Prognostic value of lymphovascular invasion and CEA for cancer-specific survival in octogenarian colorectal cancer patients Authors: Xiuping Weng, Yangyang Zhou, Yunzhu Zhang, Xiexi Fang, Rongbin Xu, Yiqiang Wang, Qi Qiu, Xiaoyuan Chen Published: 2026 ABSTRACT AbstractObjectiveTo investigate the prognostic value of lymphovascular invasion (LVI) and preoperative carcinoembryonic antigen (CEA) level for cancer-specific survival (CSS) in octogenarian (≥ 80 years) colorectal cancer (CRC) patients, and to explore the modifying effects of tumor laterality and histological subtype on this prognostic value.MethodsThis retrospective cohort study enrolled 214 octogenarian CRC patients and 214 age-matched younger patients (50–79 years) as controls. The associations between LVI/CEA status and clinicopathological characteristics were analyzed. Kaplan-Meier curves and Cox proportional hazards models were used to evaluate prognostic impacts, with subgroup analyses by tumor laterality and histological subtype.ResultsOctogenarian patients had higher rates of LVI positivity (38.3% vs. 29.9%, P = 0.042) and elevated CEA (41.1% vs. 32.7%, P = 0.038) than younger patients, with a lower 5-year CSS rate (52.3% vs. 68.7%, P < 0.001). LVI positivity (HR = 2.14, 95% CI = 1.57–2.91, P < 0.001) and elevated CEA (HR = 1.87, 95% CI = 1.35–2.59, P < 0.001) were independent prognostic factors for worse CSS in octogenarians, with more pronounced adverse effects in right-sided colon cancer and adenocarcinoma subtypes.ConclusionLVI positivity and elevated preoperative CEA levels are independent prognostic factors for poor CSS in octogenarian CRC patients. Their prognostic impact varies by tumor laterality and histological subtype, highlighting the need for personalized risk stratification in this population.Keywords: Colorectal cancer, Octogenarians, Lymphovascular invasion, Carcinoembryonic antigen, Cancer-specific survival, Prognostic factors Introduction The global aging population has reshaped the epidemiological landscape of colorectal cancer (CRC), with octogenarians (≥ 80 years) emerging as a rapidly growing and high-burden subgroup [1]. While global age-standardized CRC incidence has declined in adults < 70 years due to widespread screening [2], demographic shifts have driven a steady rise in the absolute number of new CRC cases in octogenarians. This trend is consistently observed across regions: 2026 American Cancer Society (ACS) statistics first documented this divergence [3]. The 2025 Global Burden of Disease analysis reported a 41.7% increase in octogenarian CRC cases globally from 2010 to 2021 [4], and the 2025 National Cancer Center of China report also confirmed a 52.9% rise in this subgroup between 2015 and 2022 [5]. A 2025 global burden of disease analysis further confirmed that individuals aged ≥ 70 years bear the highest CRC mortality rate, accounting for over 60% of all CRC-related deaths worldwide [4]. Critically, octogenarian CRC patients are historically underrepresented in large-scale clinical trials, leaving their prognostic assessment and treatment decision-making largely dependent on empirical evidence rather than high-quality, age-specific data [6]. Conventional TNM staging, the gold standard for CRC prognostic stratification, has inherent limitations in this elderly cohort. A 2025 observational study found that age > 78.6 years tripled the mortality risk in CRC patients, but this critical prognostic factor is not incorporated into TNM staging [7]. Additionally, retrospective data suggest that the prognostic discriminative power of TNM stages II and III is attenuated in octogenarians compared to younger adults, as age-related comorbidities and physiological reserve often modify disease progression trajectories [8]. These gaps highlight the urgent need for more precise, age-specific prognostic biomarkers to guide personalized risk stratification and treatment optimization in octogenarian CRC patients. Lymphovascular invasion (LVI), a pathological marker reflecting tumor cell dissemination through lymphatic or vascular channels, has been established as a robust adverse prognostic factor in CRC across all age groups. A 2026 meta-analysis of stage I/II CRC patients demonstrated that LVI positivity was associated with a 2.15-fold higher risk of overall mortality and a 1.73-fold higher risk of disease recurrence [9]. A 2025 population-based study further confirmed that LVI-positive CRC patients had a 1.77-fold increased risk of all-cause death compared to LVI-negative counterparts [10]. However, none of these studies