Journal of Surgical Oncology 2014;110:739–744
Histopathological Predictors for Local Recurrence in Patients with T3 and T4 Rectal Cancers Without Preoperative Chemoradiotherapy YOSHITO AKAGI, MD, PhD,* TORU HISAKA, MD, PhD, TOMOAKI MIZOBE, MD, PhD, TETSUSHI KINUGASA, MD, PhD, YUTAKA OGATA, MD, PhD, AND KAZUO SHIROUZU, MD, PhD Department of Surgery, Kurume University School of Medicine, Kurume, Japan
Background and Objectives: Identification of suitable predictors of local recurrence (LR) in patients with rectal cancer would be of clinical benefit. The aim of this study was to identify histopathological factors that could predict LR. Methods: A total of 796 stage II/III patients with pT3 and pT4 rectal cancer who did not undergo preoperative chemoradiation were enrolled. LR was defined as intra‐pelvic recurrence only. Histopathological factors related to LR were investigated. Results: LR was found in 25 patients (6.1%) with stage II and 54 patients (13.9%) with stage IIIB/IIIC. In patients with stage II, distance of mesorectal extension (DME) >4 mm (P ¼ 0.011) and positive venous invasion (P ¼ 0.035) were independent factors that predicted LR. In patients with stage IIIB/IIIC, circumferential resection margin (CRM) 1 mm (P ¼ 0.003) and positive lymphatic invasion (P ¼ 0.006) were independent factors. The cumulative 5‐year LR rate was higher (11.9%) in patients with a combination of DME > 4 mm and/or positive venous invasion for stage II (P < 0.001), and was also higher in patients with a combination of CRM1 mm and/or positive lymphatic invasion for stage IIIB and IIIC (22.2%; P < 0.002, and 34.3%; P < 0.006, respectively). Conclusions: Important histopathological predictors for LR in patients with pT3 and pT4 rectal cancer were different at each stage.
J. Surg. Oncol. 2014;110:739–744. ß 2014 Wiley Periodicals, Inc.
KEY WORDS: risk factor; distance of mesorectal extension; circumferential resection margin; vessel invasion
INTRODUCTION Local recurrence (LR) is a critical problem for patients with rectal cancer after curative surgery. Since early detection of LR is difficult, patients with LR tend to have poor outcomes. Risk factors that can predict which patients are at high risk of LR have not yet been identified. Treatment for primary tumors with total mesorectal excision (TME) and a negative circumferential resection margin (CRM) is very important at the time of initial operation, because TME can reduce LR [1,2], and because a positive CRM is related to LR in rectal cancer patients [3]. The rate of LR, however, varies widely after TME with or without adjuvant treatment [4,5], and patients with a negative CRM can still develop LR. Thus, identification of other factors related to LR would be of benefit. However, additional clinicopathological factors that can predict stage‐specific LR are required to improve clinical decision‐making with respect to postoperative adjuvant treatment planning and follow‐up. A recent multicenter study demonstrated that the distance of mesorectal extension (DME) strongly correlated with distant metastasis and LR in every stage III case [6,7]. Therefore, the goal of the present study was to assess the relationship between traditional histopathological adverse factors, including CRM and DME, and LR after curative resection. Special focus was placed on the prediction of LR in stage II/III patients with pT3 and pT4 rectal cancer based on the International Union Against Cancer‐TNM Classification of Malignant Tumors (UICC‐TNM) 7th edition classification system [8].
PATIENTS AND METHODS Between 1986 and 2007, 796 consecutive, stage II/III patients with pT3 and pT4 rectal cancer underwent curative resection without preoperative or postoperative radiotherapy (RT) at Kurume University. All patients provided their written, informed consent after receiving a full explanation of the study. All protocols in this study were approved by the local institutional review board.
ß 2014 Wiley Periodicals, Inc.
The pathological diagnosis was made using whole‐mount sections of surgical specimens. Detailed clinical and pathological information was prospectively entered into our original computerized registration system (dBASE III Ashton‐Tate, Torrance, CA). All patients with histologically proven adenocarcinoma were extracted from the prospective database for this study. There were 408 patients with stage II disease and 388 with stage III disease, all of which were determined according to the pathological TNM classification 7th edition [8], in which node‐negative patients with pT3 and pT4 rectal cancer were classified as stage II. On the other hand, stage III disease followed a more complicated algorithm that took into account the number of lymph node metastases and the depth of invasion. According to the classification, patients with stage III disease were stratified into stage IIIB and IIIC for evaluation in this study. Baseline parameters related to the patients, tumors, operative methods, and investigated variables are summarized in Table I. The histopathological type of tumor was divided into two categories: “well” (well differentiated type of tubular adenocarcinoma) and “other” (moderately to poorly differentiated type of adenocarcinoma, mucinous adenocarcinoma, or signet ring cell carcinoma). Venous invasion (V) was defined as tumor cells within a space lined by endothelial cells and smooth muscle or elastic fibers, as shown in Figure 1a [9], and was categorized as negative or positive. Lymphatic invasion (Ly) was defined as occurring only when cancer cells were floating within an endothelial‐lined space and was categorized as negative or positive [10]
*Correspondence to: Yoshito Akagi, MD, PhD, Department of Surgery, Kurume University School of Medicine, 67 Asahi‐machi, Kurume City, Fukuoka 830‐0011, Japan. Fax: þ81‐942‐34‐0709. E‐mail: [email protected] Received 5 February 2014; Accepted 15 May 2014 DOI 10.1002/jso.23678 Published online 24 June 2014 in Wiley Online Library (wileyonlinelibrary.com).
740
Akagi et al.
TABLE I. Patient and Tumor Characteristics and Details of Treatment Characteristic Age (years) Sex Preoperative CEA (ng/ml) Gross type Location from AV (cm) Size of tumor (cm) T‐classification Number of LNs retrieved Number of LN metastasis TNM stage Tumor differentiation Venous invasion (V) Lymphatic invasion (Ly) Perineural invasion (PNI) Tumor budding DME (mm) CRM (mm) Surgical procedure SSO Lateral LND ANP TME Adjuvant chemotherapy
Category
No. of patients
65/>65 Male/female 5.0/>5.0 Unknown Expansive/infiltrative 5/5–10/>10 5/>5 pT3/pT4 Median (range) 4 >1/1
455/341 532/264 453/266 77 698/98 290/235/271 389/407 481/315 29 (2–122) 102/694 408/250/78/60 408/388 494/302 627/169 613/183 691/105 111/293a 384/412 728/68
Yes/no Yes/no Yes/no Yes/no Yes/no
552/244 385/411 514/284 560/236 493/303
CEA, carcinoembryonic antigen; AV, anal verge; LN, lymph node; Well, well differentiated adenocarcinoma; Mode, moderately differentiated adenocarcinoma; Others, poorly differentiated or mucinous adenocarcinoma; DME, distance of mesorectal extension; CRM, circumferential resection margin; SSO, sphincter saving operation; LND, lymph node dissection; ANP, autonomic nerve preservation; TME, total mesorectal excision. a Unknown for 392 cases.
