Log odds of positive lymph nodes-based staging system for colorectal cancer patients with inadequate lymph nodes harvested: a potential reference for adjuvant chemotherapy

Introduction

Colorectal cancer (CRC) is a major threat to health globally due to its high morbidity and mortality (1). To date, radical surgery for CRC remains the cornerstone of the therapeutic schedule for stage I–III CRC patients. The number of lymph nodes examined (LNE), which is considered as a vital determinant to guarantee the exactitude of postoperative pathological staging, is crucial for clinicians to perform prognostic estimation and formulate a rational therapeutic option.

Based on the American Society of Clinical Oncology (ASCO) and the National Comprehensive Cancer Network (NCCN) guidelines, a minimum of 12 LNE is essential for clinicians to obtain an accurate tumor stage (2,3). Although evidence-based references are available in most countries, the number of lymph nodes harvested from the surgical specimens has usually been influenced by patient-specific, cancer-related, hospital-dependent and technique-varying factors (4-6). Due to the imperfect characteristic of positive lymph node (pN)-based nodal stage, patients with inadequate lymphadenectomy are associated with a higher incidence of stage migration, which could result in the error of the following therapeutic strategy and is not conducive to patients’ long-term survival (7). However, the number of LNE is a completely postoperative datum which surgeons could not count during the procedure and therefore, inadequate lymphadenectomy is inevitable in some patients.

Because of the blemish of pN-based classification, a modified nodal staging system is needed urgently. In recent years, log odds of positive lymph nodes (LODDS), as a novel prognostic nodal index, which has been raised and found to be related to superior ability to stratify CRC cases with inhomogeneous prognosis regardless of the scope of lymphadenectomy, might improve the accuracy of nodal stage for CRC patients, especially for those with LNE <12 (8-10).

Although the prognostic significance of LODDS in CRC has been investigated to some extent, the number of lymph nodes examined was not considered in the inclusion criteria in all previous studies, and besides, the calculated LODDS cutoff point was not completely applicable for those with inadequate lymphadenectomy. Hence, the aim of this paper was to reveal the optimal LODDS cutoff point and subsequently, construct a novel staging system by incorporating LODDS-based nodal stage and T stage to better perform prognostic assessment for patients with LNE <12. We present this article in accordance with the TRIPOD reporting checklist (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2024-910/rc).

Methods

Date source and variables

The medical records of CRC patients used in the study was downloaded and extracted from the Surveillance, Epidemiology, and End Results (SEER) database. The National Cancer Institute’s SEER database collects de-identified data including clinicopathological characteristics, cancer diagnosis, therapy and prognostic data for about 28 percent of the U.S. population and is considered as a representative cancer-related data (11). The study cohort consisted of patients diagnosed with CRC between January 2011 and December 2015. Inclusion criteria included: (I) aged ≥18 years; (II) patients with LNE <12; (III) radical resection was the first course of treatment; (IV) patients pathologically diagnosed as stage I–III CRC; (V) histological types were limited to adenocarcinoma and mucous/signet-ring cell carcinoma and (VI) CRC was the only malignancy. Exclusion criteria included: (I) patients receiving neoadjuvant therapy; (II) patients without active follow-up and (III) patients with unknown data.

In addition, to verify the performance of the nStage system constructed by the SEER set, CRC patients from The Second Affiliated Hospital of Harbin Medical University (January 2011 to December 2016) were assigned to the validation set during the research. The last follow-up visit was in December 2021. The histological grade (I–IV) in this study represented the degree of differentiation of CRC. Grade I indicates well-differentiated tumors, grade II is moderately differentiated, grade III is poorly differentiated, and grade IV is undifferentiated. Cancer-specific survival (CSS) was defined as a length of time from diagnosis to death due to CRC. The screening criteria of the Chinese set were consistent with the development set mentioned above. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethical Committee of The Second Affiliated Hospital of Harbin Medical University (No. KY2022313) and informed consent was waived due to retrospective nature of the study.