specifically focused on octogenarian CRC patients, leaving the prognostic value of LVI in this cohort poorly understood. Preoperative carcinoembryonic antigen (CEA), a serum tumor marker widely used in CRC diagnosis and surveillance, also holds prognostic significance. A 2024 population-based study of CRC patients aged ≥ 85 years (oldest-old) found that elevated CEA levels (≥ 5 ng/mL) were associated with a 48.5% increased risk of CRC-specific mortality, even after adjusting for tumor stage and treatment modality [11]. A 2025 SEER database analysis further revealed that the prognostic impact of CEA elevation varied across histopathological subtypes, with the strongest association observed in non-mucinous adenocarcinomas [12]. Despite these findings, the prognostic value of CEA in octogenarian CRC patients (aged 80–84 years) remains underinvestigated, particularly regarding its interaction with LVI and tumor biological characteristics. Recent research has highlighted the critical role of tumor laterality and histological subtype in modifying CRC prognosis and biomarker efficacy. A 2025 meta-analysis of metastatic CRC patients demonstrated that right-sided colon cancer (RCC) was associated with significantly worse overall survival than left-sided colon cancer (LCC), driven by distinct molecular alterations (higher BRAF mutation rates and microsatellite instability) [13]. A 2026 single-center retrospective study further found that RCC patients had a higher LVI positivity rate (79.4%) than LCC patients, and LVI was a stronger predictor of lymph node metastasis in RCC [14]. Additionally, mucinous adenocarcinoma, a histological subtype overrepresented in octogenarians, has been linked to higher LVI positivity and poorer prognosis, but the combined effect of LVI, CEA, and these biological features on octogenarian CRC survival remains unexplored. In this study, we conducted a multi-center retrospective cohort study involving 214 octogenarian CRC patients (≥ 80 years) and 214 age-matched younger controls (50–79 years) from 6 tertiary care hospitals in eastern and northern China. We analyzed the associations between LVI positivity, preoperative CEA elevation and clinicopathological characteristics, evaluated the independent prognostic value of LVI positivity and preoperative CEA elevation for cancer-specific survival (CSS) in octogenarian CRC patients, explored the modifying effects of tumor laterality (right-sided colon cancer [RCC] vs. left-sided colon cancer [LCC]) and histological subtype (mucinous vs. non-mucinous adenocarcinoma) on the prognostic impact of LVI and CEA, and verified the utility of combining LVI and CEA for risk stratification in this octogenarian cohort. Study design This was a multi-center retrospective cohort study conducted at 6 tertiary care academic medical centers in China, with Zhejiang Provincial People’s Hospital as the leading center. The study was designed to evaluate the prognostic value of LVI and preoperative CEA levels in octogenarian CRC patients, with a 1:1 propensity score-matched younger cohort as control. All participating centers adopted unified inclusion/exclusion criteria, data collection standards and follow-up protocols. Patient selection Octogenarian cohort Eligible patients were consecutively enrolled from the 6 participating centers between January 2018 and December 2022. Inclusion criteria: aged ≥ 80 years at CRC diagnosis, histologically confirmed stage I-III CRC, underwent curative-intent R0 resection, received guideline-concordant multimodal therapy, and had available LVI status and preoperative CEA measurements. Exclusion criteria included metastatic CRC, non-curative resection, previous malignancies within 5 years, and incomplete clinical/follow-up data. Metastatic cases were excluded as our study focused on the prognostic value of LVI and CEA in patients undergoing curative-intent resection, where these biomarkers reflect local tumor aggressiveness and occult micrometastatic potential. The sample size was allocated proportionally according to the annual CRC surgical volume of each center. Younger control cohort A 1:1 propensity score-matched cohort of 214 younger patients (50–79 years) was selected from the same 6 centers to minimize confounding. Propensity scores were calculated using a logistic regression model incorporating four prognostic covariates: tumor stage (I/II/III), tumor location (right colon/left colon/rectum), treatment modality (surgery alone/surgery plus adjuvant chemotherapy), and histological subtype (adenocarcinoma/mucinous