Postoperative recurrence of gastric cancer depends on whether the chemotherapy cycle was more than 9 cycles

Abstract We retrospectively reviewed the medical records of patients with pathologically confirmed gastric cancer/adenocarcinoma who underwent curative surgical resection follow-up within 3 years at Shanxi cancer hospital between 2002 and 2020. The clinicopathologic parameters explored included gender, age at surgery, vascular invasion, neural invasion, Tumor infiltration depth (T stage), N stage, TNM stage, chemotherapy, Lauren classification, maximum diameter of tumor, type of gastrectomy, tumor location and survival data. With a median follow-up of 29 months (range 0–36 months), the ratio of patients with recurrence was 26.80% (n = 226) and the death rate of patients was 45.31% (n = 382) in this period. According to the results of univariate analysis, gender (P = .014), age at surgery (P = .010), vascular invasion (P = .000), neural invasion (P = .000), T stage (P = .000), N stage (P = .000), TNM stage (P = .000), chemotherapy cycle (P = .000), lauren classification (P = .000), maximum diameter of tumor (P = .000), type of gastrectomy (P = .000) were independent risk factors of recurrence of follow-up within 3 years. From the multivariate analysis by logistic regression showed that TNM Stage (P = .002), chemotherapy cycle (P = .000) were risk factors of recurrence of follow-up within 3 years. Univariate analysis of survival by Kaplan–Meier showed that gender (P = .038), vascular invasion (P = .000), neural invasion (P = .000), maximum diameter of tumor (P = .000), Lauren classification (P = .000), T stage (P = .000), N stage (P = .000), TNM Stage (P = .000) and type of gastrectomy (P = .000) were key factors linked to overall survival of follow-up within 3 years. The results of the multivariate analysis by Cox regression were clearly presented that T Stage (P = .000), TNM stage (P = .001), maximum diameter of tumor (P = .001) were key factors of overall survival of follow-up within 3 years. TNM Stage, chemotherapy cycle were closely related to recurrence and of follow-up within 3 years. More than 9 cycles of chemotherapy was able to reduce the probability of recurrence. T Stage, TNM stage, maximum diameter of tumor were independent factors associated with overall survival of gastric cancer of follow-up within 3 years. For maximum diameter of tumor, the probability of death of more than 6 cm was 1.317 times less than 6 cm within 3 years of follow-up.


Introduction
Gastric cancer is the fifth most common malignant tumor in the world and the third most common cause of malignant tumorrelated death. [1] Though the overall survival (OS) of gastric cancer patients has improved with the development of standardized D2 lymphadenectomy [2] and subsequent adjuvant chemotherapy in recent years, [3,4] the long-term survival rate is still unsatisfactory. Recurrencies the main cause of disease-related death. [5] In our retrospective study, we found that the ratio of recurrence was 26.80% (n = 43) and the mortality rate was 46.07% (n = 94) within 1 year of follow-up. More than 1 year and less than 3 years, the ratio of recurrence and was 28.64% (n = 183) and the mortality rate was 45.07% (n = 288). In present study, we aimed to clarify the prognostic factors associated with recurrence after curative resection within 3 years of follow-up, which would help clinicians perform the appropriate treatment in order to improve OS.

Patients
Between May 2002 and December 2020, a total of 1715 patients with gastric cancer/adenocarcinoma after radical resection were selected from Shanxi cancer hospital, Shanxi, China. All of these patients underwent lymphadenectomy higher than D2 (complete removal of group 1 and 2 lymph nodes). All patients' clinicopathological characteristics, including gender, age at surgery, vascular invasion, neural invasion, Tumor infiltration depth (T stage), N stage, TNM stage (according to the 8th Edition of the American Joint Committee on Cancer staging manual), [6] chemotherapy, Lauren classification, maximum diameter of tumor, type of gastrectomy, tumor location and survival data, were retrospectively reviewed based on operative notes, medical records and telephone follow-ups. In our group, we selected 843 patients with follow-up within 3 years, all of these patients follow-up less than 3 years, including patients who died within 3 years and patients who alive and follow-up ended within 3 years.

