Development of a nomogram predicting perineural invasion risk and assessment of the prognostic value of perineural invasion in colon cancer: a population study based on the Surveillance, Epidemiology, and End Results database

Introduction

Colon cancer (CC) is one of the most common malignant cancers, with 1.15 million new cases reported every year and 576,858 deaths related to it (1,2). Currently, the American Joint Committee on Cancer (AJCC) tumor-node-metastasis (TNM) staging system is considered the clinical treatment guide, and TNM stage is the primary prognostic factor for CC. However, previous studies have reported that various histopathological factors, such as perineural invasion (PNI) and vascular invasion, are associated with a poor prognosis in CC patients, indicating that TNM staging is not accurate enough to predict the prognosis of all CC patients (3-5).

PNI, also called neurotropic carcinomatous spread, is defined as invasion into the neural tissue and/or nerve sheaths, which may prove to be another metastatic route (6). Growing through nerves has become a common mode of metastasis in CC, in addition to direct infiltration and dissemination through blood and lymph channels (7). In colorectal cancer (CRC), PNI is widely associated with poor outcome (8,9). According to the National Comprehensive Cancer Network (NCCN) clinical practice guidelines, PNI is also defined as a high-risk feature (10). However, despite its strong clinical importance, PNI can only be diagnosed by histopathology after surgery. Preoperative prediction of PNI in CC patients is very useful for the determination of treatment plans. Therefore, further studies are urgently needed to construct a prediction model for PNI.

At present, radical surgery combined with adjuvant chemotherapy (ChemT) is the cornerstone of treatment for locally advanced CC. However, the prognosis remains dismal (11,12). Despite radical surgery and adjuvant ChemT, nearly 30% of CC patients will have recurrence, and the prognosis remains poor (13,14). While adjuvant radiotherapy (RT) can only be considered in several clinical scenarios, such as positive margins or advanced local disease [pathologic stage T4 (pT4)] (15,16), combined chemotherapy and radiotherapy (CCRP) has been less explored in CC patients. And whether there is an influence of PNI on the outcomes of CCRP has not been reported. Thus, additional studies are needed to confirm the efficacy of CCRP in pT4 CC patients with or without PNI.

In the present study, we developed a nomogram for predicting PNI in CC patients using the population-based Surveillance, Epidemiology, and End Results (SEER) database. The effect of PNI on CC prognosis was further investigated using Kaplan-Meier analysis, a competing risk model, and the Fine-Gray multivariable regression model. Finally, we investigated the efficiency of CCRP in pT4 CC patients with or without PNI. We present this article in accordance with the TRIPOD reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-24-1030/rc).

Methods

Patient cohort

The SEER Program of the National Cancer Institute (NCI) is a cancer database based on the United States (US) population that consists of clinical data of cancer patients from 18 registries and covers approximately 26% of the US population (17). The data of CC patients in the SEER database from January 2010 to December 2015 were queried using SEER*Stat software version 8.3.9 (www.seer.cancer.gov). All procedures were performed according to approved guidelines. As SEER is an anonymized public database, there was no ethical approval needed. The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013).

The inclusion criteria included the following: (I) primary CC (histology codes: 8140–8147, 8210, 8211, 8220, 8221, 8260–8263, 8480, 8481, 8490; site codes: C18.0 and C18.2–18.9); (II) aged ≥18 years; (III) complete follow-up; and (IV) CC as the only primary malignancy. In our study, the following variables were included: race, age at diagnosis, sex, tumor size, tumor grade, tumor site, histology, T status according to AJCC 7^th edition, N status according to AJCC 7^th edition, carcinoembryonic antigen (CEA), PNI, metastasis sites (liver, lung, brain, bone), and information on therapy (surgery status, RT status, ChemT status), vital status, survival months, cause of death, and marital status.

Following preliminary selection, patients were excluded if (I) they had a second primary tumor; (II) they had an overall survival time <1 month; or (III) they had incomplete patient records.

