Abstract
Background: A sufficient surgical margin is critical for preventing re-recurrence and achieving R0 status after resection of a local recurrence of rectal cancer (LRRC). Patients and Methods: Re-recurrence-free survival was analyzed in 110 cases of LRRC according to histological type of primary lesion. The circumferential resection margin (CRM) was classified as ‘R1’ (x=0 μm), ‘R0 shortness’ (0 μm <x<2,000 μm), or ‘R0 longness’ (x≥2,000 μm). The histological change from the primary lesion to the recurrent lesion was classified as ‘No change pattern’, from well- to moderately-differentiated (W/M) to an expanding lesion; as ‘Change pattern’, from W/M to an infiltrating lesion. Results: Re-recurrence-free survival was better in ‘R0 longness’ than ‘R0 shortness’ groups and showed that ‘No change pattern’ cases had considerably better prognosis than ‘Change pattern’ cases. Conclusion: Ensuring a CRM >2,000 μm during resection of LRRC is more likely to prevent re-recurrence. Cases with poorly differentiated carcinoma from the primary lesion to the recurrent lesion tend to have poor prognoses.
It is critical to ensure a sufficient surgical margin when removing local recurrence of rectal cancer (LRRC) and to achieve R0 status to prevent re-recurrence (1). However, previous studies have only reported qualitative evaluations regarding the presence or absence of cancer tissue on denuded surfaces (2). Currently there exist no reports presenting quantitative data. In the case of primary rectal cancer, relatively large studies quantitatively described concrete measures such as the circumferential resection margin (CRM) (3, 4). Developments in appropriate surgical procedures have led to establishment of total mesorectal excision (TME) (5) and tumor-specific mesorectal excision (6), contributing to improved prognosis (7).
The surgical excision of LRRC is performed using a variety of procedures, including local resection (LR), low anterior resection (LAR), abdominoperineal resection (APR), abdominoperineal resection with sacrum resection (APRS), total pelvic exenteration (TPE), and total pelvic exenteration with sacrum resection (TPES). However, there still exists no conclusive quantitative and convincing evidence in favor of the use of one specific surgical procedure.
The present study aimed to improve resection of LRRC to prevent re-recurrence through a quantitative assessment while identifying definite histopathological risk factors for re-recurrence.
Patients and Methods
Patients' characteristics. The study included 110 LRRC patients who had previously undergone curative resection for rectal cancer at the Department of Gastroenterological Surgery at the Aichi Cancer Center Hospital in Nagoya, Japan, between January 1981 and December 2012. This was study a retrospective study. At our Hospital, this procedure is equivalent to obtaining permission from our ethics committee on human research in the event of a retrospective study (Aichi Cancer Center Ethics Committee).
Previously, radical dissection of all primary regional lymph nodes and TME were performed in all cases. No cancer cells were observed in the CRM in any of the cases. There was a median period of 679 days (range=238-2949 days) from the date of resection of the primary lesion to resection of local recurrence. The median follow-up period was 685 days (range=64-5197 days).
In almost all cases, a preoperative diagnosis for LRRC was performed by computed tomography (CT)-guided fine-needle aspiration biopsy. However, when acquisition of the biopsy was impossible, the preoperative diagnosis was performed by CT, magnetic resonance imaging, or positron emission tomography (PET). Distant recurrences were analyzed by CT or PET.
Continuous development is defined as ‘Expanding type’ (left). Discontinuous tumor development is defined as ‘Infiltrating type’ (right).
Our indications for surgical treatment of LRRC were the proximal sacral invasion until the level of the S1-S2 junction, and no invasion to the pelvic sidewall, iliac vessels, or sciatic notch (8). If unresectable distant recurrences were recognized and contraindicative, operative treatment was contraindicated. No patients diagnosed with resectable recurrence were treated with preoperative chemoradiation. The surgical procedures performed were LR, LAR, APR, APRS, TPE, and TPES.
After surgery for LRRC, 31 patients received adjuvant chemotherapy only, with oral 5-fluorouracil, 5’-doxifluridine, carmofur, xelox, or uracil-tegafur with leucovorin being the most common, for approximately 6 to 12 months. A further 21 patients were treated with adjuvant radiation only. Finally, 20 patients were treated with adjuvant chemoradiotherapy (9, 10, 11).
We conducted a review of patient hospital records to obtain clinicopathological information such as gender, age (median=58 years), primary lesion-related factors (histological type of primary lesion, lymphatic invasion, venous invasion, and presence or absence of lymph node metastasis), and recurrent lesion-related factors (surgical procedure and pre-carcinoembryonic antigen, and pre-CEA). Pre-CEA was calculated from data collected within one or two months before resection of the recurrent lesion.
Primary rectal adenocarcinomas were graded mainly based on their glandular appearance and were classified as well-, moderately, or poorly-differentiated, according to the World Health Organization histopathological classification for colon and rectum tumors (11) and the Japanese Classification of Colorectal Carcinomas (12). The histological type of the primary lesion was classified into two groups, well- and moderately-differentiated (‘W/M’) and those other than W/M (‘Others’).