(Fig. 1b). Perineural invasion (PNI) was defined as occurring when cancer cells were observed inside the perineurium and was categorized into negative and positive groups [11] (Fig. 1c). DME was directly measured on the glass slides of the pathological specimen. The distance from the outer border of the muscular layer to the deepest part of invasion was measured, and the cut‐off‐value was set to 4 mm according to a previous report [6] (Fig. 1d, D1). CRM was also pathologically measured as the closest distance from the surgical cutting line to viable cells on the glass slides (Fig. 1d, D2). Specimens were categorized as either positive CRM (distance 1 mm) or negative CRM (distance >1 mm). Cases with macroscopically positive CRM were not included in this study. Adjuvant chemotherapy was administered to some patients who had lymph node metastasis and were positive for Ly and/or V. 5‐ Fluorouracil‐based anti‐cancer agents including 1‐hexylcarbamoyl‐5‐ fluorouracil (HCFU), 50 ‐deoxy‐5‐fluorouridine (50 DFUR), or uracil‐ tegafur (UFT) were given in this period without specified protocols. 5‐Fluorouracil leucovorin (UZEL) or uracil‐tegafur‐gimeracil‐oteracil potassium (TS‐1) was preferentially used since 2000.
Follow‐Up and Oncologic Outcomes All patients were followed on an outpatient basis, and routine investigations were performed according to the rules of the JSCCR [12]. The most recent date of contact was regarded as the final date of confirmation in each case. The overall final follow‐up date for this study was the last day of March 2013. Patients were re‐staged according to the pathological TNM classification 7th edition [8]. Postoperative surveillance included measurement of serum carcinoembryonic antigen (CEA), and a rectal examination was performed every 3 months. Computed tomography (CT) and/or magnetic resonance imaging (MRI) was conducted every 6–12 months for the first 3 years, and then
Fig. 1. Histopathologic findings in the extramural fatty tissue adjacent to a colorectal tumor. a: Venous invasion is defined as tumor cells within a space lined by endothelial cells and smooth muscle or elastic fibers. b: Lymphatic invasion is defined as occurring only when cancer cells are floating within an endothelial‐lined space. c: Perineural invasion is defined as occurring when cancer cells are observed inside the perineurium. d: Measurement of mesorectal extension and circumferential resection margin is shown. D1: distance of mesorectal extension, D2: distance of circumferential resection margin. Journal of Surgical Oncology
Predictor of Local Recurrence in Rectal Cancer TABLE II. Local Recurrence According to the TNM Classification, 7th ed Characteristic
P‐value
No. of patients
No. of LR (%)
796
79 (9.9)
422 374
47 (11.1) 32 (8.6)
0.095
481 315
47 (9.8) 32 (10.2)
0.978
408 388
25 (6.1) 54 (13.9)
0.002
All patients Time period 1986–1999 (Non‐PET) 2000–2007 (PET) T stage pT3 pT4 pTNM stage II IIIB/IIIC
LR, local recurrence; PET, positron emission tomography.
annually, unless recurrence was suspected. Beginning in the year 2000, positron emission tomography (PET) was also conducted if a recurrence was suspected. The median follow‐up duration of the surviving patients was 122 months, with a range of 20–298 months. Follow‐up data for more than 60 months was 92% complete. The date and first site of recurrence were recorded for each patient. Pelvic floor or lateral lymph node metastasis was defined as intra‐pelvic LR. LR was defined as clinical or radiological tumor growth within the initial surgical field, with or without concomitant distant metastasis. For definitive diagnosis, image‐guided biopsy, salvage surgery, and/or palliative operation was performed whenever possible.
Statistical Analysis All independent variables listed in Table II were coded for statistical analysis. Statistical analysis was performed using StatView 5.0 and JMP 8.0 (SAS Institute, Inc., Cary, NC) for Windows. The Chi‐square test, Fisher’s exact test, and the Mann–Whitney U test were used to determine the statistical significance of any differences. Cox regression analysis was used to identify independent risk factors for LR. The Kaplan–Meier method was used to calculate survival rates. Any significant difference in survival was assessed using the log‐rank test. Statistical significance was indicated by a two‐tailed. P‐value < 0.05 and by 95% confidence intervals (CI).
741
and a PET group, in the year 2000, when PET became available. The LR rate was lower in the PET group than in the non‐PET group. The depth of invasion (T‐stage) was not a significant risk factor for LR (P ¼ 0.978). The rate of LR was significantly higher in patients with stage IIIB/IIIC disease (n ¼ 54; 13.9%) than in patients with stage II disease (n ¼ 25; 6.1%, P ¼ 0.002).
Risk Factors for LR Using Cox Regression Analysis Parameters listed in Table I were investigated as risk factors for LR using univariate Cox regression analysis. Variables in each stage that showed a significant association with the risk of LR on univariate Cox regression analysis were subsequently subjected to multivariate Cox regression analysis. On univariate Cox regression analysis for patients with stage II disease, the rate of LR was significantly higher in patients with positive V (odds ratio [OR], 3.6; 95% CI, 1.192–2.948; P ¼ 0.003), positive PNI (OR, 4.3; 95% CI, 1.204–2.626; P ¼ 0.004), and DME > 4 mm (OR, 11.3; 95% CI, 1.182–2.544; P ¼ 0.001). Other parameters listed in Table I were not significant risk factors for LR (data not shown). Similarly, in patients with stage IIIB/IIIC disease, moderate to poorly differentiated adenocarcinoma (OR, 2.0; 95% CI; 1.100– 3.832; P ¼ 0.013), positive Ly (OR, 2.9; 95% CI, 1.789–5.306; P ¼ 0.0001), positive V (OR, 2.5; 95% CI, 1.436–3.737; P ¼ 0.001), positive PNI (OR, 3.0; 95% CI, 1.785‐5.254; P ¼ 0.0001), DME > 4 mm (OR, 3.8; 95% CI, 1.823–3.136; P ¼ 0.0006), and CRM 1 mm (OR, 4.1; 95% CI, 2.912–5.124; P < 0.0001) were significant risk factors for LR (data not shown). These above‐mentioned significant parameters were investigated as independent predictors for LR by multivariate Cox regression analysis (Table III). DME > 4 mm was the most powerful independent predictor for LR in patients with stage II disease (HR, 3.3; 95% CI, 1.308–8.270; P ¼ 0.011), and positive V was also identified as an independent risk factor for LR (P ¼ 0.035). For patients with stage IIIB/IIIC disease, CRM 1 mm was the most powerful adverse factor for LR (HR, 2.6; 95% CI, 1.369–4.664; P ¼ 0.003). Moreover, positive Ly (HR, 2.6; 95% CI, 1.310–5.162; P ¼ 0.006) and DME > 4 mm (HR, 2.5; 95% CI, 1.123–5.722; P ¼ 0.025) were identified as independent risk factors for LR.
RESULTS
Cumulative LR Rate
LR by TNM Stage
Combining the two most powerful pathological risk factors identified on multivariate analysis could identify groups at high risk for LR in each stage: DME > 4 mm and/or positive V in stage II disease, and CRM 1 mm and/or positive Ly in stage IIIB/IIIC disease.