Postoperative follow-up
The patients were followed closely until December 2020, the average length of follow-up was 38.32 (range 1-156) months. Follow-up assessments were performed every 3 months for the first 1 years after surgery, every 6 months for 2 to 5 years, and yearly thereafter. Routine follow-up consisted of physical examination, laboratory tests, chest radiography, abdominopelvic ultrasonography, computed tomography and magnetic resonance imaging. OS was defined as the time from curative resection of gastric cancer to death or the last follow-up time. All of methods and follow-up information reviewed and approved by the Medical Ethics committee of Shanxi cancer hospital.

Statistical analysis
All data were analyzed with SPSS 23.0 (SPSS Inc., Chicago, IL) software. Categorical variables were analyzed with a Chi-Squared test and Fisher exact test. Univariate analyses were performed with the Kaplan-Meier method. Continuous data, expressed as the mean ± standard deviation (SD), were compared by Student t test. In the multivariate analysis, logistic regression analysis was used to evaluate the risk factors of recurrence and cox regression analysis was applied to identify independent risk factors associated with survival. Multiple linear regression was applied to screen out the risk factors of recurrence and stepwise regression was used to select the most factor linked to recurrence. Survival analyses and curves were established with the Kaplan-Meier method and compared with the Log rank test. P < .05 was considered statistically significant.

Clinical characteristics
Of all these patients (n = 1715), the ratio of patients with recurrence was 23.90% (n = 410) and the death rate of patients was 38.08% (n = 653). The mean age was 58.74 ± 10.04 years old. In this group, 79.53% were female (n = 1364) and 20.47% were is female (n = 351). The average of follow-up was 38.32 ± 22.50 months. The 1 -, 3 -, 5 -year survival rates of these patients were 94.14%, 69.53%, 37.71% respectively. With regard to tumor location, including upper (n = 866), middle (n = 285), lower (n = 384) of the stomach and multiple (n = 16) site. About type of gastrectomy, including proximal gastrectomy (n = 166), distal gastrectomy (n = 555), total gastrectomy (n = 992) and PPG (n = 2). The average of maximum diameter of tumor was 5.03 ± 2.60 cm. The mean of chemotherapy cycle was 9.83 ± 3.80 cycles. In these patients (n = 1715), we selected patients with follow-up within 3 years (n = 843), the ratio of patients with recurrence was 26.80% (n = 226) and the death rate of patients was 45.31% (n = 382) in this period.

Risk factors for recurrence of follow-up within 3 years
The univariate analysis of postoperative recurrence was presented in Table 1 and these data was analyzed by Chi-Squared test and Fisher exact test. According to the results of univariate analysis, gender (P = .014), age at surgery (P = .010), vascular invasion (P = .000), neural invasion (P = .000), T stage (P = .000), N stage (P = .000), TNM stage (P = .000), chemotherapy cycle (P = .000), Lauren classification (P = .000), maximum diameter of tumor (P = .000), type of gastrectomy (P = .000) were independent risk factors of recurrence of follow-up within 3 years. In addition, tumor location was not significantly different between recurrence group and no-recurrence group (P > .05).
For multiple experimental groups, T stage, N stage, TNM stage, Type of gastrectomy, Lauren classification were closely associated with recurrence, so we performed pairwise comparisons for these risk factors ( Table 2). The results of pairwise comparisons of T stage is that the difference of T1vsT3 (P = .001), T1vsT4 (P = .000), T2vsT4 (P = .020) T3vsT4 (P = .000) was significantly between each other. For N stage, the otherness between each stage was apparently (P < .05) except for N1vsN2 (P = .176). With regard to TNM Stage and Lauren classification, there were significant differences between their each stage. About the type of gastrectomy, the ratio of recurrence of proximal gastrectomy versus total gastrectomy (P = .000), distal gastrectomy vs total gastrectomy (P = .034) were different between each other. In other words, the ratio of recurrence of total gastrectomy is higher than proximal gastrectomy and distal gastrectomy.
From the multivariate analysis analyzed by logistic regression showed that TNM Stage (P = .002), chemotherapy cycle (P = .000) were risk factors of recurrence of follow-up within 3 years ( Table 3). As for N stage, N Stage and recurrence of follow-up within 3 years were positive correlation. Furthermore, the difference of N0 VS N3 (P = .029) was significant, that means the ratio of recurrence of N3 was higher than N0. Moreover, TNM Stage and recurrence of follow-up within 3 years were positive correlation, chemotherapy cycle and recurrence of follow-up within 3 years were negative correlation. Namely, regarding TNM Stage, as the stage increases, the likelihood of recurrence increases. Inadequate chemotherapy cycle may be a risk factor for recurrence of follow-up within 3 years. The value of odds ratio for chemotherapy cycle means that the probability of recurrence more than 9 cycles is 0.296 times as much as less than 9 cycles.
On the basis of Table 4, that clearly displayed that N stage (P = .004), TNM Stage (P = .002), chemotherapy cycle (P = .000) According to absolute value of beta, chemotherapy cycle has the greatest influence on recurrence of follow-up within 3 years due to the maximum absolute value of 3 of these risk factors of recurrence. The value of these 3 factors of VIF no more than 3, that means that there is no collinearity problem between these independent variables.

