Implication of lymph node staging in migration and different treatment strategies for stage T2N0M0 and T1N1M0 resected gastric cancer: a SEER population analysis

Gastric cancer (GC) is the fourth leading cause of cancer-related death worldwide [1]. Gastric cancer patients, survival can be predicted based on the 8th edition of the American Joint Committee on Cancer (AJCC) tumor–lymph node–metastasis (TNM) classification, which provides important prognostic instruction for further treatment [2]. According to the National Comprehensive Cancer Network (NCCN), adjuvant chemotherapy (CT) is recommended for stage T1N1M0 (invading the mucosa or submucosa, having 1 or 2 positive lymph nodes and no distant metastasis) GC after the curative resection [3]. On the other hand, while CT is not recommended for all stage T2N0M0 (invading the muscularis propria and having no positive lymph node and no distant metastasis) patients, it is recommended for high-risk patients with poorly differentiated or higher-grade cancer, lymphatic and/or blood vessel invasion, perineural invasion, or who are under 50 years of age. According to the Japanese Gastric Cancer Treatment Guidelines, observation alone is recommended without any adjuvant treatment (AT) after curative resection for both T1N1M0 and T2N0M0 GC [4]. Currently, two main randomized controlled trial (RCT) studies exist for AT, ACTS-GC trial [5] (included stage II-III GC) and CLASSIC trial [6] (including stage II-IIIB GC) in Asia. However, these studies do not address whether stage IB GC patients would benefit from AT following surgery. Thus, unlike stage II–III GC, there is no consensus on using AT for stage IB GC.

Kwon et al. reported a small gap in the 5-year survival between stage IB and IA GC [7]. Previous studies have found an overlapping survival curve between stage IB and stage IIA GC using the 7th/8th edition AJCC staging system [8‐10]. Furthermore, Gold et at reported significantly different disease-specific survival at 5 years between stage T1N1M0 GC and stage T2N0M0 GC which are both stage IB GC [11]. In contrast, other reports demonstrated no differences in survival between the two subgroups [12‐15]. Few reports exist on stage migration and treatment strategies of stage IB GC and the prognostic factors of stage IB remain unclear. If the similar prognosis of stages T2N0M0 and T1N1M0 were confirmed, some patients may benefit from sparing of treatment-induced toxicity. If a worse prognosis of the T2N0M0 stage is confirmed, AT would be recommended.

The primary aim of this study was to explore the differences between the stages T2N0M0 and T1N1M0 GC and to identify the necessity of AT in each stage using the Surveillance, Epidemiology and End Results (SEER) database.

We sought to seek evidence to explore the differences between T2N0M0 and T1N1M0 GC, both classified as stage IB GC, using the SEER database. Overall, T2N0M0 group had the clinical pathological characteristics comparable to T1N1M0 group except the tumor size. The larger tumors were widely distributed in T2N0M0 group, while the smaller tumors were mainly aggregated in T1N1M0 group. These findings are consistent with the study of Du et al. [14].

The T2N0M0 group had a significantly better GCSS and OS than the T1N1M0 group, although more patients received AT in T1N1M0 group. The univariate and multivariate Cox proportional hazard regression models for the stage IB GC demonstrated that T1N1M0 was an independent risk factor of worse GCSS and OS for prognosis, which strongly supports the survival analysis results. Furthermore, following the propensity score matching, the T2N0M0 group was found to have a better GCSS and OS in comparison to the T1N1M0 group suggesting that T2N0M0 exhibited better survival than T1N1M0. These results are consistent with the findings of Gold et al. in SEER database from 1988 to 2008 [10]. Conversely, other reports demonstrated a comparable survival rate between the two subgroups [12‐15]. According to the 8th edition AJCC GC staging system, it is widely accepted that the number of ELN should be at least 16 [18]. The majority of patients in Japan and Korea received extensive lymphadenectomies (D2 lymphadenectomy or greater) and had more ELN [19, 20]. However, most surgeons in the western countries perform more limited lymphadenectomies (D1 lymphadenectomy or less) because of high perioperative complications and mortality in D2 lymphadenectomy [21, 22]. The proportion of 16 or more ELN was only 39.22% after radical surgery in our study which is similar to the other study showing that more than half of the GC patients in the USA had a fewer than 16 ELN [23]. Furthermore, subgroup analysis revealed that T2N0M0 and T1N1M0 were consistent in survival when ELN were > 15, but T1N1M0 demonstrated a worse survival than T2N0M0 when ELN were ≤ 15. One of the obvious reasons to explain the survival difference between T2N0M0 and T1N1M0 is number of ELN which can cause stage migration [24, 25].

