Survival Contradiction in Stage II, IIIA, And IIIB Colon Cancer: A Surveillance, Epidemiology, and End Result-Based Analysis

There exists an inconsistency between stage and survival in the current American Joint Committee on Cancer (AJCC) staging system for colon cancer. In this study, we compared the clinicopathological characteristics and prognosis of colon cancer patients with stage II, IIIA, and IIIB disease based on the surveillance, epidemiology, and end results (SEER) database. Kaplan-Meier analysis was used to generate overall survival (OS) and cancer-specific survival (CSS) curves. The Cox regression was employed to identify risk factors. The competing risk model was completed by the cumulative incidence function and Gray's test. In the final population of 31,361 colon cancer patients, Kaplan-Meier curve analysis showed that stage IIIA had the highest OS and CSS, followed by stage IIA and IIIB, and IIB and IIC showed the worst OS and CSS. In the Cox model, the stage was proven to be an independent prognostic factor. In the competing risk model, stage IIIA colon cancer patients had the lowest 5-year cancer-specific death rate in stages II, IIIA, and IIIB. In conclusion, the prognosis of colon cancer patients in stage IIA was worse than that of patients in stage IIIA, while the survival rate of stage IIB and IIC was lower than that of stage IIIB.


Introduction
Colorectal cancer (CRC) is the second leading cause of cancer-related death in the United States. It is estimated that in 2022, approximately 151,030 people will be diagnosed with CRC, and 52,580 people will die from this disease [1]. Te tumor, node, and metastasis (TNM) stage of the American Joint Committee on Cancer (AJCC) is the most widely used CRC staging system. In the 6th edition of the AJCC staging standard, stage II was divided into IIA and IIB for the frst time, and stage III included IIIA, IIIB, and IIIC. Te concept of stage IIC was added, and the subgroups of stage III were adjusted in the subsequent 7th edition [2]. Te provisions for stages II and III in the latest 8th edition have not been changed compared with those in the 7th edition. With the deepening understanding of CRC, the specifc substage has been redivided, but the criterion for distinguishing stage II from III has always been the presence of lymph node metastasis.
Tere exist controversies about the current staging system for CRC, and some researchers have reported inconsistencies between stage and survival. For example, in some studies, the prognosis of stage IIIA rectal cancer patients was better than that of stage II, and they believed this phenomenon was related to the defciency of the number of lymph nodes harvested (LNH) [3,4]. In another study comparing 359 colon cancer patients, Kim et al. found that the oncological outcome of T4N0 patients was worse than that of T1-2N1 [5]. Because of the standardized application of total mesorectal excision (TME) and complete mesocolic excision (CME), the treatment of lymph nodes in stage III colon cancer patients was satisfactory, and the prognosis was improved [6]. However, stage II patients have a higher T stage than stage III, which means a higher tumor load. Patients with locally advanced (especially T4) tumors have higher rates of local recurrence and peritoneal metastasis. It has been reported that peritoneal metastasis occurs in 15-20% of T4 colon cancer operations and seriously afects prognosis [7].
Due to the lack of multicenter, large population retrospective studies, the diference in the prognosis of patients with stage II, IIIA, and IIIB colon cancer is still unclear. Te surveillance, epidemiology, and end results (SEER) database can provide a wealth of information about the pathology and survival of cancer patients. Terefore, in this research, we used the SEER database to compare the clinicopathological characteristics and prognosis of colon cancer patients with stage II, IIIA, and IIIB in the real world.

Data Resource.
Te SEER program, established by the National Cancer Institute, collects demographic, clinicopathological, and survival information on tumors in representative geographic regions of the United States. Te SEER database has gathered and published information about cancer incidence and survival covering approximately 34.6% of the population in the United States [8] and is the largest public cancer database in the world. We extracted the relevant data using SEER * Stat 8.3.8 software (https://seer.cancer.gov/ seerstat/). Te Ethics Committee of the First Afliated Hospital of Xian Jiaotong University exempted the review of the study because the SEER database is publicly available.

Population Selection.
According to the SEER database, we accepted the 8th edition AJCC staging system to determine the tumor stage and identifed 48,174 patients with stage II, IIIA, or IIIB primary colon cancer between 2010 and 2015. Te exclusion criteria were as follows: (1) race unknown; (2) T stage missing; (3) N stage missing; (4) no or unknown surgery; (5) endpoints missing; (6) patients with 2 or more malignant tumors; and (7) marriage unknown. After the fnal screening, a total of 31,361 patients met the criteria. Tere were 14,871 patients in stage IIA, 1,514 patients in stage IIB, 1,140 patients in stage IIC, 1,956 patients in stage IIIA, and 11,880 patients in stage IIIB. Te detailed selection process is shown in Figure 1.

Variables and Outcome.
Te following clinicopathological features were collected according to the SEER database: sex, race, age, histological type, primary site, LNH, TNM stage, radiation state, chemotherapy state, survival months, vital status, cause of death, and marital status. On the basis of the 8th AJCC staging system, T3N0M0 is defned as stage IIA, T4aN0M0 is defned as stage IIB, T4bN0M0 is defned as stage IIC, T1-2N1M0 or T1N2aM0 is defned as stage IIIA, and T3-4aN1M0, T2-3N2aM0, or T1-2N2bM0 is defned as stage IIIB. Te ascending colon, hepatic fexure of the colon, and transverse colon are considered to be the right colon. Spiral fexibility of the colon, descending colon, and sigmoid colon are regarded as the left colon. Overall survival (OS) was measured from the date of diagnosis to the date of death. Cancer-specifc survival (CSS) and cancer-specifc death (CSD) referred to the date of diagnosis to the date of death from cancer.  (Figure 3(a)) or OS (Figure 3(b)), stage IIA tended to have worse survival rates than IIIA, while the prognosis of stage IIIB was better than that of stage IIB and IIC. Te same tendency was also shown in the LNH ≥ 12 subgroup (Figure 4). Among patients with stages II, IIIA, and IIIB disease, those with stage IIIA disease had the best prognosis, while those with stage IIB and IIC disease had the lowest survival rates.