Survival analysis
From Table 5, we can clearly see that gender (P = .038), vascular invasion (P = .000), neural invasion (P = .000), maximum diameter of tumor (P = .000), Lauren classification (P = .000), T stage (P = .000), N stage (P = .000), TNM Stage (P = .000) and type of gastrectomy (P = .000) were risk factors linked to overall survival of follow-up within 3 years. The overall median survival time of this group was 29 months. The median survival time of male was 30 months and female was 24 months (  (Fig. 4). About maximum diameter of tumor of median survival time, less than 6 cm was 33 months and more than 6 cm was 24 months (Fig. 5). The median survival time of chemotherapy cycle less than 9 cycles and more than 9 cycles was similar, was 31 and 28 months, (Fig. 6). As far as Lauren classification was concerned, intestinal type, diffuse type and mixed type of median survival time was 34, 24, 31 months singly (Fig. 7). As for T stage, T1, T2, T3, and T4 of median survival time was 35, 35, 35, 24 months respectively (Fig. 8). About of N stage, the median survival time of N0, N1, N2, N3 was 35, 33, 24, 24 months severally ( Fig. 9). Concern to TNM Stage, I, II, III, IV of median survival time was 35, 35, 24, 19 months respectively (Fig. 10).
With regard to tumor location, 3 of these of median survival time was comparable, upper was 29 month, middle was 24 months, lower was 32 months (Fig. 11). As to type of gastrectomy, Table 1 The univariate analysis of postoperative recurrence and metastasis.  proximal gastrectomy, distal gastrectomy and total gastrectomy of median survival time was 33, 33, 24 months respectively (Fig. 12).

Risky factors
The results of pairwise comparisons of T stage is that the difference of T1vsT4 (P = .000), T2 versus T4 (P = .000) T3 versus T4 (P = .000) was significantly between each other. For N stage, the difference between each stage were apparently (P < .05) except for N2vsN3 (P = .086). With regard to TNM Stage, there were significant differences between their each stage besides I vs II (P = .075). As for Lauren classification, intestinal versus diffuse (P = .000), diffuse versus mixed (P = .000) were different between each other. That means that the survival rate of diffuse type were higher than intestinal type and mixed type. About tumor location, only middle versus lower (P = .030) was significantly between each other. About the type of gastrectomy, the survival rate of proximal gastrectomy vs total gastrectomy (P = .000), distal gastrectomy vs total gastrectomy (P = 0.030) were different between each type of gastrectomy. In other words, the survival rate of total gastrectomy is higher than proximal gastrectomy and distal gastrectomy. The results of the multivariate analysis analyzed by Cox regression were clearly showed in Table 7. The outcome were that T Stage (P = .000), TNM stage (P = .001), maximum diameter of tumor (P = .001) were risk factors of OS of follow-up within 3 years. Moreover, TNM Stage, maximum diameter of tumor and OS of follow-up within 3 years were positive correlation. Maximum diameter of tumor may be a risk factor for OS of follow-up within 3 years. According to the hazard ratio of maximum diameter of tumor, the probability of death of more than 6 cm was 1.317 times as ≥6 cm.