Of 796 patients, 79 (9.9%) had LR with or without distant metastasis (Table II). The cases were classified into two groups, a non‐PET group
TABLE III. Multivariate Cox Regression Analysis of Potential Risk Factors for Local Recurrence in Patients With Stage II/III Rectal Cancer Stage II (n ¼ 408) Variable
Category
Patients
HR (95% CI)
Tumor differentiation
Well Other Negative Positive Negative Positive Negative Positive 4.0 mm >4.0 mm >1.0 mm 1.0 mm
289 119 366 42 354 54 383 25 248 160 382 26
— — — — 1 2.5 (1.065–6.009) 1 2.1 (0.766–5.998) 1 3.3 (1.308–8.270) — —
Lymphatic invasion (Ly) Venous invasion (V) Perineural invasion (PNI) DME CRM
Stage IIIB/IIIC (n ¼ 388) P‐value
0.035 0.146 0.011
Patients
HR (95% CI)
205 183 247 141 273 115 308 80 136 252 346 42
1 0.8 (0.436–1.633) 1 2.6 (1.310–5.162) 1 1.1 (0.567–2.226) 1 1.6 (0.815–3.058) 1 2.5 (1.123–5.722) 1 2.6 (1.369–4.664)
P‐value
0.615 0.006 0.738 0.176 0.025 0.003
HR, hazard ratio; CI, confidence interval; AV, anal verge; Well, well differentiated adenocarcinoma; Others, moderately differentiated, poorly differentiated, or mucinous adenocarcinoma; DME, distance of mesorectal extension; CRM, circumferential resection margin.
Journal of Surgical Oncology
742
Akagi et al.
For patients with stage II disease, the cumulative 5‐year LR rate of Group B (DME > 4 mm and/or positive V; 11.9%) was significantly higher than that of Group A (both DME 4 mm and negative V; 1.6%) (Fig. 2a, HR, 7.3; 95% CI, 2.513–21.349; P ¼ 0.0007). For patients with stage IIIB disease, the cumulative 5‐year LR rate was significantly
higher in Group B (CRM 1 mm and/or positive Ly, 22.2%) than in Group A (both CRM > 1 mm and negative Ly, 9.8%) (Fig. 2b, HR, 2.9; 95% CI, 1.450–5.980; P ¼ 0.002). For patients with stage IIIC disease, the cumulative 5‐year LR rate was significantly higher in Group B (CRM 1 mm and/or positive Ly, 34.3%) than in Group A (both CRM > 1 mm and negative Ly, 8.9%) (Fig. 2c, HR, 4.7; 95% CI, 1.390– 15.820; P ¼ 0.006).
DISCUSSION
Fig. 2. Cumulative local recurrence (LR) after curative resection in each stage. DME, distance of mesorectal extension; CRM, circumferential resection margin. a: Stage II disease. The cumulative 5‐year LR rate is significantly higher in Group B with DME > 4 mm and/ or positive V than in Group A with both DME 4 mm and negative V. b: Stage IIIB disease. The cumulative 5‐year LR rate is significantly higher in Group B with CRM 1 mm and/or positive Ly than in Group A with both CRM > 1 mm and negative Ly. c: Stage IIIC disease. The cumulative 5‐year LR rate is significantly higher in Group B with CRM 1 mm and/or positive Ly than in Group A with both CRM > 1 mm and negative Ly. Journal of Surgical Oncology
The incidence rate of LR after TME with or without adjuvant therapy over the past two decades has varied widely [4,5]. Various reasons can be considered to explain this variation. The definition of LR may be different (pelvic only versus pelvic recurrence with concomitant metastatic disease) [13–15]. The length of follow‐up is also different [13]. The cause of LR may be more closely related to tumor characteristics than to surgical or technical failure [16]. Some studies that identified tumor characteristics in patients with a high follow‐up rate may have provided more reliable data regarding LR after TME alone [17,18]. Incomplete reports of histopathological findings [19], a lack of stage‐specific analysis, and insufficient numbers of patients in the cohort are also thought to be problematic. The present study used a large enough prospective patient cohort, precise definitions of histopathological variables, and conducted long‐ term follow‐up with a low dropout rate. LR in this study was defined as any tumor located within the pelvis, either alone or in conjunction with metastases at the time of diagnosis of recurrence [15]. The incidence of LR was similar to other reports of LR (7% to 10.9%) after TME alone [17,18,20]. The LR rate did not decrease significantly in the latter half when surgical and diagnostic techniques improved. However, the stage‐specific LR rate was significantly different when comparing patients with stage II and stage IIIB/IIIC disease, though the pT‐specific LR rate was not significantly different. This shows that lymph node metastasis has a major impact on outcome, since the TNM stage consists of the combination of T‐stage and lymph node metastasis grade. It is necessary to examine this in combination with other important factors such as histopathological variables that could predict LR. The results of this study demonstrated that patients at risk of LR can be clearly identified using a combination of these adverse factors in each stage. DME > 4 mm and/or positive V, which were independently associated with LR, in stage II patients and a CRM 1 mm and/or positive Ly, which were independent predictors for LR, in stage IIIB/ IIIC patients were strongly associated with LR. A previous Japanese, multi‐institutional study demonstrated that DME was a powerful independent risk factor for distant metastasis and LR [6,7], which is consistent with observations from the present study. Ly in colorectal cancer correlates well with disease stage and represents an independent prognostic factor after curative resection [21]. The presence of lymphovascular invasion, extramural V, and a positive CRM are important independent predictive factors for the development of LR [22]. Recent studies, however, have suggested that negative CRM with complete resection for locally advanced rectal cancer was not a predictor of LR, and that positive CRM was also not associated with LR [23]. We must try to take enough CRM for rectal cancer with suspected lymph node metastases. Assessment that includes a combination of factors such as the TNM classification would certainly be more accurate, even in the absence of a significant independent factor. Such patients should be candidates for adjuvant or neoadjuvant treatment. In Western countries, preoperative RT or CRT is generally given to patients with rectal cancer accompanied by T3 and/or lymph node metastasis, which is based on reports showing that RT reduced LR and overall survival when compared with surgery alone [24]. However, prediction of the actual prognosis of patients after preoperative RT is not possible, because it is difficult to evaluate the initial pathological
Predictor of Local Recurrence in Rectal Cancer diagnosis. Recent evidence suggests that patients may be over‐treated, leading to unnecessary exposure to acute and long‐term toxicity of radiation therapy [20]. Selection of patients on the basis of pathological risk factors detected prior to operation or through imaging diagnosis may help reduce rates of radiation‐induced complications [24], anal dysfunction [25], and sexual disorders [26] Neoadjuvant treatment certainly offers the advantage of downstaging and controlling micrometastasis; however, in some cases, it may interfere with the correct timing of an operation. Morodomi et al. [27] used histopathological studies to describe the relationship between Ly and tumor budding and metastasis to lymph nodes in patients with rectal cancer. Ueno et al. [28] also showed that tumor budding could be accurately estimated by transanal submucosal biopsy. Taken together, these studies suggest that it may be possible to preoperatively identify patients with lower rectal cancer who are at high risk of LR after surgery. Sufficient tissue can be obtained transanally when the tumor is located near the anus. This preoperative submucosal biopsy may provide important pathologic information. A prospective study is needed for decision‐making using this method for tumors near the anus. The high diagnostic capability of MRI for extramural invasion was also very useful both in the preoperative assessment of tumor extension and in the planning of surgical treatment [29,30]. Thus, the progress of recent image analysis enables evaluation of histological change by micro‐units. However, LR is due to not only tumor characteristics, but also surgical skill; therefore, making decisions regarding preoperative treatment selection requires a multidisciplinary approach. Postoperative adjuvant therapy was also examined, although the data are not shown. Overall, 64% of Stage III cases received adjuvant therapy. Some oral fluoropyrimidate drugs were administered during this period. No significant differences were observed in LR‐free survival between the adjuvant and the non‐adjuvant therapy groups. The type of drug used and its administration period varied from physician to physician, leading to a number of different regimens and requirements. However, the drugs permitted for the treatment of colorectal cancer at that time were limited, and thus only a small variation was observed. The present study showed that detailed pathologic characterization of surgical specimens can help guide the selection of appropriate postoperative treatment and follow‐up. These results warrant a validation study with larger data sets or prospective trials to facilitate formulation of personalized preoperative and postoperative treatment strategies.
