R1 Resection in Gastric Cancer: Marker of Tumor Biology Rather Than Surgical Failure

Introduction

Gastric cancer remains a major global health challenge and one of the leading causes of cancer-related deaths worldwide. Despite advances in multimodal therapy, long-term survival remains unsatisfactory for many patients [1,2]. Landmark randomized trials – including Medical Research Council Adjuvant Gastric Infusional Chemotherapy (MAGIC), Capecitabine Plus Oxaliplatin for Gastric Cancer after D2 Gastrectomy (CLASSIC), and fluorouracil, leucovorin, oxaliplatin, and docetaxel (FLOT-4) – have demonstrated that perioperative chemotherapy significantly improves outcomes compared with surgery alone, and these regimens are now incorporated into international guidelines as the standard of care [3–5].

Nevertheless, surgical margin status continues to be a critical determinant of outcomes, and one of the most challenging scenarios is an R1 resection. Although uncommon, R1 resections are associated with poor prognosis and are frequently linked to diffuse histology, advanced invasion, and systemic dissemination [1–4]. In a large series from MD Anderson Cancer Center, the 5-year overall survival for R1 patients was only 6%, with peritoneal dissemination as the predominant recurrence pattern [6]. These findings highlight an unresolved question: does R1 status reflect surgical inadequacy, or is it primarily a marker of aggressive tumor biology [7–9]?

Most previous studies have evaluated R1 patients within mixed R0/R1 cohorts, which limited interpretation because the overwhelming predominance of R0 cases obscured outcomes specific to margin-positive patients [9–11]. Moreover, key prognostic variables such as lymph node ratio, Lauren subtype, and signet ring histology have rarely been analyzed exclusively in R1 cohorts [10–15]. Consequently, the determinants of survival in this high-risk subgroup remain poorly defined.

In this retrospective single-center study, we analyzed clinicopathologic prognostic factors in patients with R1-resected gastric adenocarcinoma. Leveraging long-term follow-up and multivariate modeling, we sought to determine whether survival is shaped primarily by surgical limitations or by aggressive tumor biology. Our findings aim to refine postoperative risk stratification and underscore the importance of tailored adjuvant strategies in this underrecognized high-risk subgroup.

Material and Methods

STUDY DESIGN AND STATISTICAL ANALYSIS:

This retrospective single-center cohort included 28 consecutive patients who underwent R1 resection for histologically confirmed gastric adenocarcinoma between January 2010 and June 2023 at a tertiary referral hospital. Patients with R0 resection (tumor-free margins), R2 resection (macroscopic residual disease), or radiologic/intraoperative evidence of distant metastasis were excluded. Additional criteria included complete clinicopathologic data and a minimum follow-up of 6 months. R1 status was strictly defined as the presence of tumor at the inked margin on permanent paraffin sections (0 mm clearance). Margins <1–2 mm without tumor on ink were classified as R0.

All operations were performed with curative intent. Frozen section analysis was routinely attempted but proved unreliable in cases with signet ring histology, leading to margin underestimation in 10 patients. In these cases, the finding did not alter the extent of surgery but was considered in multidisciplinary discussions when tailoring adjuvant treatment. Six patients required emergency surgery for perforation, obstruction, or bleeding, which precluded optimal preoperative planning. In other cases, resection type (total vs subtotal gastrectomy or esophagogastrectomy) was determined by surgical judgment and patient factors.

HER2 status was assessed by immunohistochemistry (IHC), with in situ hybridization (ISH) for equivocal (2+) cases; positivity was defined as IHC 3+ or ISH-amplified. The Ki-67 index was measured on tumor hot spots, and the presence of a signet ring component was recorded regardless of percentage. Clinicopathologic variables were retrieved from electronic medical records and validated against operative and pathology reports. Lymph node ratio was calculated as metastatic lymph nodes/examined lymph nodes (MLN/ELN).

The primary endpoint was overall survival, defined as time from surgery to death or last follow-up. Secondary endpoints included recurrence patterns, time to recurrence (TTR), and estimated overall survival rates at 1, 3, and 5 years. Follow-up consisted of clinical evaluation every 3–6 months for 2 years, then every 6–12 months, with computed tomography (CT) imaging annually or when clinically indicated. Recurrence was confirmed radiologically or histologically. TTR was assessed only in patients who relapsed (n=14). Survival data were cross-validated with the national death registry.