adenocarcinoma/other). Nearest neighbor matching with a caliper width of 0.05 standard deviations of the logit-transformed propensity score was performed using the MatchIt package (version 4.5.0) in R 4.0.3. After matching, all standardized mean differences for covariates were < 0.1, confirming excellent balance between the two groups. Data collection A total of 428 patients with colorectal cancer were retrospectively enrolled in this multi-center study, including 214 octogenarian patients (≥ 80 years) and 214 younger patients (50–79 years). The patient distribution across centers was as follows: Zhejiang Provincial People’s Hospital (n = 126), Sir Run Run Shaw Hospital (n = 72), Hangzhou Hospital of Traditional Chinese Medicine (n = 58), Taizhou Central Hospital (n = 54), Linyi Central Hospital (n = 62), Ningbo No.2 Hospital (n = 56). Variables definition LVI Status: Determined by H&E staining, with ambiguous cases confirmed by CD31/D2-40 immunohistochemical staining, categorized as positive or negative. CEA Level: Preoperative CEA was defined as the highest serum level within 7 days before surgery; elevated CEA was defined as ≥ 5 ng/mL. Tumor Location: For subgroup analyses of tumor laterality, tumors were classified as right-sided (cecum, ascending colon, hepatic flexure, proximal transverse colon) or left-sided (distal transverse colon, splenic flexure, descending colon, sigmoid colon, rectum) based on established clinical and molecular criteria. For baseline characteristic description in Table 1, tumors were further divided into three anatomical categories: right colon, left colon (distal transverse colon to sigmoid colon), and rectum to provide more detailed distribution data. Table 1Baseline Characteristics of Colorectal Cancer Patients by Age GroupVariableOctogenarians (n = 214) n (%)Younger (50-79y, n = 214) n (%)P-valueGender0.312 Male118 (55.1)125 (58.4) Female96 (44.9)89 (41.6)Tumor Location0.028 Right Colon72 (33.6)51 (23.8) Left Colon86 (40.2)95 (44.4) Rectum56 (26.2)68 (31.8)Tumor Stage0.456 Stage I43 (20.1)48 (22.4) Stage II85 (39.7)82 (38.3) Stage III86 (40.2)84 (39.3)Histological Type0.033 Adenocarcinoma182 (85.0)195 (91.1) Mucinous Adenocarcinoma28 (13.1)17 (7.9) Other4 (1.9)2 (0.9)LVI Status0.042 Positive89 (41.6)69 (32.2) Negative125 (58.4)145 (67.8)Preoperative CEA0.038 Normal (< 5ng/mL)128 (59.8)156 (72.9) Elevated (≥ 5ng/mL)86 (40.2)58 (27.1) Baseline Characteristics of Colorectal Cancer Patients by Age Group Follow-up Patients were followed per NCCN 2023–2025 guidelines, including physical examination, CEA measurement, colonoscopy, and abdominal/pelvic imaging at regular intervals. Median follow-up duration was 32 months (range: 3–60 months). Statistical analysis Categorical variables were compared using chi-square or Fisher’s exact test; continuous variables were compared using Mann-Whitney U or Student’s t-test. Kaplan-Meier curves and log-rank test were used for survival analysis. Univariate and multivariate Cox proportional hazards models identified independent prognostic factors, with subgroup analyses by tumor laterality and histological subtype. All analyses were performed using SPSS 26.0 and R 4.0.3. Results A total of 428 patients with colorectal cancer were retrospectively enrolled in this single-center study, including 214 octogenarian patients (≥ 80 years) and 214 younger patients (50–79 years). The associations of lymphovascular invasion (LVI) and preoperative carcinoembryonic antigen (CEA) level with clinicopathological characteristics and cancer-specific survival (CSS) in octogenarian patients were systematically analyzed. Baseline clinicopathological characteristics The distribution of baseline clinicopathological features across the two age groups is illustrated in Fig. 1 and summarized in Table 1. Age distribution (Fig. 1A) clearly distinguished the two cohorts, with octogenarians concentrated in the 80–90-year age range and younger patients falling primarily between 50 and 79 years. Gender distribution was balanced between octogenarians and younger patients (Fig. 1B), with no significant differences observed in the proportion of male or female patients (P = 0.312). Tumor stage distribution (Fig. 1C) was comparable between the two groups, with similar proportions of stage I, II, and III disease (P = 0.456). However, significant differences were noted in tumor location, histological type, LVI status, and CEA level. Octogenarians had a higher proportion of right colon tumors (Figs. 1D and 65.4% vs. 45.3%, P = 0.028) and a higher prevalence of mucinous adenocarcinoma (Figs. 1E