Colorectal cancer (CRC) is the third most common malignant tumor worldwide and the leading cause of cancer mortality. The 5-year survival rate of patients with localized stage CRC is up to 90 %. However, if cancer cells spread to regional lymph nodes, it is associated with increased recurrence rates and worse outcomes for CRC patients [1]. As one of the main prognostic factors, identification of lymph node involvement by surrounding primary CRC provides important information regarding therapeutic options and intensive surveillance in clinical practice because survival benefit possibly be achieved with adjuvant chemotherapy after curative resection in these cases [2]. In current practice, histopathological examination of hematoxylin-eosin (H-E) staining on sections is the standard or routine procedure for the diagnosis of lymphatic invasion, a reliable means of detecting macrometastases. It would be challenging if a minimal number of cancer cells invade into lymph nodes, leading to the omission of micrometastases and disease understaging, which may finally affect the clinical outcomes [3]. In addition, it remains uncertain whether clusters of tumor cells or isolated tumor cells located in lymph nodes necessarily predict tumor relapse. Therefore, it is necessary to identify molecular markers with higher sensitivity for determining and monitoring lymphatic invasion of CRC.

In recent decades, various techniques have been used to identify occult micrometastases in regional lymph nodes at the molecular level. The main advantage of these strategies is that the examination of the completely homogenized lymph node tissue minimizes the potential risk of missing cancer cells [4]. Amplification of two cancer-specific RNAs (FXYD3 and KRT20) that occur in primary bladder cancer was detectable in lymph nodes. Compared with traditional histological examination, molecular analysis of lymph node micrometastases has a higher detection rate [5]. Croner et al. applied a one-step nucleic acid amplification system to detect cytokeratin 19 mRNA for evaluation of lymph node micrometastases of CRC [6]. The data resulted in an upstaging of 25.2 % of patients with initially histologically negative lymph nodes from UICC-I/II to UICC-III. The same or similar strategy has also been used to detect cancer cells in resected lymph nodes of breast cancer and lung cancer patients for clinical decision-making [7], [8]. In addition, immunohistochemistry (IHC) staining of lymph nodes has been described to improve the clinical staging of non-small cell lung cancer and predict prognosis and outcome [9]. However, the insufficient sensitivity and specificity of the current molecular markers make it difficult for clinical applications so far [10].

Epigenetic and genetic abnormalities are hallmarks of human cancers. Compared with gene mutation, DNA methylation alteration is an early event and more stable during the process of tumor development. In recent decades, epigenetic markers have been widely used for detecting, diagnosing and predicting malignant tumors. For example, estrogen receptor promoter methylation has been applied to check occult lymph node dissemination of CRC and concluded that methylation signature of estrogen receptor might be a useful marker for molecular detection of lymph node spreading and predicting local recurrence or distant metastasis [11]. However, very few studies on this topic have been documented. In addition, the presence and significance of lymphatic invasion for tumor recurrence in subjects who tested positive for methylation markers remain undefined.

Recent studies have correlated DNA methylation of many genes with CRC patients' prognosis/outcome. Among all the analyzed genes, Septin 9 was the most frequently reported hypermethylated gene in CRC. Septin 9 is a member of the septins that can form membrane diffusion barriers and play roles in cell cycle, cytokinesis and vesicle trafficking [12]. Septin 9 methylation in CRC tissues and plasma has been comprehensively reported [13], [14], [15]. According to the 2016 US recommendation statement on screening for colorectal cancer, a blood test for detecting circulating methylated Septin 9 was mentioned in that text [16]. In addition, we also identified hypermethylated Septin 9 in cervical cancer and nasopharyngeal carcinoma, which could be a promising marker for early cancer detection [17]. Therefore, we speculated that Septin 9 methylation would be of great convenience and significance in determining lymph node micrometastases of CRC. The present work aimed to evaluate the relationship between Septin 9 methylation in lymph nodes and the risk of recurrence in CRC patients treated with surgery. Our study revealed a clear association between the epigenetic maker for lymph node involvement and CRC progression.