Prediction of distant metastasis and survival of appendiceal cancer patients: a SEER population-based study

Introduction

Background

Appendiceal cancer (AC) is a rare malignancy, accounting for less than 1% of all cancers and about 0.5% of gastrointestinal tumors (1,2). ACs are heterogeneous, ranging from typical neuroendocrine tumors (NETs) to adenocarcinomas (ADs) (3). Recent work by Singh et al. (4) showed that the incidence of malignant appendiceal tumors continues to rise across different demographic groups. In clinical practice, primary appendiceal tumors often lack clear symptoms, and more than half of cases are incidentally diagnosed after acute appendicitis (5). Shaib et al. reported that in more than 50% of patients with appendiceal mucinous neoplasms, diffuse mucinous ascites is present, a condition referred to as pseudomyxoma peritonei (PMP); this presentation is associated with a worse prognosis than localized tumors without extra-appendiceal spread (6). Early symptoms are often vague and easily mistaken for other abdominal conditions. Most patients seek care for right lower abdominal pain or for abdominal or pelvic masses (7). Therefore, these tumors are most commonly discovered by chance during routine histopathological evaluation of appendectomy specimens removed for suspected acute appendicitis (8).

Rationale and knowledge gap

This incidental pattern of diagnosis creates two major challenges. First, preoperative imaging and laboratory tests often lack enough specificity to distinguish malignancy from benign inflammation, leading to delayed diagnosis (9). Second, there is no standardized post-operative assessment, which may lead to missed high-risk features associated with distant metastasis (DM). Given these difficulties, better tools are needed to predict DM and assess prognosis in AC. The tumor-node-metastasis (TNM) staging system is widely used to estimate cancer prognosis (10). However, it does not include key demographic characteristics such as age, gender, or race. This limits its ability to predict DM at diagnosis or to provide personalized survival estimates (11). In addition, Aguirre et al. noted that ACs differ clearly from other colorectal cancers in incidence, histology, and biological behavior. Their prognostic factors remain poorly defined and depend largely on tumor subtype (12). Thus, these factors should be incorporated into new prediction models to better estimate DM risk and prognosis in patients with AC.

Nomograms are visual tools based on regression models that estimate clinical outcomes and provide individualized risk predictions (13). They have clear advantages in cancer prognosis assessment. Several studies have developed nomograms to predict lymph node metastasis (LNM) in AC. These models may help guide decisions on right hemicolectomy in patients unexpectedly diagnosed with malignancy after appendectomy when lymph node status is unknown (14,15). Other studies, including that by Wang et al., identified factors linked to 5-year overall survival (OS) in AC, such as histologic type, age, race, tumor size, grade, chemotherapy, and TNM stage (16). However, few studies have focused on DM in AC. This lack of evidence leaves clinicians without reliable tools to estimate DM risk. It also highlights the need to develop nomograms specifically designed to predict DM in patients with AC. Given the substantial heterogeneity of AC and its implications for treatment and prognosis, we summarized the histologic subtype distribution of the study cohort as follows: of the 6,916 patients with AC included in this study, histologic subtypes were distributed as follows: NET, 2,874 (41.56%); mucinous adenocarcinoma (MA), 1,432 (20.71%); AD, 1,151 (16.64%); goblet cell carcinoid (GCC), 500 (7.23%); neuroendocrine carcinoma (NEC), 415 (6.00%); signet ring cell carcinoma (SRCC), 281 (4.06%); and mixed adenoneuroendocrine carcinoma (MANEC), 263 (3.80%).

Objective

This study uses data from the Surveillance, Epidemiology, and End Results (SEER) database, which covers more than 27% of the U.S. population and includes diverse demographic and clinical information. The aims are to (I) identify independent predictors of DM and poor survival in AC and (II) develop and validate nomograms that combine clinicopathological and demographic factors to predict DM risk and 1-, 3-, and 5-year survival. These tools are intended to support treatment decisions, improve follow-up strategies, and ultimately improve outcomes for patients with AC. We present this article in accordance with the TRIPOD reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1-2821/rc).

Methods

Data extraction and population inclusion

We used the Incidence-SEER Research Data, 17 Registries, November 2023 submission [2000–2021], to identify eligible cases. Information on T, N, and M stages for AC has been available in SEER since 2004. Therefore, we used SEER*Stat version 8.4.4 (https://seer.cancer.gov/seerstat/) to extract pathologically confirmed primary AC cases diagnosed between 2004 and 2021, based on “The International Classification of Diseases for Oncology, 3rd Edition (ICD-O-3) Hist/behav, malignant” and “C18.1-Appendix”.