LODDS

LODDS was defined as log [(number of positive lymph node + 0.5)/(number of negative lymph node + 0.5)] (10). In the numerator and denominator, the value of ‘0.5’ has been added to avoid singularity error. The optimal cutoff point of LODDS for prognosis was calculated using the X-tile software (12). X-tile worked by dividing the population into high, medium, and low LODDS groups for each possible cutoff value. Log-rank tests were then performed for prognostic comparison based on all possible LODDS stratifications. Finally, the optimal LODDS cutoff value is determined by selecting the highest χ² value.

Statistical analysis

All statistical analyses were performed using the R version 3.6.1 software. Patients’ basic information was summarized by number and percentage. The primary endpoint was CSS. Significant prognosticators for CRC patients with LNE <12 were determined by Cox regression model. The performance of the nStage system was appraised by concordance index (C-index), decision curve analyses (DCA) and receiver operating characteristic (ROC) curves (13,14). Kaplan-Meier curves were plotted and analyzed using the log-rank test. P<0.05 (two-sided) was deemed statistically significant.

Results

Patient characteristics

According to the screening standard, a total of 3,927 cases from the SEER database were in the development set and 243 CRC patients from our department were in the validation set. Overall, the main proportions of the patients were related to adenocarcinoma (93.4% in the development set and 90.1% in the validation set), grade I/II (85.8% and 88.5%), carcinoembryonic antigen (CEA) negative (63.8% and 70.4%), age ≥65 years (61.3% and 67.1%), and T3 stage (48.4% and 40.3%). The positive lymph node count (LNC) in the development set and validation set were 0.88±1.74 and 0.61±1.43, respectively. And the number of LNE in the development set and validation set were 7.95±2.77 and 7.40±3.81, respectively. The clinicopathological characteristics and demographics of the patients in the two sets are summarized in Table 1.

Construction of the nStage system

To begin with, the best classification of LODDS for CSS was determined (LODDS ≤−1.1, −1.1< LODDS ≤−0.2 and LODDS >−0.2) (Figure 1A). Notable difference could be found among the three LODDS cohorts according to the survival curves (P<0.001) (Figure 1B). As shown in Table S1, patients with high LODDS value were related to grade III/IV (P<0.001), mucinous/signet ring-cell carcinoma (P<0.001), positive CEA (P<0.001), higher T stage (P<0.001) and postoperative recurrence (P<0.001). Subsequently, based on the LODDS (nN0: LODDS ≤−1.1, nN1: −1.1< LODDS ≤−0.2/cancer nodule formation and nN2: LODDS >−0.2), a novel N stage (nN stage) was constructed.

Figure 1 X-tile analysis for CSS in the training cohort. (A) X-tile analysis was employed to determine the optimal cut-off point of the continuous variable for CSS in the training cohort, and the results were presented in a histogram plot to show the distribution differences between different subgroups. (B) Kaplan-Meier plot. CSS, cancer-specific survival; LODDS, log odds of positive lymph nodes.

Then, as shown in Table 2 and Figure 2, nN stage (nN0, nN1 and nN2) and T stage (T1, T2, T3, T4a and T4b) were combined into 15 subsets to make further prognostic analysis. Using T1nN0 as the reference, subgroups with similar hazard ratio (HR) value and 5-year CSS rate were put together and therefore, a novel stage (nStage) system has been built (Figure 3), including nStage A (T1nN0, T1nN1 and T2nN0), nStage B (T1nN2, T2nN1 and T3nN0), nStage C (T2nN2, T3nN1 and T4anN0), nStage D (T3nN2, T4anN1 and T4bnN0) and nStage E (T4anN2, T4bnN1 and T4bnN2). The 5-year CSS rates for nStage A, B, C, D and E were 94.4%, 83.4%, 64.6%, 45.2%, and 24.2%, respectively.

Figure 2 Kaplan-Meier curves for patients in different subgroups. CSS, cancer-specific survival; Y, year.