A total of 51,826 CC patients were included. To ensure the accuracy of the nomogram, all included CC patients were randomly divided into a development cohort (70%) and an internal validation cohort (30%). To evaluate the effect of PNI on prognosis, all patients were also divided into two groups by PNI: the PNI (−) group and PNI (+) group. The tumor sites were grouped as follows: right-side colon cancer (RCC; cecum, ascending colon, hepatic flexure and transverse colon) and left-side colon cancer (LCC; splenic flexure, descending colon and sigmoid colon). “No/unknown” RT records were considered as no RT. “No/unknown” ChemT records were considered as no ChemT.

End points

Final follow-up was conducted in November 2020, with a median follow-up of 41 months (ranging from 1 to 95 months). Overall survival was defined as the time from diagnosis to death. CC-specific death and CC-specific survival were defined as the time from the initiation of therapy to death or survival from CC.

Statistical analysis

The statistical analysis in this study was performed in SPSS 24.0 (IBM Corp, Armonk, NY, USA) or R software (Version 4.1.3, https://www.r-project.org/). The data were summarized using descriptive statistics and frequency tables. Categorical data were evaluated with the Chi-squared test. Independent risk factors for PNI in CC patients were examined by univariate and multivariate logistic regression analyses. Stepwise regression analysis was used to introduce variables into the model. Nomogram construction was performed based on the results of stepwise regression analysis. The receiver operating characteristic (ROC) curves and the corresponding area under the curve (AUC) values were used to evaluate the discrimination of the nomogram. The ROC curves, with AUC values quantified with the survival ROC package. Additionally, the calibration of the nomogram was evaluated by the corresponding calibration curves. The clinical benefits of the nomogram were assessed using decision curve analysis (DCA). Furthermore, Kaplan-Meier survival curve analysis and the log-rank test were used to determine whether there was a difference in overall survival and CC-specific survival between the PNI (−) group and the PNI (+) group. We used the R package cmprsk to identify statistical differences between CC-specific death and non-CC-specific death due to PNI by using competing risk model analysis and Gray’s test. We examined associations between factors and CC-specific death, including non-CC-specific death as a competing risk, using the Fine-Gray multivariable regression model. Statistical significance was established for a two-sided P value of 0.05.

Results

Patient characteristics

Among the 51,826 patients, 6,610 (12.75%) exhibited PNI, while 45,216 (87.25%) did not. By comparing the PNI (+) group and PNI (−) group, there were significant differences (P<0.05) in age, race, tumor site, tumor size, tumor grade, histology, T status, N status, CEA level, surgery status, RT status, ChemT status, bone metastasis, brain metastasis, liver metastasis, lung metastasis and marital status. The key information of the variables is shown in Table 1. The CC patients included in this study were randomly divided into a development cohort (n=36,280) and an internal validation cohort (n=15,546). There were no significant differences between the two groups according to the Chi-squared test (Table 2).

Identification of risk factors of the patients with PNI in the training cohort

A univariate logistic regression analysis was performed to identify the risk factors for PNI in CC patients. The results revealed that age, race, tumor site, tumor size, grade, histology, T status, N status, CEA level, bone metastasis, brain metastasis, liver metastasis, and lung metastasis were independent risk factors for PNI (Table 3). Then, multivariate logistic regression analysis was performed, and the results showed that younger age, black race, tumor in the left colon, larger tumor size, higher tumor grade, higher T status, higher N status, higher CEA levels and liver metastasis were independently associated with PNI. Conversely, age ≥80 years and a histological type of mucinous adenocarcinoma were identified as protective factors against PNI (Table 3).

Development and validation of the nomogram to predict PNI probability

To further predict the probability of PNI in CC patients, a nomogram was developed with the 11 significant factors associated with PNI in the stepwise multivariate logistic regression (Figure 1). The AUCs for the development and validation cohorts were 0.787 [95% confidence interval (CI): 0.780–0.793] and 0.781 (95% CI: 0.771–0.791), respectively (Figure 2). In addition, the calibration curves illustrated that our nomogram has strong calibration (Figure 3). Furthermore, the DCA curves also indicated that the nomogram has good clinical practicability in the development cohort and internal validation cohort (Figure 4).

