Abstract
Aim
Microscopically positive (R1) margins are associated with poorer outcomes in patients with colorectal cancer. However, the impact of subdivisions of R1 margins, be they to the primary tumour (R1 tumour) or to lymph node metastases (R1LNM), on patterns of relapse is unknown.
Methods
Patients treated for stage III colorectal cancer from 01 January 2016 to 31 December 2019 in four specialist centres were identified from the Danish national cancer registry. Patients were stratified into three groups according to margin status (R0 vs. R1 tumour vs. R1LNM). The primary outcomes were local recurrence‐free survival (LRFS), distant metastases‐free survival (DMFS) and disease‐specific survival (DSS).
Results
A total of 1,164 patients were included, with R1 margins found in 237 (20.4%). Irrespective of tumour location, R1 tumour and R1LNM margins were independent prognostic factors for systemic relapse (R1 tumour HR 1.84, CI: 1.17–2.88, p = 0.008; R1LNM HR 1.59, CI: 1.12–2.27, p = 0.009) and disease‐related death (R1 tumour HR 2.08, CI: 1.12–3.85, p = 0.020; R1LNM HR 1.84, CI: 1.12–3.02, p = 0.016). Whereas R1 tumour margins were associated with poorer 3‐year LRFS in both colon and rectum cancer, R1LNM margins only reduced LRFS in patients with rectal cancer. Patterns of relapse differed between R1 subdivisions, with R1 tumour margins more likely to affect multiple anatomical sites, with a predilection for extra‐hepatic/pulmonary metastases.
Conclusion
Subdivisions of R1 margins have a distinct impact on the oncological outcomes and patterns of disease relapse in patients with stage III colorectal cancer.
Keywords: Colon cancer, Rectal cancer, R1 resection margins
What does this paper add to the literature?
Subdivisions of R1 margins have been recently shown to have a distinct impact on overall survival in patients with colorectal cancer. However, the impact of these subdivisions on cancer‐specific survival and patterns of relapse is unknown. Here we show that R1 margins to the primary tumour or to lymph node metastases are associated with poorer cancer‐specific survival and different patterns of relapse.
INTRODUCTION
Microscopically positive (R1) resection margins are a key prognostic factor in patients with rectal cancer, being associated with higher rates of local recurrence, distant relapse, and disease‐specific death [1, 2, 3, 4]. Although less extensively investigated, there is evidence to suggest that the same is true for patients with colon cancer [5, 6]. Despite this recognized association with a poor prognosis, relatively few studies have investigated the importance of subdivisions of R1 margins, be they to the primary tumour (R1 tumour) or to metastatic lymph nodes or tumour deposits (R1LNM), to their impact on both recurrence and survival [7, 8, 9].
Variation in the definition of microscopically positive resection margins in colorectal cancer has been noted in the literature, with some groups referring to positive or negative circumferential resection margins (CRM) and others using the R classification [10]. Using the R classification system, whereby R1 resections were defined as the presence of tumour cells at or within 1 mm of the resection margin, our group recently published a study of the Danish colorectal cancer registry that demonstrated significant differences between R1 subdivisions [11]. Not only did R1 tumour and R1LNM margins have distinct prognostic factors but they were also associated with distinct impacts on overall survival in patients with stage III disease. However, as the national registry does not contain data regarding disease recurrences, this previous study was unable to determine whether patterns of relapse and oncological outcomes differ between R1 subdivisions.
A greater understanding of the impact of the nature of R1 margins on patterns of disease relapse could not only identify patients in whom more intensive treatment or surveillance may be justified but may also allow more comprehensive counselling of patients after their initial treatment. The purpose of this study was to investigate the impact of R1 subdivisions on oncological outcomes and patterns of metastasis in patients with stage III colorectal cancer.
METHODS
This was a retrospective cohort study based on data from the Danish Colorectal Cancer Group (DCCG) database and conducted according to STROBE guidelines [12]. This national cancer registry includes at least 95% of all patients diagnosed with colorectal cancer in Denmark and has been recently validated, with an overall data accuracy of >95% [13]. Patients diagnosed with Union for International Cancer Control (UICC) stage III colorectal cancer between 01/01/2016 and 31/12/2019 were identified. The impact of R1 subdivisions on overall survival in this cohort has already been described [11]. However, as the DCCG database does not include data on disease recurrence, a regional study of patients treated within the Capital Region of Denmark (Region Hovedstaden) was performed, whereby these data could be retrieved from electronic patient records. This region comprises four specialist colorectal centres (Bispebjerg, Herlev, Nordsjælland and Hvidovre hospitals). Patients undergoing a potentially curative oncological resection at one of these centres during the study period were included in the study cohort. Patients undergoing palliative operations, stents or local excisions were excluded.
Clinicopathological variables were extracted directly from the DCCG database, whereas data regarding local recurrence, distant metastases and the cause of death were retrieved from electronic patient journals. Patients were stratified into three groups according to their resection margins. Margin status was defined by pathologists at each centre according to national guidelines, with inking of the nonperitonealised and radial margins prior to sectioning of the specimen. R1 resections were defined as the presence of viable cancer cells ≤1 mm from the resection margin. These resections were then further divided into those ≤1 mm from the primary tumour (R1 tumour) and those ≤1 mm from a metastatic lymph node or tumour deposit (R1LNM). Patients were excluded from further analyses if margin status was not assessed or the type of R1 margin was not specified. R0 resections were defined as the absence of viable cancer cells ≤1 mm from the resection margin.
