Kaplan-Meier local recurrence–free survival curves for limited vs radical resection (log-rank test, P = .049).
Kaplan-Meier tumor-free survival curves for limited vs radical resection (log-rank test, P = .39).
Kaplan-Meier overall survival curves for limited vs radical resection (log-rank test, P = .16).
Ptok H, Marusch F, Meyer F, Schubert D, Koeckerling F, Gastinger I, Lippert H, Colon/Rectal Cancer (Primary Tumor) Study Group. Oncological Outcome of Local vs Radical Resection of Low-Risk pT1 Rectal Cancer. Arch Surg. 2007;142(7):649-654. doi:10.1001/archsurg.142.7.649
Despite the noninclusion of locally draining lymph nodes, limited resection of low-risk pT1 rectal cancer can achieve an adequate oncological outcome with lower morbidity and mortality compared with radical resection.
Based on the data of a prospective multicenter observational study performed from January 1, 2000, through December 31, 2001, patients with low-risk pT1 rectal cancer underwent analysis with regard to the early postoperative outcome and the oncological long-term results achieved after limited vs radical resection with curative intent.
Two hundred eighty-two hospitals of all categories.
Four hundred seventy-nine patients with low-risk pT1 rectal cancer treated for cure.
Eighty-five patients (17.7%) underwent limited excision using a conventional transanal approach and 35 (7.3%) using transanal endoscopic microsurgery. The remaining 359 (74.9%) underwent radical resection.
Main Outcome Measures
Postoperative morbidity and mortality, local recurrence rate, and tumor-free and overall survival.
In comparison with radical resection, limited resection was associated with fewer general (25.1% vs 7.5%; P<.001) and specific (22.8% vs 9.2%; P<.001) postoperative complications. After a mean follow-up of 44 months, patients who underwent limited resection had a significantly higher 5-year local tumor recurrence rate than did those who underwent radical resection (6.0% vs 2.0%; P = .049), but tumor-free survival did not differ.
Limited resection of pT1 low-risk rectal cancer can result in an oncologically acceptable outcome but must nevertheless be considered an oncological compromise compared with radical resection.
In the treatment of rectal cancer, total mesorectal excision and complete resection of the draining lymphatic tissue have significantly reduced the local recurrence rate, which has proved to be considerably lower than 10%.1,2 Radical tumor resection with total mesorectal excision is therefore considered the oncological standard in the curative treatment of low-lying rectal cancer. Morbidity and mortality rates of elective low anterior resection are reported to be 15% to 30% and 5% to 10%, respectively.3,4 An abdominoperineal resection with a definitive colostomy performed for oncological reasons is required in approximately 30% of rectal cancers, as has previously been reported by our Colon/Rectal Cancer (Primary Tumor) Study Group.5
Alternatively, early-stage tumors such as low-risk pT1 rectal cancer are increasingly being treated with limited resections. Despite the noninclusion of resection of the locally draining lymph nodes, limited resection enables an adequate oncological outcome,6 as well as lower morbidity and mortality rates than radical resection.3,7- 10 Specifically, the outcome is determined by the risk of lymphatic infiltration and established lymph node metastases. To appropriately define the low-risk cancer, criteria have been developed that permit adequate treatment in the form of a locally limited resection, ie, with no additional resection of the regional lymphatic tissue.11- 13 The probability of lymph node metastases has been calculated to be less than 2%.6,11 Apart from lower morbidity and mortality, a major advantageof limited resection is the preservation of the sphincter in low-lying rectal cancers.
Numerous published studies on the oncological outcome of local tumor excision (conventional transanal excision [TAE] and transanal endoscopic microsurgery [TEM]) of T1 rectal cancer report local recurrence rates ranging from 10% to 30%.7,14,15 These data, obtained mostly from retrospective analyses, question the oncological safety and outcome of local surgical procedures. In view of these studies, one can only speculate on the reasons for the published high local recurrence rates after limited resection, which are higher than the probability of lymphatic tumor cell invasion—the origin of extraluminal tumor recurrence. Thus, the question is whether local resection is truly oncologically adequate or merely a compromise treatment of low-risk pT1 rectal cancer in patients with a high perioperative risk that is associated with an acceptable probability of tumor recurrence and low morbidity and mortality.
Based on data obtained in a prospective, multicenter observational study, our aims were to obtain early and long-term follow-up data after local and radical resection of low-risk pT1 rectal cancer and to assess the oncological outcomes of both surgical approaches.