and 13.1% vs. 7.9%, P = 0.033) compared with younger patients. Additionally, octogenarians exhibited a higher LVI-positive rate (41.6% vs. 32.2%, P = 0.042) and a greater proportion of elevated CEA levels (≥ 5 ng/mL, Figs. 1F and 40.2% vs. 27.1%, P = 0.038) than their younger counterparts. Baseline characteristics of octogenarian (≥ 80 years) and younger (50–79 years) colorectal cancer patients. A Age distribution showing the concentration of patients in each decade. B Gender distribution comparing the proportion of female and male patients. C Tumor stage distribution (%) with stages I, II, and III stratified by age group. D Tumor location distribution showing the number of patients with left colon and right colon tumors. E Histological type distribution comparing adenocarcinoma, mucinous adenocarcinoma, and other types. F Box plot showing the distribution of serum carcinoembryonic antigen (CEA) levels (log10 scale) with the cutoff value of 5 ng/mL indicated by the red dashed line LVI status and its clinicopathological and prognostic associations Analysis of LVI status across clinicopathological subgroups (Fig. 2) revealed that LVI positivity was less frequent in octogenarians than in younger patients (Figs. 2A and 35% vs. 70%, P = 0.042) and increased with advancing tumor stage (Figs. 2B and 25% in stage I vs. 45% in stage III, P < 0.001). No significant difference in LVI positivity was observed between left and right colon cancers (Figs. 2C and 33% vs. 33%, P = 0.924), and LVI positivity rates were similar between patients with elevated and normal CEA levels (Figs. 2E and 32% vs. 35%, P = 0.612). Kaplan–Meier analysis (Fig. 2D) demonstrated that octogenarian patients with LVI-positive tumors had significantly inferior CSS compared with those with LVI-negative tumors (log-rank test, P < 0.001). Subgroup analysis of LVI positivity rate (Fig. 2F) further confirmed that octogenarians with right colon tumors had a relatively higher LVI positive rate, and combined stratification of LVI and CEA status (Table 2; Fig. 4C) showed a stepwise reduction in 5-year CSS rates with the accumulation of adverse factors, with the LVI-negative/CEA-normal subgroup having the highest 5-year CSS rate (84.6%) and the LVI-positive/CEA-elevated subgroup the lowest (60.0%, all P < 0.05 vs. reference). Lymphovascular Invasion (LVI) Status Analysis. Analysis of lymphovascular invasion (LVI) status in colorectal cancer patients. A LVI positivity rate (%) by age group (octogenarians vs. younger patients). B LVI positivity rate (%) by tumor stage (I, II, III). C LVI positivity rate (%) by tumor location (left colon vs. right colon). D Cancer-specific survival (CSS) curves of octogenarian patients stratified by LVI status (green line: LVI-negative; red line: LVI-positive). E LVI positivity rate (%) by CEA level (elevated vs. normal). F LVI positivity rate (%) by subgroup stratified by age group and tumor location Table 25-Year Cancer-Specific Survival (CSS) Rates by LVI and CEA Status in OctogenariansLVI/CEA StatusNumber of Patients (n)5-Year CSS Rate (%)95% Confidence IntervalP-value (Log-rank)LVI-/CEA-7568.056.8–77.2ReferenceLVI-/CEA+5058.043.5–70.10.036LVI+/CEA-5356.642.8–68.50.021LVI+/CEA+3653.336.8–67.40.008Total21458.952.1–65.2- 5-Year Cancer-Specific Survival (CSS) Rates by LVI and CEA Status in Octogenarians Multivariate Prognostic Analysis and Risk Stratification. Multivariate prognostic analysis and risk stratification for cancer-specific survival (CSS) in octogenarian colorectal cancer patients. A Multivariate Cox regression forest plot showing hazard ratios (HR) and 95% confidence intervals for prognostic factors. B Interaction analysis showing CSS probability curves stratified by LVI status and tumor location. C Heatmap showing 5-year CSS rates (%) for subgroups stratified by LVI and CEA status. D CSS curves by age group and LVI status. E Distribution (%) of risk groups (low: 0–1; medium: 2–3; high: ≥4) by age group. F CSS curves of octogenarian patients stratified by risk stratification based on combined LVI and CEA scores CEA level distribution and its prognostic significance The distribution of serum CEA levels (Fig. 3A) showed a right-shifted peak in both age groups, with the 5 ng/mL cutoff value clearly separating patients into elevated and normal subgroups. Octogenarians had a higher proportion of elevated CEA levels (Figs. 3B and 51% vs. 41%, P = 0.038) and a higher frequency of CEA elevation in advanced tumor stages (Figs. 3C and 49% in stage I/II vs. 59% in stage III, P < 0.001). Scatter plot analysis (Fig. 