The primary outcomes of this study were local recurrence‐free survival (LRFS), distant metastases‐free survival (DMFS) and disease‐specific survival (DSS). These oncological outcomes were defined as the time from surgery to the event or point of last follow‐up and were investigated separately for patients with colon or rectal cancers. Additional subgroup analyses were performed using the entire cohort to investigate the impact of R1LNM margins according to nodal stage. Patient who died within 90 days of operation were excluded from the survival analyses. Secondary outcomes included prognostic factors for local recurrence, distant metastases, or disease‐specific death; patterns of distant relapse; and the treatment of recurrent disease. Local recurrence was defined as recurrence at the site of the anastomosis, or in the pelvis in patients with rectal cancer. Distant recurrences were defined as recurrences at any other anatomical sites. Follow‐up was conducted according to national guidelines, with computed tomography scanning at 12 and 36 months after surgery. This study was approved by the Danish Patient Safety Authority (R‐20061154) and the Danish Data Protection Agency (P‐2020‐902).
Univariable analyses of the differences in clinicopathological variables according to margin status were performed using the Chi‐square test for categorical data and the Kruskal Wallis test for continuous data. All analyses were two‐sided and considered statistically significant with a p‐value <0.05. LRFS, DMFS and DSS were calculated using the Kaplan Meier method and compared using the log‐rank test. Univariable Cox regression analyses were performed to identify prognostic factors of these outcomes. The following factors were chosen a priori: resection margins (R0 vs. R1 tumour vs. R1LNM); tumour location (colon vs. rectum); T stage; N stage; lymph node yield (<12 vs. ≥12); tumour budding; venous invasion (defined as the presence of extramural or intramural invasion); lymphatic invasion; perineural invasion; tumour perforation; involvement of other organs; peritoneal involvement. Significant factors (p < 0.05) were then combined using backwards selection methods in multivariable analyses. The results of the Cox regression analyses are presented as hazard ratios (HR) with 95% confidence intervals (CI). All analyses were performed using SPSS version 27.0 (IBM).
RESULTS
A total of 1,174 patients were identified during the study period. Resection margin status was not available for 10 patients, leaving a cohort of 1,164 patients for further analyses. R1 margins were found in 237 patients (20.4%), with R1LNM margins accounting for the majority of these cases (163 patients, 14.0%). Of the patients with R1LNM margins, 22 (13.5%) had tumour present at the resection margin, whereas 141 (86.5%) had tumour ≤1 mm from the margin. The median follow‐up for the entire cohort was 33 months (interquartile range 22–47 months). A total of 32 (3.8%) patients with colon cancer died within 90 days of operation compared with seven (2.1%) patients with rectal cancer.
Oncological outcomes in patients with colon cancers
A total of 832 patients had colon cancers and the clinicopathological demographics of these patients are summarized in Table 1. As in our previous study [11], significant differences in several tumour characteristics were noted according to margin status. Patients with R1 tumour margins had almost exclusively T3 or T4 tumours, with the highest rates of tumour perforation or involvement of other organs, whereas patients with R1LNM margins were more likely to have right‐sided rather than left‐sided tumours. Both R1 subdivisions had higher N stage and higher rates of venous invasion, lymphatic invasion and perineural invasion when compared to patients with R0 margins. Different patterns of adjuvant therapy were also noted, with the highest rates seen in patients with R1LNM margins (86.0%) and the lowest rates seen in patients with R1 tumour margins (55.6%).
TABLE 1.
Clinicopathological demographics of patients with stage III colon cancer
| Margin status | p‐value | |||
|---|---|---|---|---|
| R0 | R1 tumour | R1LNM | ||
| Number of patients | 666 (80.0) | 51 (6.1) | 115 (13.8) | – |
| Male: female | 1.15 | 0.70 | 1.05 | 0.231 |
| Median age (IQR) | 71 (63–77) | 74 (71–81) | 72 (65–78) | 0.054 |
| Cancer location | ||||
| Right | 319 (47.9) | 29 (56.9) | 81 (70.4) | <0.001 |
| Left | 339 (50.9) | 22 (43.1) | 32 (27.8) | |
| T stage | ||||
| 1 | 39 (5.9) | 0 (0) | 3 (2.6) | <0.001 |
| 2 | 48 (7.2) | 1 (2.0) | 5 (4.3) | |
| 3 | 377 (56.6) | 11 (21.6) | 53 (46.1) | |
| 4 | 202 (30.3) | 39 (76.5) | 54 (47.0) | |
| N stage | ||||
| 1a | 239 (35.9) | 12 (23.5) | 15 (13.0) | <0.001 |
| 1b | 206 (30.9) | 9 (17.6) | 20 (17.4) | |
| 1c | 40 (6.0) | 5 (9.8) | 3 (2.6) | |
| 2a | 116 (17.4) | 14 (27.5) | 40 (34.8) | |
| 2b | 65 (9.8) | 11 (21.6) | 37 (32.2) | |
| Median lymph node yield (IQR) | 30 (22–41) | 29 (21–37) | 28 (22–39) | 0.894 |
| Tumour budding | 129 (19.4) | 13 (25.5) | 28 (24.3) | 0.104 |
| Venous invasion | 305 (45.8) | 29 (56.9) | 68 (59.1) | 0.018 |
| Lymphatic invasion | 156 (23.4) | 27 (52.9) | 42 (36.5) | 0.003 |
| Perineural invasion | 163 (24.5) | 19 (37.3) | 41 (35.7) | 0.009 |
| Tumour perforation | 16 (2.4) | 15 (29.4) | 2 (1.7) | <0.001 |
| Involvement of other organs | 41 (6.2) | 22 (43.1) | 15 (13.0) | <0.001 |
| Peritoneal involvement | 187 (28.1) | 29 (56.9) | 50 (43.5) | <0.001 |
| Adjuvant therapy | ||||
| Eligible | 535 (80.3) | 36 (70.6) | 86 (74.8) | 0.128 |
| Received | 391 (73.1) | 20 (55.6) | 74 (86.0) | 0.002 |
Note: Numbers in parentheses are percentages unless otherwise stated. Bold values are statistically significant (p < 0.05).
