A Pathologic Complete Response of Rectal Cancer to Preoperative Combined-Modality Therapy Results in Improved Oncological Outcome Compared With Those Who Achieve No Downstaging on the Basis of Preoperative Endorectal Ultrasonography
- First Online:
- Cite this article as:
- Stipa, F., Chessin, D.B., Shia, J. et al. Ann Surg Oncol (2006) 13: 1047. doi:10.1245/ASO.2006.03.053
- 208 Downloads
Preoperative combined-modality therapy (CMT) is the preferred treatment for locally advanced rectal cancer (endorectal ultrasonography [ERUS] T3–4, N1, or clinically bulky) and achieves a pathologic complete response (pCR) in 4% to 33% of patients. However, the prognostic significance of pCR remains unclear.
A prospectively collected database was queried to identify 200 patients with locally advanced disease treated from 1992 to 2002. The pCR group was defined as having no evidence of viable tumor on pathologic analysis. The no-downstaging group was defined as no difference between the pre-CMT ERUS stage and the pathologic stage. Those achieving some downstaging but not pCR were excluded. Patients were treated with CMT (5040 cGy of radiation and 5-fluorouracil–based chemotherapy) followed by surgery, and 51 (85%) in the pCR group and 129 (92%) in the no-downstaging group (P = .1) received postoperative chemotherapy. Recurrence-free survival (RFS) and overall survival (OS) were determined by using the Kaplan-Meier method.
The median follow-up was 38.6 months (range, 18.2–124.9 months). The pCR (n = 60) and control (n = 140) groups were similar in age (P = .6), sex (P = .4), distance of the tumor from the anal verge (P = .3), pre-CMT ERUS stage (P = .2), and comorbidities (P = .2). The 5-year RFS was 96% and 54% in the pCR and control groups, respectively (P < .00001); the 5-year OS was 90% and 68% (P = .009). Sphincter-preservation rates were higher in the pCR group (P = .01).
Rectal cancer patients with pCR after preoperative CMT have improved RFS, OS, and sphincter preservation compared with patients without downstaging. Because pCR seems to be associated with better outcome, an understanding of the factors governing the response to CMT should be pursued.
KeywordsRectal cancer Radiotherapy Pathologic response Combined modality therapy Oncologic outcome
Largely on the basis of a recently completed prospective, randomized trial, preoperative combined-modality therapy (CMT) with long-course external beam radiotherapy (EBRT) and 5-fluorouracil (5-FU)–based chemotherapy followed by total mesorectal excision (TME) has emerged as the optimal therapeutic approach to locally advanced (endorectal ultrasonography [ERUS] T3/4 and/or N1 or clinically bulky) rectal cancer.1 Advantages of preoperative CMT, as compared with postoperative CMT, include avoidance of irradiating a fresh anastomosis or small bowel that may become trapped in the pelvis after surgery, tumor downsizing, an increase in sphincter-preservation rates, and an improvement in local control.1 In addition, preoperative CMT leads to a pathologic complete response (pCR), defined as no viable tumor on pathologic analysis, in 4% to 33% of patients.2, 3, 4, 5 However, the effect of a pCR on the long-term oncological outcome is controversial.4, 5, 6, 7, 8
We recently reviewed our institutional experience with rectal cancer treated with preoperative CMT. On multivariate analysis, absence of lymphovascular invasion/perineural invasion, negative lymph nodes, and a pathologic tumor response >95% emerged as independent predictors of improved long-term oncological outcome.4 Our results are similar to those of several prior reports5,8,9 but contradictory to others that suggest that a pCR has no effect on long-term oncological outcome.6,7 It must be noted that some of these prior series are limited by heterogeneous chemoradiation protocols, limited surgical and pathologic standardization, small numbers of pathologic complete responders, and short follow-up. In addition, the control groups, containing patients who did not achieve a pCR, were made up of heterogeneous populations with varied extents of pathologic response.
Therefore, to better examine the effect of a pCR on the oncological outcome of rectal cancer, we sought to compare patients with a pCR with a homogeneous, clearly defined control group. To this end, we focused on patients with rectal cancer whose preoperative evaluation included ERUS. We stratified patients into two well-defined, highly distinct groups: one with patients who achieved a pCR to preoperative CMT and the other with patients with no downstaging, defined as no difference between the preoperative ERUS stage and the pathologic stage.
We identified 476 patients with rectal adenocarcinoma evaluated at our institution with preoperative ERUS and treated with preoperative CMT from 1992 to 2002. The patients were identified from the prospectively maintained Colorectal Service Database at Memorial Sloan-Kettering Cancer Center. The study population consisted of those patients who achieved either a pCR (n = 60) or no downstaging when ERUS was compared with pathologic analysis (n = 140). The Colorectal Service Database, supplemented by a comprehensive chart review, was queried to identify clinicopathologic details and long-term oncological outcomes.