Figure 3 Novel staging system.

Validation of the nStage system

According to the Cox analyses even after the adjustment for other variables, the nStage system was considered as the significant prognostic indicator for CSS (P<0.001) (Table 3). Moreover, according to the Kaplan-Meier curves classified by nStage and tumor node metastasis (TNM) stage in the development set, we could find that compared with TNM manual, the nStage system could distinguish patients with inhomogeneous survival more efficiently (Figure 4).

Figure 4 Kaplan-Meier curves stratified by AJCC staging system (A) and nStage (B). AJCC, American Joint Committee on Cancer.

In the development set, the C-indexes of the TNM stage and nStage were 0.723 [95% confidence interval (CI): 0.706–0.741] and 0.748 (95% CI: 0.730–0.766), respectively, also revealing the superiority of the nStage system compared with the conventional TNM method in discrimination ability. And in the validation set, compared to the TNM manual (C-index =0.713, 95% CI: 0.650–0.776), a higher C-index could also be detected for nStage (C-index =0.774, 95% CI: 0.723–0.825).

In the development set, the area under the curves (AUCs) at 5 years for the TNM stage and nStage were 0.759 and 0.782 (P=0.02) (Figure 5A). In the validation set, the AUCs at 5-year for the TNM stage and nStage were 0.673 and 0.768 (P=0.03) (Figure 5B).

Figure 5 Comparison of the areas under the ROC curves of the nStage and AJCC staging system for predicting 5-year cancer-specific survival in the development set (A) and validation set (B). AJCC, American Joint Committee on Cancer; AUC, area under the curve; nStage, novel stage; ROC, receiver operating characteristic; TNM, tumor node metastasis.

In the development set, the DCA at 5-year also revealed that the novel system posed more excellent benefits in guiding clinical decision than AJCC staging system (Figure 6A). And the DCA in the validation set also revealed superior net benefits regarding guiding clinical decision compared than the AJCC staging system (Figure 6B).

Figure 6 Comparison of the DCA of the nStage and the AJCC staging system in the development set (A) and validation set (B). AJCC, American Joint Committee on Cancer; DCA, decision curve analyses; nStage, novel stage; TNM, tumor node metastasis.

Prognostic significance of adjuvant chemotherapy in different nStage system

Furthermore, the prognostic significance of adjuvant chemotherapy on patients with different nStage diseases was explored in this study.

In the development set, patients who were treated with radical surgery plus adjuvant chemotherapy harbored notably better prognosis compared to those with radical surgery only in nStage C–E, while adjuvant chemotherapy had not produced any beneficial changes on CSS in nStage A–B (Figure 7A). Besides, to better determine the significance of nStage C–E in providing the reference in selecting patients who might benefit from adjuvant chemotherapy, we also analyzed the impact of adjuvant chemotherapy in nStage C–E patients who classified into TNM stage I–II (n=420), and the result showed that these population could still benefit from adjuvant chemotherapy (P=0.03) (Figure S1).

Figure 7 HR values (no chemotherapy vs. chemotherapy) for patients with different nStage tumors in the development set (A) and validation set (B). *, P<0.05; ***, P<0.001. HR, hazard ratio; ns, no significant difference; nStage, novel stage.

In the validation set, those in the nStage C–E could also benefit from adjuvant chemotherapy, while the phenomenon had not happened in nStage A–B CRC (Figure 7B).

Discussion

CRC remains a major health burden worldwide due to its high morbidity and mortality (1). The AJCC TNM staging manual is the worldwide benchmark for clinicians to judge the quality of life of patients and guide subsequent clinical treatment (15). Although this tumor staging method is widely accepted, its performance level for CRC patients with less than 12 lymph nodes detected remains to be tested.

