Oncologic Outcomes

ORIGINAL ARTICLES Oncologic Outcomes After Neoadjuvant Chemoradiation Followed by Curative Resection With Tumor-Specific Mesorectal Excision for Fixed Locally Advanced Rectal Cancer Impact of Postirradiated Pathologic Downstaging on Local Recurrence and Survival Nam Kyu Kim, MD, PhD,*† Seung Hyuk Baik, MD,*† Jin Sil Seong, MD, PhD,*§ Hoguen Kim, MD, PhD,*� Jae Kyung Roh, MD, PhD,*‡ Kang Young Lee, MD,*† Seung Kook Sohn, MD, PhD,† and Chang Hwan Cho, MD, PhD*† Objective: The purpose of this study was to determine the oncologic outcomes and clinical factors affecting survival in patients who underwent neoadjuvant chemoradiotherapy following tumor specific mesorectal excision for locally advanced, fixed rectal cancer. Summary Background Data: Neoadjuvant chemoradiation therapy has resulted in significant tumor downstaging, which enhances curative resection and subsequently improves local disease control for rectal cancer. However, oncologic outcomes, according to clin- ical factors, have not yet been fully understood in locally advanced and fixed rectal cancer. Methods: A total of 114 patients who had undergone neoadjuvant chemoradiation for advanced rectal cancer (T3 or T4 and node positive) were investigated retrospectively. Chemotherapy was ad- ministered intravenously with 5-FU and leucovorin during weeks 1 and 5 of radiotherapy. The total radiation dose was 5040 cGY in 25 fractions delivered over 5 weeks. Tumor-specific mesorectal exci- sion was done 4 to 6 weeks after the completion of neoadjuvant chemoradiation. Survival and recurrence rates, according to the pathologic stage, were evaluated. Moreover, factors affecting sur- vival were investigated. Results: The 5-year survival rates according to pathologic stage were: 100% in pathologic complete remission (n � 10), 80% in stage I (n � 23), 56.8% in stage II (n � 34), and 42.3% in stage III (n � 47) (P � 0.0000). Local, systemic, and combined recurrence rates were 11.4%, 22.8%, and 3.5%, respectively. Multivariate analysis showed that the pathologic N stage and operation method were the independent factors affecting survival rate. Conclusion: Pathologic complete remission showed excellent on- cologic outcomes, and the pathologic N stage was the most impor- tant factor for oncologic outcomes. (Ann Surg 2006;244: 1024–1030) Continued efforts have been made to improve local controland survival in rectal cancer. Postoperative adjuvant chemoradiation therapy has been recommended and has be- come the standard of care for stage II and stage III rectal adenocarcinomas.1 With the introduction of pelvic sharp dissection and tumor specific mesorectal excision, the rate of local recurrence markedly reduced to 5% to 10% for resect- able rectal cancer.2,3 Total mesorectal excision, defined as a sharp pelvic dissection under clear vision with the excision of the rectum and mesorectum within the mesorectal fascia, has been emphasized. However, local recurrence still remains a problem in locally advanced, fixed rectal cancers. The current strategy for the management of locally advanced, fixed rectal cancer is the multimodality approach. Neoadjuvant chemora- diation therapy has been regarded as a good modality for increasing resectability, decreasing the rate of locoregional recurrence, and improving survival rate. In neoadjuvant chemoradiation therapy, the clinical implications of this modality for unresectable, fixed rectal cancer were: tumor size reduction; increased tumor mobil- ity, which enhances sphincter saving curative resection; and histopathologic downstaging, which is more important for long-term oncologic outcomes. A variable range of tumor responses has been observed from a complete re- sponse to nonresponse. Pathologic complete remission was reported to range from 8% to 29%, depending on the stage From the *Colorectal Cancer Clinic, Severance Hospital, Yonsei University Medical Center, and the Departments of †Surgery, ‡Medical