Survival after local excision for rectal cancer: a population-based overview of clinical practice and outcome

Introduction

Rectal cancer is diagnosed in approximately 4200 patients annually in the Netherlands. Due to colorectal cancer screening programs, the incidence and rate of patients with early rectal cancer will increase [1]. The cornerstone of curative treatment for rectal cancer is total mesorectal excision (TME) [2]. The high-quality oncological clearance of tumor and regional lymph nodes must be balanced against the risk of postoperative morbidity and mortality [3].

Local excision (LE) proved to be oncologically safe in selected T1 rectal cancer and is considered standard therapy and an attractive alternative for major surgery [4–7]. However, LE for ≥ pT2 cancer may lead to unacceptable high local recurrence rates, with a significant decrease in survival compared to TME [8,9].

Several studies addressed the increase in LE for rectal cancers, yet, few studies report long-term follow-up data specified on this approach [10–12]. This nationwide study investigated the number of patients who underwent LE only or primary TME. Moreover, overall survival (OS) and relative survival (RS) after LE only for all T-stages were compared to results of completion TME (cTME, e.g. TME within 6 months of prior LE) and primary TME surgery (without neoadjuvant or adjuvant therapy).

Materials and methods

Data were extracted from the Netherlands Cancer Registry (NCR), a nationwide population-based registry including all newly diagnosed malignancies. NCR data on patient characteristics, tumor characteristics and treatment are collected from hospital patient files and coded according to a national manual, e.g. to the International Classification of Diseases for Oncology (ICD-O) and stage according to the TNM classification [13,14].

Patient selection

We selected patients (≥18 years) with stage pT1 to pT4, cN0/x, cM0/x rectal and rectosigmoid adenocarcinoma (International Classification of Diseases for Oncology, third edition, C199, C209), diagnosed between 2005 and 2015. Patients were not treated with neoadjuvant or adjuvant therapy. LE included TEM, TAE, TEO and TAMIS procedures and cTME included low anterior resection (LAR), Hartmann and abdominoperineal excision (APE). The vital status of all patients was obtained by linking the NCR to the Municipal Personal Records Database. Follow-up was completed until 31 January 2017.

Statistical analysis

Continuous variables are presented as median ± range. Categorical variables are presented as frequency with percentages. Differences in patient and tumor characteristics between groups were tested using the χ² test.

RS was used as an estimation of disease-specific survival. It reflects the survival of cancer patients, adjusted for the survival of the general population with the same age and gender distributions. RS is calculated as the ratio of the observed rates for cancer patients to the expected rates for the general population using the Ederer method [15].

Five-year OS was calculated as the percentage of patients alive at 5 years after the date of (first) surgery.

Follow-up time was calculated from the date of diagnosis to death or alive up to the last date of follow-up. Values of p < .05 were considered statistically significant. Statistical Package for Social Sciences (SPSS, Chicago, IL, USA, version 25.0) was used to prepare the database and for analysis. RS was analyzed using STATA (version 14.2).

Results

According to the NCR, 46,877 patients were diagnosed with rectal and rectosigmoidal cancer. In 1090 patients, a LE was performed, of which 144 patients underwent a cTME. Thus, in 946 patients, LE was the only treatment for rectal cancer. A total of 5101 patients with cT1-4N0/x were identified in whom a primary TME was performed.

Patient and tumor characteristics for all three groups (LE only, cTME and primary TME) are given in Table 1. In the cTME group, the patients were significantly younger than in the LE only and primary TME group (p < .001). There is a marked increase in LE and primary TME over the years.

Table 1. Patient and tumor characteristics.

	LE only	LE followed by cTME	Primary TME	p value
N	946	144	5101
Male:female	570 (60.3%): 376 (39.9%)	81 (56.3%):63 (43.8%)	3073 (60.2%):2029 (39.8%)	NS
Age, median (range)	71 (35–95)	68 (30–89)	72 (20–99)	.001^a
Year incidence, n
2005	53	7	498
2006	53	10	542
2007	74	5	488
2008	67	9	389
2009	66	5	317
2010	64	6	253
2011	66	8	265
2012	95	9	261
2013	90	13	267
2014	147	26	833
2015	171	46	988
Pathological T-stage
pT1	753 (79.6%)	37 (25.7%)	770 (15.1%)
pT2	167 (17.7%)	78 (54.2%)	1832 (35.9%)
pT3	26 (2.7%)	28 (19.4%)	2315 (45.4%)
pT4	–	1 (0.7%)	184 (3.6%)
Pathological N-stage
pN0	61 (6.4%)	117 (81.3%)	3945 (77.3%)
pN1	4 (0.4%)	18 (12.5 %)	786 (15.5%)
pN2	–	1 (0.7%)	283 (5.5%)
pNx	881 (93.1%)	8 (5.6%)	87 (1.7%)
Residual tumor
R0	771 (81.1%)	141 (97.9%)	4812 (94.3%)
R1	57 (6.0%)	2 (1.4%)	123 (2.4%)
R2	6 (0.6%)	–	48 (0.9%)
Rx	112 (11.8%)	1 (0.7%)	119 (2.3%)
Follow-up in months, median (range)	43 (0–145)	33 (0–142)	35 (0–145)

Note: LE: local excision; cTME: completion total mesorectal excision; TME: total mesorectal excision; IQR: interquartile range; NS: not significant.