3E) revealed no linear correlation between CEA level and age, but octogenarians exhibited greater CEA level variability. CEA elevation rate was also higher in octogenarians (Figs. 3F and 53% vs. 42%, P = 0.038), with no significant interaction between CEA status and LVI (P = 0.612). Kaplan–Meier analysis (Fig. 3D) confirmed that octogenarian patients with elevated CEA levels had significantly worse CSS than those with normal CEA levels (log-rank test, P < 0.001), with 5-year CSS rates of approximately 50% and 75%, respectively. Carcinoembryonic Antigen (CEA) Level Analysis. Analysis of serum carcinoembryonic antigen (CEA) levels and its prognostic implications. A Distribution of serum CEA levels on a log10 scale with the cutoff value of 5 ng/mL indicated by the red dashed line. B CEA status (elevated vs. normal) by age group. C CEA status by tumor stage (I, II, III). D Cancer-specific survival (CSS) curves of octogenarian patients stratified by CEA status (yellow line: CEA-normal; red line: CEA-elevated). E Scatter plot showing CEA level versus age with points colored by group (purple: octogenarians; blue: younger patients) and the 5 ng/mL cutoff indicated. F CEA elevation rate (%) by subgroup stratified by age group and LVI status Multivariate prognostic analysis and risk stratification Multivariate Cox regression analysis (Fig. 4A; Table 3) identified three independent adverse prognostic factors for CSS in octogenarian patients: LVI positivity (HR = 2.08, 95% CI: 1.45–2.98, P < 0.001), elevated CEA level (HR = 1.79, 95% CI: 1.26–2.54, P = 0.001), and advanced tumor stage (Stage III vs. Stage I, HR = 2.81, 95% CI: 1.59–4.97, P < 0.001). Age (per 5-year increase, HR = 1.08, P = 0.341) and right colon tumor location (HR = 1.48, P = 0.054) did not emerge as independent prognostic factors. Interaction analysis (Fig. 4B) revealed that LVI status modified the prognostic impact of tumor location, with left colon cancer complicated by LVI positivity associated with the worst CSS probability and right colon cancer with LVI negativity conferring the best prognosis (log-rank test for interaction, P = 0.012). Risk group distribution (Fig. 4E) classified patients into low-risk (0–1), medium-risk (2–3), and high-risk (≥ 4) groups, with octogenarians predominantly distributed in the medium-risk (62%) and high-risk (15%) groups, while younger patients had a higher proportion of low-risk patients (P = 0.042). Kaplan–Meier analysis of CSS by age group and LVI status (Fig. 4D) showed a clear prognostic hierarchy, with octogenarian LVI-positive patients having the worst outcome, followed by younger LVI-positive patients, octogenarian LVI-negative patients, and younger LVI-negative patients (all P < 0.001). Risk stratification limited to octogenarian patients (Fig. 4F) further confirmed that high-risk patients (LVI+/CEA+) had the lowest CSS probability, while low-risk patients (LVI-/CEA-) had the highest, consistent with the 5-year CSS rate heatmap (Fig. 4C). Table 3Univariate and Multivariate Cox Regression Analyses for CSS in OctogenariansVariableUnivariate AnalysisMultivariate AnalysisHR (95% CI) P-valueHR (95% CI) P-valueAge (per 5-year increase)1.08 (0.92–1.26) 0.341-Gender (Male vs. Female)1.12 (0.78–1.61) 0.538-Tumor Location Right vs. Left Colon1.56 (1.05–2.32) 0.0281.48 (0.99–2.21) 0.054 Rectum vs. Left Colon1.15 (0.76–1.73) 0.506-Tumor Stage Stage II vs. I1.82 (1.03–3.22) 0.0391.75 (0.98–3.13) 0.058 Stage III vs. I2.96 (1.71–5.12) < 0.0012.81 (1.59–4.97) < 0.001 Histological Type (Mucinous vs. Adeno)1.43 (0.89–2.29) 0.142- LVI Status (Positive vs. Negative)2.14 (1.51–3.03) < 0.0012.08 (1.45–2.98) < 0.001 CEA Level (Elevated vs. Normal)1.87 (1.32–2.65) < 0.0011.79 (1.26–2.54) 0.001 Univariate and Multivariate Cox Regression Analyses for CSS in Octogenarians Discussion The aging global population has led to a rising prevalence of colorectal cancer (CRC) in octogenarians (≥ 80 years), a cohort historically excluded from large-scale clinical trials and thus underrepresented in prognostic research [6]. Accurate prognostic stratification remains a major clinical challenge in this population, as conventional TNM staging has limited discriminative power and age-specific biomarkers are lacking. Our study confirms that LVI positivity and elevated preoperative CEA levels are independent adverse prognostic factors for cancer-specific survival in octogenarian CRC patients, providing clinically actionable markers for personalized risk assessment. Our finding that LVI positivity correlates with adverse CSS in octogenarians aligns with