Diagnostic confirmation was obtained from the SEER Variable Diagnostic Confirmation. The vast majority of cases were microscopically confirmed by positive histology, with only a very small proportion confirmed by exfoliative cytology or by radiography without microscopic confirmation.

Since the publicly available SEER database is composed of open-access and anonymous data, this study was exempt from ethical review. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Patients were included if they met all of the following criteria: (I) age 18 years or older at diagnosis; (II) only one primary malignant tumor; (III) diagnosis of AC (ICD-O-3 site code C18.1); and (IV) complete baseline clinical data, metastatic status, and survival information, including survival status and duration. Patients were excluded if any of the following applied: (I) diagnosis based only on autopsy or death certificate; (II) missing survival time or survival time of less than 1 month; (III) more than one primary malignancy; (IV) pathologically confirmed T0 tumors or carcinoma in situ; (V) unknown cause of death; or (VI) receipt of preoperative chemotherapy or radiotherapy.

Variable exhibition and outcomes

The following variables were collected, including age at diagnosis, tumor size, race (White, American Indian/Alaska Native, Asian/Pacific Islander, and Black), gender (female and male), marital status (married, divorced/widowed/single/separated, and other), grade (grade I, grade II, grade III, and grade IV), T stage (T1, T2, T3, and T4), N stage (N0, N1, and N2), DM (no/unknown and yes), radiation treatment (no/unknown and yes), chemotherapy (no/unknown and yes), surgery (no/unknown, local tumor destruction, and resection), survival status (alive, dead of cancer, and dead of other), and survival months. Age and tumor size were categorized using X-tile software in the training cohort. X-tile provides outcome-oriented, data-driven cutpoint selection based on log-rank statistics. The cutpoints used in this study were 30 and 50 years for age (18–30, 31–50, ≥51 years) and 30 mm (3 cm) for tumor size (≤3 vs. >3 cm) (Figure S1). Tumor histology was classified using ICD-O-3 codes. Code 8240 was defined as NET, 8243 as GCC, 8248 as MA, 8140 as AD, 8490 as SRCC, 8246 as NEC, and 8244 as MANEC. The main outcomes were DM, OS, and cancer-specific survival (CSS). DM status was recorded at the time of initial diagnosis. M stage was obtained from the SEER-derived American Joint Committee on Cancer (AJCC) M category at diagnosis (M0 vs. M1). Peritoneal metastasis is classified as M1 when recorded, although it may not be separately identifiable as “peritoneal” in the public SEER site-specific metastasis fields. OS was defined as the time from AC diagnosis to death from any cause or last follow-up. CSS was defined as the time from AC diagnosis to death caused by AC.

Construction and validation of the nomogram model

The “caret” package in R was used to randomly split patients into training and validation cohorts at a 7:3 ratio. The training cohort was used to develop and internally validate the nomogram, while the validation cohort was employed to external validation.

In the training cohort, we explored factors associated with DM using univariate logistic regression. Variables with P<0.05 in the univariable analysis were considered statistically significant and were subsequently entered into the multivariable logistic regression model. For OS and CSS, we applied univariate and multivariable Cox proportional hazards models. Variables with P<0.05 in univariate analyses were included in the multivariable models. Factors that remained significant were used to construct nomograms to estimate the risk of DM and to predict 1-, 3-, and 5-year OS and CSS.

Nomograms were created using the “nomogram” function from the “rms” package. Model performance was assessed in terms of discrimination and calibration. Discrimination was measured using the concordance index (C-index) from the “survival” package and the area under the curve (AUC) from the “pROC” package. Higher AUC values indicated better predictive accuracy. Calibration curves were used to compare predicted outcomes with observed results. Decision curve analysis (DCA) was performed to evaluate the clinical value of the models. All assessments were carried out in both the training and validation cohorts.

Statistical analysis

To facilitate analysis and interpretation, continuous variables were categorized using X-tile software. This approach allowed a clearer assessment of their association with outcomes. Baseline characteristics of the training and validation cohorts were summarized as counts and percentages. Differences between the two cohorts were assessed using the chi-square test. Survival differences between high- and low-risk groups defined by the prognostic models were evaluated using Kaplan-Meier (KM) curves and the log-rank test. We also developed a web-based dynamic nomogram using the “DynNom” package to help clinicians estimate DM risk, OS, and CSS in patients with AC. To assess possible selection bias, standardized mean differences (SMDs) were calculated to compare included patients with those excluded because of missing data. All statistical analyses and figures were generated using R version 4.4.2 (https://www.r-project.org/), with relevant packages such as rms, survival, and pROC. A two-sided P<0.05 was considered statistically significant.