^a Patients in the LE group and primary TME group significantly older than in cTME group.

In 790 patients following LE, a pT1 rectal carcinoma was diagnosed. In 37 patients, a cTME was performed. Of those, six had an explanatory reason (poorly differentiated tumor; unclear margin) but for the remaining 31 patients, the rationale for cTME could not be obtained from the database.

In 300 patients following LE, a pT2–4 was diagnosed. In 193 patients, a cTME was omitted. In this group, age was significantly higher compared to those patients in whom a cTME was performed (77 vs. 68 years, p < .001).

Five-year RS rates for pT1 rectal cancer were comparable for all procedures. For pT2 tumors, OS was significantly worse in the LE only group (60% vs. 80.1%, p < .001) in cTME and 75.4% (p < .001) in TME patients; however, RS was comparable for all procedures. RS was worse for pT3–4 patients (Table 2).

Table 2. Differences in 5-year relative survival between patients, who underwent LE only versus patients who underwent completion surgery (cTME) or primary TME, split by pathological T-stage.

	LE only		cTME		TME
	N	5y RS (95% CI)	N	5y RS (95% CI)	N	5y RS (95% CI)
pT1	753	93.4 (88.7–97.5)	37	85.1 (60.5–98.8)	770	95.0 (90.4–98.9)
pT2	167	88.2 (73.9–100.43)	78	89.5 (70.8–100.33)	1832	92.9 (89.8–95.8)
pT3–4	26	20.4 (3.7–51.2)	29	73.4 (32.3–96.2)	2499	74.9 (72.0–77.7)

Note: LE: local excision; cTME: completion total mesorectal excision; TME: (primary) total mesorectal excision; 5y RS: 5-year relative survival; CI: confidence interval.

Discussion

This retrospective nationwide data analysis shows that LE only for early staged (pT1–2) rectal cancer without neoadjuvant treatment is an acceptable and oncological safe treatment with 5-year RS rates comparable to TME surgery. Patients who underwent cTME had similar survival to both LE and TME only pT1–2 patients [16]. In contrast, RS after LE only in ≥pT3 invasive rectal cancer is worse compared to patients who underwent primary or cTME.

As in this study, a previous study from the US also showed an increase in LE as treatment for pT1 rectal cancer but also for more invasive tumors [10]. Especially in the elderly, an increase in LE was observed, which is probably due to the fact that surgical morbidity after standard TME surgery is not insignificant and a reason for exploring less extensive treatment methods [8,17].

Although OS was significantly worse in patients treated with LE only in case of pT2 rectal cancer compared to cTME, RS was similar to patients who underwent cTME or TME only. This is different from previous studies were only OS was usually presented [18–20]. RS has the advantage of correcting for the expected survival. Since patients who underwent a LE population were elderly, this might have contributed to the good RS [21]. Patients in whom a cTME was omitted were significantly older; it is very plausible that these patients were also more fragile. Increasing age, increased burden of comorbidities and perioperative surgical morbidity might also contribute to the decision to withdraw from cTME.

In high-risk pT1 rectal cancer (tumor >3 cm, poor differentiation, tumor budding, lymph and vascular invasion), local recurrences occur more often, which is a reason to perform cTME in pT1 tumors [5,22,23]. These histopathological criteria are crucial in adequate patient selection for rectum preserving options. In the present study, 37 patients with a pT1 tumor underwent cTME, possibly due to having a high-risk pT1 tumor. Importantly, RS in this presumed high-risk pT1 tumors treated with cTME was equal those with a primary TME or LE. Histological prognostic features for lymph node metastases were not reported in the NCR. It should, therefore, be emphasized that this study does not unambiguously demonstrate that pT2 rectal cancer can generally be treated safely with LE. However, in the current study, the decision not to perform cTME in pT1 and pT2 cancers did not lead to a decrease in RS.

In the present study, there was also a marked increase in the number of patients who underwent primary TME without neoadjuvant treatment. This is probably due to the improvement of pre-operative imaging using MRI and to adjustment of the guidelines in the Netherlands in which short course radiotherapy was no longer advised in clinically node-negative patients from 2012 [24,25]. The implementation of a nationwide screening program for colorectal cancer is another reason for the relatively steep increase in 2014 and 2015 [17,26].

Several authors reported worsened outcome following cTME after prior LE [27,28]. In this study, cTME did not lead to different RS compared to primary TME for all pT stages. It should be taken into account that cTME was only performed in 144 patients, which is a relatively small group.

Thriving to refrain from radical surgery, other treatment algorithms are the current subject of studies. In the TESAR trial high-risk pT1 and low-risk pT2 rectal cancer patients, treated by LE, are randomized between adjuvant chemoradiotherapy or cTME [29]. Whether this adjuvant treatment strategy will lead to better outcome has to be awaited.

Although the present study is a large and nationwide representation of daily practice in the Netherlands, it has some limitations. The NCR database lacks information of histopathological factors, co-morbidity or frailty and no data are available regarding local recurrence. As such, biases could play a role in the results described in this study.

In conclusion, over the years, an increase is seen in LE only and primary TME for rectal cancer in the Netherlands. LE only seems an oncological safe treatment option for patients with pT1 and pT2 tumors with similar long-term results to TME surgery.

Disclosure statement

The authors report no relationships that could be construed as a conflict of interest.