Figure 1 Nomogram for predicting the probability of PNI in CC patients. RCC, right-side colon cancer; LCC, left-side colon cancer; CEA, carcinoembryonic antigen; PNI, perineural invasion; CC, colon cancer.

Figure 2 The ROC curves of nomogram for predicting PNI in the development cohort (A) and validation cohort (B). AUC, area under the curve; ROC, receiver operating characteristic; PNI, perineural invasion.

Figure 3 The calibration curves of the nomogram for predicting PNI in the development cohort (A) and validation cohort (B). ROC, receiver operating characteristic; PNI, perineural invasion.

Figure 4 The DCA curves of the nomogram for predicting PNI in the development cohort (A) and validation cohort (B). DCA, decision curve analysis; PNI, perineural invasion.

Kaplan-Meier survival analysis

After analyzing the risk factors for PNI in patients with CC, overall survival and CC-specific survival in CC patients were explored using Kaplan-Meier analysis. Among all 51,826 patients included in this study, 19,165 (36.98%) died. The cumulative incidence rate of CC-specific death was 24.69% (12,794/51,826), and the cumulative non-CC-specific death incidence rate was 12.29% (6,371/51,826). Compared with patients in the PNI (−) group, patients in the PNI (+) group had worse overall survival and CC-specific survival (Figure 5).

Figure 5 Kaplan-Meier survival analysis for PNI (+) and PNI (−) CC patients. (A) Overall survival curves for the PNI (+) and PNI (−) group. (B) CC-specific survival curves for the PNI (+) and PNI (−) CC patients. PNI, perineural invasion; CC, colon cancer.

Competing risk model of CC-specific death and non-CC-specific death

A total of 19,165 patients who died were included in our cohort, of which 66.76% (12,794/19,165) were CC-specific death and 33.24% (6,371/19,165) were non-CC-specific death. Compared with patients in the PNI (−) group, patients in the PNI (+) group had a higher cumulative CC-specific death rate (Gray’s test, P<0.001), and there was no significant difference in the cumulative non-CC-specific death rate between the two groups (Gray’s test, P=0.63) (Figure 6).

Figure 6 Cumulative incidence of CCSD and non-CCSD in PNI (+) and PNI (−) CC patients. PNI, perineural invasion; CCSD, colon cancer-specific death; CC, colon cancer.

Multivariable competing risk analysis of survival

The Fine-Gray multivariable regression model was constructed to investigate independent prognostic factors of CC-specific death (Table 4). The results revealed that patients with PNI had a higher CC-specific death rate [hazard ratio (HR) =1.243; 95% CI: 1.183–1.305; P<0.001] than patients without PNI. Compared with the corresponding subgroups, patients with female sex, tumor in the left colon, and no RT tended to have significantly lower CC-specific death (P<0.05). Conversely, patients with older age, black race, highly differentiated tumors, mucinous adenocarcinoma and signet ring cell carcinoma, higher T status, higher N status, elevated CEA levels, bone metastasis, liver metastasis, lung metastasis, no surgery, no ChemT tended to have significantly higher CC-specific death (P<0.05). Additionally, age, race, sex, tumor site, histology, AJCC T status, CEA levels, liver metastasis, lung metastasis, marital status, ChemT status and RT status were also related to non-CC-specific death (P<0.05).

Competing risk model of the effect of different treatment methods after surgery on the survival of pT4 CC patients

Since adjuvant RT is not traditionally used for early CC, we then explored the efficiency of adjuvant therapy on the prognosis of pT4 patients with or without PNI. A total of 5,832 patients with pT4 stage disease who died were included in this analysis, of which 32.61% (1,901/5,830) were in the PNI (+) subgroup, and 67.39% (3,929/5,830) were in the PNI (−) subgroup. As shown in Figure 7, PNI (+) pT4 CC patients treated with CCRP or ChemT had a better cumulative CC-specific death incidence (Gray’s test, P<0.001) than those in the non-therapy group (Figure 7A). In contrast, patients in the CCRP group had a better cumulative CC-specific death incidence (Gray’s test, P<0.001) than did the ChemT group and non-therapy group in PNI (−) pT4 CC patients (Figure 7B).

