Abstract
Background and Aims:
In a randomized trial the effect of short-term preoperative radio therapy and postoperative chemotherapy was studied in patients undergoing total mesorectal excision (TME) for clinically resectable rectal cancer. The primary endpoint was overall survival. The secondary endpoints published herein were the incidence of postoperative complications and adverse events with perioperative adjuvant therapy.
Material and Methods:
In 1995–2002, 278 eligible patients with stage II and stage III rectal cancer were randomly assigned to TME alone (surgery group) or to preoperative 25Gy radiotherapy in 5 fractions and postoperative 5-fluorouracil and leucovorin chemotherapy in addition (RT+CTgroup).
Results:
Anastomotic leakage rate did not significantly differ between the surgery and the RT + CT group, 20.6% vs. 27.4%. Postoperative infections (15.5 vs. 26.2%, p = 0.037) and perineal wound dehiscence (15.9 vs. 38.5%, p = 0.045) were more common after radiotherapy. Grade 3–5 adverse events were uncommon with preoperative radiotherapy (one, 0.7% with reversible lumbar plexopathy) and postoperative chemotherapy (hematologic in 10.8%, with one septic death, and gastrointestinal in 4.8%).
Conclusions:
Perioperative adjuvant therapy was generally well tolerated and did not lead to an increase in serious surgical complications. Wound infections and perineal wound dehiscence were more common in irradiated patients.
Keywords
INTRODUCTION
Treatment of clinically resectable rectal cancer has significantly changed since the introduction of Total Mesorectal Excision (TME). Conventional blunt dissection technique has been replaced by sharp dissection along the mesorectal fascia preserving intact mesorectal tissue around the rectum. TME has been associated with 4–13% local recurrence rate compared with 30 to 45% with the earlier techniques (1–4). TME is now considered the standard surgical treatment of resectable rectal cancer worldwide.
Even after introduction of TME surgery local recurrence has remained a clinical problem causing significant suffering. Postoperative chemoradiation has been proven efficient in reducing local recurrence and improving survival (5). Preoperative radiotherapy has reduced local recurrences by 50–60% in large randomized trials and the effect was also confirmed in two meta-analysis of 22 and 14 trials respectively (6–9). A survival benefit has been shown in only one trial and in a meta-analysis of 14 randomized trials (7, 8). These randomized trials were, however, conducted with the conventional surgical technique and the effects in conjunction with TME surgery were unknown at that time. Adjuvant postoperative chemotherapy after TME surgery had not been studied either.
In one of these randomized trials preoperative radiation was associated with an increase in postoperative complications and mortality (10). There were also concerns about the safety of low anastomoses after preoperative radiotherapy and removal of mesorectal fat. This together with the low local recurrence rates seen after TME alone indicated a re-evaluation of preoperative radiotherapy and postoperative adjuvant chemotherapy in resectable rectal cancer. In the present prospective, multicentre, randomized trial TME alone (surgery group) was compared with TME combined with short-term preoperative radiotherapy and postoperative 5-fluorouracil based chemotherapy (RT + CT group) in patients with resectable stage II–III rectal cancer. This paper presents the acute postoperative complications and treatment-related side-effects. Survival and local recurrence rates will be published later.
MATERIAL AND METHODS
STUDY DESIGN
Patients with clinical stage II and III rectal cancer were randomized either to TME alone (surgery group) or TME combined with short-term preoperative radiotherapy and postoperative adjuvant chemotherapy (RT + CT group). Patient flow chart is presented in Fig. 1.
Flow-chart of the study.
SAMPLE SIZE CALCULATION
The primary endpoint of this study was overall survival. With 278 randomized patients the trial was designed to detect 15% increase in survival rate, from 70% to 85% using power of 80%, and alpha level of 5%. The calculated accrual time was five years assuming 10% loss of patients during follow-up.
ELIGIBILITY
Patients, aged 18–80 years, with clinical stage II or III primary adenocarcinoma of the rectum were eligible. The upper border of the rectum was defined as 15 cm from the anus. Patients with obstructive tumours or tumours invading adjacent organs were excluded as well as patients with a diagnosis of another cancer less than five years previously or with a diagnosis of ulcerative colitis, Crohn's disease, familial polyposis or hereditary nonpolypotic colorectal carcinoma.
RANDOMIZATION
Patients were randomized after informed consent at the Finnish Cancer Registry. A true randomization was used to avoid the possibility that treating units might have guessed the treatment arm of the last patients of the block. The randomization function gave unfortunately by chance a slightly lower number of patients in the adjuvant group, 134 vs. 150.
