JSUR Journal

Introduction

Gastric cancer is the fifth most frequently diagnosed cancer and the 3^rd leading cause of cancer-related deaths worldwide, contributing to over 768.000 deaths in 2020 [1]. Its incidence has a wide geographic variation, with a decreasing incidence over the years in Western Europe and the United States, and a high incidence in Japan and Korea, where it is the most diagnosed cancer in males [1].

Globally, gastric cancer has a poor prognosis since it’s mostly diagnosed in an advanced stage.[1] Patients treated with curative intent have a 5-year survival rate of 70% for stage I resected disease but less than 30% for stage IIB and beyond [2]. Despite the improvement of outcomes in advanced gastric cancer related to the use of perioperative chemotherapy, surgical resection remains the only potentially curative treatment. [3] Surgical technique, Resection Margins (RM) and the extent of lymphadenectomy are important factors for patient outcomes [3]. Complete resection with negative margins (R0) is the primary goal of surgery in gastric cancer [1] Defining the minimum adequate RM to ensure negative margins has proven difficult, with several studies showing conflicting results [4-6]. The aim of this study was to determine the association between margin status and RM distance to the tumour and margin status and recurrence, Disease Free Survival (DFS) and Overall Survival (OS).

Material and Methods

This retrospective study included patients with histologically proven gastric or esophagogastric junction cancer, followed during a 10-year period (January 2011 - December 2020), who underwent surgery with curative intent between January 2011 and December 2015, in a tertiary center. Exclusion criteria were lymphadenectomy other than D2, R2 resection, histological diagnosis different from adenocarcinoma, Type I adenocarcinoma of the esophagogastric junction according to the Siewert classification [7] and follow up less than 5 years. The following demographic, clinical and pathological variables were recorded for each patient and inserted in a computerized database: age, sex, clinical TNM classification, neoadjuvant chemotherapy, type of gastrectomy, multivisceral resection, margin status, margin distance, tumour location, tumour diameter, tumour grade, T category, N category, Lauren pattern, WHO classification, lymphatic, vascular and perineural infiltration (LVI and PNI), pathological TNM classification, timing and site of recurrence, RFS and OS. Information was extracted from the clinical records. In addition, histological confirmation analysis of some of the specimens was performed to complete and standardize values. For this study, the minimum follow-up was 5 years or until death to determine RFS and OS.

Proximal and distal Resection Margin (RM) were defined as the shortest distance between the tumour, at the most proximal or distal end, to each resection line macroscopically, measured on formalin-fixed surgical specimens by pathologists. A negative RM (R0 resection) was defined as the complete absence of both macroscopic and microscopic tumour involvement at the resection line on histopathology assessment. A microscopically positive RM (R1 resection) was defined as the presence of viable singular tumour cells or cell aggregates at the line of resection on histopathology assessment, in the absence of visible tumour involvement of the resection line. Cancer staging was based on the American Joint Committee on Cancer TNM classification, 8^th edition. [8] Time to recurrence was calculated from the date of surgery to the date of recurrence diagnosis. Recurrence was classified as locoregional (anastomosis site, remaining stomach, gastric bed, regional lymph nodes, adjacent organ or paraaortic lymph node), hematogenous (distant organs), peritoneal (peritoneal seeding or Krukenberg’s tumour) and mixed. OS rate was computed as the percentage of patients in the study who were alive at the 1-, 3- and 5 year-mark after surgery. RFS rate was calculated as the percentage of patients in the study who were alive and without recurrence at the 1-, 3- and 5 year-mark.

Statistical Analysis

Categorical data are presented as the percent proportion; continuous data are presented as the mean ± Standard Deviation (SD) or median ± Interquartile Range (IQR) depending on its distribution. To compare differences in categorical data between two groups, Fisher’s exact test or the Chi-Square test was used, as appropriated. The Unpaired T Student t-test for independent samples was used to compare continuous data between two groups. Survival and recurrence analysis was made using the Kaplan-Meier Method (LogRank test) for categorical data and Univariate Cox Regression for continuous data. All variables found to be statistically significant were subsequently included in a multivariate Cox regression model to determine their weight predicting recurrence or survival. When both N category and pTNM classification were statistically significant, only pTNM classification was maintained. Hazard Ratio (HR) and 95% confidence intervals (95% CI) were presented. All analyses were two-sided, and statistical significance was defined as p<0.05. Statistical analyses were performed with the IBM SPSS Advance Statistics 28.0 package.