The histological type of the central lesion of recurrence was classified into two groups, the ‘Expanding type’ (continuous development of the main tumor) and ‘Infiltrating type’ (discontinuous tumor development with mesenteric implants, and lymphatic, venous, and perineural invasion) (Figure 1).
The CRM of recurrence (radial margin), defined as the minimal distance (x) from the greatest depth of tumor invasion to the denuded surface, was estimated by microscopy (median, 2,000 μm) and classified into three groups, ‘R1’ (x=0 μm), ‘R0 shortness’ (0 μm <x<2,000 μm), and ‘R0 longness’ (x ≥2,000 μm) (Figure 2).
The histological change from the primary lesion to the recurrent lesion was classified into three groups, ‘No change pattern’, a change from W/M to an expanding type; ‘Change pattern’, a change from W/M to an infiltrating type, and ‘Other pattern’, a change from any type but W/M to an infiltrating type.
CRM of recurrence, the minimal distance (x) from the greatest depth of tumor invasion divided by the denuded surface. The left image depicts ‘R0 short’ (0 μm <x<2,000 μm) and the right image, ‘R1’ (x=0 μm).
Statistical analysis. Univariate and multivariate stepwise logistic regression analyses were performed to identify factors influencing local re-recurrence-free survival, re-recurrence-free survival, and overall survival. The log-rank test was used to evaluate differences in local re-recurrence-free survival, re-recurrence-free survival, and overall survival rates between groups. Statistical significance was set at p<0.05.
Results
Table I shows the patients' characteristics, there were 64 males/46 females, 55 patients more than 58 years old/55 patients less than 57 years old. The primary lesion-related factors were 104 patients in W/M/6 patients in Others in the histological type of primary lesion, 55 patients in present lymphatic invasion/43 patients in absent lymphatic invasion (12 patients: unknown), 75 patients in present venous invasion/23 patients in absent venous invasion (12 patients: unknown), and 57 patients in present lymph node metastasis/47 patients in absent lymph node metastasis (6 patients: unknown). The recurrent lesion-related factors were 23 patients in LR/14 patients in LAR/20 patients in APR/8 patients in APRS/28 patients in TPE/17 patients in TPES regarding surgical procedures, 59 patients in Expanding-type/51 patients in Infiltrating-type regarding the histological type, 44 patients in R0 longness/43 patients in R0 shortness/23 patients in R1 regarding the CRM of recurrence, 36 patients in No change pattern/48 patients in Change pattern/6 patients in Other pattern regarding the histological change from the primary lesion to the recurrent lesion, and 62 patients in <5.0/44 patients in ≥5.0 (6 patients: unknown) in pre-CEA (ng/ml).
We performed univariate and multivariate analyses to identify clinicopathological factors correlating with local re-recurrence-free survival rates, re-recurrence-free survival rates, and overall survival rates (Table I). No significant differences were observed in local re-recurrence-free survival rates with respect to gender, age, histological type of primary lesion, lymphatic invasion, venous invasion, presence or absence of lymph node dissection, surgical procedure, histological type in the central lesion of recurrence, and pre-CEA. There were significant differences in local re-recurrence-free survival rates with varying CRMs of recurrence (all strata, p=0.023). There was a significant difference in re-recurrence-free survival rates with varying histological type of the central lesion of recurrence (all strata, p=0.038). There was a significant difference in overall survival with varying histological type of the primary lesion (p=0.025), lymphatic invasion (p=0.015), CRM of recurrence (all strata, p=0.033), and histological change from the primary lesion to recurrent lesion (all strata, p=0.040). A multivariate analysis did not identify independent variables correlating with overall survival.
Prognosis of patients included in the study.
Figure 3 shows a detailed prognosis using the CRM of recurrence. There were significant differences in local re-recurrence-free survival rates between ‘R0 longness’ and ‘R1’ (p=0.007).
There were significant differences in the overall survival rate between ‘R0 longness’ and ‘R1’ (p=0.036) and between ‘R0 longness’ and ‘R1’ (p=0.010).
Although not significantly significant, re-recurrence-free survival rates were higher in ‘R0 longness’ than ‘R0 shortness’ cases (p=0.053). And not significantly significant, local re-recurrence-free survival rates were slightly higher in ‘R0 longness’ than ‘R0 shortness’ cases (p=0.296).
Figure 4 shows a detailed prognosis using the histological change from the primary lesion to the recurrent lesion. There was a significant difference in local re-recurrence-free survival rates between ‘No change pattern’ and ‘Change pattern’ cases (p=0.042). There was a significant difference in re-recurrence-free survival rates between ‘No change pattern’ and ‘Change pattern’ cases (p=0.024). Though not significantly significant, the overall survival rate was higher in ‘No change pattern’ cases than in ‘Change pattern’ cases (p=0.222).