Figure 7 Competing risk analysis of the effect of different treatment methods after surgery on the survival in pT4 CC patients. (A) Cumulative incidence of CCSD and non-CCSD in PNI (+) CC patients of different treatments. (B) Cumulative incidence of CCSD and non-CCSD in PNI (−) CC patients of different treatments. CCSD, colon cancer-specific death; ChemT, chemotherapy; CCRP, combined chemotherapy and radiotherapy; pT4, pathologic stage T4; CC, colon cancer; PNI, perineural invasion.

Discussion

In the present study, based on the analysis of a cohort of 51,826 patients with CC in the SEER database from January 2010 to December 2015, we developed a nomogram to predict PNI based on the clinical features of patients, and we also validated that PNI could significantly reduce CC-specific survival for CC patients. pT4 CC patients without PNI treated with CCRP had a lower CC-specific death rate than patients with ChemT or non-therapy. The nomogram in this study demonstrated a high degree of discrimination and calibration, and this study is the first real-world study to examine the impact of PNI on CC patients by using the Fine-Gray multivariable regression model.

PNI was reported for the first time in 1800s, and its prognostic value has gradually attracted attention since the 1970s (18). PNI is described as tumor invasion in, around and through the nerves (6). According to a previous study, 33% of patients with CRC exhibit PNI (19). In this study, the prevalence of PNI in CC patients was approximately up to 12.75%. Previous studies have reported that PNI is associated with more severe clinicopathologic characteristics (7,10,20). Our study found that PNI was significantly correlated with tumor site, tumor size, tumor grade, tumor histology, T status, N status, CEA levels, bone metastasis, brain metastasis, liver metastasis and lung metastasis.

PNI is related to poor prognosis for CRC patients, and early identification of PNI in patients with CRC is necessary for the development of individualized clinical therapeutic strategies (9,21). Currently, PNI in CC can only be assessed by histopathological examination after surgery (22). Therefore, it is essential to construct a prediction model for PNI. As a clinical tool, nomograms have become widely used as prediction models in recent years (23,24). Huang et al. selected clinical features by least absolute shrinkage and selection operator (LASSO) logistic regression analysis and constructed a nomogram based on a population in a single center to predict PNI status in CRC patients with AUCs of 0.704 and 0.692 in the training and validation cohorts, respectively (25). In our study, to construct a nomogram for predicting PNI in CC patients from the SEER database, significant clinical features were selected using stepwise logistic regression. The nomogram was constructed using eleven features: age, race, tumor grade, tumor site, histology, T status, N status, CEA levels, liver metastasis, bone metastasis and tumor size. The nomogram demonstrated good discrimination and calibration with AUCs of 0.787 and 0.781 in the training cohort and validation cohort, respectively. The difference between the two nomograms could be mainly attributed to the different features used for model construction. In our study, more clinical features were included, and tumor stage was the main feature of the model. Thus, we think that the nomogram has certain application value in the clinic for developing therapeutic strategies.

Researchers have found that PNI is related to a poor prognosis in CRC patients (7,21,26). PNI has been included in the AJCC cancer staging system as a negative prognostic factor, and PNI has been presented in pathology reports for years (27). Our study validated the prognostic value of PNI based on the data of CC patients from the SEER database. Kaplan-Meier curve analysis indicated that patients in the PNI (+) group had poorer overall survival and CC-specific survival than patients in the PNI (−) group, which is consistent with findings from past studies (7,28). Tu et al. found that PNI was a negative prognostic factor for CC in stage I (HR =1.590; 95% CI: 0.951–2.658) and stage II (HR =1.607; 95% CI: 1.426–1.812) (7). In this study, we demonstrated that the PNI (+) group had a higher CC-specific death rate (HR =1.243; 95% CI: 1.183–1.305) than the PNI (−) group, and there was no statistically difference in the non-CC-specific death rate between the two groups. The difference in HR between the two studies could be attributed to the fact that the occurrence of competitive events might affect CC-specific survival and preclude the primary event of CC-specific death. Compared to Fine-Gray regression, Cox regression overestimated the cumulative event rate and yielded a higher hazard ratio. Additionally, the differences might be occurred because of the different stage of tumors. Tu et al. performed their study mainly in CC patients in stage I and stage II, while in our study, CC patients in all stages were included (7).