Results

Patient enrollment and characteristics

After applying the inclusion and exclusion criteria, a total of 6,916 patients with AC and complete data were included in the analysis. After random allocation, 4,844 patients were assigned to the training cohort and 2,072 to the validation cohort (Figure 1). The cohort had a median survival of 38 [interquartile range (IQR), 17–72] months. As summarized in Table 1, most patients were White and female, and more than half were married. The most common histological subtype was NET (41.56%), followed by MA (20.71%) and AD (16.64%). ICD-O-3 differentiation grades were 59.36%, 25.72%, 13.74%, and 1.19% for grades I–IV, respectively. According to the AJCC 8th edition, T stages were 42.52% (T1), 8.14% (T2), 19.78% (T3), and 29.55% (T4); N stages were 81.87% (N0), 12.88% (N1), and 5.25% (N2); and M stages were 82.14% (M0) and 17.86% (M1). The majority of patients were treated with surgery (91.12%) while chemotherapy and radiation therapy were less commonly employed, at 26.56% and 1.13%, respectively. Detailed clinicopathological characteristics are shown in Table 1.

Figure 1 Flowchart of the process of data selection. SEER, Surveillance, Epidemiology, and End Results.

Baseline characteristics were comparable between the training and validation cohorts, with no statistically significant differences across variables (all P>0.05; Table 1). To assess potential selection bias due to missing data, SMDs were also calculated to compare included patients with those excluded for missing information (Table S1). Most variables showed small differences between included and excluded patients, whereas larger differences were observed for surgery status and year of diagnosis.

Univariate and multivariate logistic analyses of risk factors for DM in AC

Univariate logistic analysis showed that marital status, gender, race, age, tumor size, grade, T stage, N stage, and histological type were associated with DM in AC (all P<0.05, Table 2). Multivariable results showed that male sex was linked to a lower risk of DM compared with female sex [odds ratio (OR) = 0.52; 95% confidence interval (CI) 0.42–0.63; P<0.001]. Tumor size greater than 3 cm increased the risk of DM (OR =1.45; 95% CI: 1.18–1.79; P<0.001). Advanced local and nodal disease showed strong associations with presence of DM, including T4 stage (OR =6.88; 95% CI: 3.90–12.13; P<0.001) and N1 (OR =2.40; 95% CI: 1.86–3.09) or N2 stage (OR =3.37; 95% CI: 2.38–4.76), both P<0.001. Several histological subtypes were strongly associated with DM. These included MA (OR =23.97), SRCC (OR =22.61), MANEC (OR =13.62), AD (OR =9.51), and GCC (OR =6.39), all P<0.001. Patients who were unmarried (divorced, widowed, single, or separated) had a lower risk of DM than married patients (OR =0.65; 95% CI: 0.53–0.80; P<0.001).

Development and validation of a prediction model for DM in AC

A nomogram was constructed to estimate the probability of DM in patients with AC (Figure 2A). Each predictor was assigned a score based on individual patient characteristics, and the total score was then used to estimate DM risk.

Figure 2 Development and validation of the DM nomogram. (A) Nomogram for predicting DM in AC. (B,E) ROC curves for the training and validation cohorts. (C,F) Calibration curves for the training and validation cohorts. (D,G) Decision curve analyses for the training and validation cohorts. AC, appendiceal cancer; AD, adenocarcinoma; AUC, area under the curve; CI, confidence interval; D/W/S/S, divorced/widowed/single/separated; DM, distant metastasis; GCC, goblet cell carcinoid; MA, mucinous adenocarcinoma; MANEC, mixed adenoneuroendocrine carcinoma; N, node; NEC, neuroendocrine carcinoma; NET, neuroendocrine tumor; ROC, receiver operating characteristic; SRCC, signet ring cell carcinoma; T, tumor.

In the training cohort, the ROC curve, calibration curve, and DCA are shown in Figure 2B-2D, respectively. In the validation cohort, the corresponding ROC curve, calibration curve, and DCA are shown in Figure 2E-2G, respectively. The AUC was 0.916 (95% CI: 0.908–0.925) in the training cohort and 0.905 (95% CI: 0.890–0.919) in the validation cohort, indicating strong discriminatory ability. Calibration curves showed close agreement between predicted and observed outcomes in both cohorts. DCA showed that the model provided net clinical benefit across a wide range of threshold probabilities.Overall, the nomogram showed good accuracy and consistency.