Discussion
Most studies had noted that curative resection for gastric cancer focused largely on early gastric cancer or advanced gastric cancer. With regard to early gastric cancer, there were many articles on the risk factors of recurrence of early gastric cancer in Japan, Korea, and China. Hee Jun Choi [7] reported that among the 8753 patients of T1N0M0 gastric cancer 95 patients (1.1%) experienced tumor Table 3 Multivariate analysis of postoperative recurrence and metastasis analysed by logistic regression.  In another cohort study in Korea, [8] age over 65 years, male gender, stage IB, lymphovascular invasion, perineural invasion, and elevated level of carcinoembryonic antigen were independent poor prognostic factors for recurrence-free survival of stage I gastric cancer. In Japan, research revealed that age≥70 years and lymphatic and/or venous invasion were independent prognostic factors for poor recurrence-free survival of T1N+ or T2-3N0 gastric cancer. [9] Another study in Japan proved a location in the upper third of the stomach, tumor size of ≥30 mm, undifferentiated adenocarcinoma and clinical tumor depth were identified as independent risk factors for T1N0 gastric cancer. [10] In China, from 734 early gastric carcinoma radical resections found that Independent risk factors for lymph node metastasis in early gastric carcinoma include tumor size ≥3.0 cm, submucosa invasion to a depth more than 200 mm, moderate/poor differentiation, lymphovascular invasion and tumor necrosis. [11] Baesed on a joint study of China and the United States found that 7% (76) of our 1,058 patients from the United States (n = 414) and China (n = 644) recurred. Liver (43%) was the most common site of recurrence in both countries (US: 24%, China: 52%), followed by peritoneum              Tumor size (P = .001), depth of invasion (P = .010), histological type (P = .022) and lymphovascular invasion (P = .001) were independently associated with recurrence. [12] As far as advanced gastric cancer, many researcher explored the related factors of recurrence of advanced gastric cancer. In Japan, a retrospective study [13] showed that the 5-year relapse-free survival (RFS) rates of patients with Stage III gastric cancer were 42.0%. Univariate and multivariate analyses for RFS revealed that venous invasion was an independent factor predicting a shorter RFS. In China, another retrospective study [14] displayed that tumors located at upper, middle third, or mixed, a positive lymph node ratio ≥0.335, pTNM stage III, lymphocyte count < 1.5 Â 109/L, postoperative infection complications and adjuvant chemotherapy <6 cycles were all independent predictors for early recurrence after curative resection of stage II/III gastric cancer. The other 1 investigation in China [15] about prognostic factors and recurrence patterns in T4 gastric cancer patients after curative resection indicated that age, location of tumor and intraoperative blood loss were independent prognostic factors for overall survival. After a median follow-up of 25.87 months, a total of 109 (43.8%) patients suffered from recurrence, and 90 patients had been observed specific recurrence sites, among which peritoneal metastasis was the most common recurrence pattern, 59.0% for T4a and 88.3% for T4b, respectively. Another research hotspot concerning recurrence of gastric cancer concentrated upon early recurrence. In China, through a total of 149 patients with recurrence of gastric cancer/ adenocarcinoma of the esophagogastric junction after curative resection proved that perineural invasion, postoperative chemotherapy and postoperative complications were independent factors associated with early recurrence after curative resection. The survival analysis showed that perineural invasion (P = .011) and postoperative complications (P = .007) were independent prognostic factors. [15] Another study in China by 417 gastric cancer patients exhibited that there was no significant difference in recurrence patterns between early and late recurrence (P < .05 each). For pT1 stage gastric cancer, tumor size (P = .011) and pN stage (P = .048) were associated with early recurrence of gastric tumors. Patient age, pT stage, pN stage, Lauren histotype, lymphovascular invasion, intraoperative chemotherapy, and postoperative chemotherapy were independent predictors of early recurrence in patients with pT2-4a stage gastric cancer. [16] A retrospective study in Japan [17] via 96 patients with pStage III found that lymph node metastasis≥14, CA19-9≥37IU/mL and blood loss ≥445 mL were independent risk factors for early Recurrence after curative gastrectomy in pStage III GC. Another research in Japan showed that [18] early recurrence was associated with a high lmph node ratio (P = .0020) and high CA19-9 levels (P = .0415). The other factors were not significantly associated with early recurrence. Lymph node ratio≥0.15 and CA19-9≥37 U/mL were effective surrogate markers for predicting early recurrence.