Continuous variables were summarized as medians (range) and categorical data as frequencies (%). Univariate Cox regression identified variables associated with overall survival. Variables with P<0.10 in univariate analysis or deemed clinically relevant were considered for multivariable modeling. To comply with the events-per-variable principle, no more than 3 covariates were entered simultaneously. Because metastatic lymph node count and lymph node ratio both reflect nodal burden and are strongly correlated, only one of them was included in a given model to minimize the risk of collinearity. The proportional hazards assumption was verified with Schoenfeld residuals. Continuous variables (eg, Ki-67, MLN count) were kept continuous to preserve statistical power. All variables were complete for all 28 patients; therefore, no imputation was required. Kaplan-Meier estimates were used for overall survival and survival rates at 12, 36, and 60 months. Given the small cohort size, subgroup log-rank comparisons were avoided to minimize the risk of type II error; emphasis was placed on hazard ratios (HR) and 95% confidence intervals from Cox models. Analyses were performed with SPSS v25.0 (IBM Corp., Armonk, NY, USA), and results were cross-checked in Excel. Two-sided P<0.05 was considered significant.

ETHICAL APPROVAL:

This study was approved by the Ethics Committee of Van Training and Research Hospital (GOKAEK/2025-05-14; July 4, 2025). The requirement for informed consent was waived because of the retrospective design. All procedures complied with the Declaration of Helsinki (1964) and its later amendments.

USE OF ARTIFICIAL INTELLIGENCE TOOLS:

AI-based assistance (ChatGPT-4o, OpenAI, May 2024 version) was used solely for language refinement and style improvement. All scientific content, data interpretation, and statistical analyses were generated and verified by the authors. The final version was critically reviewed and approved by all authors. The use of AI was fully compliant with journal transparency and integrity standards.

Results

PATIENT AND TUMOR CHARACTERISTICS:

A total of 28 patients underwent R1 resection for gastric adenocarcinoma. The median age was 63 years (range 44–86 years), and 60.7% were male. Diffuse-type histology was the predominant subtype (present in 60.7% of patients), and a signet ring component was observed in 35.7% of cases. Most tumors were moderately or poorly differentiated, with a median Ki-67 index of 32.5% (Table 1).

PATHOLOGIC AND TREATMENT FINDINGS:

At diagnosis, 67.9% had clinical stage III disease and 71.4% had nodal involvement. Twelve patients (42.9%) received neoadjuvant chemotherapy, most commonly with FLOT or FOLFOX; 6 achieved partial response and 6 had stable disease. Emergency surgery was required in 6 patients due to perforation, bleeding, or obstruction, while intraoperative frozen section analysis underestimated margin status in 10 patients harboring signet ring histology. Surgical approach and resection extent varied according to intraoperative findings and the surgeon’s judgment (Table 2).

RECURRENCE PATTERNS AND SURVIVAL OUTCOMES:

After a median follow-up of 82.0 months (range 23.0–157.5 months), recurrence occurred in 14 patients (50.0%) – all of whom ultimately died. Among the 14 patients without recurrence, 5 deaths occurred from non-cancer-related causes (mainly cardiovascular events and comorbidity-related complications), while 9 patients remained alive at the last follow-up.

The most frequent recurrence site was peritoneal dissemination (9 patients; 32.1%), followed by distant nodal metastases (5 patients; 17.9%), liver metastases (3 patients; 10.7%), lung metastases (2 patients; 7.1%), and bone metastases (2 patients; 7.1%). Multiple patients developed multi-organ relapse, and ovarian involvement was noted in 2 women (7.1%).

The estimated overall survival rates at 1, 3, and 5 years were 89.3%, 39.3%, and 21.4%, respectively. Median TTR was 11.8 months (range 4.2–23.7 months) (Figures 1, 2).

PROGNOSTIC ANALYSES:

In accordance with the events-per-variable principle, no more than 3 covariates were included in the multivariate model. Based on the strongest univariate associations and clinical plausibility, pathological stage, signet ring histology, and lymph node ratio were retained in the final model. Because metastatic lymph node count and lymph node ratio represent overlapping measures of nodal burden and were highly correlated, only one of them (lymph node ratio) was included. In this analysis, both pathological stage (HR 13.28, 95% CI 2.76–63.89, P=0.001) and signet ring histology (HR 6.00, 95% CI 1.32–27.22, P=0.020) remained independent predictors of poor overall survival, whereas lymph node ratio demonstrated borderline significance (HR 35.77, 95% CI 0.98–1308.06, P=0.051) (Table 3).