Cox Proportional Hazard Model.
To further explore the infuence of multiple factors on CSS, we used the Cox proportional risk model to evaluate risk factors and protective factors (  Figure 5, after excluding the efect of non-CSD on survival, stage IIIA had the lowest 5-year CSD among all stages analyzed (8.21%, Gray's test, P < 0.001). In stage IIA, 11.59% of patients died in 5 years. Te cumulative CSD rates of stage IIB (28.34%, Gray's test, P < 0.001) and IIC (32.64%, Gray's test, P < 0.001) were both higher than that of IIIB (25.57%, Gray's test, P < 0.001).

Discussion
In this study, we evaluated the survival of 31,361 patients with stage II, IIIA, and IIIB colon cancer and found a contradiction between the prognosis and stage. Stage IIIA had the lowest mortality rate, and stage IIIB had better survival than IIB and IIC. Te Kaplan-Meier method, logrank test, and Cox regression are widely used in survival analysis. However, these statistical methods only involve one type of endpoint. When multiple endpoints exist, competition events and events of interest form a competition relationship, which makes Kaplan-Meier analysis and Cox regression overestimate the probability of events of interest or even obtain the opposite conclusion [9]. Terefore, a competing risk model is needed to provide a better clinical prediction [10]. After Gray's test, when the risk of death from other causes was removed, we still obtained a similar survival tendency, which further confrmed our conclusion. To our knowledge, this is the frst and largest populationbased study to determine the relationship between survival and stage in patients with stage II, IIIA, and IIIB colon cancer.
It has been reported that the survival rate of rectal cancer patients in stage IIA is lower than that of patients with stage IIIA when there is not enough lymph node collection, and this phenomenon may be related to stage migration [4]. Stage migration refers to when the number of lymph nodes examined is insufcient; it may be misdiagnosed as "lymph    Evidence-Based Complementary and Alternative Medicine node-negative" and reduce the stage, thus lacking necessary adjuvant treatment and resulting in poor prognosis [11][12][13]. Terefore, in our study, the population was divided into LNH ≥ 12 and LNH < 12 subgroups. Te results showed that even in the LNH < 12 subgroup, the prognosis of the IIIA stage was better than that of the II and IIIB stage, which indicated that there was no signifcant correlation between this survival contradiction and the number of LNH in colon cancer.
Although the AJCC staging system of CRC has been constantly modifed, the over-emphasis on the N stage has always existed in various versions [14][15][16]. For a long time, as long as lymph node metastasis is detected, it will be cataloged into stage III, regardless of the size of the primary tumor. It has been confrmed that the T stage has a greater proportion of infuence than the N stage in CRC [17]. Terefore, considering the contradiction between colon cancer stage and survival in the 8th edition, the AJCC staging system needs to reconsider the weight of the T and N stages and redefne the classifcation criteria of colon cancer stages II and III.   Evidence-Based Complementary and Alternative Medicine Te priority of chemotherapy in stage III is considered to be a possible reason for the better prognosis of stage III compared with T4N0 colon cancer patients [18,19]. Stage III patients are often recommended to have a higher rate of adjuvant chemotherapy, and postoperative adjuvant systemic therapy can signifcantly improve the prognosis [20,21]. With the emergence of various new drugs, as well as the wide application of immunotherapy and targeted therapy [22][23][24][25], the prognosis of stage III patients is becoming increasingly ideal. In our cohort, stage IIIA had the highest chemotherapy acceptance rate (67.7%), and the percentage of patients undergoing chemotherapy in stage IIIB (62.4%) was also higher than that in stages IIB (34.3%) and IIC (46.8%). Terefore, the better prognosis of stage III may be attributed to the fact that most patients with stage III colon cancer received adjuvant chemotherapy. Whether patients with stage II colon cancer should receive standard chemotherapy deserves further discussion in the latest guidelines.
Te diference in biological characteristics may be the internal mechanism of the survival contradiction [26][27][28]. Higher T staging not only means a larger primary tumor but also refects stronger aggressiveness [29]. At the same time, the ratio of microsatellite instability [30][31][32] and perineural invasion in T4N0 colon cancer was signifcantly increased [33], suggesting that stage IIB and stage IIC tumors may have diferent biological behaviors compared with stage III.
Tere are some limitations in our study. First, as a retrospective study, the inevitable selection bias should not be ignored. Second, the SEER database does not include information on specifc cancer-related biomarkers, which is very important for prognosis prediction. Tird, we only studied the data from 2010 to 2015, which is a relatively short follow-up period, and a longer follow-up time is needed to confrm our results in the future.

Conclusions
Te prognosis of colon cancer patients in stage IIA was worse than that of patients in stage IIIA, while the survival rate of colon cancer patients with stage IIB and IIC was lower than that of patients in stage IIIB, and this contradiction between survival and stage was not related to insufcient LNH. In the future, we may need to evaluate and modify the AJCC stage of colon cancer according to the prognostic information to determine a more accurate and suitable new staging system. Tis is a retrospective study, and large prospective studies are needed to verify our results.

Data Availability
Te datasets generated during and/or analyzed during the current study are available in the SEER repository [https:// seer.cancer.gov/seerstat/].

Ethical Approval
Te Ethics Committee of the First Afliated Hospital of Xian Jiaotong University exempted the review of the study because the SEER database is publicly available.