Results

Ninety-six patients were included in the study, 56 (58.3%) of which were male, with a mean age of 70±12 years old. Most patients (n=45 (46.9%)) had stage III disease, 18 (18.8%) were treated with neoadjuvant chemotherapy, 55 (57.2%) patients had tumours located at the antrum and 54 (56.3%) underwent subtotal gastrectomy. Positive margins (R1) were present in 5 (5.2%) of the patients: 1 with proximal margin invasion (a patient with a tumour located on the cardia who underwent total gastrectomy), 3 with distal margin invasion (1 patient with a tumour located on the antrum who underwent total gastrectomy, 1 patient with an antrum tumour who underwent subtotal gastrectomy with multivisceral resection and 1 patient with a tumour located on the gastric body who underwent total gastrectomy) and 1 with lymphatic invasion on the distal margin. Since only one patient presented R1 due to proximal positive margin (distance: 3mm), it was not possible to test the association between proximal margin distance and margin status. Mean proximal R0 distance was 41±40mm. Mean distal margin distance in R0 patients was 41±29mm and in R1 due to distal margin involvement was 27±12mm. No statistical difference was found between the medians of R0 and R1 patients (p=0,520). Table 1 resumes demographic, clinical and histopathological characteristics of positive and negative margin groups. The distribution is homogeneous for most characteristics except for Lauren pattern, WHO classification, pTNM classification and recurrence.

*Significance value of the model; BMI: Body Mass Index; LVI: Lymphovascular invasion; PNI: Perineural infiltration; WHO: World Health Organization.

Table 1: Demographic, clinical and histopathological characteristics of positive and negative margin groups and associated with recurrence.

Recurrence Analysis

The median follow-up was 56.5 months (IQR: 18-63.5 months); during this period, 44 deaths (45.8%) and 33 recurrences (34%) were registered, presenting mainly as peritoneal (36%) or locoregional (33%) disease. Table 1 summarizes the univariate analysis of demographic, clinical and histopathological associations with recurrence.

Mean proximal margin distance was 41±26mm in patients without recurrence and 41±25mm in patients with recurrence. There was no statistical difference between recurrence depending on proximal margin distance (p=0.907). Patients were also divided in 4 groups depending on proximal margin distance (<2cm, ≥2cm and <3cm, ≥3cm and <5cm and ≥5cm); no statistical difference between groups was found relating to recurrence (p=0.387).

Median distal margin distance was 35±43mm in patients without recurrence and 25±38mm in patients with recurrence. There was no statistical difference between median distal margin distance between patients with or without recurrence (p=0.819). Tumour location (p=0.007), margin status (p=<0.001), pN (p<0.001), pTNM classification (p<0.001), OMS classification (p<0.001), Lauren classification (p<0.001), tumour grade (p=0.031), PNI (p=<0.001) and LVI (p=0.032) had a statistically significant association with recurrence.

Cox regression analysis was employed to study the weight of each variable when predicting recurrence. Table 2 shows the Cox regression coefficient, Wald test, HR and 95% CI for each predictor. Tumor location, positive margins and PNI were associated with recurrence. Tumours located in the fundus (p=0.017; HR: 0.017 (95% CI: 0.001-0.49)), body (p=0.030; HR: 0.056 (95% CI: 0.004-0.75)) and antrum (p=0.026; HR: 0.050 (95% CI: 0.004-0.70)) were less associated with recurrence than tumours located in the cardia. Patients with positive margins (p=0.035; HR: 3.85 (95% CI: 1.10 - 13.45)), had four times more recurrence than the remaining patients. Perineural infiltration (p=0.013; HR: 5.43 (95% CI: 1.43-2062)) was five times more associated with recurrence than patients without PNI.

LVI: Lymphovascular invasion; PNI: Perineural infiltration; WHO: World Health Organization

Table 2: Cox regression predicting recurrence using tumour location, margin status, tumour stage, WHO classification, Lauren classification, tumour grade, LVI and PNI.

Survival Analysis

The 1-, 3-, and 5-year OS rate of the entire series were 80.1%, 59.4% and 54.2% (Figure 1), respectively, while the 1-, 3- and 5-year were RFS 70.8%, 54.2% and 51.0% (Figure 2). Table 3 resumes the univariate analysis of demographic, clinical and histopathological associations with OS. Mean proximal margin distance was 40±26mm in alive patients in 5-years OS and 41±25mm in deceased patients in OS. There was no statistical difference between proximal margin distance between both groups (p=0.969). After further division in 4 groups depending on proximal margin distance as previously described, proximal margin distance remained without statistically significant difference between groups (p=0.401). Median distal margin distance was 40±37mm in alive patients in OS and 40±44mm in deceased patients. There was no statistical difference between median distal margin distance in patients with or without recurrence (p=0,935). Age (p=0.002), Margin status (p=<0.001), pN (p=0.007), pTNM stage (p=<0.001), PNI (p=0.024) and recurrence (p=<0.001) were associated with OS. Cox regression analysis was employed to study the weight of each variable when predicting OS.