Oncology, §Radiation Oncology, and �Pathology, Yonsei University College of Medicine, Seoul, Korea. Reprints: Nam Kyu Kim, MD, PhD, Department of Surgery, Division of Colorectal Surgery, Yonsei University College of Medicine, Seoul, Korea. E-mail: namkyuk@yumc.yonsei.ac.kr. Copyright © 2006 by Lippincott Williams & Wilkins ISSN: 0003-4932/06/24406-1024 DOI: 10.1097/01.sla.0000225360.99257.73 Annals of Surgery • Volume 244, Number 6, December 20061024 at presentation, regimen of chemotherapy, and dose of radiation administered.4–7 The status of downstaged pathologic stage after neoad- juvant chemoradiation was an important factor for oncologic outcomes. However, to date, few studies have investigated oncologic outcomes according to postirradiated, postoperative pathologic stages and clinical factors affecting survival in locally advanced and fixed rectal cancer. Therefore, the purpose of this study was to determine oncologic outcomes according to patho- logic stages and the clinical factors affecting survival in patients who underwent neoadjuvant chemoradiotherapy fol- lowing curative resection for locally advanced, fixed rectal cancer. MATERIALS AND METHODS Eligibility Between January 1989 and December 2000, 142 pa- tients diagnosed with locally advanced, unresectable carci- noma of the rectum were treated at Severance Hospital’s Department of Surgery at Yonsei University with a regimen of combined preoperative chemotherapy and radiotherapy, followed by surgical resection. The clinicopathologic data of these patients were accumulated prospectively and among these patients, a total of 114 patients who underwent curative resection were investigated retrospectively; 28 patients were excluded because positive circumferential resection margin was reported in 26 patients and positive distal resection margin was reported in 2 patients. The patients were confirmed as having an adenocarci- noma by sigmoidoscopic biopsy (and staged as T3 or T4), and regional lymph node enlargement by transrectal ultra- sonography (TRUS) and pelvic MRI. Most of the patients showed fixation to the rectal wall and/or invasion to the surrounding pelvic organs. Bulky and/or tethered tumors were staged by TRUS and pelvic MRI as T3 to T4 and showed enlarged lymph nodes (marginally or unresectable rectal cancer by digital rectal examination). The patients between January 1989 and December 1993 were staged by using TRUS only, and the patients between 1994 and 2000 were staged by using both TRUS and MRI. In the later period between 1994 and 2000, we excluded the patients who were staged lower than T3 to T4 and N� by only one of both examinations. All patients were evaluated with CT scan for evaluation of distant metastasis. Neoadjuvant Chemoradiotherapy Chemotherapy was administered intravenously with 5-fluorouracil, 425 mg/m2 per day and leucovorin, 20 mg/m2 per day during weeks 1 and 5 of radiotherapy. Radiation treatment was administered with a 6 MV/10 MV dual photon linear accelerator using a 4-field box technique. The total radiation dosage was 5040 cGY in 25 fractions delivered over 5 weeks. The radiation field was as follows: upper margin was 1.5 cm above the sacral promontory (L5 level), lateral margin was 1.5 cm laterally from the pelvic lateral wall, and lower margin was 3 cm below the lower margin of the tumor. Surgical Resection Surgery was performed 4 to 6 weeks after the comple- tion of chemoradiation. The method of operation was tumor specific mesorectal excision with pelvic autonomic nerve preservation. Tumor specific mesorectal excision was defined as the surgical method, which was transaction of the meso- rectum with the surrounding mesorectum enclosed by the rectal proper fascia at 4 cm distal from the lower edge of the rectum. Thus, all surgery was performed using sharp pelvic dissection under direct vision along the plane of the rectal proper fascia. Tumor-specific mesorectal excision was per- formed according