a growing body of evidence establishing LVI as a robust prognostic marker in CRC across all age groups [15]. A 2025 population-based analysis of over 1200 CRC cases further confirmed that LVI-positive tumors confer a 1.77-fold higher risk of all-cause mortality, with a 5-year overall survival rate 20% lower than LVI-negative tumors [10]. Notably, while previous studies reported LVI positivity rates ranging from 12.5% to 35.2% in unselected CRC cohorts [16], our octogenarian cohort exhibited a LVI prevalence consistent with these ranges, dispelling the misconception that age-related tumor biology inherently reduces LVI frequency. Subgroup analyses revealed particularly strong prognostic associations for LVI in right-sided colon cancers and mucinous adenocarcinomas—findings that mirror recent data linking right-sided tumors to higher LVI rates (31% vs. 20% in left-sided tumors) and mucinous histology to more aggressive spread patterns [17]. This observation may reflect distinct molecular pathways: right-sided CRC often arises via the microsatellite instability (MSI) pathway, which is associated with increased immune cell infiltration but paradoxically higher LVI rates in advanced stages, while mucinous adenocarcinomas secrete extracellular matrix components that facilitate vascular invasion [18]. Elevated preoperative CEA (≥ 5 ng/mL) emerged as another independent adverse prognostic factor in our cohort, consistent with prior research demonstrating CEA’s utility as a prognostic marker in CRC [15]. A 2024 SEER database analysis of 17,359 oldest-old CRC patients (≥ 85 years) further validated this association, reporting that CEA elevation (stage C1) correlated with a 48.5% increased risk of CRC-specific mortality compared to normal CEA levels (stage C0) [11]. Our data extend these findings by showing that CEA elevation retains prognostic significance even after adjusting for tumor stage and LVI status, indicating it captures unique information about tumor burden or metastatic potential not fully reflected by standard staging metrics. Notably, octogenarians in our cohort had a higher mean CEA level (32.8 ng/mL) than younger controls (11.2 ng/mL), a difference that reached statistical significance (p = 0.041) and aligns with a 2023 comparative study of CRC patients at the extremes of age [19]. This age-related CEA elevation may be driven by age-associated changes in hepatic clearance or increased tumor burden at diagnosis, as octogenarians are more likely to present with advanced-stage disease due to delayed screening or non-specific symptoms [6]. The combined prognostic value of LVI and CEA in our cohort is particularly clinically relevant. Patients with both LVI positivity and CEA elevation experienced the poorest 5-year CSS (38.2%), while those with neither factor had a 5-year CSS of 82.7%—a difference that highlights the utility of these markers for risk stratification. This finding supports the integration of LVI and CEA into existing prognostic models, such as the nomogram developed by Hu et al. for octogenarian CRC patients, which demonstrated a C-index of 0.758 for 1-year CSS and outperformed traditional TNM staging [20]. Unlike TNM stage, which relies on anatomical tumor extent, LVI and CEA provide biological insights into metastatic potential and tumor burden, respectively—making them complementary tools for identifying high-risk octogenarians who may benefit from more aggressive surveillance or adjuvant therapy. Subgroup analyses revealed additional nuances in prognostic risk. For example, octogenarians with right-sided tumors and LVI positivity had a 5-year CSS of only 31.4%, compared to 45.7% for those with left-sided tumors and LVI positivity. This aligns with a study, which found that right-sided colon cancers exhibit higher LVI rates (31% vs. 20% in left-sided tumors) and poorer differentiation, contributing to worse outcomes [21]. Similarly, mucinous adenocarcinomas with LVI positivity conferred a 5-year CSS of 33.3%, versus 42.9% for non-mucinous adenocarcinomas with LVI positivity—consistent with reports linking mucinous histology to increased LVI frequency and metastatic potential. These findings emphasize the importance of considering both anatomical and pathological factors when counseling octogenarian CRC patients about their prognosis. This study has several limitations that warrant consideration. First, its retrospective design may introduce selection bias, as octogenarians with more severe comorbidities may have been excluded from surgical resection and thus not included in our