Identification of independent prognostic factors for OS and CSS

In the training cohort, univariate analysis identified marital status, sex, race, age, tumor size, grade, TNM stage, surgery, radiation, chemotherapy, and histology as factors associated with OS. After multivariable adjustment, marital status, age, grade, TNM stage, chemotherapy, and histology remained independent predictors (Table 3).

Patients aged ≥51 years had a higher risk of death compared with those aged 18–30 years [hazard ratio (HR) =2.30; 95% CI: 1.46–3.61; P<0.001]. Poor differentiation remained a strong predictor, with grade III (HR =2.29) and grade IV tumors (HR =1.89) showing worse outcomes. DM was also independently associated with poor OS (HR =2.91; P<0.001). SRCC carried a particularly high risk (HR =4.18). Chemotherapy was protective (HR =0.73), while unmarried status was linked to higher mortality (HR =1.37).

For CSS, similar predictors were identified. High-grade tumors, advanced TNM stage (T3–T4, N1–N2, and M1), and aggressive histology, especially SRCC (HR =9.17), were associated with worse CSS. Chemotherapy (HR =0.82) and surgical resection (HR =0.65) both showed protective effects (Table 4). Based on these findings, nomograms were developed to predict 1-, 3-, and 5-year OS and CSS (Figure 3A,3B).

Figure 3 Nomograms for predicting OS and CSS in patients with AC. (A) Nomogram for predicting 1-, 3-, and 5-year OS. (B) Nomogram for predicting 1-, 3-, and 5-year CSS. AC, appendiceal cancer; AD, adenocarcinoma; CSS, cancer-specific survival; GCC, goblet cell carcinoid; M, metastasis; MA, mucinous adenocarcinoma; MANEC, mixed adenoneuroendocrine carcinoma; N, node; NEC, neuroendocrine carcinoma; NET, neuroendocrine tumor; OS, overall survival; SRCC, signet ring cell carcinoma; T, tumor.

Validation of the prediction models of the nomogram

The OS and CSS nomograms showed strong performance in both cohorts. In the training cohort, the C-index was 0.850 for OS and 0.886 for CSS. In the validation cohort, the corresponding values were 0.857 and 0.886. ROC analyses confirmed robust predictive accuracy across time points (Figure 4A-4D). Calibration plots showed close agreement between predicted and observed outcomes (Figure 4E-4H). DCA supported the clinical usefulness of both models (Figure 4I-4L).

Figure 4 Performance of the OS and CSS nomograms in patients with AC. (A-D) ROC curves for OS and CSS in the training and validation cohorts. (E-H) Calibration curves for 1-, 3-, and 5-year OS and CSS in the training and validation cohorts. (I-L) DCA for OS and CSS in the training and validation cohorts. AC, appendiceal cancer; AUC, area under the curve; CI, confidence interval; CSS, cancer-specific survival; DCA, decision curve analyses; OS, overall survival; ROC, receiver operating characteristic.

To support clinical use, web-based dynamic versions of the nomograms were developed. These tools allow clinicians to enter patient-specific variables and obtain individualized estimates of DM risk, OS, and CSS. The online tools are available for DM (https://wecan.shinyapps.io/DM_nomogram/), OS (https://wecan.shinyapps.io/OS_nomogram/), and CSS (https://wecan.shinyapps.io/CSS_nomogram/). In these web-based dynamic nomograms, clinicians can directly input clinical variables to generate individualized and accurate predictions of DM risk and long-term survival outcomes for patients with AC.

Risk stratification of the DM nomogram

Total risk scores were calculated for patients in the training cohort using the DM nomogram. X-Tile software was used to determine that the optimal cutoff value was 0.6 to divide patients in the cohort into high-risk and low-risk groups. Patients with scores below 0.6 were classified as low risk (n=536), while those with scores above 0.6 were classified as high risk (n=152). KM survival curves and log-rank tests were employed to compare OS and CSS in the validation cohort. The patients in the high-risk group had shorter survival times (Figure 5). These findings indicate that the DM nomogram score is closely associated with patient prognosis.