Figure 1: Overall Survival (Kaplan-Meier curves).

Figure 2: Recurrence Free Survival (Kaplan-Meier curves).

* Significance value of the model; LVI: Lymphovascular invasion; PNI: Perineural infiltration; WHO: World Health Organization.

Table 3: Demographic, clinical and histopathological associations with OS.

Table 4 shows the COX regression coefficient, Wald test, OR and 95% CI for each predictor. Age, positive margins, and recurrence were associated with OS. Increasing age (p=0.005; HR: 1,044 (95% CI: 1.01-1.08)) was associated with more mortality. Positive margins (p=0.048; HR: 2.92 (95% CI: 1.01 - 8.44)) and recurrence (p=<0001; HR: 6,02 (95% CI: 2,87 - 12,6)) had an almost 3-fold and 6-fold increase of mortality, respectively. Table 5 resumes the univariate analysis of demographic, clinical and histopathological associations with RFS. Mean proximal margin distance was 41±26mm in RFS patients and 41±25mm in patients with recurrence or deceased. There was no statistical difference between proximal margin distance between groups (p=0.977). After division in 4 groups depending on proximal margin distance (same as previously explained), the variable remained without statistical significance (p=0.394). Median distal margin distance was 30±31mm in event-free patients and 30±47mm in patients with recurrence or deceased. There was no statistical difference between median distal margin distance in patients with or without recurrence (p=0.795). Age (p=0.015), positive margins (p=<0.001), pTNM stage (p=<0.001), tumour location (p=0.012), OMS classification (p=0.033), Lauren classification (p=0.026), LVI (p=0.040), PNI (p=0.004) and pN (p=<0.001) were associated with RFS.

Table 4: Cox regression predicting overall survival from recurrence, PNI, margin status, pTNM stage and Lauren classification.

* Significance value of the model; LVI: Lymphovascular invasion; PNI: Perineural infiltration; WHO: World Health Organization.

Table 5: Demographic, clinical and histopathological associations with RFS.

Cox regression analysis was used to study the weight of each variable when predicting RFS. The model including these variables had no statistical significance (p=1). After testing several models, the most significant model contained age, margin status and pTNM stage. Increasing age (p=0.018; HR: 1.032 (95% CI: 1.01-1.06)) was associated with more mortality. Positive margins (p=0.011; HR: 3.76 (95% CI: 1.36 - 10.41)) and advance stage (III-IV) (p=0.007; HR: 2.35 (95% CI: 1.26 - 4.35)) had an almost 4-fold and 2-fold decrease of RFS, respectively.

Discussion

The adequate resection margin has been a question of debate for years. A margin of at least 6cm was firstly believed to be necessary [4], but recent studies have showed that shorter margins can achieve similar prognostic results. The 2018 Japanese Guidelines recommend a proximal margin of 2, 3 or 5 cm depending on tumour location, growth pattern and depth of invasion [9]. NCCN Guidelines are less clear about the ideal margin, stating that the primary goal of surgery is complete resection with negative microscopic margins (R0) along with lymphadenectomy, with T4b tumours requiring en bloc resection of involved structures [1]. ESMO Guidelines recommend a proximal margin of ≥3cm for tumours with an expansive growth pattern (including intestinal histotypes) and ≥5cm for those with an infiltrative growth pattern (including poorly cohesive/diffuse histotypes); if this margin cannot be assured, frozen section is recommended [10]. These differences between recommendations show the heterogeneous results described in literature. Disparities between the association of positive margins and survival have been reported, with some studies claiming no association [11-13] while others report worst survival if positive margins. [5,14-18] Articles reporting that this relation depends on the tumour stage can also be found, with reports maintaining this association only in advanced gastric cancer (≥T3 or N+) [19-21] or the opposite, claiming that margin status only affects prognosis in early disease with negative nodes [22-26]. The results obtained in this study are concordant with some of the literature, with positive margins remaining an independent risk factor for recurrence, OS and RFS, while distal and proximal margin distance per se has no association with neither of those.

The impact of positive margins in OS remains even when we add recurrence to our model, maintaining its independent weight, decreasing the OS almost 3-fold and RFS almost 4-fold. As such, the effort to assure negative margins whenever possible must be made. Although our results fail to corroborate the association between margin distance and negative margins, some authors have proposed a secure margin distance. [3-5] Whenever possible, frozen section should also be used to confirm negative margins [27]. Although it has some limitations, such as not being routinely available and being time and resource consuming, it still represents the safest way to ensure negative margins. [5] Intraoperative frozen section also aids decision making during the surgical procedure. If margins are positive, resection may be extended, or the procedure altered from a subtotal to a total gastrectomy.