to the tumor level. In upper rectal cancers, the mesorectum was excised 4 cm distal from the lower edge of the tumor. Total mesorectal excision was performed in the middle and distal rectal cancer. Curative resection defined no gross residual disease, and distal and circumferential resec- tion margins were negative in standard hematoxylin and eosin stain by histopathologic examination. Pathologic Evaluation The patients were staged according to the 6th UICC pTNM staging system after the final histopathologic exami- nation. Pathologic complete remission (pCR) was defined as the absence of residual microscopic tumors upon histopatho- logic examination. Pathologic evaluation was performed by one patholo- gist (H.K.), who was unaware of the clinical or radiologic findings. Postirradiated resection specimens were evaluated for depth of tumor penetration, lymph node metastases, and differentiation. Postoperative adjuvant systemic chemother- apy, consisting of 400 to 425 mg/m2 of 5-fluorouracil plus 20 mg/m2 leucovorin for 5 days, was administered every 28 days for 6 cycles in all enrolled patients. Follow-up Evaluation All patients have been closely followed by the sur- geons, medical oncologists, and radiation oncologists in the Colorectal Cancer Clinic at Severance Hospital. All patients were followed up every 3 months for 2 years after surgery. Clinical examination, measurement of serum carcinoembry- onic antigen levels, and chest x-ray were performed during each follow-up. After 2 years, patients were followed up every 6 months. Abdominopelvic CT and whole body bone scan were checked every year. Recurrences were documented by clinical and/or radiologic assessment and categorized as a local, systemic, or combined (local plus systemic) recurrence. Statistical Analyses Statistical analyses were performed using the statistical software package SPSS version 11.5 (SPSS, Chicago, IL). Survival and disease-free survival were calculated using the Kaplan-Meier method, and prognostic factors were compared with the log-rank test. Multivariate survival analysis was with the Cox proportional hazards mode, entering pathologic N stage, pathologic T stage, operation method, preoperative carcinoembryonic antigen, and histologic type. Cox propor- tional hazard model was done by a forward stepwise selection of variables, and a P value of 0.1 was adopted as the limit for Annals of Surgery • Volume 244, Number 6, December 2006 Neoadjuvant Chemoradiation for Rectal Cancer © 2006 Lippincott Williams & Wilkins 1025 inclusion of a covariant. Values of P � 0.05 were considered statistically significant. RESULTS Patient Characteristics Patients included 80 males and 34 females with a median age of 55.0 years (range, 26–81 years). Median follow-up period was 39.4 months (range, 2.2–151.6 months), and the rate of follow-up was 93%. The distribution of tumor location was: upper (n � 20, 17.5%), middle (n � 38, 33.3%), and lower (n � 56, 49.1%). The type of surgery performed was: abdominoperineal resection in 38 patients (33.3%), low anterior resection in 54 patients (47.3%), Hart- mann’s procedure in 17 patients (14.9%), ultra-low anterior resection with coloanal anastomosis in 2 patients (1.8%), and total pelvic exenteration in 3 patients (2.6%) (Table 1). Postoperative Complications Eleven patients had postoperative wound complications (9.6%). Anastomotic leakage was noted in 3 patients (2.6%) and stenosis in 1 patient (0.9%). Intestinal obstruction devel- oped in 3 patients (2.6%) (Table 2). Tumor Response Margin free resections without microscopic disease at the radial or pelvic side wall margin were achieved in all 114 patients. Pathologic complete response (pCR) of the primary rectal cancer was observed in 10 patients (8.8%). Three patients (2.6%) were downstaged to pT1N0M0, 20 patients (17.5%) to pT2N0M0, and 3 patients (2.6%) to pT2N1-2M0. Thirty-one patients (27.2%) were downstaged to stage II (T3N0M0). Forty-three