From January 1, 2000, through December 31, 2004, patients with a colorectal carcinoma were enrolled in a prospective, multicenter observational study. Participation was voluntary and depended on the patients' acceptance that their demographic, tumor- and treatment-specific, and follow-up data would be used in anonymous form.
All of the included patients had a low-risk pT1 rectal carcinoma treated with curative intent by local (TAE or TEM) or radical resection. The data collected were analyzed with regard to postoperative and oncological long-term outcomes after radical vs limited surgery.
In accordance with the national recommendations for the treatment of rectal cancer, in addition to primary radical anterior or abdominoperineal resection, a further treatment option is available for low-risk pT1 carcinoma, namely local TAE. Local TAE can be considered oncologically adequate (ie, having no need for further surgical treatment) in the case of a suspected rectal adenoma that cannot be removed endoscopically. Local TAE is also sufficient for a mucosal or submucosal T1 rectal carcinoma diagnosed by means of endosonography and biopsy findings that has undergone complete transanal full-wall-thickness excision and that has been diagnosed by means of histological workup as a completely resected poorly or moderately differentiated pT1 carcinoma (G1/2) with no invasion of lymphatic vessels and with a diameter of less than 3 cm. Otherwise, immediate radical resection of the rectum should be applied.
The decision to perform a primary local or a radical resection was made by the respective surgeon. For the analysis of the data, patients were selected from the resulting data pool of all patients with rectal carcinoma who after radical or limited resection were shown histopathologically to have a completely removed low-risk tumor (pT1G1/2L0R0) and thus could be considered to have received curative treatment. Distant metastasis from the rectal carcinoma must have been excluded at least by results of chest radiography and an abdominal ultrasonographic examination.
The patients were divided into groups undergoing TAE, TEM, or radical tumor resection (anterior/low anterior resection or abdominoperineal resection). The groups were examined and compared for age, sex, American Society of Anesthesiologists physical status classification, and tumor location. Tumors were divided into the following 4 groups on the basis of the distance of the distal tumor margin from the anal verge as measured with a rigid rectoscope: less than 4.0, 4.0 to 7.9, 8.0 to 11.9, and 12.0 to 16.0 cm.
With regard to the early postoperative outcome and oncological long-term results, patients undergoing local tumor excision (TAE and TEM) were compared with those undergoing radical resection. The 2 groups of patients were compared in terms of the following variables: perioperative morbidity and mortality, local recurrence rate, and tumor-free and overall survival. Morbidity was characterized by general and specific postoperative complications, as summarized in Table 1. The postoperative mortality was limited to the duration of patient care at the hospital (hospital mortality). The follow-up database contained the following data: date of the last visit of the patient, life status (alive or dead), and evidence of local recurrence and/or distant metastases.
The study was conducted in accordance with the Declaration of Helsinki on Biomedical Research, and written informed consent was obtained from every patient enrolled in the study. Because participation was voluntary (written informed consent), data were recorded on an anonymous basis, and the observational study had no influence on the therapeutic procedure applied, no approval by an ethics committee was required.
Continuous variables were compared using 1-way analysis of variance and are presented as mean and 95% confidence interval or as median and range. Categorical variables were compared with the χ2 test. Survival was analyzed using the Kaplan-Meier method; the patient groups were compared using the log-rank test. To identify the variables with a significant effect on tumor recurrence and tumor-free and overall survival, the multivariate Cox proportional hazards model was used. A P < .05 was considered statistically significant.
From January 1, 2000, through December 31, 2001, 6886 patients with rectal cancer were enrolled in the study, 652 (9.5%) of whom had a pT1 rectal cancer. In 479 of these 652 patients (73.5%), a low-risk pT1 rectal cancer was diagnosed on the basis of the defined inclusion criteria. In 85 of the 479 patients (17.7%), a conventional TAE was performed; 35 patients (7.3%) were treated with TEM and 359 patients (74.9%) underwent radical resection. In 62 of the 359 patients undergoing radical resection (17.3%), an abdominoperineal resection was performed.
Table 2 shows the patient characteristics of the 3 treatment groups. Significant differences were found in tumors located above the anal verge (P < .001). The patients undergoing radical resection had significantly more tumors located higher than 8 cm above the anal verge. For the additional categories, no significant differences were seen (age, P = .08; American Society of Anesthesiologists physical status classification, P = .71; sex, P = .55).