Although authors differ in opinions regarding the advantage of extending the resection during surgery [12], or even which patients benefit from re-excision [15,28,29], the evidence is piling towards avoiding positive margins.

The association between cardia tumours and recurrence found in this study has also been previously found, [30,31] with some authors describing a worse OS and RFS for upper third tumours [32,33]. Several limitations can be pointed to this study. It is a retrospective, single-center observational study, that considers a period of 10 years, during which changes in the clinical management of gastric cancer and in surgical technique occurred. The advances in perioperative treatment due to the MAGIC [34] and then FLOT [35] trials aren’t reflected by our population, as demonstrated by the low number of patients submitted to perioperative treatment. Our limited sample size and the low number of R1 patients can also compromise some of the conclusions of this study. The main advantage is the 5-year minimum follow-up, permitting a survival analysis of both OS and RFS. Still, more studies are needed with larger, multicentric samples, to validate the findings presented in this paper.

Conclusion

R0 resection is an important prognostic factor and should be achieved during gastric cancer surgery with curative intent, but it is difficult to predict this R0 resection based solely on the macroscopic resection margin distance.

References

1. Nicole McMillian N, Lenora Pluchino MA, Ajani JA, Chair V, Bentrem DJ, et al (2022) NCCN Guidelines - Gastric Cancer 20 :167-192.
2. Petrillo A, Smyth EC (2020) Multimodality treatment for localized gastric cancer: state of the art and new insights. Curr Opin Oncol 32 : 347-355.
3. Mocan L (2021) Surgical management of gastric cancer: A systematic review. Journal of Clinical Medicine. MDPI. 10: 2557.
4. Bozzetti F, Bonfanti G, Bufalino R, Menotti V, Persano S, et al (1982) Adequacy of Margins of Resection in Gastrectomy for Cancer. Ann Surg 685-690.
5. Bissolati M, Desio M, Rosa F, Rausei S, Marrelli D, et al (2017) Risk factor analysis for involvement of resection margins in gastric and esophagogastric junction cancer: an Italian multicenter study. Gastric Cancer. 20 : 70-82.
6. Kim JH, Park SS, Kim J, Boo YJ, Kim SJ, et al (2006) Surgical Outcomes for Gastric Cancer in the Upper Third of the Stomach. World J Surg 30 : 1870-1876.
7. Siewert JR, Stein HJ (1998) Classification of adenocarcinoma of the oesophagogastric junction. British Journal of Surgery 85 : 1457-1459.
8. Amin MB, Edge SB, Greene FL, Byrd DR, Brookland RK, et al (2018) AJCC Cancer Staging Manual [Internet]. Springer International Publishing.
9. Japanese Gastric Cancer Association. (2021) Japanese gastric cancer treatment guidelines 2018 (5th edition). Gastric Cancer. Springer Japan 24 : 1-21.
10. Lordick F, Carneiro F, Cascinu S, Fleitas T, Haustermans K, et al (2022) Gastric cancer: ESMO Clinical Practice Guideline for diagnosis, treatment and follow-up. Annals of Oncology [Internet] 33.
11. Postlewait LM, Maithel SK (2016) The importance of surgical margins in gastric cancer. Vol. 113, Journal of Surgical Oncology. John Wiley and Sons Inc 277-282.
12. Squires MH, Kooby DA, Pawlik TM, Weber SM, Poultsides G, et al (2014) Utility of the Proximal Margin Frozen Section for Resection of Gastric Adenocarcinoma: A 7-Institution Study of the US Gastric Cancer Collaborative. Ann Surg Oncol 21: 4202-4210.
13. Lee CM, Jee YS, Lee JH, Son SY, Ahn SH, et al (2014) Length of negative resection margin does not affect local recurrence and survival in the patients with gastric cancer. World J Gastroenterol 20 :10518-10524.
14. Woo JW, Ryu KW, Park JY, Eom BW, Kim MJ, et al (2014) Prognostic impact of microscopic tumor involved resection margin in advanced gastric cancer patients after gastric resection. World J Surg 38 : 439-446.
15. Nagata T, Ichikawa D, Komatsu S, Inoue K, Shiozaki A, et al (2011) Prognostic impact of microscopic positive margin in gastric cancer patients. J Surg Oncol 104 : 592-597.