The AT strategies differed between T2N0M0 and T1N1M0 which may affect the survival. On one hand, AT demonstrated better survival than SA in GCSS but not in OS for T2N0M0. On the other hand, AT demonstrated significantly better survival than SA in both GCSS and OS for T1N1M0 which was consistent with the study of Jabo et al. [26]. Furthermore, AT was found to yield better GCSS and OS outcome than SA after case–control matched analysis for T1N1M0. However, AT demonstrated a better OS than SA for T2N0M0 before matched; this advantage was dismissed after matched. Therefore, T2N0M0 GC may not benefit from AT based on the similar GCSS and OS results after matched.

To further understand which specific treatment modalities might lead to the differences in survival, especially for stage T1N1M0 GC, AT was divided into two subgroups: CT and CRT. In the USA, CRT using 5-fluorouracil-based regimen is recommended after D1 gastrectomy for T1N1M0 GC. In Europe, CT is recommended after D2 gastrectomy for all stages except for T1N0M0 GC [27]. In contrast, the Japanese Gastric Cancer Treatment Guidelines recommend observation alone after curative resection for T1N1M0 and T2N0M0 GC [3]. Overall, there has been no consensus on the use of AT in T1N1M0 and T2N0M0 GC. In our study, CRT subgroup showed the same survival when compared to CT subgroup in both T2N0M0 and T1N1M0 GC. Similarly, Kim et al. reported the CT and CRT had the similar survival in T1N1M0 GC [28]. However, another SEER-based data showed that the CRT demonstrated a better survival in comparison to SA or CT in both T2N0M0 and T1N1M0 GC [29]. While their SEER results of this study were based on the 6th edition AJCC GC staging, our results were based on the 8th edition AJCC GC staging. Our results clearly showed that RT was not required after radical surgery in either the T2N0M0 or the T1N1M0 GC. The CT demonstrated a significantly better GCSS and OS than the SA for T1N1M0 GC. While lack of CT was not an independent risk factor affecting prognosis for stage IB and T2N0M0, it was an independent risk factor for T1N1M0 highlighting the profound influence of CT on the prognosis of T1N1M0 [30].

Although T2N0M0 and T1N1M0 are both stage IB according to the 8th edition AJCC classification, they have significantly different prognosis and treatment strategies, especially when ELN were ≤ 15. We should pay more attention to the improvement of surgical quality to obtain adequate lymph nodes to avoid stage migration.

Our study has limitations, based on the retrospective nature of the study. First, factors such as the specific surgical methods (D2 or D1), specific CT information (regimen and duration), lymphatic and/or blood vessel invasion, perineural invasion were not available in SEER database. Secondly, due to the inability to match 1: 1: 1 for lack of technology, we cannot match SA: CT: CRT with 1: 1: 1 using the propensity score matching. Thirdly, because all the patients included in our paper were surgery performed, we could only acquire the major clinical data and the TNM stage based on pathological results; so, our conclusions were only applied to pTNM stage but not clinical TNM stage (cTNM) or postneoadjuvant TNM stage (ypTNM). Moreover, the number of patients from the three subgroups would decline sharply after matched. However, based on our existing results, we can also explain the effects of specific AT methods on survival because of large sample size. Moreover, we only prove that T2N0M0 may not benefit from AT and that T1N1M0 requires CT. But in T2N0M0 patients with high-risk factors, AT may be necessary and in T1N1M0 patients with low-risk factors, CT may be avoided to prevent toxicity and discomfort. Finally, these findings need to be further validated in non-US cohorts.

In conclusion, stage T2N0M0 GC has a better survival than stage T1N1M0 GC when ELN were ≤ 15. After radical surgery, T2N0M0 GC may not benefit from AT, while T1N1M0 GC requires CT but may not require RT. However, more than 15 ELN are essential to avoid stage migration. A prospective and RCT is required for further confirmation of these results.