CONCLUSIONS In conclusion, DME, CRM, positive Ly, and positive V are independent risk factors for LR in patients who undergo curative resection for rectal cancer without preoperative CRT. Moreover, a combination of these factors can identify a group of patients who are at high risk of LR. This information may help guide stage‐specific treatment decisions and allow stratification of patients for neoadjuvant or adjuvant therapy.
REFERENCES 1. Heald RJ, Moran BJ, Ryall RD, et al.: Rectal cancer: The Basingstoke experience of total mesorectal excision, 1978–1997. Arch Surg 1998;133:894–899. 2. Enker WE, Thaler HT, Cranor ML, et al.: Total mesorectal excision in the operative treatment of carcinoma of the rectum. J Am Coll Surg 1995;181:335–346. 3. Hall NR, Finan PJ, al‐Jaberi T, et al.: Circumferential margin involvement after mesorectal excision of rectal cancer with curative intent. Predictor of survival but not local recurrence? Dis Colon Rectum 1998;41:979–983. Journal of Surgical Oncology
743
4. MacCall JL, Cox MR, Wattchow DA: Analysis of local recurrence rate after surgery alone for rectal cancer. Int J Colorect Dis 1995; 10:126–132. 5. Enríquez‐Navascués JM, Borda N, Lizerazu A, et al.: Patterns of local recurrence in rectal cancer after a multidisciplinary approach. World J Gastroenterol 2011;17:1674–1684. 6. Shirouzu K, Akagi Y, Fujita S, et al.: Clinical significance of the mesorectal extension of rectal cancer: A Japanese multi‐institutional study. Ann Surg 2011;253:704–710. 7. Akagi Y, Shirouzu K, Fujita S, et al.: Predicting oncologic outcomes by stratifying mesorectal extension in patients with pT3 rectal cancer: A Japanese multi‐institutional study. Int J Cancer 2012;131: 1220–1227. 8. Sobin LH, Gaspodarowicz MK, Wittekind C: TNM Classification of Malignant Tumors, 7th edition. New York, NY: Wiley— Blackwell; 2009. 9. Shirouzu K, Isomoto H, Kakegawa T, et al.: A prospective clinicopathological study of venous invasion in colorectal cancer. Am J Surg 1991;162:216–222. 10. Shirouzu K, Isomoto H, Morodomi T, et al.: Carcinomatous lymphatic permeation.Prognostic significance in patients with rectal carcinoma—A long term prospective study. Cancer 1995;75:4–10. 11. Shirouzu K, Isomoto H, Kakegawa T: Prognostic evaluation of perineural invasion in rectal cancer. Am J Surg 1993;165:233–237. 12. Japanese Society for Cancer of the Colon and Rectum. Japanese classification of colorectal carcinoma, Second English Edition. Tokyo, Japan: Kanehara & Co.; 2009. 13. Carlsson U, Lasson Å, Fkelund G: Recurrence rates after curative surgery for rectal carcinoma, with special reference to their accuracy. Dis Colon Rectum 1987;30:431–434. 14. Abulafi AM, Williams NS: Local recurrence of colorectal cancer: The problem, mechanism, management and adjuvant therapy. Br J Surg 1994;81:7–19. 15. Marsh PJ, James RD, Schofield PF: Definition of local recurrence after surgery for rectal carcinoma. Br J Surg 1995;82:465–468. 16. Di Cataldo A, Scilletta B, Latino R, et al.: The surgeon as a prognostic factor in the surgical treatment of rectal cancer. Surg Oncol 2007;1:S53–S56. 17. Zaheer S, Pemberton JH, Farouk R, et al.: Surgical treatment of adenocarcinoma of the rectum. Ann Surg 1998;227:800–811. 18. Peeters KC, Marijnen CA, Nagtegaal ID, et al.: The TME trial after a median follow‐up of 6 years: Increased local control but no survival benefit in irradiated patients with resectable rectal carcinoma. Ann Surg 2007;246:693–701. 19. Quirke P, Durdey P, Dixson MF, et al.: Local recurrence of rectal adenocarcinoma due to inadequate surgical resection. Histopathological study of internal tumour spread and surgical excision. Lancet 1986;328:996–999. 20. Frasson M, Garcia‐Granero E, Roda D, et al.: Preoperative chemoradiation may not always be needed for patients with T3 and T2Nþ rectal cancer. Cancer 2011;117:3118–3125. 21. Akagi Y, Adachi Y, Ohchi T, et al.: Prognostic impact of lymphatic invasion of colorectal cancer: A single‐center analysis of 1,616 patients over 24 years. Anticancer Res 2013;33:2965–2970. 22. Dresen RC, Peters EE, Rutten HJ, et al.: Local recurrence in rectal cancer can be predicted by histopathological factors. Eur J Surg Oncol 2009;35:1071–1077. 23. Reshef A, Lavery I, Kiran RP: Factors associated with oncologic outcomes after abdominoperineal resection compared with restorative resection for low rectal cancer: Patient‐ and tumor‐related or technical factors only? Dis Colon Rectum 2012;55:51–58. 24. Kapiteijn E, Marijnen CA, Nagtegaal ID, et al.: Preoperative radiotherapy combined with total mesorectal excision for resectable rectal cancer. N Engl J Med 2001;345:638–646. 25. Hassan I, Larson DW, Wolff BG, et al.: Impact of pelvic radiotherapy on morbidity and durability of sphincter preservation after coloanal anastomosis for rectal cancers. Dis Colon Rectum 2008;51:32–37. 26. Bonnel C, Parc YR, Pocard M, et al.: Effects of preoperative radiotherapy for primary resectable rectal adenocarcinoma on male sexual and urinary function. Dis Colon Rectum 2002;45: 934–939.
744
Akagi et al.
27. Morodomi T, Isomoto H, Shirouzu K, et al.: An index for estimating the probability of lymph node metastasis in rectal cancers. Lymph node metastasis and histopathology of actively invasive regions of cancer. Cancer 1989;63:539–543. 28. Ueno H, Mochizuki H, Shinto E, et al.: Histologic indices in biopsy specimens for estimating the probability of extended local spread in patients with rectal carcinoma. Cancer 2002;94:2882–2891.
Journal of Surgical Oncology
29. Beets‐Tan RG, Beets GL, Vliegen RF, et al.: Accuracy of magnetic resonance imaging in prediction of tumour‐free resection margin in rectal cancer surgery. Lancet 2001;357: 497–504. 30. Salerno GV, Daniels IR, Moran BJ, et al.: Magnetic resonance imaging prediction of an involved surgical resection margin in low rectal cancer. Dis Colon Rectum 2009;52:632–639.