To begin with, the current nodal stage which is based on the number of positive lymph nodes in the specimen is inevitably affected by the degree of lymph node dissection during surgery, specimen processing and pathological evaluation, which may be the main reason for the limited application of this staging. The ASCO and NCCN guidelines indicate examination of at least 12 lymph nodes are necessary to ensure a reliable postoperative pathological stage. Due to the strict dependence of current nodal stage on the status of lymph adenectomy, patients with less than 12 lymph nodes detected are correlated with higher rates of stage migration, which could lead to the error of next treatment strategy and deprive patients who might benefit from adjuvant therapy. Besides, patients within the same AJCC stage present heterogeneous prognosis as a result of the variability in tumor biology and clinicopathological features and it could further increase the instability of prognostic judgement (16). In addition, the gradient monotonicity of the existing TNM manual for CRC causes the prognostic paradox between stage IIIA and stage II cases, especially for those with inadequate lymph node harvested, therefore failing to provide individualized survival identification to clinicians as we expect (17,18).

Recently, the prognostic significance of LODDS in CRC has started to be explored globally. Many studies suggested that compared with traditional N staging, LODDS-based nodal staging could decrease bias and is more suitable for clinicians to predict survival (8-10). Zhang et al. found that LODDS, as an important prognostic indicator for CRC in spite of nodal stage, tumor site, number of retrieved lymph node, can be a complementary tool to the current TNM staging system, improving its predictive potential, and enhancing its predictive sensitivity and comprehensiveness (8). Ozawa et al. revealed that LODDS is considered a significant prognostic factor in stage IV CRC cases who have receiving radical surgery. In addition, this nodal indicator can be utilized to stratify metastatic CRC patients (9). In this study, we limited the inclusion criteria to LNE <12 for patients, which keeps our research-cohort much more specific. Moreover, the best LODDS cutoff point for this special population was calculate through X-tile analysis. Consistent with previous studies, LODDS was determined as an integral indicator in the prognostic analyses. And subsequently, LODDS-based nodal classification was incorporated into the construction of the novel tumor stage system. The C-index of the TNM system is lower than the nStage classification in the training cohort, and so does the validation. Besides, the ROC curves and DCA showed the superior power of the nStage system compared with the AJCC system, which indicated that the novel staging manual outperformed the AJCC TNM staging in the aspect of discrimination and clinical utility. That was to say, the nStage system demonstrated a superiority over the AJCC stage manual, especially for CRC patients with inadequate LNE. However, it is important to note that our study does not recommend less lymph node dissection. The current standard of examining at least 12 lymph nodes is crucial for accurate staging. Our nStage system is developed to better stratify patients when the number of lymph nodes examined is less than 12, not to support inadequate lymphadenectomy.

Neoadjuvant chemoradiotherapy is still the standard treatment strategy for locally advanced CRC, which can effectively improve the postoperative quality of life of patients. However, compared with patients who only received surgical treatment, there was a significant decrease in the number of lymph nodes detected for patients who received preoperative treatment, and only 20% of patients could meet the sampling requirements of the standard 12 lymph nodes (19). Therefore, the current 12 lymph node detection criteria are not suitable for patients receiving neoadjuvant therapy. Up to now, there is no final criterion on the minimum LNC of patients receiving neoadjuvant therapy, and this criterion may further affect ypN (20). Hence, although neoadjuvant therapy is of great significance for locally advanced CRC, cases receiving preoperative therapy were not included in the study.

Although adjuvant chemotherapy is deemed as a highly beneficial therapeutic choice for CRC patients at the highest risk for disease recurrence, there are still some considerable controversies on whether some patients should be treated with adjuvant chemotherapy (21-23). In this study, we found that adjuvant chemotherapy could provide inequable prognostic effects on patients with nStage A–B and C–E lesions. However, the proposed result could only provide a potential reference in determining patients who might benefit from adjuvant chemotherapy, but not guidance. In other words, since the significance of adjuvant chemotherapy for stage III cases has been established clearly, patients with stage III diseases but classified into nStage A–B should be also treated with chemotherapy. While for some patients who were not in the consideration of receiving chemotherapy before but classified into nStage C–E, clinicians could re-think the potential benefits of chemotherapy on the population and make optimal therapeutic plans for these patients. Currently, confirmation of the high-risk factors and postoperative pathological classification are still the key determinants of the application of adjuvant chemotherapy.