16. Jiang Z, Liu C, Cai Z, Shen C, Yin Y, et al (2021) Impact of Surgical Margin Status on Survival in Gastric Cancer: A Systematic Review and Meta-Analysis. Cancer Control 28: 10732748211043665.
17. Wang J, Liu J, Zhang G, Kong D (2017) Individualized proximal margin correlates with outcomes in gastric cancers with radical gastrectomy. Tumor Biology 39: 1010428317711032.
18. Maspero M, Sposito C, Benedetti A, Virdis M, di Bartolomeo M, et al (2022) Impact of Surgical Margins on Overall Survival after Gastrectomy for Gastric Cancer: A Validation of Japanese Gastric Cancer Association Guidelines on a Western Series. Ann Surg Oncol 29 : 3096-3108.
19. Bickenbach KA, Gonen M, Strong V, Brennan MF, Coit DG (2013) Association of positive transection margins with gastric cancer survival and local recurrence. Ann Surg Oncol 20: 2663-2668.
20. Morgagni P, la Barba G, Colciago E, Vittimberga G, Ercolani G (2018) Resection line involvement after gastric cancer treatment: handle with care. Vol. 70, Updates in Surgery. Springer-Verlag Italia s.r.l 70: 213-223.
21. Nakamura K, Ueyama T, Yao T, Xuan X, Ambe K, et al (1992) Pathology and Prognosis of Gastric Carcinoma Findings in 10,000 Patients Who Underwent Primary Gastrectomy. Cancer 70 : 1030-1037.
22. Luo J, Jiang Y, Chen X, Chen Y, Lopsang Gurung J, et al (2020) Prognostic value and nomograms of proximal margin distance in gastric cancer with radical distal gastrectomy. Chinese Journal of Cancer Research 32: 186-196
23. Byoung CC, Hei CJ, Hye JC, Sun YR, Woo JH, et al. (2007) Prognostic impact of resection margin involvement after extended (D2/D3) gastrectomy for advanced gastric cancer: A 15-year experience at a single Institute. J Surg Oncol 95 : 461-468.
24. Sun Z, Li D ming, Wang Z ning, Huang B jun, Xu Y, et al. (2009) Prognostic Significance of Microscopic Positive Margins for Gastric Cancer Patients with Potentially Curative Resection. Ann Surg Oncol 16 : 3028-3037.
25. Aurello 2014.
26. Kim S, Karpeh M, Klimstra D, Leung D, Brennan M (1999) Effect of microscopic resection line disease on gastric cancer survival. Journal of Gastrointestinal Surgery 33:24-33.
27. Kumazu Y, Hayashi T, Yoshikawa T, Yamada T, Hara K, et al (2020) Risk factors analysis and stratification for microscopically positive resection margin in gastric cancer patients. BMC Surg 20: 95.
28. Cascinu S, Giordani P, Catalano V, Agostinelli R, Catalano G (1999) Resection-line Involvement in Gastric Cancer Patients Undergoing Curative Resections: Implications for Clinical Management [Internet] 29 : 291-293.
29. Kim A, Kim BS, Yook JH, Kim BS (2020) Optimal proximal resection margin distance for gastrectomy in advanced gastric cancer. World J Gastroenterol 26 : 2232-2246.
30. Kim JH, Lee HH, Seo HS, Jung YJ, Park CH (2018) Borrmann Type 1 Cancer is Associated with a High Recurrence Rate in Locally Advanced Gastric Cancer. Ann Surg Oncol 25 : 2044-2052.
31. Kunisaki C, Shimada H, Ono H, Otsuka Y, Matsuda G, et al (2006) Comparison of Results of Surgery in the Upper Third and More Distal Stomach. Journal of Gastrointestinal Surgery 10 : 718-726.
32. Ichikawa D (2015) Clinicopathological characteristics of clinical early gastric cancer in the upper-third stomach. World J Gastroenterol 21: 12851-12856.
33. Matsuda T, Ohashi M, Tsujiura M, Ida S, Kumagai K, et al (2020) Shorter Survival of Patients with Upper-Third Gastric Cancer Preoperatively Diagnosed as Stage IA Compared with Those with Middle to Lower Lesions. Ann Surg Oncol 27 : 276-283.
34. Cunningham D, Allum WH, Stenning SP, Thompson JN, van de Velde CJ, et al (2006) Perioperative Chemotherapy versus Surgery Alone for Resectable Gastroesophageal Cancer From the Departments of Medicine 355: 11-20.
35. Sah BK, Zhang B, Zhang H, Li J, Yuan F, Ma T, et al  (2020) Neoadjuvant FLOT versus SOX phase II randomized clinical trial for patients with locally advanced gastric cancer. Nat Commun 11: 6093.