An intraoperative complication was observed in a single patient in each of the limited resection groups (TAE, 1.2%; TEM, 2.9%), whereas, in the radical resection group, 24 patients (6.7%) experienced intraoperative complications, but no statistically significant differences were seen (P = .10). During the postoperative course, there was a significantly lower complication rate in the TAE and TEM groups than in the radical resection group. The rate of general postoperative complications in the TAE, TEM, and radical resection groups were as follows: 8 patients (9.4%), 1 (2.9%), and 90 (25.1%) (P<.001), respectively; the distribution of specific postoperative complications in the TAE, TEM, and radical resection groups was 11 patients (12.9%), none (0%), and 82 (22.8%), respectively (P<.001). There were no treatment-associated deaths in any of the groups. Table 3 shows a comparison of the early postoperative results in the limited and radical resection groups.
The median postoperative hospital stay was significantly longer in the radical resection group (median, 15 days; range, 4-90 days) than in the combined TAE and TEM groups (median, 8 days; range, 1-15 days) (P<.001). No difference in hospital stay was seen between the TAE and TEM groups (P = .96).
Of the 479 patients, 417 agreed to undergo follow-up investigations. Data on the follow-up of 377 of the 417 patients (90.4%) were obtained. The median follow-up time was 44 months (range, 1-65 months). There was no significant difference between the patient groups with regard to rate and time of the follow-up (Table 4). Overall, 38 of the 377 patients (10.1%) died, including 4 individuals (1.1%) because of tumor recurrence.
During follow-up, 9 of the 377 patients (2.4%) developed local recurrences. After limited tumor resection (TAE and TEM), local recurrences were found in 5 of 99 patients (5.1%); after radical resection, local recurrences were found in 4 of 278 subjects (1.4%) (Table 4). Local recurrences were found after a median of 19.1 (range, 4.5-45.7) months. The mean local recurrence-free survival was 61 (95% confidence interval, 59-63) months in the limited resection groups and 64 (95% confidence interval, 64-65) months in the radical resection group. The actuarial 5-year local recurrence rate was 6.0% and 2.0% in the limited resection and radical resection groups, respectively. Kaplan-Meier analysis and the log-rank test results indicated a significant difference (P = .049; Figure 1).
Overall, distant metastases after surgical treatment of rectal cancer with curative intent were diagnosed in 15 of the 377 patients (4.0%). In patients undergoing limited resections, distant metastases were found in 4 of 99 cases (4.0%) and after radical resection in 11 of 278 patients (4.0%), with no significant difference (P = .91).
Distant metastases were diagnosed after a median of 35.1 (range, 3.6-53.2) months. Kaplan-Meier analysis showed no significant difference between the limited (TAE and TEM) and radical resection treatment groups in terms of metastasis-free survival (log-rank test, P = .76).
Overall, tumor recurrence was seen in 23 of 377 patients (6.1%), including in 9 patients (2.4%) as local recurrence only, in 14 (3.7%) as distant metastasis only, and in 1 (0.3%) as both. The analysis of tumor-free survival did not show any significant difference between limited and radical resections (Figure 2; log-rank test, P = .39). The actuarial tumor-free 5-year survival rate was 91.4% in patients undergoing limited resection and 92.3% in those undergoing radical resection. There was no difference in the overall survival of the patient groups (Figure 3; log-rank test, P = .16).
The effect of age, American Society of Anesthesiologists physical status classification, histopathologic grading (G1/2), tumor distance from the anal verge, and surgical procedure on overall and disease-free survival and the local recurrence rate was determined by multivariate Cox proportional hazards model. None of the variables tested had a significant effect on tumor-free survival. Similarly, none of the variables tested by multivariate analysis had a significant effect on metastasis-free survival. In contrast, the Cox proportional hazards model identified a significant effect of surgical procedure (limited vs radical resection, P = .03) and age (P < .001) on the local recurrence rate. Furthermore, age (P < .001) and American Society of Anesthesiologists physical status classification (P < .001) had a significant effect on overall survival.
In local resection of low-risk pT1 rectal cancer, the locoregional draining lymphatic tissue is not removed, and this may become the source of local recurrences if tumor cells have invaded the lymphatic vessels. Other investigators16 have shown that the risk of local recurrence is comparable for lymphovascular and lymph node infiltration. The probability of lymphovascular invasion by tumor cells correlates with the depth of tumor infiltration and has been identified as an independent risk factor for local recurrence,17 which has been reported to range from 0% for infiltration of the upper third of the submucosa to 27% for infiltration of the lower third of the submucosa.11,15,18 The risk of lymph node metastasis for all pT1 rectal cancers ranges from 9% to 15%, irrespective of lymphovascular invasion.12,19- 22 In the present observational study, no data on the depth of submucosal tumor infiltration were available.