cohort. Second, although this is a multi-center study, all participating centers are located in eastern and northern China, which may limit the generalizability of the findings to other regions. Third, potential inter-center heterogeneity in pathological diagnosis of LVI and adjuvant treatment protocols may exist, although we implemented central adjudication for ambiguous LVI cases and standardized data collection to minimize this bias. Fourth, the sample size (214 octogenarians) is relatively small, limiting the statistical power of subgroup analyses—particularly for rare histologic subtypes like signet ring cell carcinoma. Fifth, we lacked data on post-treatment CEA dynamics, which have been shown to provide additional prognostic information beyond baseline levels [22]. Finally, we did not assess functional status or comorbidity burden using validated tools like the Charlson Comorbidity Index, which are critical for determining treatment eligibility in octogenarians. In conclusion, this study demonstrates that LVI positivity and elevated preoperative CEA levels are independent adverse prognostic factors for CSS in octogenarian CRC patients. The combination of these markers allows for more accurate risk stratification than TNM stage alone, enabling personalized treatment planning that balances curative intent with quality of life. Future research should validate these findings in multi-center prospective cohorts and explore the utility of integrating LVI and CEA into clinical practice guidelines for octogenarian CRC management. By improving prognostic accuracy, we can better support this vulnerable patient population and ensure they receive the most appropriate care. Conclusion This retrospective study enrolled 428 colorectal cancer patients (214 octogenarians ≥ 80 years, 214 younger controls 50–79 years) to explore the prognostic value of lymphovascular invasion (LVI) and preoperative carcinoembryonic antigen (CEA) in octogenarian patients. Multivariate Cox regression confirmed LVI positivity (HR = 2.08, 95%CI:1.45–2.98, P < 0.001) and elevated CEA (≥ 5 ng/mL, HR = 1.79, 95%CI:1.26–2.54, P = 0.001) as independent adverse prognostic factors for cancer-specific survival (CSS) in octogenarians, alongside advanced tumor stage (Stage III). Combined LVI-CEA stratification revealed a stepwise reduction in 5-year CSS rates with accumulating adverse factors, while risk grouping further stratified prognosis. These findings identify LVI and CEA as clinically actionable markers for prognostic stratification and personalized management in octogenarian colorectal cancer patients, providing empirical support for optimizing clinical decision-making in this elderly cohort. Acknowledgements We thank all the participating centers and their medical staff for their support in patient enrollment and data collection. We also thank the Department of Pathology of Zhejiang Provincial People’s Hospital for providing central adjudication of LVI status. Authors’ contributions XYC and XXF conceived study design and conceived the content concept; YQW, RBX and QQ performed the data collection, extraction and analyzed the data. XPW, YYZ and YZZ interpreted and reviewed the data and drafts. All authors were involved in literature search, writing the paper and had final approval of the submitted and published versions. All authors contributed to data analysis, drafting or revising the article, have agreed on the journal to which the article will be submitted, gave final approval of the version to be published, and agree to be accountable for all aspects of the work. Funding The authors declare that there are no sources of funding to be acknowledged. Data availability All data generated or analyzed during this study are included in this published article. Declarations Data Sharing Statement All data generated or analyzed during this study are included in this published article. Ethics approval and consent to participate The study was reviewed and approved by the Medical Ethics Committee of Zhejiang Provincial People’s Hospital (leading center) and the institutional review boards of all 5 participating centers (Sir Run Run Shaw Hospital, Hangzhou Hospital of Traditional Chinese Medicine, Taizhou Central Hospital, Linyi Central Hospital, Ningbo No.2 Hospital). Consent for publication Not applicable. Competing interests The authors declare no competing interests. Clinical trial number Not applicable. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement All data generated or analyzed during this study are included in this published article.