PATIENTS
In year 1995, TME was just gaining acceptance as a primary operative technique in rectal cancer. To reduce the effect of the learning phase, only five university hospitals (Helsinki, Tampere, Kuopio, Oulu, and Turku) and maximum two surgeons per site participated in study.
During years 1995–2001, 284 patients were enrolled:165 at Helsinki, 41 at Kuopio, 36 at Tampere, 30 at Oulu, and 12 at Turku University Hospitals. Of these 150 were randomized to the surgery-alone group and 134 to the adjuvant group. After randomization, six patients were ineligible due to erroneous diagnosis, two in the surgery alone group and four in the adjuvant group (Fig. 1). One had no tumour left after biopsy, two tumours were inflammatory non-neoplastic and three tumours were metastases from other cancers.
PREOPERATIVE EVALUATION
Colonoscopy confirmed the adenocarcinoma and excluded synchronous colon tumours. The height of the tumour in the rectum was measured with a rigid rectoscope. Local evaluation of tumour spread was made either with endoanal ultrasound (149 patients), with computer tomography (14 patients), magnetic resonance imaging (7 patients) or only by clinical evaluation of tumour mobility (108 patients). Systemic spread was identified with upper abdominal ultrasound and chest x-ray. Whole body computer tomography was used during the last years of the trial. Carcinoembryonal antigen (CEA) was measured preoperatively.
PREOPERATIVE RADIOTHERAPY
Patients received preoperative radiotherapy in 5 Gy fractions on five consecutive days. The interval between the last day of radiotherapy and surgery was one week or less. Radiation field included a 5 cm craniocaudal margin and a 3 cm lateral and anteroposterior margin. Radiotherapy was given from three fields for 92 (72.4% of 127) and from four fields for 35 patients (27.6%).
SURGERY
TME was performed as first described by Heald with high ligation of the inferior mesenterial artery and vein (1). Professor Heald also personally demonstrated the technique in one of the first cases of the study. The use of a diverting colostomy or ileostomy was recommended. In 1996 the use of a J-pouch at the distal end of the colon was also recommended and use of diverting stoma preferred. Sphincter-saving anterior resection was used for all tumours situated in the middle or upper part of the rectum. The choice between anterior resection and abdominoperineal excision for low (<7 cm) tumours was made during the operation also considering the pelvic anatomy. The operation was considered curative when neither macroscopic residual tumour nor metastases remained after surgery. Routine pathologic evaluation did not include circumferential margin evaluation during the early days of the study.
POST-OPERATIVE CHEMOTHERAPY
In the RT + CT group, postoperative chemotherapy was initiated within 3 to 10 weeks from surgery in patients with stage II–III tumours in the final pathology report. The chemotherapy consisted of six cycles of bolus injections of 5-fluorouracil (5-FU 350 mg/m2) and leucovorin (20 mg/m2) for 5 consecutive days every 4 weeks. In five cases the start of chemotherapy was delayed 12 weeks or more after the operation because of anastomotic leakage (n = 2) or unknown reasons (n = 3).
The National Cancer Institute Common Toxicity Criteria for adverse events (CTCAE) version 2.0 was used to score chemotherapy and acute radiation toxicities (11).
FOLLOW-UP
Patients with a diverting stoma had a sigmoidoscopy 6 weeks after the operation before closure of the stoma. For all patients sigmoidoscopy was performed every 6 months for 2 years. Thereafter sigmoidoscopy was done annually for 5 years. At 2 and 5 years after surgery, colonoscopy replaced sigmoidoscopy. CEA level was measured every 3 months for 2 years and then annually ad 5 years. If an increased CEA level was detected total body computer tomography and colonoscopy were performed.
PROTOCOL COMPLIANCE
Protocol compliance is presented in the patient flow chart in Fig. 1. Of the 130 eligible patients randomized to the adjuvant group, preoperative radiotherapy was delivered to 127 patients (97.7%). Scheduled 5 × 5 Gy was delivered to 99.2% and there was one radiotherapy protocol deviation (4 × 5Gy).
In the surgery alone group, four patients received postoperative chemoradiation; one with stage II due to tumour perforation and three with stage III cancer. One stage III patient with susceptible nodules on peritoneal surface was given chemotherapy although biopsies were benign. Stage IV patients were treated outside the study protocol.
In the RT + CT group, chemotherapy was omitted in 10 of 79 patients with stage II or III disease, due to anastomotic dehiscence (n = 4), heart condition (n = 1), withdrawn consent or unknown reasons (n = 5). Adjuvant chemotherapy was given to 10 of 29 patients with a stage I tumour against the study protocol.