Abbreviation: IQR, interquartile range.
Oncological outcomes in patients with colon cancer were found to differ significantly according to margin status. R1 tumour margins were associated with significantly poorer 3‐year LRFS (76.2% [95% CI: 68.5–83.9]) when compared with R1LNM (97.9% [95% CI: 96.4–99.4]) or R0 margins (95.3% [95% CI: 94.4–96.2], p < 0.001) (Figure 1A). However, both R1 subdivisions were associated with significantly poorer 3‐year DMFS (R1tum – 58.3% [95% CI: 50.8–65.8]; R1LNM – 57.3% [95% CI: 52.0–62.6]; R0–78.9% [95% CI: 77.2–80.6], p < 0.001) when compared to patients with R0 margins (Figure 1B). Corresponding reductions in 3‐year DSS were also noted, with the worst outcomes seen in patients with R1 tumour margins (R1tum–64.8% [95% CI: 56.9–72.7]; R1LNM – 76.2% [95% CI: 71.7–80.7]; R0–88.6% [95% CI: 87.2–90.0], p < 0.001) (Figure 1C).
FIGURE 1.
Oncological outcomes of patients with stage III colon cancer stratified by margin status. (A) Local recurrence‐free survival, p < 0.001. (B) Distant metastases‐free survival, p < 0.001. (C) Disease specific survival, p < 0.001
Oncological outcomes in patients with rectal cancers
A total of 332 patients had rectal cancers and the clinicopathological demographics of these patients are summarized in Table 2. In this subgroup, patients with R1 tumour margins had exclusively T3 or T4 tumours and the highest proportions of low tumours, tumour perforation and involvement of other organs. No differences in the use of neoadjuvant or adjuvant therapies were noted according to margin status.
TABLE 2.
Clinicopathological demographics of patients with stage III rectal cancer
| Margin status | p‐value | |||
|---|---|---|---|---|
| R0 | R1 tumour | R1 LNM | ||
| Number of patients | 261 (78.6) | 23 (6.9) | 48 (14.5) | ‐ |
| Male: female | 1.61 | 0.77 | 1.18 | 0.172 |
| Median age (IQR) | 68 (60–74) | 68 (55–72) | 66 (60–75) | 0.776 |
| Tumour height | ||||
| Low | 38 (14.6) | 10 (43.5) | 9 (18.8) | 0.004 |
| Middle | 103 (39.5) | 10 (43.5) | 21 (43.8) | |
| High | 113 (43.3) | 3 (13.0) | 17 (35.4) | |
| Missing | 7 (2.7) | 0 (0) | 1 (2.1) | |
| T stage | ||||
| 1 | 17 (6.5) | 0 (0) | 5 (10.4) | 0.097 |
| 2 | 52 (19.9) | 0 (0) | 9 (18.8) | |
| 3 | 159 (60.9) | 17 (73.9) | 27 (56.3) | |
| 4 | 33 (12.6) | 6 (26.1) | 7 (14.6) | |
| N stage | ||||
| 1a | 95 (36.4) | 6 (26.1) | 14 (29.2) | 0.231 |
| 1b | 81 (31.0) | 4 (17.4) | 15 (31.3) | |
| 1c | 17 (6.5) | 3 (13.0) | 2 (4.2) | |
| 2a | 38 (14.6) | 7 (30.4) | 7 (14.6) | |
| 2b | 30 (11.5) | 3 (13.0) | 10 (20.8) | |
| Neoadjuvant therapy | ||||
| None | 41 (15.7) | 7 (30.4) | 9 (18.8) | 0.182 |
| RT | 2 (4.9) | 1 (14.3) | 2 (22.2) | 0.253 |
| CT | 35 (85.4) | 4 (57.1) | 6 (66.7) | |
| CRT | 4 (9.8) | 2 (28.6) | 1 (11.1) | |
| Median lymph node yield (IQR) | 32 (23–43) | 23 (17–35) | 31 (21–42) | 0.142 |
| Tumour budding | 53 (20.3) | 6 (26.1) | 7 (14.6) | 0.529 |
| Venous invasion | 147 (56.3) | 12 (52.2) | 27 (56.3) | 0.921 |
| Lymphatic invasion | 43 (16.5) | 4 (17.4) | 14 (29.2) | 0.174 |
| Perineural invasion | 70 (26.8) | 13 (56.5) | 12 (25.0) | 0.013 |
| Tumour perforation | 4 (1.5) | 5 (21.7) | 0 (0) | <0.001 |
| Involvement of adjacent organs | 6 (2.3) | 4 (17.4) | 1 (2.1) | <0.001 |
| Peritoneal involvement | 29 (11.1) | 1 (4.3) | 6 (12.5) | 0.559 |
| Adjuvant therapy | ||||
| Eligible | 176 (67.4) | 13 (56.5) | 32 (66.7) | 0.568 |
| Received | 143 (81.3) | 12 (92.3) | 23 (71.9) | 0.254 |
Note: Numbers in parentheses are percentages unless otherwise stated. Bold values are statistically significant (p < 0.05).