ERUS was performed by using a 10-MHz probe (type 1850; Bruel and Kjaer Medical, Naerum, Denmark). Rectal wall penetration (uT stage) was assessed according to the criteria of Hildebrandt et al.10 Hypoechoic circular structures >4 mm were considered to be malignant lymph nodes (uN positive). Hyperechoic nodes or nodes with central hyperechogenicity were considered benign. Echogenicity was assessed according to the criteria of Beynon et al.11
Patients received preoperative EBRT according to previously published techniques.12 A multiple-field technique was used, and the lateral fields were blocked as much as possible. EBRT was delivered in 26 fractions of 180 cGy followed by a 360-cGy boost to the primary tumor, for a total dose of 5040 cGy. The median dose of EBRT was 5040 cGy (range, 4500–6000 cGy).
5-FU–based chemotherapy was delivered to 181 patients (91%). 5-FU was delivered as a bolus in 153 cases (85%) and as a continuous infusion in 28 cases (15%). The most common protocol for bolus chemotherapy was 5-FU (325 mg/m2/day) with leucovorin (20 mg/m2/day) given for two cycles of five consecutive days on weeks 1 (days 1–5) and 5 (days 29–33) of radiotherapy. The most common protocol for continuous infusion chemotherapy was 5-FU (225 mg/m2/day) for a 6-week continuous cycle.
Irinotecan was delivered to six patients (3%). Irinotecan was most commonly administered as a bolus infusion of 10 mg/m2/day for a 4-week continuous cycle. Thirteen patients (6%) received no preoperative chemotherapy.
Radical Rectal Resection
After a median interval of 46 days (range, 30–591 days) after completion of EBRT, patients under went radical rectal resection according to the principles of TME. Five patients had an interval of >100 days from completion of EBRT to surgery: the reasons included patient preference, delay because of treatment of hepatitis C, treatment of an abdominal aortic aneurysm, and a failed attempt at local therapy (two patients).
TME was performed as previously described,13,14 with dissection in the avascular plane between the visceral fascia of the rectum and parietal fascia of the pelvis. For middle and low rectal cancers, the entire mesorectum was mobilized and resected as an intact unit. For high rectal cancers, the mesorectum was divided at a right angle to the bowel, 5 cm distal to the mucosal edge of the tumor.
In our institution, postoperative 5-FU–based chemotherapy is recommended for patients with rectal cancer treated with preoperative CMT. Postoperative chemotherapy was delivered to 51 patients (85%) in the pCR group and 129 patients (92%) in the no-downstaging group (P = .1). All patients in the pCR group who received postoperative chemotherapy were treated with 5-FU–based regimens, and 125 patients (97%) in the no-downstaging group who received postoperative chemotherapy were treated with 5-FU–based regimens. Three patients (2%) were treated with postoperative irinotecan, and one patient (1%) was treated with postoperative capecitabine. Postoperative chemotherapy was not delivered to 9 patients (15%) in the pCR group and 11 patients (8%) in the no-downstaging group for a variety of clinical reasons.
Standard pathologic tumor staging of the resected specimen was performed according to the AJCC Cancer Staging Manual (5th edition).15 By using hematoxylin and eosin staining and light-microscopic analysis, areas of tumor treatment response were characterized by the replacement of neoplastic glands with loosely collagenized fibrous tissue and scattered chronic inflammatory cells: a pathologic observation we have previously reported.16,17 The pCR group was defined by no evidence of viable tumor cells on pathologic analysis. The no-downstaging group was defined as no difference between the pre-CMT ERUS stage and pathologic stage.
Long-Term Oncological Outcome
Follow-up information was determined from the prospectively collected Colorectal Service Database and supplemented by a comprehensive chart review. Local recurrence was defined as clinical, radiological, or pathologic evidence of tumor recurrence in the pelvis, perineum, or anastomosis. Distant recurrence was defined as clinical, radiological, or pathologic evidence of tumor recurrence in any other site. The time to last follow-up, local recurrence, or death was measured from the time of radical rectal resection.
Statistical analysis was performed by using SPSS (version 11.0; SPSS Inc., Chicago, IL). Proportions were compared by using Fisher’s exact test, and means were compared by using a t-test. Recurrence-free survival (RFS) and overall survival (OS) were estimated by using the Kaplan-Meier method, and differences between survival curves were determined by using the log-rank test. A P value of <.05 was considered statistically significant. The study was reviewed and approved by the Memorial Sloan-Kettering Cancer Center Institutional Review Board.
Characteristics of the Study Population
Characteristics of the patient population
Median age, y (range)
Tumor DAV (cm)
LAR: 9 (15)
LAR: 52 (37)
LAR/CAA: 42 (70)
LAR/CAA: 52 (37)
APR: 5 (9)
APR: 34 (24)
TAE: 3 (5)
TAE: 1 (1)
TPC: 1 (1)
TPC: 1 (1)
Positive circumferential margins
Major postoperative complicationc
Minor postoperative complicationsd
Use of adjuvant chemotherapy
Pre–combined-modality therapy staging by endorectal ultrasonography (ERUS)
pCR, n (%)
No downstaging, n (%)
Fifty-five patients (91%) in the pCR group underwent a sphincter-preserving resection, whereas 106 patients (76%) in the no-downstaging group had a sphincter-preserving resection. The sphincter-preservation rate was significantly higher in the pCR group (P = .01). Reciprocally, the abdominoperineal resection (and permanent colostomy) rate was significantly higher in the no-downstaging group, despite the fact that the median tumor distance from the anal verge was similar between the groups (pCR, 6.8 cm; no downstaging, 6.5 cm).