The recommendation to use local resection for the treatment of rectal cancer applies only to low-risk pT1 rectal cancers that have a less than 2% probability of lymph node metastasis and that are defined as carcinomatous tumor growth maximally infiltrating the submucosa and having good or moderate cell differentiation (G1/2) with no evidence of lymphovascular invasion of tumor cells (pL0).11
In addition to the low-risk cancer criterion, additional factors for assessing the risk of lymph node metastasis such as hemangioinvasion, tumor cell dissemination at the tumor invasion front (budding), and submucosal depth of infiltration are discussed in the literature. These factors are associated with a higher risk for local recurrences and distant metastases and have a negative effect on survival.11,19,23- 26 However, these criteria have not yet been incorporated within the definition of low-risk carcinoma.
The present study analyzed the postoperative oncological outcome after limited vs radical resection of rectal cancer meeting the defined criteria of a low-risk tumor (pT1G1/2L0R0). The radical resection group also included those lesions in which results of histological analysis showed lymph node metastases but no invasion of the lymphatic vessels (33/379 [8.7%]), thus fulfilling the criteria for local resection.
A comparison of the oncological results of limited and radical resection at a median follow-up of 44 months (3.7 years) showed a significantly higher 5-year local recurrence rate for limited resection (6.0% vs 2.0%; P = .049). In contrast, no significant difference between the patient groups was found in tumor-free and overall survival, which may be owing to the small number of local recurrences compared with the overall number of tumor recurrences.
Although the local recurrence rate of 6.0% after limited resection is considered acceptable, and although it is comparable with the figures reported by other authors,7,14,15,27- 29 in our study it was significantly higher than that of the radical resection group (P = .049). Furthermore, the Cox proportional hazards model showed that the surgical procedure (limited vs radical resection) is an independent variable for the local tumor recurrence rate. Local resection must be considered an oncological compromise because other study groups have also reported that the local recurrence rate tends to be higher than with radical resection of pT1 rectal cancer.28,30,31 A possible reason for this may be the diagnostic uncertainty in assessing lymphovascular invasion at the histological workup of the surgical specimen. Even after histopathologic and immunohistochemical evaluation with definitive exclusion of lymphatic vessel invasion, there remains a residual risk of lymph node metastases of more than 2% owing to failure to detect such an invasion. In the present study, a node-positive L0 evaluation was made in 33 of 379 patients (8.7%) undergoing radical resection. This means that lymphovascular invasion cannot be identified, as others also reported. Meyer et al16 found lymph node metastases in 18.1% of colorectal surgical specimens after radical resection but no lymphangioinvasion. Although the careful histopathologic workup of the locally excised tumor specimen can exclude a lymphovascular invasion with greater certainty than in the resected rectal specimen, it may be assumed that a rate of missed lymphangioinvasion of about 5% to 10% is the reason for the higher local recurrence after limited resection. Thus, the local recurrence rate was significantly higher in the limited resection groups.
The advantage of limited surgical intervention is the appreciably lower morbidity and mortality rates compared with radical resection.3,8,9,28,29,32- 36 In the present investigation, limited resection was associated with a similar rate of intraoperative complications but significantly fewer general (P < .001) and intervention-specific postoperative complications (P < .001). In 17.3% of the patients in the radical resection group, low-lying rectal cancer required an abdominoperineal resection with permanent colostomy. Therefore, the slightly but statistically significant higher risk of local recurrence after local resection of low-risk rectal cancer needs to be weighed against the significantly higher postoperative morbidity after radical resection.
To achieve an acceptable oncological outcome after limited resection, appropriate patient selection is essential. Pretherapeutic diagnostic procedures such as endorectal ultrasonography and histopathologic workup of the tumor biopsy specimen are of limited value. For local staging, endorectal ultrasonography combined with considerable experience is reliably accurate in assessing the tumor category (including the submucosal depth of infiltration) when high-frequency ultrasound probes are used.37- 39 However, in the routine clinical setting, the diagnostic potential of endorectal ultrasonography cannot be fully realized.40 At the same time, evaluation in the tumor specimen (forceps biopsy) of histopathologic features of relevance for the prognosis is not representative of the tumor as a whole. Local excision of the preoperatively assessed mucosal or submucosal T1 cancer can therefore be considered merely a total excision biopsy that, assuming a high-risk cancer is excluded, determines further therapeutic measures. Should a high-risk cancer be found, radical resection must be performed directly. Hahnloser et al41 have shown that this can be considered an oncologically adequate approach. In a matched-paired analysis, they retrospectively investigated patients with high-risk pT1 rectal cancer who underwent primary radical resection (n = 78) or local tumor excision and a subsequent radical resection within 30 days (n = 37). There were no significant differences in local recurrence rates and the rate of metachronous distant metastases, although in 19% of the patients with initially local tumor resection, residual tumor tissue was detectable. In the case of later manifestation of local recurrence through follow-up investigations, salvage resection results in a significant impairment of the prognosis.42 This underscores the importance of appropriate patient selection based on histopathologic variables for considerably reducing the risk of local recurrence. In the case of multimorbid patients at a high risk owing to radical resection, or of patients who reject the procedure, local excision with curative intent plus adjuvant irradiation is a therapeutic option.8,27 Evidence of the effectiveness of this approach can be seen even for higher tumor categories.9,27,36 However, at present, a relevant general recommendation cannot be made.