Abbreviation: IQR, interquartile range.
Oncological outcomes were again found to differ significantly according to margin status. Both R1 subdivisions were associated with poorer 3‐year LRFS when compared to patients with R0 margins, although the worst outcomes were seen in patients with R1 tumour margins (R1tum ‐ 68.9% [95% CI: 57.9–79.9]; R1LNM – 82.0% [95% CI: 75.3–88.7]; R0–93.4% [95% CI: 91.6–95.2], p < 0.001) (Figure 2A). Patients with R1 tumour or R1LNM margins had similar reductions in 3‐year DMFS (R1tum – 56.8% [95% CI: 45.8–67.8]; R1LNM – 52.9% [95% CI: 45.3–60.5]; R0–74.0% [95% CI: 71.1–76.9], p < 0.001) (Figure 2B). Whilst both R1 tumour and R1LNM margins were associated with reduced 3‐year DSS when compared to patients with R0 margins, this outcome did not differ between R1 subdivisions (R1tum – 75.2% [95% CI: 65.5–84.9]; R1LNM – 76.9% [95% CI: 69.9–83.9]; R0–88.5% [95% CI: 86.2–90.8], p < 0.001).
FIGURE 2.
Oncological outcomes of patients with stage III rectal cancer stratified by margin status. (A) Local recurrence‐free survival, p < 0.001. (B) Distant metastases‐free survival, p < 0.001. (C) Disease specific survival, p < 0.001
Impact of R1LNM margins according to N stage
In order to investigate whether the presence or absence of R1LNM margins provides additional prognostic information to nodal staging, subgroup analyses were performed comparing DMFS (Figure S1) and DSS (Figure S2) in patients with R0 or R1LNM margins. The presence of R1LNM margins was associated with significantly poorer 3‐year DMFS when compared to R0 margins in both patients with N1 or N2 nodal disease (N1: R0–81.9% [95% CI: 80.3–83.5], R1LNM – 64.8% [95% CI: 58.1–71.5], p = 0.003; N2: R0–65.5% [95% CI: 62.3–68.7], R1LNM – 49.5% [95% CI: 43.8–55.2], p = 0.017). R1LNM margins were also associated with poorer 3‐year DSS in patients with N1 nodal disease (R0–92.8% [95% CI: 90.8–94.0], R1LNM – 83.8% [95% CI: 78.6–89.0], p = 0.001). No statistically significant difference in DSS was noted in patients with N2 nodal disease (R0–77.7% [95% CI: 74.8–80.6], R1LNM–71.7% [95% CI: 66.5–76.9], p = 0.165), although these curves appear to be diverging at 3 years follow‐up.
Prognostic factors for oncological outcomes
Univariable and multivariable analyses were performed to identify independently prognostic factors for local recurrence (Table 3), distant metastases (Table 4), and disease‐specific death (Table 5). R1 tumour margins (HR 4.06 [1.89–8.72], p < 0.001) were found to be independently associated with local recurrence, alongside rectum cancers, N stage, venous invasion, lymphatic invasion, tumour perforation and peritoneal involvement. No significant independent association between R1LNM margins and local recurrence was identified. However, both R1 tumour (HR 1.84 [1.17–2.88], p = 0.008) and R1LNM margins (HR 1.59 [1.12–2.27], p = 0.009) were found to be independently associated with distant metastases, alongside N stage, perineural invasion and peritoneal involvement. Both R1 subdivisions were also independently associated with disease‐specific death (R1 tumour HR 2.08 [1.12–3.85], p = 0.020; R1LNM HR 1.84 [1.12–3.02], p = 0.016), alongside T stage, N stage and lymphatic invasion.
TABLE 3.