DATA COLLECTION AND STATISTICS
The ethics committees of participating hospitals approved the study. The basic database including patient identification, date of randomization, and treatment group was created automatically in the database of the Finnish Cancer Registry as part of the randomization process. None of the research group members had the right to make changes to these data. Completed datasheets concerning the operation and postoperative follow-up went directly to the principal investigators. All original data were monitored by one of the investigators (SS).
Data was analysed on intention to treat-basis. Data are presented as mean ± SD, median (range) and absolute values and percentages as appropriate. Continuous variables were compared with Student's t-test and categorical variables with Fisher's exact test. Wilcoxon-Mann-Whitney test was used for categorical data with ordered categories. The level of p < 0.05 was considered statistically significant for all tests. SPSS 12 (SPSS Inc., Chicago, IL, USA) was used for the analyses.
RESULTS
PATIENTS
The groups were well balanced (Table 1). The only difference between groups was tumour size, which was 0.6 cm smaller in the RT + CT group compared to surgery group (p = 0.019).
Patient and tumour characteristics.
ADVERSE EFFECTS OF RADIOTHERAPY
Preoperative 5 × 5 Gy radiotherapy was generally well tolerated (Table 2). Grade 1 to 3 adverse events was seen in 16 out of 127 patients (12.6%). Mild to moderate diarrhoea was seen in 3.9%, nausea in 3.9% and urinary irritation in 3.2% of patients.
Adverse effects of radio- and chemotherapy.
Nine patients (7.1%) suffered from mild pain (4.7%) or parestesia (2.4%) in the perineal region. One patient developed reversible lumbar plexopathy with weakness and numbness in the lower extremities 3 months after operation. These symptoms were mistaken as a local recurrence, and the patient received further radiotherapy at a dose of 30.9 Gy to the pelvic region at 4 months postoperatively. Symptoms progressed for 6 months and then faded off.
SURGERY
There were no differences in type of operation, length of hospital stay, operative time or blood loss (Table 1). The operation was considered curative, i.e., no metastases or macroscopic residual tumour left, in 127 (85.8% of the 148) patients in the surgery group and 112 (86.2% of the 130) patients in the RT + CT group (p = 1.00).
POSTOPERATIVE COMPLICATIONS
Three patients died within 30 days postoperatively in the surgery group and one in the radiotherapy group. Postoperative mortality (days 0–79) was 2.7%, (four patients died) in the surgery group and 3.1% (four) in the adjuvant group (p = 0.74). Four of these eight deaths were caused by anastomotic leakage, two in each group. Three patients died of acute myocardial ischemia and one of mechanical ileus (Table 3).
Postoperative complications.
Values in parenthesis are total number of operations performed
The incidence of anastomotic leakage was 20.6% in the surgery group and 27.4% in the RT+CT group (p = 0.305, Table 3). These figures include both clinical leakages (n = 33) and sinuses detected in routine sigmoidoscopy 6 weeks after surgery (n = 11). In the surgery group there were 18 symptomatic leakages out of 21 detected, and 11 were reoperated. The number of symptomatic leakages in the RT and CT group was 15 out of 23, and 5 cases needed reoperation.
The height of the inferior margin of the tumour had no effect on the leakage rate: the number of leakages was 20 out of 57 in the lower third of the rectum, 17 out of 66 in the middle third and 7 out of 19 in the upper third (p = 0.655). The leakage was treated with antibiotics alone in 26 cases, of which 24 had a temporary faecal diversion. Of the 16 patients who underwent re-operation for leakage, 10 had no temporary diversion. Among patients with anastomotic leakage, a permanent colostomy had to be constructed in seven of 14 patients without a temporary diversion (50.0%) and in three out of 30 (10.0%) with a diversion.
Wound infections were more common in the RT + CT group, 11 of 130 (8.5%) and four of 148 (2.7%; p = 0.036). Operation type had no effect on the wound infection rate (data not shown). The total number of infectious complications was also higher in the RT + CT group compared to the surgery group, 26.2% vs. 15.5%, respectively (p = 0.037). However, the number of serious infectious complications, e.g. sepsis, intra-abdominal abscess or pneumonia, did not differ.
The amount of perineal dehiscence after abdominoperineal excision was higher in the RT + CT group, 15 of 39 (38.6%) compared to the surgery group 7 of 44 (15.9%; p = 0.045). The incidences of other postoperative complications were not different between the groups (Table 3).
ADVERSE EFFECTS OF CHEMOTHERAPY
The chemotherapy was given as planned in 69 of 79 (87.3%) of the patients. The adverse effects were mostly mild (Table 2). Grade 3–5 hematologic was seen in 10.8% and gastrointestinal adverse events in 4.8% of patients. One chemotherapy related death occurred. A 55-year-old woman with no co-morbidity died of febrile neutropenia after the first 5-day cycle of chemotherapy. Twenty-three patients had a dose-reduction due to side-effects.