Currently, CC can be treated with a combination of surgery, RT, ChemT and targeted therapy (29,30). The role of RT for the treatment of CC remains controversial (12,31). In our study, after multivariable competing risk analysis, CC patients who underwent surgery and ChemT had a lower cumulative CC-specific death rate than others, which is consistent with previous studies (22,32). Some previous studies reported that additional RT after surgery conferred a survival benefit in T4 CC patients (15,33), while other studies reported that RT had little effect on prognosis (31,34). In this study, the results indicated that CC patients who underwent RT had a higher cumulative CC-specific death rate than patients without RT. This may be due to selection bias and the preference of patients with RT who may have a higher tumor stage and a poorer prognosis. Therefore, we further explored the impact of adjuvant therapy on the prognosis of pT4 stage patients with or without PNI. The results showed that patients with CCRP had lower cumulative CC-specific death incidence than patients treated with ChemT or non-therapy in the PNI (−) pT4 CC patients, and there was no statistically difference between the CCRP group and the ChemT group in PNI (+) pT4 CC patients. Collectively, these findings indicate that CC patients can benefit from surgery and ChemT, and pT4 patients without PNI might benefit from CCRP.

Inevitably, this study has also some notable limitations that need to be addressed. First, as a retrospective study design, unregistered confounding factors and selection bias cannot be avoided, and study external validity may be limited. Second, it is possible that risk estimates could be influenced by potential confounders, such as family history, comorbidities, surgery methods, or the exact scope of lymph node dissection. Third, new biomarkers have been widely reported to have prognostic values in CC recently (35-37), but they could not be included in this study because of missing related data. Finally, our model has internal validation, but does not have external validation to confirm its accuracy. An external multicenter dataset needs to be used to validate our results. Although the data from SEER database indicates that RT was conducted after radical resection, yet we are unable to determine the sequence of RT and ChemT. New clinical studies should be conducted to see if CCRP after surgery can improve prognostic of pT4 CC patients without PNI.

Conclusions

Our study develops a nomogram for predicting the risk of PNI in patients with CC before operation. Our study also demonstrates that the presence of PNI is independently associated with an increase in the CC-specific death rate and could be considered a reliable prognostic factor in CC patients. Finally, our study indicates that pT4 patients without PNI might benefit from CCRP after surgery.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-24-1030/rc

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-24-1030/prf

Funding: This work was supported by Shaanxi Province Science Foundation for Youths (No. 2024JC-YBQN-0805).

Conflicts of Interest: Both authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-24-1030/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. The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013).