Figure 5 Prognosis of AC. (A) KM analysis of OS in high- and low-risk patients. (B) KM analysis of CSS in high- and low-risk patients. AC, appendiceal cancer; CSS, cancer-specific survival; KM, Kaplan-Meier; OS, overall survival.

Discussion

Primary AC is rare and is often found incidentally during appendectomy for acute appendicitis. Because AC is uncommon and there are no dedicated screening programmes, early-stage disease is easy to miss. Minhas et al. reported that 38.4% of 3,447 patients with AC in SEER [2010–2015] had metastases at diagnosis, most often to the peritoneum and liver (17). In our study, 1,235 of 6,916 patients (17.8%) diagnosed between 2004 and 2021 had metastatic disease at presentation. The high incidence of metastasis at diagnosis represents a major threat to survival. Therefore, timely recognition and intervention of DM in AC are of utmost importance. Several nomograms have been developed for AC. For instance, Liu et al. constructed a model to estimate OS in appendectomy patients (18), and Wang et al. developed a nomogram to predict the risk of LNM (14). Still, these tools did not estimate DM probability, and they did not combine multiple outcomes such as OS and CSS. Some earlier studies examined DM-related factors in AC, but they usually focused on selected histological subtypes or a single endpoint. Few offered a simple visual tool that clinicians can use. To address these gaps, we used a large population-based cohort and built nomograms that predict DM, OS, and CSS across histological subtypes using routinely available clinical variables.

In this study, we developed and validated three nomograms, each aimed at one endpoint. included marital status, sex, tumor size, histology, T stage, and N stage. The OS nomogram included age at diagnosis, race, histology, grade, T stage, N stage, M stage, and chemotherapy. The CSS nomogram included histology, grade, T stage, N stage, M stage, surgery, and chemotherapy. All three nomograms showed good predictive performance, with high C-index values. Calibration curves in both cohorts showed close agreement between predicted and observed outcomes. DCA also supported clinical usefulness. We also used the DM nomogram scores to group patients into low- and high-risk categories with X-tile, and KM curves showed clear separation between these groups.

From a clinical perspective, the DM nomogram helps identify patients who may need closer surveillance or more intensive treatment. The OS and CSS nomograms provide individual survival estimates that can help guide treatment choices and follow-up. In this way, routine clinicopathological data can be turned into practical risk estimates.

Histology, T stage, and N stage were retained in all three models, which underlines their importance for both metastasis risk and survival. Higher T and N stages were linked to greater DM risk. Deeper invasion and nodal spread can increase the chance of dissemination through vascular or lymphatic routes (19). We also found that T, N, and M stages were independent predictors of survival. M stage had the strongest influence on OS, while T stage was most informative for CSS. Histological subtype had the largest influence in each nomogram, which fits with the well-known biological diversity of AC. Among subtypes, SRCC contributed most to DM risk and showed the worst survival. SRCC is a poorly differentiated and aggressive form of AD. It is defined by signet-ring morphology and may invade early by disrupting the basement membrane, which can promote spread (20). MA and MANEC also showed higher DM risk and worse outcomes. MANEC has both glandular and neuroendocrine components, which may allow spread through lymphatic and blood pathways (21). MA often grows more slowly, but it is frequently diagnosed at a late stage. This may reflect the rarity of AC and the lack of early clinical suspicion, which can delay diagnosis and treatment (22). AD and GCC showed intermediate metastatic risk and prognosis. AD represents a heterogeneous group of tumors, whose behavior varies depending on tumor size, grade, and stage (23). GCC often follows a slower course, which may relate to its partial neuroendocrine features (24). NEC and NET had the lowest DM risk and the best survival. These tumors are often well differentiated, grow slowly, and are commonly diagnosed at early stages, with more than half detected at stage I (25). Even when metastasis is present, NET can still be associated with long survival because the disease often progresses slowly (26). Together, these findings reinforce that subtype biology shapes both metastatic behaviour and long-term outcomes. This difference needs to be recognised when clinicians stratify risk and plan treatment.

Beyond histology and stage, the DM nomogram also captured other factors linked to metastasis. Tumor size remained an independent predictor. Tumors larger than 3 cm carried higher DM risk, which is in line with meta-analytic evidence that size above 2 cm is associated with increased DM risk (27). We also found that married patients and female patients were more likely to present with DM. The reasons for higher metastatic rates in women are not clear. One explanation relates to anatomy. Unlike in males, the female peritoneal cavity is not fully closed because the fallopian tubes open into the peritoneal space. This pathway could allow tumor cells to spread more widely within the peritoneum (28). Hormonal factors may also play a role. Oestrogen can affect tumour biology and immune responses in several cancers, which may influence metastatic potential (29).