Univariable and multivariable analyses of prognostic factors for local recurrence
| Univariable | Multivariable | |||
|---|---|---|---|---|
| HR (95% CI) | p‐value | HR (95% CI) | p‐value | |
| Resection margins | ||||
| R0 | Reference | – | Reference | – |
| R1 tumour | 5.17 (2.88–9.30) | <0.001 | 4.06 (1.89–8.72) | < 0.001 |
| R1LNM | 1.33 (0.65–2.73) | 0.431 | 0.90 (0.37–2.00) | 0.822 |
| Cancer type | ||||
| Colon | Reference | – | Reference | – |
| Rectum | 1.87 (1.16–3.02) | 0.013 | 3.11 (1.68–5.74) | < 0.001 |
| T stage | ||||
| 1 | 0.23 (0.05–0.94) | 0.041 | * | * |
| 2 | 0.13 (0.03–0.54) | 0.005 | * | * |
| 3 | 0.33 (0.20–0.54) | <0.001 | * | * |
| 4 | Reference | – | Reference | – |
| N stage | ||||
| 1a | Reference | – | Reference | – |
| 1b | 1.30 (0.63–2.66) | 0.476 | 1.49 (0.62–3.56) | 0.375 |
| 1c | 0.83 (0.19–3.65) | 0.804 | 0.96 (0.21–4.50) | 0.961 |
| 2a | 3.05 (1.58–5.90) | <0.001 | 2.84 (1.24–6.54) | 0.014 |
| 2b | 2.36 (1.09–5.11) | 0.029 | 1.06 (0.39–2.89) | 0.917 |
| Lymph node yield <12 | 3.64 (1.15–11.61) | 0.068 | – | – |
| Tumour budding | 0.98 (0.54–1.80) | 0.957 | – | – |
| Venous invasion | 2.31 (1.38–3.86) | 0.001 | 2.43 (1.24–4.78) | 0.010 |
| Lymphatic invasion | 2.11 (1.24–3.59) | 0.006 | 1.89 (1.01–3.52) | 0.045 |
| Perineural invasion | 2.45 (1.52–3.94) | <0.001 | * | * |
| Tumour perforation | 5.80 (2.86–11.75) | <0.001 | 3.15 (1.31–7.59) | 0.030 |
| Involvement of other organs | 2.73 (1.43–5.21) | 0.002 | * | * |
| Peritoneal involvement | 3.20 (1.99–5.15) | <0.001 | 2.06 (1.08–3.93) | 0.027 |
Note: Bold values are statistically significant (p < 0.05).
Not significant on multivariable analysis.
TABLE 4.
Univariable and multivariable analyses of prognostic factors for distant metastases
| Univariable | Multivariable | |||
|---|---|---|---|---|
| HR (95% CI) | p‐value | HR (95% CI) | p‐value | |
| Resection margins | ||||
| R0 | Reference | – | Reference | – |
| R1 tumour | 2.09 (1.40–3.10) | <0.001 | 1.84 (1.17–2.88) | 0.008 |
| R1LNM | 2.15 (1.61–2.85) | <0.001 | 1.59 (1.12–2.27) | 0.009 |
| Cancer type | ||||
| Colon | Reference | – | Reference | – |
| Rectum | 1.23 (0.96–1.57) | 0.099 | – | – |
| T stage | ||||
| 1 | 0.25 (0.12–0.51) | <0.001 | * | * |
| 2 | 0.30 (0.18–0.50) | <0.001 | * | * |
| 3 | 0.51 (0.40–0.64) | <0.001 | * | * |
| 4 | Reference | – | Reference | – |
| N stage | ||||
| 1a | Reference | – | Reference | – |
| 1b | 1.25 (0.89–1.76) | 0.190 | 1.15 (0.77–1.73) | 0.495 |
| 1c | 1.58 (0.92–2.69) | 0.095 | 1.54 (0.88–2.71) | 0.130 |
| 2a | 2.00 (1.43–2.81) | <0.001 | 1.78 (1.18–2.69) | 0.006 |
| 2b | 3.43 (2.45–4.80) | <0.001 | 2.02 (1.31–3.11) | 0.001 |
| Lymph node yield < 12 | 2.71 (1.40–5.27) | 0.003 | 2.98 (1.44–6.16) | 0.003 |
| Tumour budding | 1.04 (0.78–1.39) | 0.772 | – | – |
| Venous invasion | 1.81 (1.43–2.30) | <0.001 | * | * |
| Lymphatic invasion | 1.67 (1.28–2.18) | <0.001 | * | * |
| Perineural invasion | 2.22 (1.76–2.81) | <0.001 | 1.64 (1.22–2.20) | <0.001 |
| Tumour perforation | 2.53 (1.59–4.03) | <0.001 | * | * |
| Involvement of other organs | 1.33 (0.89–1.99) | 0.164 | – | – |
| Peritoneal involvement | 2.32 (1.84–2.92) | <0.001 | 1.48 (1.10–1.98) | 0.009 |
Note: Bold values are statistically significant (p < 0.05).
Not significant on multivariable analysis.
TABLE 5.
Univariable and multivariable analyses of prognostic factors for disease‐specific death
| Univariable | Multivariable | |||
|---|---|---|---|---|
| HR (95% CI) | p‐value | HR (95% CI) | p‐value | |
| Resection margins | ||||
| R0 | Reference | – | Reference | – |
| R1 tumour | 3.16 (1.97–5.07) | <0.001 | 2.08 (1.12–3.85) | 0.020 |
| R1LNM | 2.48 (1.68–3.65) | <0.001 | 1.84 (1.12–3.02) | 0.016 |
| Cancer type | ||||
| Colon | Reference | – | Reference | – |
| Rectum | 1.01 (0.71–1.44) | 0.947 | – | – |
| T stage | ||||
| 1 | 0.10 (0.02–0.39) | 0.001 | 0.43 (0.10–1.81) | 0.247 |
| 2 | 0.12 (0.04–0.33) | <0.001 | 0.21 (0.05–0.90) | 0.035 |
| 3 | 0.36 (0.26–0.50) | <0.001 | 0.50 (0.32–0.78) | 0.002 |
| 4 | Reference | – | Reference | – |
| N stage | ||||
| 1a | Reference | – | Reference | – |
| 1b | 1.36 (0.81–2.26) | 0.244 | 1.56 (0.78–3.10) | 0.209 |
| 1c | 1.14 (0.44–2.97) | 0.783 | 1.39 (0.50–3.90) | 0.529 |
| 2a | 2.63 (1.62–4.28) | <0.001 | 2.36 (1.21–4.59) | 0.012 |
| 2b | 4.95 (3.08–7.94) | <0.001 | 2.67 (1.35–5.26) | 0.005 |
| Lymph node yield < 12 | 1.65 (0.53–5.17) | 0.003 | * | * |
| Tumour budding | 0.90 (0.60–1.35) | 0.615 | – | – |
| Venous invasion | 1.99 (1.42–2.78) | <0.001 | * | * |
| Lymphatic invasion | 2.65 (1.80–3.89) | <0.001 | 1.76 (1.13–2.74) | 0.012 |
| Perineural invasion | 2.26 (1.64–3.12) | <0.001 | * | * |
| Tumour perforation | 4.07 (2.38–6.95) | <0.001 | * | * |
| Involvement of other organs | 1.76 (1.08–2.89) | 0.024 | * | * |
| Peritoneal involvement | 3.39 (2.46–4.66) | <0.001 | * | * |
Note: Bold values are statistically significant (p < 0.05).