Risky factors
In short, little attention paid to middle and long term recurrence of gastric cancer. Moreover, the sample capacity of current study of risk factors for recurrence is small, in our present study, the 1-year survival rate was 94.14%, the 3-year survival rate was 69.53% and the 5-year survival rate was 37.71% respectively. We noticed that from 1 year to 3 years, the survival rate dropped nearly a quarter. What is more, from 3 year to 5 years, the survival rate dropped nearly a half. If we attached much importance to insure chemotherapy cycle adequately performance, will the overall survival rate may be significantly improved. At the same time, maximum diameter of tumor can be reduced through complete chemotherapy implement. Finally, overall survival ratio of follow-up within 3 years may be enhanced by this measure.
Futhermore, studies on the relationship between chemotherapy and recurrence of gastric cancer as following. A study in Japan focused on comparison of recurrence patterns between S-1 single drug and simple operation groups. [19] The result was that overall and recurrence-free survival were better for the S-1 adjuvant group. In the surgery alone group, carcinoembryonic antigen ≥5 ng/mL, total gastrectomy, vessel invasion, pT4, and stage 3 were identified as significant prognostic factors. In striking contrast, macroscopic tumor size ≥50 mm was the only significant prognostic factor for the S-1 adjuvant group. Another study in Japan took notice of 34 patients underwent curative conversion surgery. [20] The result was that in 17 (50%) patients, with a median time to recurrence of 22 months (range = 1-98 months). In 9 (53%) patients with recurrence, the pattern was consistent with their initial metastatic disease. Initial clinical T4b disease was the only significant independent risk factor affecting recurrence-free survival. The other 1 study in Japan through Li and Zhao Medicine (2022) 101:5 Medicine 396 patients proved the relationship between S-1 monotherapy and the timing and sites of recurrence. [21] The results were that the 1-, 3-and 5-year RFS rates were 67.2%, 23.0% and 5.7%, respectively. Local recurrence, lymph node involvement and peritoneal and hematogenous metastases were found in 6, 25, 63, and 42 patients, respectively. Local recurrence and lymph node metastasis plateaued 3 years after gastrectomy. Peritoneal and hematogenous metastasis increased within 5 years after surgery.
In patients with hematogenous metastasis, the number of liver metastases plateaued but increased in others. A study in Korea enrolled 130 patients who had undergone an R0 resection and had completed 6 cycles of adjuvant chemotherapy. [22] The lymph node ratio (LNR), which was defined as the number of metastatic lymph nodes divided by the retrieved lymph nodes, was a significant risk factor in the lymph node-positive group (P < .01). Baseline CA19-9 level was a risk factor in the lymph nodenegative group (P = .01). We noticed that the sample studies on the relationship between chemotherapy and recurrence of gastric cancer was small. What is more, we need more samples and timing supervision to confirm the relationship between chemotherapy and recurrence of gastric cancer. At the same time, identify the effect of follow-up time on recurrence.

Conclusions
TNM Stage, chemotherapy cycle were closely related to recurrence of follow-up within 3 years. More than 9 cycles of chemotherapy was able to reduce the probability of recurrence.
T Stage, TNM stage, maximum diameter of tumor were independent factors associated with overall survival of gastric cancer of follow-up within 3 years. About maximum diameter of tumor, the probability of death of more than 6 cm was 1.317 times less than 6 cm after surgery. Li and Zhao Medicine (2022) 101:5 www.md-journal.com