Histopathological Predictors for Local Recurrence in Patients with T3 and T4 Rectal Cancers Without Preoperative Chemoradiotherapy YOSHITO AKAGI, MD, PhD,* TORU HISAKA, MD, PhD, TOMOAKI MIZOBE, MD, PhD, TETSUSHI KINUGASA, MD, PhD, YUTAKA OGATA, MD, PhD, AND KAZUO SHIROUZU, MD, PhD Department of Surgery, Kurume University School of Medicine, Kurume, Japan
Background and Objectives: Identification of suitable predictors of local recurrence (LR) in patients with rectal cancer would be of clinical benefit. The aim of this study was to identify histopathological factors that could predict LR. Methods: A total of 796 stage II/III patients with pT3 and pT4 rectal cancer who did not undergo preoperative chemoradiation were enrolled. LR was defined as intra‐pelvic recurrence only. Histopathological factors related to LR were investigated. Results: LR was found in 25 patients (6.1%) with stage II and 54 patients (13.9%) with stage IIIB/IIIC. In patients with stage II, distance of mesorectal extension (DME) >4 mm (P ¼ 0.011) and positive venous invasion (P ¼ 0.035) were independent factors that predicted LR. In patients with stage IIIB/IIIC, circumferential resection margin (CRM) 1 mm (P ¼ 0.003) and positive lymphatic invasion (P ¼ 0.006) were independent factors. The cumulative 5‐year LR rate was higher (11.9%) in patients with a combination of DME > 4 mm and/or positive venous invasion for stage II (P < 0.001), and was also higher in patients with a combination of CRM1 mm and/or positive lymphatic invasion for stage IIIB and IIIC (22.2%; P < 0.002, and 34.3%; P < 0.006, respectively). Conclusions: Important histopathological predictors for LR in patients with pT3 and pT4 rectal cancer were different at each stage.
J. Surg. Oncol. 2014;110:739–744. ß 2014 Wiley Periodicals, Inc.
KEY WORDS: risk factor; distance of mesorectal extension; circumferential resection margin; vessel invasion
INTRODUCTION Local recurrence (LR) is a critical problem for patients with rectal cancer after curative surgery. Since early detection of LR is difficult, patients with LR tend to have poor outcomes. Risk factors that can predict which patients are at high risk of LR have not yet been identified. Treatment for primary tumors with total mesorectal excision (TME) and a negative circumferential resection margin (CRM) is very important at the time of initial operation, because TME can reduce LR [1,2], and because a positive CRM is related to LR in rectal cancer patients [3]. The rate of LR, however, varies widely after TME with or without adjuvant treatment [4,5], and patients with a negative CRM can still develop LR. Thus, identification of other factors related to LR would be of benefit. However, additional clinicopathological factors that can predict stage‐specific LR are required to improve clinical decision‐making with respect to postoperative adjuvant treatment planning and follow‐up. A recent multicenter study demonstrated that the distance of mesorectal extension (DME) strongly correlated with distant metastasis and LR in every stage III case [6,7]. Therefore, the goal of the present study was to assess the relationship between traditional histopathological adverse factors, including CRM and DME, and LR after curative resection. Special focus was placed on the prediction of LR in stage II/III patients with pT3 and pT4 rectal cancer based on the International Union Against Cancer‐TNM Classification of Malignant Tumors (UICC‐TNM) 7th edition classification system [8].
PATIENTS AND METHODS Between 1986 and 2007, 796 consecutive, stage II/III patients with pT3 and pT4 rectal cancer underwent curative resection without preoperative or postoperative radiotherapy (RT) at Kurume University. All patients provided their written, informed consent after receiving a full explanation of the study. All protocols in this study were approved by the local institutional review board.
ß 2014 Wiley Periodicals, Inc.
The pathological diagnosis was made using whole‐mount sections of surgical specimens. Detailed clinical and pathological information was prospectively entered into our original computerized registration system (dBASE III Ashton‐Tate, Torrance, CA). All patients with histologically proven adenocarcinoma were extracted from the prospective database for this study. There were 408 patients with stage II disease and 388 with stage III disease, all of which were determined according to the pathological TNM classification 7th edition [8], in which node‐negative patients with pT3 and pT4 rectal cancer were classified as stage II. On the other hand, stage III disease followed a more complicated algorithm that took into account the number of lymph node metastases and the depth of invasion. According to the classification, patients with stage III disease were stratified into stage IIIB and IIIC for evaluation in this study. Baseline parameters related to the patients, tumors, operative methods, and investigated variables are summarized in Table I. The histopathological type of tumor was divided into two categories: “well” (well differentiated type of tubular adenocarcinoma) and “other” (moderately to poorly differentiated type of adenocarcinoma, mucinous adenocarcinoma, or signet ring cell carcinoma). Venous invasion (V) was defined as tumor cells within a space lined by endothelial cells and smooth muscle or elastic fibers, as shown in Figure 1a [9], and was categorized as negative or positive. Lymphatic invasion (Ly) was defined as occurring only when cancer cells were floating within an endothelial‐lined space and was categorized as negative or positive [10]
*Correspondence to: Yoshito Akagi, MD, PhD, Department of Surgery, Kurume University School of Medicine, 67 Asahi‐machi, Kurume City, Fukuoka 830‐0011, Japan. Fax: þ81‐942‐34‐0709. E‐mail: [email protected] Received 5 February 2014; Accepted 15 May 2014 DOI 10.1002/jso.23678 Published online 24 June 2014 in Wiley Online Library (wileyonlinelibrary.com).
740
Akagi et al.
TABLE I. Patient and Tumor Characteristics and Details of Treatment Characteristic Age (years) Sex Preoperative CEA (ng/ml) Gross type Location from AV (cm) Size of tumor (cm) T‐classification Number of LNs retrieved Number of LN metastasis TNM stage Tumor differentiation Venous invasion (V) Lymphatic invasion (Ly) Perineural invasion (PNI) Tumor budding DME (mm) CRM (mm) Surgical procedure SSO Lateral LND ANP TME Adjuvant chemotherapy
Category
No. of patients
65/>65 Male/female 5.0/>5.0 Unknown Expansive/infiltrative 5/5–10/>10 5/>5 pT3/pT4 Median (range) 4 >1/1
455/341 532/264 453/266 77 698/98 290/235/271 389/407 481/315 29 (2–122) 102/694 408/250/78/60 408/388 494/302 627/169 613/183 691/105 111/293a 384/412 728/68
Yes/no Yes/no Yes/no Yes/no Yes/no
552/244 385/411 514/284 560/236 493/303
CEA, carcinoembryonic antigen; AV, anal verge; LN, lymph node; Well, well differentiated adenocarcinoma; Mode, moderately differentiated adenocarcinoma; Others, poorly differentiated or mucinous adenocarcinoma; DME, distance of mesorectal extension; CRM, circumferential resection margin; SSO, sphincter saving operation; LND, lymph node dissection; ANP, autonomic nerve preservation; TME, total mesorectal excision. a Unknown for 392 cases.