Recurrence and RFS
The median follow-up for the entire study population was 38.6 months (range, 18.2–124.9 months). With a median follow-up of 41.7 months in the pCR group, two patients (3%) developed rectal cancer recurrence. Both recurrences were distant metastases in the lungs, occurring 11 and 19 months after rectal resection.
Local and distant recurrence in study population
pCR, n (%)
No downstaging, n (%)
Local recurrence only
Distant recurrence only
Local and distant recurrence
Our study demonstrates that relative to those who experience no downstaging, patients with rectal cancer who achieve a pCR to preoperative CMT have an improved RFS and OS. Although local failure was not significantly different between the pCR (no local failure) and no-downstaging (3% local failure rate) groups, distant failure was significantly lower in the pCR group. Therefore, it seems that the response of the primary rectal cancer may serve as a marker of effectiveness of systemic control and long-term oncological outcome. Another notable finding is that patients who achieve a pCR are significantly more likely to have a sphincter-preserving rectal resection. This is consistent with a recently reported large, prospective trial comparing patients with rectal cancer treated with preoperative CMT or postoperative CMT.1 The important conclusion from this information is that in select patients with rectal cancer who achieve a pCR after preoperative CMT, abdominoperineal resection with permanent colostomy may be avoided without compromising the long-term oncological outcome.
Some limitations to this study deserve mention. First, we did not include patients with a partial response in our analysis. We intentionally chose to compare two highly distinct groups to determine whether a pCR results in a significantly improved outcome relative to patients with a poor response, which we defined as “no downstaging.” We recognize that a significant portion of the treated population of rectal cancer patients—namely, partial responders—have been excluded from analysis. However, because the extent of response in the partial responders varied greatly, stratification of this heterogeneous group as a function of the extent of response was not possible. In addition, although we previously reported that a >95% pathologic response was associated with an improved oncological outcome, we were not able to perform this subset analysis because the pathologic data concerning extent of response as a continuous variable were not obtainable from this data set. To address this concern in a prospective manner, we recently implemented a technique for assessing the extent of rectal cancer response to preoperative CMT as a continuous variable. We plan to use these pathologic response data to more accurately stratify patients with rectal cancer who achieve a partial response to preoperative CMT. In addition, even though most of our patients (91%) were treated with 5-FU–based chemotherapy and the median radiation dose was 5040 cGy, some heterogeneity exists in the preoperative chemoradiation delivered to our study population and is a limitation of our study. Finally, there was no difference in the local failure rates between the pCR and no-downstaging groups. This may be due to our overall low rate of local failure or may indicate that the response to preoperative CMT may not be a marker of an increased risk for local failure, as it is at the systemic level.
Although a pCR may be associated with a significantly improved long-term oncological outcome and an increased sphincter-preservation rate, only 4% to 33% of patients with rectal cancer treated with preoperative CMT will actually achieve a pCR.2, 3, 4, 5 Potential methods to improve the pCR rate include using alternative chemotherapeutic agents and immunological agents, as well as altering the radiation-delivery protocol. Chemotherapeutic agents such as oxaliplatin and irinotecan and the immunological agents cetuximab and bevacizumab have shown promise in the postoperative treatment of colorectal cancer.18, 19, 20, 21, 22 These agents, and others, may provide for increased rectal cancer pathologic response rates. However, there have been limited data on their use in preoperative regimens,23 and prospective, randomized studies demonstrating an improvement in response with acceptable toxicity are needed before their use can be recommended. In addition, although some have suggested that altering the interval from completion of radiation to surgery may increase pCR rates,24 our previous work suggests that this approach deserves further investigation.
Recently, we and others have evaluated cell-cycle mediators such as p27, p21, p53, cyclin D1, Ki-67, bcl-2, and K-ras as potential molecular markers for the rectal cancer response to preoperative CMT.25, 26, 27, 28, 29 In addition, pharmacological manipulation of the cell cycle has been evaluated as a means to increase the pathologic response of other cancers.30,31 Although this is promising, it is clear that a better understanding of cell-cycle control is required to develop therapeutic approaches to increase the pathologic response of rectal cancer to preoperative CMT.
In summary, patients with rectal cancer who achieve a pCR to preoperative CMT can anticipate a 5-year RFS of 96% and a 5-year OS of 90%. In addition, our results suggest that patients with rectal cancer who achieve a pCR have a significantly improved RFS, OS, and sphincter-preservation rate when compared with patients without any downstaging. The improved oncological outcome in patients with rectal cancer who achieve a pCR is likely related to their significantly decreased rate of distant failure when compared with patients with no downstaging, because local failure was not significantly different between our study groups. To further improve oncological outcomes and sphincter-preservation rates in patients with locally advanced rectal cancer, the molecular mechanisms governing the rectal cancer response to preoperative CMT need to be explored.
Supported by National Institutes of Health grant RO1 82534-01 (J.G.G.).