Not significant on multivariable analysis.
Patterns of disease relapse
At the time of writing, 318 patients had developed recurrent disease, with the patterns of relapse shown in Figure 3 and summarized in Table S1. A greater proportion of patients with R1 tumour margins developed isolated local recurrences in the absence of distant metastases, whilst patients with R1LNM had the greatest proportion of isolated distant metastases in the absence of local recurrence. Patients with R1LNM margins were less likely to present with a single metastatic lesion when compared to those with R1 tumour or R0 margins, whereas patients with R1 tumour margins were more likely to present with metastases affecting multiple anatomical sites. Of the patients with metastases affecting a single anatomical site, those with R1 tumour margins had the lowest proportion of lung metastases (6.3%) and the highest proportion of metastases at other anatomical sites (37.5%), although this was not statistically significant. In keeping with this evidence of increased burden of disease, the proportion of patients undergoing potentially curative treatment of systemic metastases was lower in patients with R1 margins, although this was not statistically significant. Patients with either R1 tumour or R1 LNM margins had a shorter median survival after the diagnosis of their first metastasis when compared to patients with R0 margins, again this was not statistically significant (17.0 vs. 18.0 vs. 23.0 months, p = 0.139) (Figure S3).
FIGURE 3.
Patterns of relapse in patients with stage III colorectal cancer. (A) The distribution of local recurrences (LR) and distant metastases (DM) in patients developing recurrent disease. (B) The percentage of patients presenting with single or multiple metastatic lesions (C) The percentage of patients presenting with metastases effecting single or multiple anatomical sites
DISCUSSION
This study builds on the results of our previous publication and demonstrates that not only do R1 subdivisions have a negative effect on overall survival in patients with stage III colorectal cancer but also that they have a distinct impact on oncological outcomes and patterns of relapse [11]. Furthermore, the impact of R1 resection on oncological outcomes appears to differ according to both the nature of R1 resection and tumour location. Although both R1 subdivisions were associated with increased risks of distant relapse compared with R0 resection regardless of tumour location, R1 tumour margins were associated with a greater risk of disease‐specific death when compared to R1LNM margins but only in patients with colon cancers. With regards to local recurrence, whereas R1 tumour margins were associated with increased risk regardless of tumour location, R1LNM margins were only associated with an increased risk in patients with rectal cancers.
The extent to which R1 resection margins represent a technical failure at the time of initial surgery as opposed to being a surrogate for more aggressive disease biology remains a matter of debate. The current study provides several lines of evidence to support the latter theory. First, although patients with R1 tumour margins had higher rates of isolated local recurrences, the most common mode of failure in both R1 subdivisions was isolated systemic relapse. This is highly suggestive of the presence of subclinical micrometastatic disease at the time of surgery and, in the absence of concomitant local recurrence, is unlikely to be due to residual disease left within the operative field. Second, R1 tumour and R1LNM margins were associated with several tumour characteristics suggestive of more aggressive disease including higher T and N stage, lymphatic invasion, perineural invasion and venous invasion. Third and finally, a nonsignificant trend towards shorter survival after the diagnosis of first metastasis was found in both R1 subdivisions, which is again suggestive of more aggressive cancer biology.
This study also provides evidence that while both may be surrogates for more aggressive disease, R1 subdivisions may represent distinct biological entities. Significant differences were observed not only in several pathological characteristics between these subgroups but also in their patterns of disease relapse. Metastatic disease following R1 tumour margins was more likely to affect multiple organs and, although not statistically significant, showed a predilection for extra‐hepatic and extra‐pulmonary metastases. In contrast, whilst metastatic disease from R1LNM margins was more likely to affect a single anatomical site, predominantly in the form of hepatic or pulmonary metastases, this group was less likely to present with a single metastatic lesion. The mechanisms responsible for these differences are unknown and further studies are underway to determine if a biological basis for these behavioural differences can be identified.