(Fig. 1b). Perineural invasion (PNI) was defined as occurring when cancer cells were observed inside the perineurium and was categorized into negative and positive groups [11] (Fig. 1c). DME was directly measured on the glass slides of the pathological specimen. The distance from the outer border of the muscular layer to the deepest part of invasion was measured, and the cut‐off‐value was set to 4 mm according to a previous report [6] (Fig. 1d, D1). CRM was also pathologically measured as the closest distance from the surgical cutting line to viable cells on the glass slides (Fig. 1d, D2). Specimens were categorized as either positive CRM (distance 1 mm) or negative CRM (distance >1 mm). Cases with macroscopically positive CRM were not included in this study. Adjuvant chemotherapy was administered to some patients who had lymph node metastasis and were positive for Ly and/or V. 5‐ Fluorouracil‐based anti‐cancer agents including 1‐hexylcarbamoyl‐5‐ fluorouracil (HCFU), 50 ‐deoxy‐5‐fluorouridine (50 DFUR), or uracil‐ tegafur (UFT) were given in this period without specified protocols. 5‐Fluorouracil leucovorin (UZEL) or uracil‐tegafur‐gimeracil‐oteracil potassium (TS‐1) was preferentially used since 2000.
Follow‐Up and Oncologic Outcomes All patients were followed on an outpatient basis, and routine investigations were performed according to the rules of the JSCCR [12]. The most recent date of contact was regarded as the final date of confirmation in each case. The overall final follow‐up date for this study was the last day of March 2013. Patients were re‐staged according to the pathological TNM classification 7th edition [8]. Postoperative surveillance included measurement of serum carcinoembryonic antigen (CEA), and a rectal examination was performed every 3 months. Computed tomography (CT) and/or magnetic resonance imaging (MRI) was conducted every 6–12 months for the first 3 years, and then
Fig. 1. Histopathologic findings in the extramural fatty tissue adjacent to a colorectal tumor. a: Venous invasion is defined as tumor cells within a space lined by endothelial cells and smooth muscle or elastic fibers. b: Lymphatic invasion is defined as occurring only when cancer cells are floating within an endothelial‐lined space. c: Perineural invasion is defined as occurring when cancer cells are observed inside the perineurium. d: Measurement of mesorectal extension and circumferential resection margin is shown. D1: distance of mesorectal extension, D2: distance of circumferential resection margin. Journal of Surgical Oncology
Predictor of Local Recurrence in Rectal Cancer TABLE II. Local Recurrence According to the TNM Classification, 7th ed Characteristic
P‐value
No. of patients
No. of LR (%)
796
79 (9.9)
422 374
47 (11.1) 32 (8.6)
0.095
481 315
47 (9.8) 32 (10.2)
0.978
408 388
25 (6.1) 54 (13.9)
0.002
All patients Time period 1986–1999 (Non‐PET) 2000–2007 (PET) T stage pT3 pT4 pTNM stage II IIIB/IIIC
LR, local recurrence; PET, positron emission tomography.
annually, unless recurrence was suspected. Beginning in the year 2000, positron emission tomography (PET) was also conducted if a recurrence was suspected. The median follow‐up duration of the surviving patients was 122 months, with a range of 20–298 months. Follow‐up data for more than 60 months was 92% complete. The date and first site of recurrence were recorded for each patient. Pelvic floor or lateral lymph node metastasis was defined as intra‐pelvic LR. LR was defined as clinical or radiological tumor growth within the initial surgical field, with or without concomitant distant metastasis. For definitive diagnosis, image‐guided biopsy, salvage surgery, and/or palliative operation was performed whenever possible.
Statistical Analysis All independent variables listed in Table II were coded for statistical analysis. Statistical analysis was performed using StatView 5.0 and JMP 8.0 (SAS Institute, Inc., Cary, NC) for Windows. The Chi‐square test, Fisher’s exact test, and the Mann–Whitney U test were used to determine the statistical significance of any differences. Cox regression analysis was used to identify independent risk factors for LR. The Kaplan–Meier method was used to calculate survival rates. Any significant difference in survival was assessed using the log‐rank test. Statistical significance was indicated by a two‐tailed. P‐value < 0.05 and by 95% confidence intervals (CI).
741
and a PET group, in the year 2000, when PET became available. The LR rate was lower in the PET group than in the non‐PET group. The depth of invasion (T‐stage) was not a significant risk factor for LR (P ¼ 0.978). The rate of LR was significantly higher in patients with stage IIIB/IIIC disease (n ¼ 54; 13.9%) than in patients with stage II disease (n ¼ 25; 6.1%, P ¼ 0.002).
Risk Factors for LR Using Cox Regression Analysis Parameters listed in Table I were investigated as risk factors for LR using univariate Cox regression analysis. Variables in each stage that showed a significant association with the risk of LR on univariate Cox regression analysis were subsequently subjected to multivariate Cox regression analysis. On univariate Cox regression analysis for patients with stage II disease, the rate of LR was significantly higher in patients with positive V (odds ratio [OR], 3.6; 95% CI, 1.192–2.948; P ¼ 0.003), positive PNI (OR, 4.3; 95% CI, 1.204–2.626; P ¼ 0.004), and DME > 4 mm (OR, 11.3; 95% CI, 1.182–2.544; P ¼ 0.001). Other parameters listed in Table I were not significant risk factors for LR (data not shown). Similarly, in patients with stage IIIB/IIIC disease, moderate to poorly differentiated adenocarcinoma (OR, 2.0; 95% CI; 1.100– 3.832; P ¼ 0.013), positive Ly (OR, 2.9; 95% CI, 1.789–5.306; P ¼ 0.0001), positive V (OR, 2.5; 95% CI, 1.436–3.737; P ¼ 0.001), positive PNI (OR, 3.0; 95% CI, 1.785‐5.254; P ¼ 0.0001), DME > 4 mm (OR, 3.8; 95% CI, 1.823–3.136; P ¼ 0.0006), and CRM 1 mm (OR, 4.1; 95% CI, 2.912–5.124; P < 0.0001) were significant risk factors for LR (data not shown). These above‐mentioned significant parameters were investigated as independent predictors for LR by multivariate Cox regression analysis (Table III). DME > 4 mm was the most powerful independent predictor for LR in patients with stage II disease (HR, 3.3; 95% CI, 1.308–8.270; P ¼ 0.011), and positive V was also identified as an independent risk factor for LR (P ¼ 0.035). For patients with stage IIIB/IIIC disease, CRM 1 mm was the most powerful adverse factor for LR (HR, 2.6; 95% CI, 1.369–4.664; P ¼ 0.003). Moreover, positive Ly (HR, 2.6; 95% CI, 1.310–5.162; P ¼ 0.006) and DME > 4 mm (HR, 2.5; 95% CI, 1.123–5.722; P ¼ 0.025) were identified as independent risk factors for LR.
RESULTS
Cumulative LR Rate
LR by TNM Stage
Combining the two most powerful pathological risk factors identified on multivariate analysis could identify groups at high risk for LR in each stage: DME > 4 mm and/or positive V in stage II disease, and CRM 1 mm and/or positive Ly in stage IIIB/IIIC disease.