Still, the authors point out several deficiencies in this study. First of all, although X-tile analysis was used to achieve the optimal LODDS cutoff point in this paper, a unified cutoff value of LODDS is still lacking and it would restrict the clinical transformation of the nStage system. Second, there are inevitably observer variation and confusion bias due to the retrospective nature of the study and it was expected to be confirmed by further perspective clinical research. Besides, the chemotherapy administration was based on the existing clinical records. Future prospective research can standardize the chemotherapy administration criteria to better explore its impact on different nStage groups and reduce potential treatment selection bias. Third, the LODDS calculation and nStage system development may appear complex. But we anticipate that with the help of modern technological tools and educational resources, they can be effectively integrated into clinical practice. Future work will focus on creating user-friendly guides for clinicians to facilitate the adoption of this new staging system. Fourth, while we used a validation set from a different institution, we recognize the need for further external validation. Future research will focus on collaborating with multiple institutions to obtain independent external validation cohorts, which will enhance the generalizability of our nStage system. Furthermore, some potential prognosticators such as the genetic and molecular markers, details of surgical procedures and adjuvant therapy are not available from the SEER database.

Conclusions

In conclusion, this current study constructed a novel staging system for CRC cases with inadequate LNE by incorporating T stage and LODDS-based nodal stage, which demonstrated the superiority over the AJCC staging system regrading to prediction capacity. Moreover, the proposed nStage classification might also provide clinicians with a potential reference in identifying patients who might benefit from adjuvant chemotherapy.

Acknowledgments

We would like to thank the staff of the SEER database for their kind work in data collection and delivery.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2024-910/rc

Data Sharing Statement: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2024-910/dss

Peer Review File: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2024-910/prf

Funding: This study was supported by Haiyan Foundation of Harbin Medical University Cancer Hospital (Grant No. JJZD2024-12).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2024-910/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 and its subsequent amendments. The study was approved by the Ethical Committee of The Second Affiliated Hospital of Harbin Medical University (No. KY2022313) and informed consent was waived due to retrospective nature of the study.