The OS and CSS nomograms also included predictors tied to long-term survival. Tumor grade remained important for both outcomes. Grade 3 or 4 disease was linked to poorer prognosis in our cohort. This aligns with findings from Asare et al., who reported that mucinous tumors had better outcomes than non-mucinous tumors at early stages. Still, in stage IV disease, poorly differentiated mucinous tumors had very poor survival, which suggests that grade modifies prognosis within subtypes (30). At a molecular level, high-grade appendiceal tumors show enrichment in pathways linked to aggressive growth and progression, including TNFα signalling via NFκB, MYC, P53, E2F, and KRAS-related gene sets (31).

Treatment factors also mattered. We classified surgery as no surgery, local tumor destruction, or resection. Resection, especially more extensive operations such as right hemicolectomy or combined resection involving adjacent organs, was linked to better outcomes. This may be because more extensive resection is more likely to achieve negative margins,which can reduce residual disease and potentially lower the risk of recurrence (32). In our analysis, surgery was independently associated with CSS but not OS. Prior work supports stage- and subtype-specific strategies. Marks et al. reported that right hemicolectomy was associated improved survival in stage II disease and stage I non-MA, whereas for early-stage appendiceal NETs, appendectomy is typically performed in early, low-risk disease, which generally has a low risk of metastasis and an excellent prognosis, and may therefore be adequate (33). Guzman et al. found no survival advantage to colectomy over appendectomy for appendiceal NETs; the apparent differences in survival are largely driven by tumor stage/grade and patient selection (34). These findings suggest that surgical planning should consider age, overall fitness, subtype, stage, and grade.

Chemotherapy was associated with better OS and CSS in our cohort. Chemotherapy can benefit the prognosis of stage I–III AC patients regardless of histological type, but in stage IV, the effect of systemic chemotherapy varied depending on tumor histology and grade. Specifically, well-differentiated MA did not experience any improvement in survival from systemic chemotherapy (30,35). This may be due to differences in tumor cell biology, such as differences in gene expression profiles or drug-resistance mechanisms among different histological types.

Age at diagnosis and race were independent predictors of OS but not CSS. Among patients diagnosed with appendiceal NETs, young individuals were less likely to present with advanced disease, which may help explain why survival worsens with age (36). We also observed worse OS among Black patients, followed by Asian or Pacific Islander patients, White patients, and American Indian or Alaska Native patients. Although this pattern has not been confirmed specifically in AC, similar disparities are reported in other cancers. These differences may reflect unequal access to care, biological variation, or both (37).

Although the nomograms demonstrated good internal and external validation, several limitations should be acknowledged. First, due to the limited availability of biomarker data [such as carcinoembryonic antigen (CEA), carbohydrate antigen 19-9 (CA19-9), and cancer antigen 125 (CA125)] before 2010 in the SEER database, these variables were not included in our analysis to ensure a sufficient number of patients. Second, the SEER database primarily records information at the time of initial diagnosis, which precluded the capture of metachronous DM occurring during follow-up. Third, this retrospective study reflects US population, so findings may not fully apply to other settings. Additionally, selection bias related to missing data cannot be fully excluded. Given the rarity of AC, performing such validation in real-world settings remains a significant challenge. Future studies should use broader datasets and add further prognostic factors to strengthen and refine these models.

Conclusions

In this population-based study, we developed and validated three nomograms to predict DM, OS, and CSS in patients with AC using routinely available clinical and demographic variables from the SEER database. These models showed excellent discrimination, good calibration, and meaningful clinical utility, and the DM-based risk stratification clearly separated patients with different survival outcomes. We further built web-based dynamic nomograms to provide convenient, individualized risk estimation in clinical practice. These tools may support more precise prognostic assessment and treatment planning for patients with AC.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1-2821/rc