Of 796 patients, 79 (9.9%) had LR with or without distant metastasis (Table II). The cases were classified into two groups, a non‐PET group
TABLE III. Multivariate Cox Regression Analysis of Potential Risk Factors for Local Recurrence in Patients With Stage II/III Rectal Cancer Stage II (n ¼ 408) Variable
Category
Patients
HR (95% CI)
Tumor differentiation
Well Other Negative Positive Negative Positive Negative Positive 4.0 mm >4.0 mm >1.0 mm 1.0 mm
289 119 366 42 354 54 383 25 248 160 382 26
— — — — 1 2.5 (1.065–6.009) 1 2.1 (0.766–5.998) 1 3.3 (1.308–8.270) — —
Lymphatic invasion (Ly) Venous invasion (V) Perineural invasion (PNI) DME CRM
Stage IIIB/IIIC (n ¼ 388) P‐value
0.035 0.146 0.011
Patients
HR (95% CI)
205 183 247 141 273 115 308 80 136 252 346 42
1 0.8 (0.436–1.633) 1 2.6 (1.310–5.162) 1 1.1 (0.567–2.226) 1 1.6 (0.815–3.058) 1 2.5 (1.123–5.722) 1 2.6 (1.369–4.664)
P‐value
0.615 0.006 0.738 0.176 0.025 0.003
HR, hazard ratio; CI, confidence interval; AV, anal verge; Well, well differentiated adenocarcinoma; Others, moderately differentiated, poorly differentiated, or mucinous adenocarcinoma; DME, distance of mesorectal extension; CRM, circumferential resection margin.
Journal of Surgical Oncology
742
Akagi et al.
For patients with stage II disease, the cumulative 5‐year LR rate of Group B (DME > 4 mm and/or positive V; 11.9%) was significantly higher than that of Group A (both DME 4 mm and negative V; 1.6%) (Fig. 2a, HR, 7.3; 95% CI, 2.513–21.349; P ¼ 0.0007). For patients with stage IIIB disease, the cumulative 5‐year LR rate was significantly
higher in Group B (CRM 1 mm and/or positive Ly, 22.2%) than in Group A (both CRM > 1 mm and negative Ly, 9.8%) (Fig. 2b, HR, 2.9; 95% CI, 1.450–5.980; P ¼ 0.002). For patients with stage IIIC disease, the cumulative 5‐year LR rate was significantly higher in Group B (CRM 1 mm and/or positive Ly, 34.3%) than in Group A (both CRM > 1 mm and negative Ly, 8.9%) (Fig. 2c, HR, 4.7; 95% CI, 1.390– 15.820; P ¼ 0.006).
DISCUSSION
Fig. 2. Cumulative local recurrence (LR) after curative resection in each stage. DME, distance of mesorectal extension; CRM, circumferential resection margin. a: Stage II disease. The cumulative 5‐year LR rate is significantly higher in Group B with DME > 4 mm and/ or positive V than in Group A with both DME 4 mm and negative V. b: Stage IIIB disease. The cumulative 5‐year LR rate is significantly higher in Group B with CRM 1 mm and/or positive Ly than in Group A with both CRM > 1 mm and negative Ly. c: Stage IIIC disease. The cumulative 5‐year LR rate is significantly higher in Group B with CRM 1 mm and/or positive Ly than in Group A with both CRM > 1 mm and negative Ly. Journal of Surgical Oncology
The incidence rate of LR after TME with or without adjuvant therapy over the past two decades has varied widely [4,5]. Various reasons can be considered to explain this variation. The definition of LR may be different (pelvic only versus pelvic recurrence with concomitant metastatic disease) [13–15]. The length of follow‐up is also different [13]. The cause of LR may be more closely related to tumor characteristics than to surgical or technical failure [16]. Some studies that identified tumor characteristics in patients with a high follow‐up rate may have provided more reliable data regarding LR after TME alone [17,18]. Incomplete reports of histopathological findings [19], a lack of stage‐specific analysis, and insufficient numbers of patients in the cohort are also thought to be problematic. The present study used a large enough prospective patient cohort, precise definitions of histopathological variables, and conducted long‐ term follow‐up with a low dropout rate. LR in this study was defined as any tumor located within the pelvis, either alone or in conjunction with metastases at the time of diagnosis of recurrence [15]. The incidence of LR was similar to other reports of LR (7% to 10.9%) after TME alone [17,18,20]. The LR rate did not decrease significantly in the latter half when surgical and diagnostic techniques improved. However, the stage‐specific LR rate was significantly different when comparing patients with stage II and stage IIIB/IIIC disease, though the pT‐specific LR rate was not significantly different. This shows that lymph node metastasis has a major impact on outcome, since the TNM stage consists of the combination of T‐stage and lymph node metastasis grade. It is necessary to examine this in combination with other important factors such as histopathological variables that could predict LR. The results of this study demonstrated that patients at risk of LR can be clearly identified using a combination of these adverse factors in each stage. DME > 4 mm and/or positive V, which were independently associated with LR, in stage II patients and a CRM 1 mm and/or positive Ly, which were independent predictors for LR, in stage IIIB/ IIIC patients were strongly associated with LR. A previous Japanese, multi‐institutional study demonstrated that DME was a powerful independent risk factor for distant metastasis and LR [6,7], which is consistent with observations from the present study. Ly in colorectal cancer correlates well with disease stage and represents an independent prognostic factor after curative resection [21]. The presence of lymphovascular invasion, extramural V, and a positive CRM are important independent predictive factors for the development of LR [22]. Recent studies, however, have suggested that negative CRM with complete resection for locally advanced rectal cancer was not a predictor of LR, and that positive CRM was also not associated with LR [23]. We must try to take enough CRM for rectal cancer with suspected lymph node metastases. Assessment that includes a combination of factors such as the TNM classification would certainly be more accurate, even in the absence of a significant independent factor. Such patients should be candidates for adjuvant or neoadjuvant treatment. In Western countries, preoperative RT or CRT is generally given to patients with rectal cancer accompanied by T3 and/or lymph node metastasis, which is based on reports showing that RT reduced LR and overall survival when compared with surgery alone [24]. However, prediction of the actual prognosis of patients after preoperative RT is not possible, because it is difficult to evaluate the initial pathological
Predictor of Local Recurrence in Rectal Cancer diagnosis. Recent evidence suggests that patients may be over‐treated, leading to unnecessary exposure to acute and long‐term toxicity of radiation therapy [20]. Selection of patients on the basis of pathological risk factors detected prior to operation or through imaging diagnosis may help reduce rates of radiation‐induced complications [24], anal dysfunction [25], and sexual disorders [26] Neoadjuvant treatment certainly offers the advantage of downstaging and controlling micrometastasis; however, in some cases, it may interfere with the correct timing of an operation. Morodomi et al. [27] used histopathological studies to describe the relationship between Ly and tumor budding and metastasis to lymph nodes in patients with rectal cancer. Ueno et al. [28] also showed that tumor budding could be accurately estimated by transanal submucosal biopsy. Taken together, these studies suggest that it may be possible to preoperatively identify patients with lower rectal cancer who are at high risk of LR after surgery. Sufficient tissue can be obtained transanally when the tumor is located near the anus. This preoperative submucosal biopsy may provide important pathologic information. A prospective study is needed for decision‐making using this method for tumors near the anus. The high diagnostic capability of MRI for extramural invasion was also very useful both in the preoperative assessment of tumor extension and in the planning of surgical treatment [29,30]. Thus, the progress of recent image analysis enables evaluation of histological change by micro‐units. However, LR is due to not only tumor characteristics, but also surgical skill; therefore, making decisions regarding preoperative treatment selection requires a multidisciplinary approach. Postoperative adjuvant therapy was also examined, although the data are not shown. Overall, 64% of Stage III cases received adjuvant therapy. Some oral fluoropyrimidate drugs were administered during this period. No significant differences were observed in LR‐free survival between the adjuvant and the non‐adjuvant therapy groups. The type of drug used and its administration period varied from physician to physician, leading to a number of different regimens and requirements. However, the drugs permitted for the treatment of colorectal cancer at that time were limited, and thus only a small variation was observed. The present study showed that detailed pathologic characterization of surgical specimens can help guide the selection of appropriate postoperative treatment and follow‐up. These results warrant a validation study with larger data sets or prospective trials to facilitate formulation of personalized preoperative and postoperative treatment strategies.
