| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Article |
1 Department of Surgical Oncology, Antoni van Leeuwenhoek Hospital/The Netherlands Cancer Institute, Amsterdam, The Netherlands
2 Department of Radiotherapy, Antoni van Leeuwenhoek Hospital/The Netherlands Cancer Institute, Amsterdam, The Netherlands
3 Department of Surgical Oncology, University Hospital, P.O. Box 1352, 7111 10, Herakleion, Greece
Correspondence: Address correspondence and reprint requests to: Eelco de Bree, MD; E-mail: debree{at}edu.uoc.gr
| ABSTRACT |
|---|
|
|
|---|
Methods: The data of 83 patients primarily treated by RT with curative intent throughout a 20-year period in our institute were analyzed regarding the need for colostomy.
Results: Totally, 28 patients (34%) required creation of a colostomy after primary RT for local failure or treatment-related complications during a mean follow-up period of 39 months. The 3-year actuarial colostomy-free rate was 59% (mean 85 ± 9 months). Early stage disease, low T-score and absence of infiltration in adjacent organs were associated with a reduced need for colostomy in univariate analysis. In multivariate analysis only T-score was an independent variable in predicting prolonged colostomy-free interval. In this study, no statistically significant differences were noted for gender, age, nodal status, total radiation dose, radiation boost and concurrent chemotherapy.
Conclusions: In approximately one-third of the patients treated by anal sphincter saving management with curative aimed primary RT, the creation of a colostomy appeared to be necessary for RT complications and local treatment failure. Therefore, patients should be well informed regarding the considerable risk of need for colostomy after RT for anal cancer.
Key Words: Anal cancer Radiotherapy Local failure Complications Colostomy
| INTRODUCTION |
|---|
|
|
|---|
Its treatment has undergone substantial reassessment over the past two decades. Whereas historically abdominoperineal resection (APR) of the anal canal and distal rectum with formation of a permanent end-colostomy was the standard treatment for anal cancer, during the last decades many patients have been treated successfully by radiotherapy (RT) in an attempt to save the anal sphincter function. With this treatment option colostomy can frequently be avoided, while survival is not impaired.2 Moreover, long-term quality of life after primary RT for anal cancer appears to be more than acceptable, with the exception of diarrhea and perhaps sexual function.3,4 However, surgery frequently seems to be necessary after primary RT, either for residual disease, local recurrence or RT-related complications. With most studies generally being concerned with the effectiveness of RT, there has been little attention focused on the need for colostomy after RT failure. In the present study, we examined how frequently and for the reason why colostomy was required after primary RT for anal cancer in our institution during a 20-year period.
| PATIENTS AND METHODS |
|---|
|
|
|---|
|
Treatment failure, expressed by the observation of residual local disease and the occurrence of local recurrence, prevalence of RT-related complications requiring surgery and the number of patients eventually having some kind of colostomy were recorded.
Overall survival, disease-specific survival, disease-free survival, local control interval over time and colostomy-free interval were estimated using KaplanMeier curves. Overall survival was defined as the delay between the initiation of treatment until death, disregarding the cause of death. Since many patients were high-aged, we also studied disease-specific survival, defined as the delay between the beginning of treatment until death due to disease. Patients who died of other cause were censored at the time of death. Disease-free survival was defined as the interval to disease recurrence or death after initial response to RT, with patients demonstrating residual disease after RT being scored as having no disease-free interval at all. Local control interval was defined as the one between treatment initiation and local relapse; patients with local residual disease after RT were considered local control failures at the beginning of RT. Colostomy-free interval was conventionally defined as the time until the first colostomy. All colostomies were taken as failure in the present analysis. An intentionally temporary colostomy created before RT was considered as a colostomy-free interval of 0 month. Patients who died without colostomy and free of disease have not been considered as treatment failure and were censored on the date of death.
Outcome was calculated for the entire group of patients and univariate subanalysis was performed for various potential risk factors including gender, age (< vs.
60,
vs. >65,
vs. >70, < vs.
75 years), clinical AJCC tumor stage, T-score (according to AJCC staging system), infiltration of adjacent organs, total radiation dose (< vs.
50, < vs.
60, < vs.