patients (37.7%) were found to be stage III (T3N1-2M0), 3 patients (2.6%) in T4N0M0, and 1 patient (0.9%) in T4N1-2M0 (Table 3). Five-Year Survival and 5-Year Disease-Free Survival Rates The 5-year survival rate was 58.1%. The 5-year dis- ease-free survival was 52%. Five-year survival rates accord- ing to pathologic stage were: 100% in pathologic complete remission (pCR) (n � 10), 80% in stage I (n � 23), 56.8% in stage II (n � 34), and 42.3% in stage III (n � 47) (P � 0.0000) (Fig. 1). The 5-year disease-free survival rates, ac- cording to pathologic stage, were: 100% in pCR, 72.3% in stage I, 49.7% in stage II, and 33.6% in stage III (P� 0.0002) (Fig. 2). The 5-year disease-free survival rate, according to pathologic T stage, pathologic N stage, and operation meth- ods, is shown in Figures 3 to 5. Pattern of Recurrence Overall recurrence was noted in 44 patients (38.6%). The 5-year local recurrence rate was 0% in pCR, 9.6% in stage I, 10.3% in stage II, and 24% in stage III (P� 0.226) (Fig. 6). The 5-year systemic recurrence rate was 0% in pCR, 10.1% in stage I, 22.3% in stage II, and 42.8% in stage III (P � 0.009). The 5-year local and systemic recurrence was 0% in pCR and stage I, 4.0% in stage II, and 7.2% in stage III (P � 0.437). TABLE 1. Patient Characteristics (n � 114) No. (%) Sex Male 80 (70.2) Female 34 (29.8) Median age (yr) (range) 55.0 (26–81) Median follow-up period (mo) 39.4 (2.2–151.6) Follow-up rate (%) 93 Tumor level Upper 20 (17.5) Middle 38 (33.3) Lower 56 (49.1) Type of surgery Abdominoperineal resection 38 (33.3) Low anterior resection 54 (47.3) Hartmann’s procedure 17 (14.9) Ultra low anterior resection with coloanal anastomosis 2 (1.8) Total pelvic exenteration 3 (2.6) Pathologic stage Complete response 10 (8.8) Stage I 23 (20.2) Stage II 34 (29.8) Stage III 47 (41.2) TABLE 2. Postoperative Morbidity (n � 114) Morbidity No. (%) Wound infection 11 (9.6) Anastomotic leakage 3 (2.6) Intestinal obstruction 3 (2.6) Intraabdominal abscess 1 (0.9) Anastomotic stricture 1 (0.9) TABLE 3. Correlation of pT Stage and pN Stage (n � 114) Stage No. (%) Total No. (%) pT0 pN0 10 (8.8) 10 (8.8) pN1-2 0 (0) pT1 pN0 3 (2.6) 3 (2.6) pN1-2 0 (0) pT2 pN0 20 (17.5) 23 (20.2) pN1-2 3 (2.6) pT3 pN0 31 (27.2) 74 (64.9) pN1-2 43 (37.7) pT4 pN0 3 (2.6) 3 (3.5) pN1-2 1 (0.9) Kim et al Annals of Surgery • Volume 244, Number 6, December 2006 © 2006 Lippincott Williams & Wilkins1026 Prognostic Factors Affecting Survival Univariate analyses of factors affecting survival showed perioperative serum carcinoembryonic antigen level and pathologic T and N stages to be significant factors. Operation method and histologic differentiation were margin- ally significant factors (Table 4). Multivariate analysis of these factors showed that pathologic N stage and operation method were the independent factors affecting survival rates (Table 5). DISCUSSION Neoadjuvant multimodality therapy for advanced rectal adenocarcinoma continues to gain wide acceptance. Short course preoperative radiotherapy treatment combined total mesorectal excision for resectable rectal cancer improved local control more than TME alone. However, this therapy did not show any survival benefits.8 In addition, neoadjuvant chemoradiotherapy has been applied for resectable rectal cancer cases to improve anal sphincter preservation. When rectal cancer is fixed at the time of presentation, recurrence after surgical resection has been reported to range from 50% to 70% and is characterized by poor survival rates and early distant metastasis.9–11 Neoadjuvant chemoradiation therapy is now a well-established treatment of locally advanced middle to lower rectal cancer. In locally advanced rectal cancer, this multimodality therapy has been known to im- prove resectability, local control, and overall survival bene- fits. The ultimate goal of this treatment is to obtain tumor downstaging and tumor volume reduction which is related to curative surgical resection