Discussion
In primary rectal cancer, the CRM is advocated to be an important prognostic factor. Quirke et al. (3) reported that the CRM is an important prognostic factor in primary rectal cancer and Bernstein et al. (12) reported that a CRM of ≤2 mm confers a poorer prognosis in primary rectal cancer.
However, very few studies have reported the histopathology of LRRC. Almost all previous studies were qualitative evaluations and, in particular, many report on the denuded surfaces of LRRC. However, many previous reports compared R1 and R0 cases, with very little attention to quantitative evaluations.
Cases were previously classified into two qualitative categories, R0 and R1. R1 cases on denuded surfaces have poor prognosis. Uehara et al. (13) reported that local re-recurrence-free survival rates were significantly worse in R1 cases than R0 cases. Ike et al. (14) reported that overall survival rates were significantly poorer in R1 cases compared to R0 cases. Likewise, Akasu et al. (2) classified the denuded surface of LRRC into three patterns, ‘Microscopic-negative margin’, ‘Microscopic-positive margin’, and ‘Gross-positive residual margin’. These factors also significantly influence the overall survival rate. Cases with ‘Microscopic-positive margin’ and ‘Gross-positive residual margin’ have a poorer prognosis than cases with ‘Microscopic negative margin’. However, neither of these reports were qualitative. To the best of our knowledge, our study provides the first quantitative data on the CRM to prevent local re-recurrence in rectal cancers. In particular, this is the first report of the value of measuring the CRM in the recurrent lesion from the greatest depth of tumor invasion to the denuded surface. The CRM was classified into three groups, ‘R1’ (x=0 μm), ‘R0 shortness’ (0 μm<x<2,000 μm), and ‘R0 longness’ (x≥2,000 μm). This classification appears to be superior to surgical procedures that only ensure R0 on the denuded surface, that were inadequate in preventing local re-recurrence. Procedures ensuring a CRM >2,000 μm prevent local re-recurrence more adequately and effectively. Notably, both local re-recurrences and all re-recurrences, including distant events, can be significantly inhibited by ensuring a CRM >2,000 μm. Although the exact mechanism behind this effect is unknown, our findings will make an important contribution to clinical prognosis. Accordingly, if LRRC is strongly adherent to the surrounding organs, aggressive surgical procedures such as TPE and TPES must be selected to ensure a CRM >2,000 μm.
The histopathology of recurrent lesions of rectal cancer has not been fully investigated. Umemura et al. (15) identified three qualitative patterns of tumor infiltration in LRRC, ‘Expanding type’, ‘Infiltrating type’, and ‘Intermediate type’. Moreover the presence or absence of ‘Distant isolated’ cancer cells was estimated by immunohistochemistry. The authors concluded that the ‘Infiltrating type’ and presence of ‘Distant isolated’ cancer cells correlated with local and distant re-recurrence. Akasu et al. (2) classified the infiltration of LRRC into three qualitative patterns, ‘Solitary expanding’, ‘Multiple expanding’, and ‘Diffuse infiltrating’. These were macroscopic, rather than microscopic, growth patterns. Cases with ‘Multiple expanding’ and ‘Diffuse infiltrating’ lesions showed poorer prognosis than cases with ‘Solitary expanding’ lesions.
To our knowledge, our study provides the first detailed description of changes in histopathology from the primary lesion to the recurrent lesion. This information is necessary to prevent local re-recurrence in rectal cancer. Further investigations are essential in order to understand the impact of preoperative therapy on histopathology.
Conclusion
Ensuring a CRM >2,000 μm during resection of LRRC is more likely to prevent re-recurrence. The surgical procedure must, therefore, be selected with great caution. Cases of poorly-differentiated carcinoma from the primary lesion to the recurrent lesion have poor prognoses and neoadjuvant treatment must be considered.
Detailed prognosis with CRM of recurrence. Local re-recurrence-free survival rates, re-recurrence-free survival rates, and overall survival rates are given in descending order. Although not significantly significant, re-recurrence-free survival rates were higher in ‘R0 longness’ than ‘R0 shortness’ cases (p=0.053). And not significantly significant, local re-recurrence-free survival rates were slightly higher in ‘R0 longness’ than ‘R0 shortness’ cases (p=0.296).
Detailed prognosis with histological change from the primary lesion to the recurrent lesion. There was a significant difference in local re-recurrence-free survival rates between ‘No change pattern’ and ‘Change pattern’ cases (p=0.042). There was a significant difference in re-recurrence-free survival rates between ‘No change pattern’ and ‘Change pattern’ cases (p=0.024).
Acknowledgements
All Authors have no conflict of interest and no financial and personal relationships with other people or organizations that could potentially and inappropriately influence (bias) this work and conclusions.
- Received August 14, 2015.
- Revision received September 12, 2015.
- Accepted September 15, 2015.
- Copyright© 2015 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