These findings emphasize that poor outcomes after R1 resection are primarily driven by aggressive tumor biology rather than surgical technique.

Discussion

Microscopically positive margins (R1) after gastrectomy remain one of the most unfavorable scenarios in gastric cancer. While traditionally attributed to surgical shortcomings, accumulating evidence indicates that R1 status is more often a marker of biologically aggressive disease. In our cohort, the 5-year overall survival was only 21.4%, and recurrence occurred in 50% of patients – all of whom subsequently died during follow-up. Our results are consistent with prior series, including those by Wang et al (5-year overall survival: 13.4%) and Liang et al (5-year overall survival: 13.3%), as well as the meta-analysis by Jiang et al, which confirmed R1 resection as an independent predictor of poor prognosis (HR 1.67) [16–18].

Aggressive histologic subtypes appear to drive this biology. In our series, 35.7% of patients harbored a signet ring component, which emerged as an independent adverse factor for survival. Prior studies have similarly linked R1 resections to diffuse-type, poorly differentiated, or linitis plastica tumors, often accompanied by advanced T/N stage, vascular invasion, and large tumor size [6,19,20]. These findings reinforce the concept that margin positivity reflects infiltrative growth patterns rather than technical failure.

Lymph node burden has also been recognized as a key prognostic determinant. Although examined lymph node (ELN) count was not independently significant in our cohort, lymph node ratio demonstrated a borderline effect (P=0.051). This trend parallels studies by Liu, Shannon, and Erstad, which consistently showed lymph node ratio to outperform both ELN and absolute metastatic node count in prognostic discrimination [21–23]. Our median ELN retrieval of 30 aligns with recommendations advocating at least 25–30 nodes for accurate staging and potential survival benefit [24–26].

Recurrence patterns in our series further support the systemic nature of disease at the time of R1 resection. Peritoneal dissemination predominated, followed by distant nodal, hepatic, pulmonary, and osseous metastases; ovarian involvement occurred in 2 women. Notably, all relapses were fatal, with a median time to recurrence of only 11.8 months. Comparable studies by Wang, Hirata, and Gaspar-Figueiredo have also highlighted early, multi-organ progression after R1 resection [6,16,27].

The role of adjuvant intensification in this setting remains controversial. Some series have suggested improved outcomes with chemoradiotherapy in selected subgroups, particularly those with limited nodal burden [28–30]. Others, however, have reported no consistent survival benefit, underscoring the biological disadvantage conferred by R1 status, which conventional regimens may not overcome. Similarly, while re-resection is theoretically recommended when feasible, in practice it is often precluded by emergency presentations, advanced invasion, or prohibitive morbidity. In our cohort, 6 patients underwent urgent surgery; others were deemed unfit for extended procedures, exemplifying the real-world barriers to secondary resection.

Institutional expertise and surgical strategy also play an important role in margin outcomes. Van der Werf et al reported higher R1 rates in low-volume centers, while diffuse histology and advanced T stage were major predictors of margin positivity [31]. In addition, frozen section analysis – although routinely attempted – remains unreliable in diffuse or signet ring carcinomas, as reflected in our 10 cases where margins were underestimated, consistent with the observations of Guo and van der Werf [31,32].

Our study has limitations, including its small sample size, retrospective design, and treatment heterogeneity. Cancer-specific survival could not be reliably assessed, and evolving treatment protocols over the 2010–2023 period may have introduced era effects. Although our findings suggest that adverse outcomes after R1 resection are primarily driven by tumor biology rather than surgical inadequacy, this interpretation is constrained by the single-cohort design. The possibility that technical or institutional factors may also contribute to prognosis cannot be fully excluded. Nevertheless, the long median follow-up of 82 months provides robust outcome estimates and strengthens the validity of our observations. Importantly, this cohort exclusively comprised R1-resected patients – an underrepresented subgroup in the literature – thereby enabling a focused analysis of prognostic determinants.

Conclusions

R1 resection in gastric cancer appears to reflect aggressive tumor biology rather than surgical inadequacy. Signet ring histology and advanced pathological stage were independent predictors of poor survival, while lymph node ratio demonstrated borderline prognostic significance. These findings suggest that margin positivity is primarily a marker of unfavorable tumor behavior, although surgical refinement remains important for minimizing the risk of positive margins. Overall, our results emphasize the need for refined postoperative risk stratification and support the development of biology-driven, risk-adapted adjuvant strategies, rather than attributing adverse outcomes solely to technical shortcomings.