The findings of this study should have a significant impact on the clinical management of patients with stage III colorectal cancers. The first should be the recognition of R1LNM margins as a poor prognostic factor in these patients. Whilst the association of R1 tumour margins is well recognised, the same cannot be said of R1LNM margins [14]. Our data demonstrate that R1LNM margins have an almost identical impact on DMFS as R1 tumour margins in patients with colon or rectal cancers. In colon cancer, whilst the DSS of patients with R1LNM margins was not as poor as those with R1 tumour margins, it is still substantially reduced when compared to those with R0 resections. The second impact of this study should be on the selection of the type and duration of adjuvant therapy, particularly for patients with colon cancer. Current guidelines recommend that a shorter course of the CAPOX‐regime may be offered to patients who are judged to be at low risk of relapse according to their T and N stage [15, 16]. Given that both R1 tumour and R1LNM margins were found to be independently associated with systemic relapse and disease‐related death, a strong argument can be made for the consideration of these factors when determining the type and duration of adjuvant therapy. A further consideration is whether margin status should be considered as a basis for subgroup analyses in future clinical trials. Neither the original studies of doublet chemotherapy nor those included in the IDEA collaboration included data regarding resection margin status of patients [16, 17, 18, 19]. The results of the current study provide strong evidence that patients with colon cancers and R1 margins represent a high‐risk group that should be considered as distinct from those with R0 margins and warrant further investigation.
This study may also have implications for the surveillance of patients with stage III colorectal cancer. Given the increased risks of relapse in patients with R1 margins, an argument could be made for adopting a more intensive approach to surveillance in these patients. There is evidence that more intensive surveillance results in both an increase in the diagnosis of asymptomatic recurrences and the proportion of patients who are amenable to potentially curative re‐operation [20]. However, this was not accompanied by any improvements in cancer‐specific survival, indicating that intensive surveillance may only lead to higher rates of intervention without improving the final outcome. A further cause for caution in adopting a more intensive surveillance in patients with R1 margins is their reduced tendency to present with oligometastatic disease. Whether defined as the total number of metastatic lesions or as the number of anatomical sites affected, patients with R1 margins were found to be less likely to present with oligometastatic disease than patients with R0 resections, which was associated with a statistically nonsignificant reduction in the proportion of patients treated with curative intent [21]. Despite these concerns, novel surveillance techniques, such as the use of circulating tumour DNA, have the potential to stratify patients more precisely and dynamically according to their risk of relapse, which in turn may allow subsequent interventions to be targeted to those who stand to benefit most [22].
The authors recognize the limitations of this study. In order to determine the impact of R1 margins in the context of other histopathological factors, only patients treated from 2016 onwards were included, when comprehensive pathological data were first coupled to the DCCG database. Consequently, the follow‐up of the cohort is relatively short (33 months), and it would be of interest to investigate if the differences in oncological outcomes are maintained at 5 year follow‐up. An additional limitation is the size of some of the subgroups, particularly with regard to the patterns and treatment of metastatic disease. Although trends were identified, a larger cohort may have been better equipped to define differences in these outcomes between R1 subdivisions. Furthermore, the limited number of patients and events prohibited the investigation of site‐specific prognostic factors for oncological outcomes. Finally, regarding R1LNM margins, the DCCG database does not differentiate between positive margins to lymph node metastases and tumour deposits, and as such we were unable to determine if outcomes differ between these further subdivisions of R1LNM margins.
In conclusion, subdivisions of R1 resections are associated with different patterns of relapse and impacts on oncological outcomes in patients with stage III colorectal cancer. Further research to investigate if there is a biological basis for these differences is warranted.
Funding information
None to declare.
CONFLICT OF INTEREST
None to declare.
ETHICAL APPROVAL
This study was approved by the Danish Patient Safety Authority (R‐20061154) and the Danish Data Protection Agency (P‐2020‐902).
AUTHOR CONTRIBUTIONS
Study conception: Henry G. Smith, Daniel M. Skovgaards, Deepthi Chiranth, and Nis H. Schlesinger. Data analysis and interpretation: Henry G. Smith, Daniel M. Skovgaards, and Nis H. Schlesinger. Manuscript drafting: Henry G. Smith. Manuscript revision: Daniel M. Skovgaards, Deepthi Chiranth, and Nis H. Schlesinger. Final approval: Henry G. Smith, Daniel M. Skovgaards, Deepthi Chiranth, and Nis H. Schlesinger.