CONCLUSIONS In conclusion, DME, CRM, positive Ly, and positive V are independent risk factors for LR in patients who undergo curative resection for rectal cancer without preoperative CRT. Moreover, a combination of these factors can identify a group of patients who are at high risk of LR. This information may help guide stage‐specific treatment decisions and allow stratification of patients for neoadjuvant or adjuvant therapy.
REFERENCES 1. Heald RJ, Moran BJ, Ryall RD, et al.: Rectal cancer: The Basingstoke experience of total mesorectal excision, 1978–1997. Arch Surg 1998;133:894–899. 2. Enker WE, Thaler HT, Cranor ML, et al.: Total mesorectal excision in the operative treatment of carcinoma of the rectum. J Am Coll Surg 1995;181:335–346. 3. Hall NR, Finan PJ, al‐Jaberi T, et al.: Circumferential margin involvement after mesorectal excision of rectal cancer with curative intent. Predictor of survival but not local recurrence? Dis Colon Rectum 1998;41:979–983. Journal of Surgical Oncology
743
4. MacCall JL, Cox MR, Wattchow DA: Analysis of local recurrence rate after surgery alone for rectal cancer. Int J Colorect Dis 1995; 10:126–132. 5. Enríquez‐Navascués JM, Borda N, Lizerazu A, et al.: Patterns of local recurrence in rectal cancer after a multidisciplinary approach. World J Gastroenterol 2011;17:1674–1684. 6. Shirouzu K, Akagi Y, Fujita S, et al.: Clinical significance of the mesorectal extension of rectal cancer: A Japanese multi‐institutional study. Ann Surg 2011;253:704–710. 7. Akagi Y, Shirouzu K, Fujita S, et al.: Predicting oncologic outcomes by stratifying mesorectal extension in patients with pT3 rectal cancer: A Japanese multi‐institutional study. Int J Cancer 2012;131: 1220–1227. 8. Sobin LH, Gaspodarowicz MK, Wittekind C: TNM Classification of Malignant Tumors, 7th edition. New York, NY: Wiley— Blackwell; 2009. 9. Shirouzu K, Isomoto H, Kakegawa T, et al.: A prospective clinicopathological study of venous invasion in colorectal cancer. Am J Surg 1991;162:216–222. 10. Shirouzu K, Isomoto H, Morodomi T, et al.: Carcinomatous lymphatic permeation.Prognostic significance in patients with rectal carcinoma—A long term prospective study. Cancer 1995;75:4–10. 11. Shirouzu K, Isomoto H, Kakegawa T: Prognostic evaluation of perineural invasion in rectal cancer. Am J Surg 1993;165:233–237. 12. Japanese Society for Cancer of the Colon and Rectum. Japanese classification of colorectal carcinoma, Second English Edition. Tokyo, Japan: Kanehara & Co.; 2009. 13. Carlsson U, Lasson Å, Fkelund G: Recurrence rates after curative surgery for rectal carcinoma, with special reference to their accuracy. Dis Colon Rectum 1987;30:431–434. 14. Abulafi AM, Williams NS: Local recurrence of colorectal cancer: The problem, mechanism, management and adjuvant therapy. Br J Surg 1994;81:7–19. 15. Marsh PJ, James RD, Schofield PF: Definition of local recurrence after surgery for rectal carcinoma. Br J Surg 1995;82:465–468. 16. Di Cataldo A, Scilletta B, Latino R, et al.: The surgeon as a prognostic factor in the surgical treatment of rectal cancer. Surg Oncol 2007;1:S53–S56. 17. Zaheer S, Pemberton JH, Farouk R, et al.: Surgical treatment of adenocarcinoma of the rectum. Ann Surg 1998;227:800–811. 18. Peeters KC, Marijnen CA, Nagtegaal ID, et al.: The TME trial after a median follow‐up of 6 years: Increased local control but no survival benefit in irradiated patients with resectable rectal carcinoma. Ann Surg 2007;246:693–701. 19. Quirke P, Durdey P, Dixson MF, et al.: Local recurrence of rectal adenocarcinoma due to inadequate surgical resection. Histopathological study of internal tumour spread and surgical excision. Lancet 1986;328:996–999. 20. Frasson M, Garcia‐Granero E, Roda D, et al.: Preoperative chemoradiation may not always be needed for patients with T3 and T2Nþ rectal cancer. Cancer 2011;117:3118–3125. 21. Akagi Y, Adachi Y, Ohchi T, et al.: Prognostic impact of lymphatic invasion of colorectal cancer: A single‐center analysis of 1,616 patients over 24 years. Anticancer Res 2013;33:2965–2970. 22. Dresen RC, Peters EE, Rutten HJ, et al.: Local recurrence in rectal cancer can be predicted by histopathological factors. Eur J Surg Oncol 2009;35:1071–1077. 23. Reshef A, Lavery I, Kiran RP: Factors associated with oncologic outcomes after abdominoperineal resection compared with restorative resection for low rectal cancer: Patient‐ and tumor‐related or technical factors only? Dis Colon Rectum 2012;55:51–58. 24. Kapiteijn E, Marijnen CA, Nagtegaal ID, et al.: Preoperative radiotherapy combined with total mesorectal excision for resectable rectal cancer. N Engl J Med 2001;345:638–646. 25. Hassan I, Larson DW, Wolff BG, et al.: Impact of pelvic radiotherapy on morbidity and durability of sphincter preservation after coloanal anastomosis for rectal cancers. Dis Colon Rectum 2008;51:32–37. 26. Bonnel C, Parc YR, Pocard M, et al.: Effects of preoperative radiotherapy for primary resectable rectal adenocarcinoma on male sexual and urinary function. Dis Colon Rectum 2002;45: 934–939.
744
Akagi et al.
27. Morodomi T, Isomoto H, Shirouzu K, et al.: An index for estimating the probability of lymph node metastasis in rectal cancers. Lymph node metastasis and histopathology of actively invasive regions of cancer. Cancer 1989;63:539–543. 28. Ueno H, Mochizuki H, Shinto E, et al.: Histologic indices in biopsy specimens for estimating the probability of extended local spread in patients with rectal carcinoma. Cancer 2002;94:2882–2891.
Journal of Surgical Oncology
29. Beets‐Tan RG, Beets GL, Vliegen RF, et al.: Accuracy of magnetic resonance imaging in prediction of tumour‐free resection margin in rectal cancer surgery. Lancet 2001;357: 497–504. 30. Salerno GV, Daniels IR, Moran BJ, et al.: Magnetic resonance imaging prediction of an involved surgical resection margin in low rectal cancer. Dis Colon Rectum 2009;52:632–639.