As the data presented herein show, a lymph node metastasis rate of approximately 10% is most likely when the established low-risk criteria for rectal cancer are applied to patient selection. This means that additional criteria, such as tumor cell dissemination on the invasion front (budding) and the depth of submucosal infiltration, should be considered for the definite exclusion of high-risk cancer.
When interpreting the data, it must be remembered that they were collected in a multicenter observational study and referred only to primary treatment and tumor recurrence. Data on the treatment of tumor recurrences were not obtained. Although an adequate statement as to the appearance of recurrences after primary treatment can be made, statements on overall survival are possible only with certain reservations because overall survival in the 2 groups might be influenced by possible salvage procedures. Furthermore, some 15% of the patients whose primary data were recorded subsequently refused permission for the use of their follow-up data and their data could not be included in the follow-up analysis.
No data are available to determine whether the different surgical procedures in the participating departments were performed by different groups of surgeons. An analysis of the data, however, failed to show any preference of individual departments for one procedure or the other. The indications for limited or radical resection were based on the national recommendations for the treatment of rectal cancer. Furthermore, patient selection was strictly in accordance with low-risk criteria, to ensure that the patient groups were comparable.
A limited resection of low-risk rectal cancer in the form of a conventional TAE or TEM may be considered oncologically acceptable because, in the present study, it was associated with a local tumor recurrence rate of only 6%. However, this approach seems to be considered an oncological compromise in definitive treatment of rectal cancer. On the one hand, limited resection results in a lower intervention-dependent morbidity and mortality; on the other, there is a higher risk of local recurrence because histomorphological and immunohistochemical investigations cannot provide absolute certainty in excluding a lymphatic tumor cell invasion.
Optimized preoperative local tumor staging with a subtle evaluation of the submucosal depth of infiltration and a strictly applied indication taking account of additional prognostic variables might substantially improve the surgical outcome. It remains to be seen whether adjuvant or neoadjuvant treatment, as applied by a number of authors who perform limited resection of tumors at more advanced stages than T1,9 can be generally recommended in low-risk pT1 rectal cancer to compensate for possible diagnostic deficits and achieve less surgical invasiveness. At present, such an approach may be a reasonable alternative treatment option in the multimorbid patient.
Correspondence: Henry Ptok, MD, Department of Surgery, University Hospital Magdeburg, Leipziger Strasse 44, D-39120 Magdeburg, Germany (firstname.lastname@example.org).
Accepted for Publication: April 7, 2006.
Author Contributions:Study concept and design: Ptok, Marusch, Meyer, Schubert, Koeckerling, Gastinger, and Lippert. Acquisition of data: Ptok, Marusch, Meyer, Schubert, Koeckerling, Gastinger, and Lippert. Analysis and interpretation of data: Ptok, Marusch, Meyer, Schubert, Koeckerling, Gastinger, and Lippert. Drafting of the manuscript: Ptok, Meyer, and Schubert. Critical revision of the manuscript for important intellectual content: Marusch, Koeckerling, Gastinger, and Lippert. Statistical analysis: Ptok. Study supervision: Marusch, Koeckerling, Gastinger, and Lippert.
Colon/Rectal Cancer (Primary Tumor) Study Group:Scientific leadership: Hans Lippert, PhD, and Ingo Gastinger, PhD, Otto-von-Guericke University Magdeburg, Magdeburg, Germany. Scientific advisory board: Ferdinand Koeckerling, PhD, Hanover-Hospital, Hanover, Germany; Jürgen Läuter, PhD, and Albert Rössner, PhD, Otto-von-Guericke University Magdeburg; and Thomas Manger, PhD, SRH Wald-Klinikum Gera gGmbH, Gera, Germany. A complete list of the participating hospitals was published in Chirurg. 2003;74(4):341-351.
Financial Disclosure: None reported.