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. Sung H, Ferlay J, Siegel RL, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin 2021;71:209-49. [Crossref] [PubMed]
2. Siegel RL, Miller KD, Fuchs HE, et al. Cancer statistics, 2022. CA Cancer J Clin 2022;72:7-33. [Crossref] [PubMed]
3. Cho SS, Park JW, Kang GH, et al. Prognostic Impact of Extramural Lymphatic, Vascular, and Perineural Invasion in Stage II Colon Cancer: A Comparison With Intramural Invasion. Dis Colon Rectum 2023;66:366-73. [Crossref] [PubMed]
4. Nikberg M, Chabok A, Letocha H, et al. Lymphovascular and perineural invasion in stage II rectal cancer: a report from the Swedish colorectal cancer registry. Acta Oncol 2016;55:1418-24. [Crossref] [PubMed]
5. Parnaby CN, Scott NW, Ramsay G, et al. Prognostic value of lymph node ratio and extramural vascular invasion on survival for patients undergoing curative colon cancer resection. Br J Cancer 2015;113:212-9. [Crossref] [PubMed]
6. Liu Z, Luo C, Chen X, et al. Noninvasive prediction of perineural invasion in intrahepatic cholangiocarcinoma by clinicoradiological features and computed tomography radiomics based on interpretable machine learning: a multicenter cohort study. Int J Surg 2024;110:1039-51. [Crossref] [PubMed]
7. Tu J, Yao Z, Wu W, et al. Perineural Invasion Is a Strong Prognostic Factor but Not a Predictive Factor of Response to Adjuvant Chemotherapy in Node-Negative Colon Cancer. Front Oncol 2021;11:663154. [Crossref] [PubMed]
8. Cao Y, Deng S, Yan L, et al. Perineural invasion is associated with poor prognosis of colorectal cancer: a retrospective cohort study. Int J Colorectal Dis 2020;35:1067-75. [Crossref] [PubMed]
9. Wang H, Huo R, He K, et al. Perineural invasion in colorectal cancer: mechanisms of action and clinical relevance. Cell Oncol (Dordr) 2024;47:1-17. [Crossref] [PubMed]
10. Benson AB, Venook AP, Adam M, et al. Colon Cancer, Version 3.2024, NCCN Clinical Practice Guidelines in Oncology. J Natl Compr Canc Netw 2024;22:e240029. [Crossref] [PubMed]
11. Agas RAF, Co LBA, Sogono PG, et al. Assessing the Effect of Radiotherapy in Addition to Surgery in Colon Adenocarcinomas: a Systematic Review and Meta-analysis of Contemporary Evidence. J Gastrointest Cancer 2020;51:445-60. [Crossref] [PubMed]
12. Lin Y, Wang L, Shao L, et al. Prognostic analysis and beneficiary identification of adjuvant external beam radiotherapy for stage pT4b sigmoid colon cancer. Sci Rep 2021;11:11782. [Crossref] [PubMed]
13. Oki E, Ando K, Taniguchi H, et al. Sustainable Clinical Development of Adjuvant Chemotherapy for Colon Cancer. Ann Gastroenterol Surg 2022;6:37-45. [Crossref] [PubMed]
14. Wu CY, Ye K. Risk factors for the prognosis of colon cancer. World J Gastrointest Oncol 2024;16:3738-40. [Crossref] [PubMed]
15. Wegner RE, Abel S, Monga D, et al. Utilization of Adjuvant Radiotherapy for Resected Colon Cancer and Its Effect on Outcome. Ann Surg Oncol 2020;27:825-32. [Crossref] [PubMed]
16. Zhang H, Li Y, Xia F, et al. Study protocol of short-course radiotherapy combined with CAPOX and PD-1 inhibitor for locally advanced colon cancer: a randomised, prospective, multicentre, phase II trial (TORCH-C). BMJ Open 2024;14:e079442. [Crossref] [PubMed]
17. Wingo PA, Jamison PM, Hiatt RA, et al. Building the infrastructure for nationwide cancer surveillance and control--a comparison between the National Program of Cancer Registries (NPCR) and the Surveillance, Epidemiology, and End Results (SEER) Program (United States). Cancer Causes Control 2003;14:175-93. [Crossref] [PubMed]
18. Liebig C, Ayala G, Wilks J, et al. Perineural invasion is an independent predictor of outcome in colorectal cancer. J Clin Oncol 2009;27:5131-7. [Crossref] [PubMed]