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1-2821/prf

Funding: This work was supported by the Clinical Medical Research Special Fund Project of Zhejiang Medical Association (No. 2025ZYC-A73).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1-2821/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Shaib WL, Zakka K, Staley C 3rd, et al. Blood-Based Next-Generation Sequencing Analysis of Appendiceal Cancers. Oncologist 2020;25:414-21. [Crossref] [PubMed]
2. Shaib WL, Assi R, Shamseddine A, et al. Appendiceal Mucinous Neoplasms: Diagnosis and Management. Oncologist 2017;22:1107-16. [Crossref] [PubMed]
3. Lietzén E, Grönroos JM, Mecklin JP, et al. Appendiceal neoplasm risk associated with complicated acute appendicitis-a population based study. Int J Colorectal Dis 2019;34:39-46. [Crossref] [PubMed]
4. Singh H, Koomson AS, Decker KM, et al. Continued increasing incidence of malignant appendiceal tumors in Canada and the United States: A population-based study. Cancer 2020;126:2206-16. [Crossref] [PubMed]
5. Teixeira FJR Jr, Couto Netto SDD, Akaishi EH, et al. Acute appendicitis, inflammatory appendiceal mass and the risk of a hidden malignant tumor: a systematic review of the literature. World J Emerg Surg 2017;12:12. [Crossref] [PubMed]
6. Shaib WL, Martin LK, Choi M, et al. Hyperthermic Intraperitoneal Chemotherapy Following Cytoreductive Surgery Improves Outcome in Patients With Primary Appendiceal Mucinous Adenocarcinoma: A Pooled Analysis From Three Tertiary Care Centers. Oncologist 2015;20:907-14. [Crossref] [PubMed]
7. Yan Q, Zheng W, Luo H, et al. Incidence and survival trends for appendiceal mucinous adenocarcinoma: an analysis of 3237 patients in the Surveillance, Epidemiology, and End Results database. Future Oncol 2019;15:3945-61. [Crossref] [PubMed]
8. Muñoz de Nova JL, Hernando J, Sampedro Núñez M, et al. Management of incidentally discovered appendiceal neuroendocrine tumors after an appendicectomy. World J Gastroenterol 2022;28:1304-14. [Crossref] [PubMed]
9. Dohner E, Kierdorf FJ, Langer R, et al. Retrospective analysis of the incidence of appendiceal neoplasm and malignancy in patients treated for suspected acute appendicitis. BMC Surg 2024;24:121. [Crossref] [PubMed]
10. Gironda DJ, Erali RA, Forsythe SD, et al. The Genomic Topography of Appendiceal Cancers: Our Current Understanding, Clinical Perspectives, and Future Directions. Cancers (Basel) 2025;17:3275. [Crossref] [PubMed]
11. Guo K, Lou Y, Zheng S. Predictors of Distant Metastasis and Prognosis in Newly Diagnosed T1 Intrahepatic Cholangiocarcinoma. Biomed Res Int 2023;2023:6638755. [Crossref] [PubMed]
12. Aguirre N, Chung SK, Foote MB, et al. Predictors of Recurrence in Nonmetastatic Appendiceal Epithelial Cancers: An Updated Single-Center Experience Over 25 Years. Ann Surg Oncol 2025;32:695-702. [Crossref] [PubMed]
13. Wang X, Lu J, Song Z, et al. From past to future: Bibliometric analysis of global research productivity on nomogram (2000-2021). Front Public Health 2022;10:997713. [Crossref] [PubMed]
14. Wang D, Liu C, Yan T, et al. A Nomogram for Predicting Lymph Nodal Metastases in Patients with Appendiceal Cancers: An Analysis of SEER Database. J Invest Surg 2021;34:924-30. [Crossref] [PubMed]
15. Daskalakis K, Alexandraki K, Kassi E, et al. The risk of lymph node metastases and their impact on survival in patients with appendiceal neuroendocrine neoplasms: a systematic review and meta-analysis of adult and paediatric patients. Endocrine 2020;67:20-34. [Crossref] [PubMed]
16. Wang Y, Geng Y, Hu W. Survival and prognostic factors for postoperative primary appendiceal cancer: a retrospective cohort study based on the Surveillance, Epidemiology, and End Results database. J Gastrointest Oncol 2022;13:1433-43. [Crossref] [PubMed]
17. Minhas A, Hendrickson J, Minhas SA. Frequency and Risk Factors for Metastasis in Newly Diagnosed Appendiceal Carcinoma. Cureus 2021;13:e16341. [PubMed]