Supporting information
Figure S1
Figure S2
Figure S3
Table S1
Smith HG, Skovgaards DM, Chiranth D, Schlesinger NH. The impact of subdivisions of microscopically positive (R1) margins on patterns of relapse in stage III colorectal cancer – A retrospective cohort study. Colorectal Dis. 2022;24:828–837. 10.1111/codi.16121
DATA AVAILABILITY STATEMENT
The data is available from the authors upon reasonable request
REFERENCES
- 1. Bernstein TE, Endreseth BH, Romundstad P, Wibe A. Norwegian Colorectal Cancer G. Circumferential resection margin as a prognostic factor in rectal cancer. Br J Surg. 2009;96(11):1348–57. [DOI] [PubMed] [Google Scholar]
- 2. Nagtegaal ID, Marijnen CA, Kranenbarg EK, van de Velde CJ, van Krieken JH, Pathology Review C, et al. Circumferential margin involvement is still an important predictor of local recurrence in rectal carcinoma: not one millimeter but two millimeters is the limit. Am J Surg Pathol. 2002;26(3):350–7. [DOI] [PubMed] [Google Scholar]
- 3. Wibe A, Rendedal PR, Svensson E, Norstein J, Eide TJ, Myrvold HE, et al. Prognostic significance of the circumferential resection margin following total mesorectal excision for rectal cancer. Br J Surg. 2002;89(3):327–34. [DOI] [PubMed] [Google Scholar]
- 4. Birbeck KF, Macklin CP, Tiffin NJ, Parsons W, Dixon MF, Mapstone NP, et al. Rates of circumferential resection margin involvement vary between surgeons and predict outcomes in rectal cancer surgery. Ann Surg. 2002;235(4):449–57. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Bateman AC, Carr NJ, Warren BF. The retroperitoneal surface in distal caecal and proximal ascending colon carcinoma: the Cinderella surgical margin? J Clin Pathol. 2005;58(4):426–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Scott N, Jamali A, Verbeke C, Ambrose NS, Botterill ID, Jayne DG. Retroperitoneal margin involvement by adenocarcinoma of the caecum and ascending colon: what does it mean? Colorectal Dis. 2008;10(3):289–93. [DOI] [PubMed] [Google Scholar]
- 7. Khan MA, Hakeem AR, Scott N, Saunders RN. Significance of R1 resection margin in colon cancer resections in the modern era. Colorectal Dis. 2015;17(11):943–53. [DOI] [PubMed] [Google Scholar]
- 8. Patel A, Green N, Sarmah P, Langman G, Chandrakumaran K, Youssef H. The clinical significance of a pathologically positive lymph node at the circumferential resection margin in rectal cancer. Tech Coloproctol. 2019;23(2):151–9. [DOI] [PubMed] [Google Scholar]
- 9. Suarez J, Goicoetxea A, Gomez ML, Jimenez G, Llanos MC, Jimenez J, et al. Impact of specific modes of circumferential resection margin involvement in rectal cancer local recurrence: a retrospective study. J Surg Oncol. 2018;118(7):1122–8. [DOI] [PubMed] [Google Scholar]
- 10. Wittekind C, Compton C, Quirke P, Nagtegaal I, Merkel S, Hermanek P, et al. A uniform residual tumor (R) classification: integration of the R classification and the circumferential margin status. Cancer. 2009;115(15):3483–8. [DOI] [PubMed] [Google Scholar]
- 11. Smith HG, Chiranth D, Mortensen CE, Schlesinger NH. The significance of subdivisions of microscopically positive (R1) margins in colorectal cancer: a retrospective study of a national cancer registry. Colorectal Dis. 2021;24:197–209. [DOI] [PubMed] [Google Scholar]
- 12. von Elm E, Altman DG, Egger M, Pocock SJ, Gotzsche PC, Vandenbroucke JP, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. J Clin Epidemiol. 2008;61(4):344–9. [DOI] [PubMed] [Google Scholar]
- 13. Klein MF, Gogenur I, Ingeholm P, Njor SH, Iversen LH, Emmertsen KJ, et al. Validation of the Danish Colorectal Cancer Group (DCCG.dk) database ‐ On behalf of the Danish Colorectal Cancer Group. Colorectal Dis. 2020;22(12):2057–67. [DOI] [PubMed] [Google Scholar]
- 14. Detering R, Rutgers MLW, Bemelman WA, Hompes R, Tanis PJ. Prognostic importance of circumferential resection margin in the era of evolving surgical and multidisciplinary treatment of rectal cancer: a systematic review and meta‐analysis. Surgery. 2021;170(2):412–31. [DOI] [PubMed] [Google Scholar]
- 15. Argiles G, Tabernero J, Labianca R, Hochhauser D, Salazar R, Iveson T, et al. Localised colon cancer: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow‐up. Ann Oncol. 2020;31(10):1291–305. [DOI] [PubMed] [Google Scholar]
- 16. Grothey A, Sobrero AF, Shields AF, Yoshino T, Paul J, Taieb J, et al. Duration of adjuvant chemotherapy for stage III colon cancer. N Engl J Med. 2018;378(13):1177–88. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Andre T, Boni C, Navarro M, Tabernero J, Hickish T, Topham C, et al. Improved overall survival with oxaliplatin, fluorouracil, and leucovorin as adjuvant treatment in stage II or III colon cancer in the MOSAIC trial. J Clin Oncol. 2009;27(19):3109–16. [DOI] [PubMed] [Google Scholar]
- 18. Haller DG, Tabernero J, Maroun J, de Braud F, Price T, Van Cutsem E, et al. Capecitabine plus oxaliplatin compared with fluorouracil and folinic acid as adjuvant therapy for stage III colon cancer. J Clin Oncol. 2011;29(11):1465–71. [DOI] [PubMed] [Google Scholar]
- 19. Kuebler JP, Wieand HS, O'Connell MJ, Smith RE, Colangelo LH, Yothers G, et al. Oxaliplatin combined with weekly bolus fluorouracil and leucovorin as surgical adjuvant chemotherapy for stage II and III colon cancer: results from NSABP C‐07. J Clin Oncol. 2007;25(16):2198–204. [DOI] [PubMed] [Google Scholar]
- 20. Tjandra JJ, Chan MK. Follow‐up after curative resection of colorectal cancer: a meta‐analysis. Dis Colon Rectum. 2007;50(11):1783–99. [DOI] [PubMed] [Google Scholar]
- 21. Treasure T. Oligometastatic cancer: an entity, a useful concept, or a therapeutic opportunity? J R Soc Med. 2012;105(6):242–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Tie J, Cohen JD, Wang Y, Christie M, Simons K, Lee M, et al. Circulating tumor DNA analyses as markers of recurrence risk and benefit of adjuvant therapy for stage III colon cancer. JAMA Oncol. 2019;5(12):1710–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Figure S1
Figure S2
Figure S3
Table S1
Data Availability Statement
The data is available from the authors upon reasonable request