DISCUSSION
The initial concern with preoperative radiotherapy, especially in conjunction with TME surgery, has been the acute toxicity and increase in surgical complications. In this randomized, multicentre study we observed similar operation time, blood loss, and length of hospital stay between the surgery and RT + CT groups. Most importantly preoperative radiotherapy did not increase postoperative in-hospital or 30-day mortality. Since the initiation of this study several randomized trials on preoperative short-term radiotherapy have been published. In-hospital mortality did not increase in the Swedish Rectal Cancer Trial with preoperative radiotherapy before conventional surgery (4% with RT and 3% with surgery alone) (7) nor in the Dutch TME Trial with short-term preoperative radiotherapy and TME (4% with RT and 3.3% with surgery alone) (7, 12). The first concerns may have been attributable to the radiotherapy technique which has evolved considerably over time. In this study modern radiotherapy is used and radiotherapy is given from 3–4 portals with the upper border of the radiotherapy fields at the L5 level.
Preoperative radiotherapy does not significantly impact anastomotic leakage rate in this study (27.4% in adjuvant vs. 20.6% in surgery group). This is in accordance with the Dutch TME-trial and earlier reports from trials with conventional surgery (6, 13, 14). The anastomotic leakage rate reported here is relatively high in both groups, but the figures include both clinically symptomatic leakages and those detected in routine postoperative sigmoidoscopy performed 6 weeks after surgery. The rate of leakages detected without routine sigmoidoscopy was 17.7% and the rate of leakages requiring operative treatment 7.5%. These figures are in line with earlier reports for symptomatic leakages from 8.6 to 19% although variability in the definition and diagnosis of leakage makes cross trial comparisons difficult (13, 15–18).
We did not find any differences in leakage rate according to the height of the tumour, either. Low lying tumours have been reported to be a significant risk factor for leakage and a protective stoma is often recommended in this group of patients (16, 18, 19). In the Dutch TME-trial there was also no correlation between leakage rate and tumour location (13). We measured the height of inferior margin of tumour from anal verge as was done in the Dutch study. The height of the anastomosis is thus lower and may partly explain these conflicting results. Based on our results omitting stoma cannot be recommended even with the tumours in the mid-and upper rectum. In patients with a protective stoma most leakages could be treated conservatively with antibiotics without re-laparotomies. Unfortunately serious leakages with a fatal outcome could not be avoided.
Infectious complications, other than anastomotic leakage, were significantly more common in the RT + CT group than in the surgery group, 26.2% vs. 15.5%, respectively. The difference was due to wound infections being more common after radiotherapy. This is in agreement with earlier studies (6, 15), as is the increased perineal wound dehiscence after radiotherapy, 15–36% vs. 6–18%, (6, 13, 15).
In accordance with earlier results most patients tolerated hypo-fractioned preoperative radiotherapy well. Acute adverse effects were scarce and mostly mild. One patient developed lumbosacral plexopathy 3 months postoperatively. In the absence of any other aetiology, plexopathy must be regarded as a complication of preoperative radiotherapy. Six of the 127 patients had mild pain and 3 had paraesthesia during radiotherapy. Acute lumbosacral plexopathy after 5 × 5 Gy radiotherapy preceding rectal cancer surgery was first reported in 1996 (20). After that, a shift from a three-beam to a four-beam radiation technique was made, and the upper border of the radiation field was adjusted at the level of L5. The Dutch TME trial showed no long-lasting neurological symptoms, although 53 patients (10%) reported some degree of pain (13).
Bolus 5-fluorouracil and leucovorin is associated with severe and life-threatening gastrointestinal and haematological toxicity (21, 22). Severe neutropenia and sepsis leading to death have been reported and explained by idiosyncrasy i.e. 5-fluorouracil degrading enzyme deficiency (23, 24). In our study, postoperative chemotherapy led to one fatal case of febrile neutropenia after the first dose of chemotherapy in a patient with no co-morbidities. Generally the toxicity profile was milder than in previous publications, probably due to the lower 5-fluorouracil dose (350 mg/m2 in concurrent study vs. 425 mg/m2 in the original Mayo regimen) (21, 22).
The results of our study show that short-term preoperative radiotherapy does not lead to an increase in serious acute complications following surgery. However, infections and perineal wound dehiscence are more common in irradiated patients. We did not analyse the effect of radiotherapy on functional outcome after anterior resection. Others have shown impaired urinary, sexual and anal function after preoperative radiotherapy and also a rise in the number of secondary cancers after irradiation (25–27). Future analyses of local recurrence rate and long-term survival will give additional insight into the patient selection for adjuvant therapy.