18. Liu T, Mi J, Wang Y, et al. Establishment and validation of the survival prediction risk model for appendiceal cancer. Front Med (Lausanne) 2022;9:1022595. [Crossref] [PubMed]
19. Pourfarhangi KE, De La Hoz EC, Cohen AR, et al. Contact guidance is cell cycle-dependent. APL Bioeng 2018;2:031904. [Crossref] [PubMed]
20. Sato O, Tsuchikawa T, Yamada T, et al. Metastatic mixed adenoneuroendocrine carcinoma of the liver successfully resected by hepatic trisectionectomy following chemotherapy: A case report. Clin Case Rep 2019;7:491-6. [Crossref] [PubMed]
21. Zhao B, Wang J, Ye H, et al. Mixed adenoneuroendocrine carcinoma originating from the appendix and colorectum: a comparative analysis of a large population-based database. Tech Coloproctol 2023;27:1017-24. [Crossref] [PubMed]
22. McClelland PH, Gregory SN, Nah SK, et al. Predicting Survival in Mucinous Adenocarcinoma of the Appendix: Demographics, Disease Presentation, and Treatment Methodology. Ann Surg Oncol 2024;31:6237-51. [Crossref] [PubMed]
23. AlMasri SS, Hammad AY, Singhi AD, et al. Appendectomy Is Oncologically Equivalent to Right Hemicolectomy for Well-Differentiated T1 Appendiceal Adenocarcinoma. Dis Colon Rectum 2023;66:67-74. [Crossref] [PubMed]
24. Wang Y, Shahabi A, Loeffler A. Appendiceal Goblet Cell Adenocarcinoma: A Historically Informed Reading of 6 Cases. Arch Pathol Lab Med 2022;146:1402-11. [Crossref] [PubMed]
25. Onyemkpa C, Davis A, McLeod M, et al. Typical carcinoids, goblet cell carcinoids, mixed adenoneuroendocrine carcinomas, neuroendocrine carcinomas and adenocarcinomas of the appendix: a comparative analysis of survival profile and predictors. J Gastrointest Oncol 2019;10:300-6. [Crossref] [PubMed]
26. Zheng C, Al Shabeeb R, Shah D, et al. Slow and Steady: A Slowly Progressing Neuroendocrine Tumor. ACG Case Rep J 2023;10:e01147. [Crossref] [PubMed]
27. Köhler F, Matthes N, Rosenfeldt M, et al. Neoplasms of the Appendix. Dtsch Arztebl Int 2023;120:519-25. [PubMed]
28. Solass W, Struller F, Horvath P, et al. Morphology of the peritoneal cavity and pathophysiological consequences. Pleura Peritoneum 2016;1:193-201. [Crossref] [PubMed]
29. Yagyu T, Miki H, Kikawa Y, et al. Appendiceal Metastasis of Breast Cancer: A Case Report and a Literature Review. Cureus 2024;16:e57929. [Crossref] [PubMed]
30. Asare EA, Compton CC, Hanna NN, et al. The impact of stage, grade, and mucinous histology on the efficacy of systemic chemotherapy in adenocarcinomas of the appendix: Analysis of the National Cancer Data Base. Cancer 2016;122:213-21. [Crossref] [PubMed]
31. Gu Y, Yousef AM, Chowdhury S, et al. Abstract 5794: Gene expression analysis of appendiceal adenocarcinomas highlights tumor heterogeneity between high and low-grade. Cancer Res 2022;82:5794. [Crossref]
32. Mo S, Zhou Z, Ying Z, et al. Epidemiology of and prognostic factors for appendiceal carcinomas: a retrospective, population-based study. Int J Colorectal Dis 2019;34:1915-24. [Crossref] [PubMed]
33. Marks VA, Kerekes D, Butensky S, et al. Role of colectomy in the management of appendiceal tumors: a retrospective cohort study. BMC Gastroenterol 2023;23:398. [Crossref] [PubMed]
34. Guzman C, Boddhula S, Panneerselvam N, et al. Appendiceal Carcinoid Tumors: Is There a Survival Advantage to Colectomy over Appendectomy? J Gastrointest Surg 2020;24:1149-57. [Crossref] [PubMed]
35. Kolla BC, Petersen A, Chengappa M, et al. Impact of adjuvant chemotherapy on outcomes in appendiceal cancer. Cancer Med 2020;9:3400-6. [Crossref] [PubMed]
36. Gibbs T, Washington MK, Eng C, et al. Histologic and Racial/Ethnic Patterns of Appendiceal Cancer among Young Patients. Cancer Epidemiol Biomarkers Prev 2021;30:1149-55. [Crossref] [PubMed]
37. Management of Endometrial Intraepithelial Neoplasia or Atypical Endometrial Hyperplasia. ACOG Clinical Consensus No. 5. Obstet Gynecol 2023;142:735-44. [Crossref] [PubMed]