Development and validation of nomogram for predicting early recurrence after radical gastrectomy of gastric cancer

Gastric cancer (GC) ranks as the fifth most prevalent malignancy globally, with the highest incidence recorded in East Asia, and generally, it exhibits a higher prevalence in men than in women [1]. Although the worldwide incidence of gastric cancer has displayed a significant decline in recent years [2], it remains the third leading cause of cancer-related mortality [3]. With the advancement of medical technology and treatment model, the current treatment method for GC is the comprehensive management of surgery combined with chemotherapy. Combined treatment regimens have demonstrated a substantial enhancement in overall survival when compared to surgery alone [4, 5]. While most patients can achieve improved medical remission through individualized treatment, those who experience postoperative gastric cancer recurrence often face unsatisfactory therapeutic outcomes, with a notably reduced overall postoperative survival rate compared to patients without recurrence [6, 7]. Postoperative recurrence stands as a significant contributor to the dismal prognosis and heightened mortality in gastric cancer cases. In China, over 60% of patients with advanced gastric cancer encounter recurrence and metastasis following surgical intervention [8].

The recurrence patterns of gastric cancer are primarily categorized into abdominal metastasis, distant metastasis or hematogenous metastasis, and local recurrence, each exhibiting distinct recurrence timeframes [9, 10]. Most instances of postoperative gastric cancer recurrence occur within the first 2 years [11, 12], and carry a high short-term mortality rate [13]. Therefore, early diagnosis and treatment intervention for postoperative gastric cancer recurrence become pivotal for improving prognosis. Presently, the prognosis of gastric cancer patients primarily relies on the TNM staging system [14]. However, an increasing number of scholars advocate that a nomogram represents a more effective tool for predicting tumour progression and guiding clinical decision-making [15, 16]. Considering the challenging survival rates after gastric cancer recurrence, several foreign studies have explored risk factors for gastric cancer recurrence and the development of nomograms for predicting disease-free survival (DFS) [17‐19]. For instance, Tonello et al. [17] established a nomogram to predict DFS, incorporating the Lauren classification and lymph node ratio (LNR). In a retrospective study by Spolverato et al. [18], it was determined that tumour site, depth of invasion, LNR, and lymphovascular invasion (LVI) were independent risk factors for recurrence. However, most of the predictive studies originate from Europe and the Americas, with limited research on the development of nomograms for predicting early recurrence in China.

In this study, our objective is to construct an effective prediction model for early gastric cancer recurrence following surgery and investigate the factors influencing the prognosis of patients with recurrent gastric cancer.

While the incidence of gastric cancer has shown a decreasing trend in China over the years, it remains a prevalent disease impacting public health [20]. Recurrence and metastasis following radical gastrectomy are critical factors contributing to the unfavorable prognosis of patients [6]. Notably, in China, limited attention has been devoted to the timing of gastric cancer recurrence, and available recurrence data often rely on small sample sizes [21, 22]. Hence, this study aims to investigate the recurrence patterns of gastric cancer in our country and establish a risk model for early recurrence prediction. The resulting nomogram not only identifies high-risk groups for gastric cancer recurrence but also offers valuable guidance to clinicians in devising effective follow-up strategies.

Comparing our patient data with those of the United States and South Korea, we observed lower rates of loco-regional recurrence (25.7% ~ 44.1% vs. 9.0%) [9, 23]. In a multicenter retrospective study conducted by Dan et al. [8], peritoneal metastasis rates in China were recorded at 13.7%. In contrast, our study revealed a significantly higher rate of abdominal metastases (32.2%), aligning more closely with the rates observed in South Korea and the United States (26.0% ~ 45.9%) [9, 23, 24]. It is worth noting that a substantial portion of Dan’s study data was from 15 years ago, possibly due to limited imaging sensitivity and the infrequent use of laparoscopic exploration at that time [25]. Our study, in contrast, predominantly comprises cases from the last decade, thus offering a more contemporary depiction of the gastric cancer recurrence landscape in our country. Previous research has established that early recurrence primarily involves hematogenous metastasis, whereas late recurrence predominantly manifests as loco-regional recurrence [9, 24, 26]. Our study results corroborate this finding, with the abdominal cavity (n = 168, 32.2%) emerging as the primary site for postoperative gastric cancer metastasis. Specifically, early recurrence is primarily associated with abdominal metastasis (n = 127, 33.9%), while late recurrence is more frequently observed in other anatomical sites (n = 49, 33.5%). Furthermore, our constructed nomogram highlights the significance of metastatic location as one of the key predictive factors. Independent risk factors for early recurrence include abdominal metastasis, liver metastasis, and retroperitoneal lymph node metastasis, with liver metastasis bearing the highest risk coefficient among them.