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. Siegel RL, Miller KD, Goding Sauer A, et al. Colorectal cancer statistics, 2020. CA Cancer J Clin 2020;70:145-64. [Crossref] [PubMed]
2. Benson AB, Venook AP, Al-Hawary MM, et al. NCCN Guidelines Insights: Colon Cancer, Version 2.2018. J Natl Compr Canc Netw 2018;16:359-69. [Crossref] [PubMed]
3. Benson AB, Venook AP, Al-Hawary MM, et al. Rectal Cancer, Version 2.2018, NCCN Clinical Practice Guidelines in Oncology. J Natl Compr Canc Netw 2018;16:874-901. [Crossref] [PubMed]
4. Polignano F, Henderson N, Alishahi SM, et al. Laparoscopic colectomy for cancer and adequate lymphadenectomy: association between survival and number of lymph nodes. Surg Endosc 2006;20:996-7. [Crossref] [PubMed]
5. West NP, Hohenberger W, Weber K, et al. Complete mesocolic excision with central vascular ligation produces an oncologically superior specimen compared with standard surgery for carcinoma of the colon. J Clin Oncol 2010;28:272-8. [Crossref] [PubMed]
6. Dorrance HR, Docherty GM, O'Dwyer PJ. Effect of surgeon specialty interest on patient outcome after potentially curative colorectal cancer surgery. Dis Colon Rectum 2000;43:492-8. [Crossref] [PubMed]
7. Compton CC, Greene FL. The staging of colorectal cancer: 2004 and beyond. CA Cancer J Clin 2004;54:295-308. [Crossref] [PubMed]
8. Zhang QW, Zhang CH, Pan YB, et al. Prognosis of colorectal cancer patients is associated with the novel log odds of positive lymph nodes scheme: derivation and external validation. J Cancer 2020;11:1702-11. [Crossref] [PubMed]
9. Ozawa T, Ishihara S, Sunami E, et al. Log odds of positive lymph nodes as a prognostic indicator in stage IV colorectal cancer patients undergoing curative resection. J Surg Oncol 2015;111:465-71. [Crossref] [PubMed]
10. Ogawa S, Itabashi M, Bamba Y, et al. Superior prognosis stratification for stage III colon cancer using log odds of positive lymph nodes (LODDS) compared to TNM stage classification: the Japanese study group for postoperative follow-up of colorectal cancer. Oncotarget 2020;11:3144-52. [Crossref] [PubMed]
11. Hankey BF, Ries LA, Edwards BK. The surveillance, epidemiology, and end results program: a national resource. Cancer Epidemiol Biomarkers Prev 1999;8:1117-21. [PubMed]
12. Camp RL, Dolled-Filhart M, Rimm DL. X-tile: a new bio-informatics tool for biomarker assessment and outcome-based cut-point optimization. Clin Cancer Res 2004;10:7252-9. [Crossref] [PubMed]
13. Vickers AJ, Elkin EB. Decision curve analysis: a novel method for evaluating prediction models. Med Decis Making 2006;26:565-74. [Crossref] [PubMed]
14. Wolbers M, Koller MT, Witteman JC, et al. Prognostic models with competing risks: methods and application to coronary risk prediction. Epidemiology 2009;20:555-61. [Crossref] [PubMed]
15. Edge SB, Compton CC. The American Joint Committee on Cancer: the 7th edition of the AJCC cancer staging manual and the future of TNM. Ann Surg Oncol 2010;17:1471-4.
16. Hari DM, Leung AM, Lee JH, et al. AJCC Cancer Staging Manual 7th edition criteria for colon cancer: do the complex modifications improve prognostic assessment? J Am Coll Surg 2013;217:181-90.
17. Huang B, Feng Y, Ni M, et al. Heterogeneous survival between stage IIA and stage IIIA colon cancer when different numbers of lymph nodes are harvested. ANZ J Surg 2018;88:316-21. [Crossref] [PubMed]
18. Huang B, Mo S, Zhu L, et al. The survival and clinicopathological differences between patients with stage IIIA and stage II rectal cancer: An analysis of 12,036 patients in the SEER database. Oncotarget 2016;7:79787-96. [Crossref] [PubMed]
19. Baxter NN, Morris AM, Rothenberger DA, et al. Impact of preoperative radiation for rectal cancer on subsequent lymph node evaluation: a population-based analysis. Int J Radiat Oncol Biol Phys 2005;61:426-31. [Crossref] [PubMed]
20. Gao P, Song Y, Yang Y, et al. What Is the Minimum Number of Examined Lymph Nodes After Neoadjuvant Therapy in Rectal Cancer? J Gastrointest Surg 2018;22:1068-76. [Crossref] [PubMed]
21. Huang LB, Yang TH, Yang L, et al. Clinical efficacy of adjuvant chemotherapy in the treatment of pT4 stage II colorectal cancer with defective mismatch repair status: A protocol for systematic review and meta-analysis. Medicine (Baltimore) 2020;99:e20693. [Crossref] [PubMed]
22. Jalaeikhoo H, Zokaasadi M, Khajeh-Mehrizi A, et al. Effectiveness of adjuvant chemotherapy in patients with Stage II colorectal cancer: A multicenter retrospective study. J Res Med Sci 2019;24:39. [Crossref] [PubMed]
23. Nitsche U, Stöss C, Friess H. Effect of Adjuvant Chemotherapy on Elderly Colorectal Cancer Patients: Lack of Evidence. Gastrointest Tumors 2017;4:11-9. [Crossref] [PubMed]