saudi J Anaesth 2019;13(Suppl 1): S31–4. ArticlePubMedPMC
- 19. Boardman LA, Vilar E, You YN, Samadder J. AGA clinical practice update on young adult-onset colorectal cancer diagnosis and management: expert review. Clin Gastroenterol Hepatol 2020;18:2415–24. Article
- 20. Kalady MF. Rectal cancer in young patients: time to take notice. Cancer 2018;124:3474–5. ArticlePubMedPDF
- 21. Yuan HY, Li Y, Yang GL, Bei DJ, Wang K. Study on the causes of local recurrence of rectal cancer after curative resection:analysis of 213 cases. World J Gastroenterol 1998;4:527–9. ArticlePubMedPMC
- 22. Farhat W, Azzaza M, Mizouni A, Ammar H, Ben Ltaifa M, Lagha S, et al. Factors predicting recurrence after curative resection for rectal cancer: a 16-year study. World J Surg Oncol 2019;17:173.ArticlePDF
- 23. Okamura R, Itatani Y, Fujita Y, Hoshino N, Okumura S, Nishiyama K, et al. Postoperative recurrence in locally advanced rectal cancer: how does neoadjuvant treatment affect recurrence pattern? World J Surg Oncol 2023;21:247.ArticlePubMedPMCPDF
- 24. Pilipshen SJ, Heilweil M, Quan SH, Sternberg SS, Enker WE. Patterns of pelvic recurrence following definitive resections of rectal cancer. Cancer 1984;53:1354–62. ArticlePubMed
- 25. Yu TK, Bhosale PR, Crane CH, Iyer RB, Skibber JM, Rodriguez-Bigas MA, et al. Patterns of locoregional recurrence after surgery and radiotherapy or chemoradiation for rectal cancer. Int J Radiat Oncol Biol Phys 2008;71:1175–80. ArticlePubMedPMC
- 26. Stornes T, Wibe A, Romundstad PR, Endreseth BH. Outcomes of rectal cancer treatment: influence of age? Int J Colorectal Dis 2014;29:825–34. ArticlePubMedPDF
- 27. Dinaux AM, Leijssen LG, Bordeianou LG, Kunitake H, Berger DL. Rectal cancer in patients under 50 years of age. J Gastrointest Surg 2017;21:1898–905. ArticlePubMedPDF
- 28. Zhang Y, Yan L, Wu Y, Xu M, Liu X, Guan G. Worse treatment response to neoadjuvant chemoradiotherapy in young patients with locally advanced rectal cancer. BMC Cancer 2020;20:854.ArticlePDF
Citations
Citations to this article as recorded by