Numerous studies have consistently demonstrated that the lymph node metastasis as an independent predictor of early recurrence [11, 22, 23, 27]. Our research led us to the conclusion that the risk of early recurrence significantly escalates when postoperative lymph node pathological staging reaches N3. In line with our findings, Kim et al. [28] identified age, LNR, and serous infiltration as potential predictors of recurrence. However, our prediction model surpasses their established nomogram in performance for predicting DFS (CI 0.74 vs. 0.71) and accounts for a broader range of influencing factors. Furthermore, our study supports the notion that serous infiltration is a pivotal risk factor for postoperative peritoneal recurrence and metastasis in patients with gastric cancer, a finding consistent with several previous studies [22, 27]. For patients with serous infiltration, multiple studies have reported that combining D2 radical surgery with hyperthermic intraperitoneal chemotherapy effectively reduces the incidence of postoperative peritoneal metastasis without causing a significant increase in perioperative complications [29, 30]. Nevertheless, it’s worth noting that there is currently no universally accepted standard regarding population selection, treatment plans, drug dosages, perfusion frequencies, temperature control, and other aspects of prophylactic hyperthermic intraperitoneal chemotherapy. Consequently, further clinical trials are essential to provide definitive answers to these critical questions. Previous research has stressed the significance of tumour markers associated with gastric cancer in the monitoring of recurrence [31]. Hence, this study took into account the impact of preoperative tumour markers. The developed nomogram integrates the tumour marker CA19-9, underscoring its predictive value in identifying high-risk recurrence patients. Consequently, we advocate the routine assessment of preoperative tumour markers in individuals diagnosed with gastric cancer.

Furthermore, exploring factors influencing OS in GC patients is crucial. Our research underscores that the most critical factor affecting postoperative survival rates in gastric cancer recurrence patients is early relapse. This viewpoint aligns with the findings of Kodera et al. [32], who assessed the impact of routine follow-up on postoperative survival. Their study indicated that while close monitoring aids in early recurrence detection and significantly improves survival rates once recurrence is identified, it doesn’t bring substantial overall benefits to postoperative survival. This is primarily due to early relapses and a shorter DFS time. Many previous studies have demonstrated the impact of neoadjuvant and adjuvant chemotherapy on recurrence and survival in gastric cancer patients [33, 34]. However, in contrast to previous reports, we did not find an association between chemotherapy and DFS. In a multifactor analysis, neoadjuvant chemotherapy did not significantly impact postoperative OS, but adjuvant chemotherapy showed significant benefits in improving OS. Subsequently, we obtained some interesting results in a subgroup survival analysis of postoperative adjuvant chemotherapy. Patients experiencing early relapse may have better survival or longer survival periods when receiving adjuvant chemotherapy. Regarding TNM staging, stage III patients showed a more significant improvement in prognosis with postoperative adjuvant chemotherapy. For patients with late relapses and stage I/II patients, the presence or absence of adjuvant chemotherapy did not show a significant difference in prognosis. These findings suggest that in the advanced disease stage, adjuvant chemotherapy may play a more critical role in disease management, emphasizing the importance of postoperative adjuvant chemotherapy in advanced disease. Currently, there have been many large clinical studies investigating the ability of postoperative adjuvant chemotherapy to extend the survival period of late-stage patients [35, 36]. Our study included a larger number of patients and considered the influence of other clinical and pathological factors, providing valuable insights. Furthermore, among various modes of recurrence, peritoneal metastasis demonstrated poorer postoperative OS compared to other recurrence patterns, as previously reported [37]. Peritoneal metastasis in gastric cancer is typically considered an advanced-stage condition often accompanied by complications like ascites, intestinal adhesions, and even obstructions, significantly impacting patients’ quality of life. Our study results serve as a warning to clinicians for preventing peritoneal recurrence and guiding subsequent treatments.

The eighth edition of the TNM classification remains pivotal for postoperative treatment and follow-up planning [14]. However, a growing body of research emphasizes alternative factors and tools for predicting postoperative recurrence risk and improving survival [12, 17, 19]. Our carefully crafted clinical prediction model encompasses a comprehensive set of factors and has shown robust performance in subsequent validation. In terms of clinical decision curve results, when compared to the traditional TNM staging system, our predictive model serves as a promising resource that can enhance informed decision-making in clinical practice and patient care. Its ability to integrate other factors and demonstrate robust performance in validation underscores its potential significance in optimizing patient prognosis.

Nevertheless, there are some limitations that should be noted. Firstly, it’s based on looking back at past data, with some patient data missing and possible biases in the way patients were selected to participate in the study. Secondly, this research only used data from one hospital, so the people in the study might not be very diverse. Lastly, while this study analyzed factors affecting OS in GC recurrence patients and considered adjuvant chemotherapy, it did not create a tool to predict OS after radical gastrectomy.

The nomogram in our study can accurately predict the early postoperative recurrence rate of gastric cancer patients, and identify those patients at high risk of postoperative early recurrence, potentially promoting highly tailored patient management. This study not only improve people’s understanding of early recurrence of gastric cancer in Chinese or Asian populations, but also has certain guiding significance for the first line clinicians to make treatment plan and follow-up plan.