19. Liebig C, Ayala G, Wilks JA, et al. Perineural invasion in cancer: a review of the literature. Cancer 2009;115:3379-91. [Crossref] [PubMed]
20. van Wyk HC, Going J, Horgan P, et al. The role of perineural invasion in predicting survival in patients with primary operable colorectal cancer: A systematic review. Crit Rev Oncol Hematol 2017;112:11-20. [Crossref] [PubMed]
21. Knijn N, Mogk SC, Teerenstra S, et al. Perineural Invasion is a Strong Prognostic Factor in Colorectal Cancer: A Systematic Review. Am J Surg Pathol 2016;40:103-12. [Crossref] [PubMed]
22. Leijssen LGJ, Dinaux AM, Taylor MS, et al. Perineural Invasion Is a Prognostic but not a Predictive Factor in Nonmetastatic Colon Cancer. Dis Colon Rectum 2019;62:1212-21. [Crossref] [PubMed]
23. Park SY. Nomogram: An analogue tool to deliver digital knowledge. J Thorac Cardiovasc Surg 2018;155:1793. [Crossref] [PubMed]
24. Iasonos A, Schrag D, Raj GV, et al. How to build and interpret a nomogram for cancer prognosis. J Clin Oncol 2008;26:1364-70. [Crossref] [PubMed]
25. Huang X, Liu J, Wu G, et al. Development and Validation of a Nomogram for Preoperative Prediction of Perineural Invasion in Colorectal Cancer. Med Sci Monit 2019;25:1709-17. [Crossref] [PubMed]
26. Yang Y, Huang X, Sun J, et al. Prognostic value of perineural invasion in colorectal cancer: a meta-analysis. J Gastrointest Surg 2015;19:1113-22. [Crossref] [PubMed]
27. Compton C, Fenoglio-Preiser CM, Pettigrew N, et al. American Joint Committee on Cancer Prognostic Factors Consensus Conference: Colorectal Working Group. Cancer 2000;88:1739-57. [Crossref] [PubMed]
28. Alotaibi AM, Lee JL, Kim J, et al. Prognostic and Oncologic Significance of Perineural Invasion in Sporadic Colorectal Cancer. Ann Surg Oncol 2017;24:1626-34. [Crossref] [PubMed]
29. Long X, Deng Z, Li G, et al. Identification of critical genes to predict recurrence and death in colon cancer: integrating gene expression and bioinformatics analysis. Cancer Cell Int 2018;18:139. [Crossref] [PubMed]
30. Yuan Y, Chen J, Wang J, et al. Development and Clinical Validation of a Novel 4-Gene Prognostic Signature Predicting Survival in Colorectal Cancer. Front Oncol 2020;10:595. [Crossref] [PubMed]
31. Sebastian NT, Tan Y, Miller ED, et al. Surgery with and without adjuvant radiotherapy is associated with similar survival in T4 colon cancer. Colorectal Dis 2020;22:779-89. [Crossref] [PubMed]
32. Cohen AM. Surgical considerations in patients with cancer of the colon and rectum. Semin Oncol 1991;18:381-7.
33. Huang Y, Gu X, Ge K, et al. The survival benefit of adjuvant radiotherapy for pathological T4N2M0 colon cancer in the Modern Chemotherapy Era: evidence from the SEER database 2004-2015. Artif Cells Nanomed Biotechnol 2020;48:834-40. [Crossref] [PubMed]
34. Xiao J, Caan BJ, Cespedes Feliciano EM, et al. Association of Low Muscle Mass and Low Muscle Radiodensity With Morbidity and Mortality for Colon Cancer Surgery. JAMA Surg 2020;155:942-9. [Crossref] [PubMed]
35. Barras D, Missiaglia E, Wirapati P, et al. BRAF V600E Mutant Colorectal Cancer Subtypes Based on Gene Expression. Clin Cancer Res 2017;23:104-15. [Crossref] [PubMed]
36. Xu J, Zhang R, Peng Q, et al. The profile and prognostic value of circulating lymphocyte subsets in metastatic colon cancer. Int Immunopharmacol 2023;117:109848. [Crossref] [PubMed]
37. Lv J, Li W, Wang X, et al. Identification of MKI67, TPR, and TCHH Mutations as Prognostic Biomarkers for Patients With Defective Mismatch Repair Colon Cancer Stage II/III. Dis Colon Rectum 2023;66:1481-91. [Crossref] [PubMed]