and long-term favorable oncologic FIGURE 1. Five-year survival rates according to stage (n � 114). FIGURE 2. Five-year disease-free survival rates according to stage (n � 114). FIGURE 3. Five-year disease-free survival rates according to pT stage (n � 114). FIGURE 4. Five-year disease-free survival rates according to node metastasis (n � 114). Annals of Surgery • Volume 244, Number 6, December 2006 Neoadjuvant Chemoradiation for Rectal Cancer © 2006 Lippincott Williams & Wilkins 1027 outcomes. Interestingly, we observed that more than a 50% tumor volume reduction shows no correlation with T and N staging.12 Therefore, improved resectability does not confer a pathologic downstaging effect. Some studies have reported on the oncologic outcomes regarding local control and dis- ease-free survival in resectable rectal cancer patients who received preoperative chemoradiation therapy.13–17 However, there have been few series regarding oncologic outcomes for patients with marginally resectable or unresectable rectal cancers diagnosed by digital rectal examination and assessed as stage T3 to T4 with multiple enlarged lymph nodes by transrectal ultrasonography and pelvic MRI.9,11 In this study, preoperative staging of the rectal cancer was performed by digital rectal examination, transrectal ultrasonography, and MRI. We considered all patients enrolled in our study to be homogenously categorized as fixed locally advanced rectal cancer staged at least T3 plus N positive. FIGURE 5. Five-year survival according to operation method. LAR, low anterior resection and ultra low anterior resection with coloanal anastomosis; APR, abdominoperineal resection and total pelvic exenteration (n � 114). FIGURE 6. Five-year local recurrence rate according to stage (n � 114). TABLE 4. Univariate Analysis of Prognostic Factors for 5-Year Survival (n � 114) No. 5-Year Survival (%) P Sex 0.214 Male 80 63.2 Female 34 54.1 Age 0.567 �55 yr 58 63.4 �55 yr 56 57.6 Histologic type 0.070 Well differentiated 7 66.7 Moderately differentiated 84 68.1 Poorly differentiated 13 47.2 Mucinous, others 10 22.5 Location of tumor 0.259 Upper 20 58.8 Mid 38 66.6 Lower 56 53.7 Pathologic N stage �0.001 N0 67 73.8 N1 29 47.6 N2 18 31.5 Pathologic T stage 0.005 T0 10 100 T1 3 100 T2 23 80.2 T3 74 49.5 T4 4 25.0 Operation method 0.069 LAR 41 77.4 APR 56 60.1 Hartmann’s procedure 17 33.0 Preoperative CEA �5 ng/mL 26 48.8 0.028 �5 ng/mL 88 64.1 LAR indicates low anterior resection and ultra low anterior resection with coloanal anastomosis; APR, abdominoperineal resection and total pelvic exenteration. TABLE 5. Multivariate Analysis of Prognostic Factors for 5-Year Survival (n � 114) RR 95% CI P Pathologic N stage �0.001 N0 1.000 N1 2.252 0.960–4.655 N2 6.495 3.698–21.231 Operation method 0.017 LAR 1.000 APR 2.866 1.265–6.494 Hartmann’s procedure 2.963 1.229–7.143 LAR indicates low anterior resection and ultra low anterior resection with coloanal anastomosis; APR, abdominoperineal resection and total pelvic exenteration. Kim et al Annals of Surgery • Volume 244, Number 6, December 2006 © 2006 Lippincott Williams & Wilkins1028 A previous study showed that the accuracy of TRUS and MRI for assessing the depth of invasion ranged from 81% to 95%.18,19 The accuracy of these methods to assess meta- static lymph nodes ranged from 63% to 87%.20,21 In our study, TRUS staging has been made since January 1989 and MRI staging has been made since January 1994. The period of this study was 11 years. Thus, there is little doubt that the reliability of the initial study groups staging may have varied with current techniques. However, in 1985, Hidebrandt and Feifel22 reported that accuracy of preoperative staging of rectal cancer TRUS was 92%. Moreover, Beynon et al20 reported the accuracy of TRUS was 91% in 1986. In 1989, accuracy of preoperative assessment of mesorectal lymph node involvement in rectal cancer was 83%.21 In 1999, Kim et al19 reported the overall accuracy