65 Gy), addition of radiation boost, type of radiation boost (interstitial versus external beam radiation) and use of concurrent chemotherapy. The significance of differences in local tumor control, incidence of RT-complications and number of patients eventually requiring colostomy between subgroups was statistically analyzed with the Fishers exact test. Differences in local-tumor-free and colostomy-free interval between subgroups were univariately examined with the log-rank test, while multivariate analysis was performed with Coxs regression model for bivalent prognostic factors, which had a P value of less than 0.2 in univariate analysis. P values of less than 0.05 were considered statistically significant.
| RESULTS |
|---|
|
|
|---|
Overall Need for Colostomy
Subsequent surgery was performed in 31 of 83 patients (37%) treated by curative intended RT for residual disease, local recurrence or treatment complications. In 28 patients (34%) creation of a colostomy was necessary. In two cases recurrent and in another two cases residual tumor were initially locally excised, but in one of the latter cases local recurrence required APR almost 1 year later. APR or creation of a diverting colostomy were performed after a median period of 6 months (262) after the initiation of RT. Thirteen patients obtained a diverting colostomy, of whom three underwent APR later. In total, eventually 18 patients underwent APR. Ultimately, 30 patients (36%) had some kind of colostomy at the end of the studys follow-up period.
Residual Disease
Residual disease was observed in 18 cases (22%). APR was performed in eight cases, local excision in two cases and diverting colostomy only in one case because of irresectability. Surgery was performed 47 months (median 5) after initiation of RT. One of the patients with residual disease treated with local excision required APR later due to local recurrence. One patient who underwent APR already had a prophylactic colostomy. The remaining seven patients were not operated for various reasons, including irresectability, (progressive) meta-static disease, poor general condition and patients refusal, and died after a median period of 6 months (317) following initiation of RT. Two of them had already a prophylactic colostomy created before RT.
Local Recurrence
Local recurrence developed in eight patients (10%) after a median period of 9 months (480) after initiation of RT. APR was performed in five patients, of whom three had already received colostomy for RT complications, while two cases could be managed with LE. In one case with an indication for APR, the patient refused further treatment.
RT Complications Requiring Colostomy
Diverting colostomy or APR was performed in 16 cases (19%) due to RT complications, including anal ulcer (11), rectovaginal fistula (3), anal canal stenosis (1) and a combination of stenosis and fecal incontinence (1). The median period from initiation of RT to surgery was 8 months (262). APR was performed in four cases and diverting colostomy in the remaining 12. In one patient a diverting colostomy was created 28 months after a prophylactic colostomy was turned down. Two patients were relieved from their colostomy 8 and 14 months later, while three patients who required diverting colostomy for radiation ulcer required 2, 7 and 14 months later an APR for local recurrence. The remaining patients retained their diverting colostomy. One patient died of diarrhea and septic shock during RT treatment (counted as died with local residual tumor on day 0).
Initially Temporary Colostomy
Four of the 7 patients retained their initially temporary colostomy created before RT because of progressive (1) and recurrent disease (2) or enduring prior to RT existing anal ulcer (1). One patient with a prophylactic colostomy underwent APR for residual disease. The remaining two patients were relieved from their colostomy 3 and 4 months after the initiation of RT, but one of them later underwent a diverting colostomy for a late radiation ulcer.
Risk Factors for Local Treatment Failure, Complications and Colostomy Creation
Gender and Age
The incidence of local treatment failure, the frequency of complications requiring colostomy and the need for colostomy were not statistically different for both sexes and various age groups. Neither local-tumor-free interval nor colostomy-free interval differed between these groups of patients.
Clinical Tumor Stage
Stage I disease was associated with a decreased risk of local treatment failure and colostomy creation in comparison with stage III disease (P = .025 and P = .017, respectively). When compared with stage II disease, stage I anal cancer was associated with a statistically significant reduced risk of colostomy creation (P = .041), but not of local treatment failure (P = 0.11). Stage I disease was also associated with prolonged local-tumor-free and colostomy-free interval when compared with stage II and III (P = .043 and P = .049, respectively, Fig. 1
). No statistically significant differences were found between stage II and stage III disease regarding these outcome parameters. No correlation between disease stage and rate of complications requiring colostomy was observed.
|
|
Nodal Involvement
No association was found between lymph node involvement and the incidence of local treatment failure, the frequency of RT complications requiring colostomy, the colostomy rate, and time for local treatment failure and colostomy creation.
Total Radiotherapy Dose
No statistically significant relation could be identified between total radiation dose and local treatment failure rate, incidence of complications requiring colostomy, overall need for colostomy, duration of local tumor control and colostomy-free interval.
Radiation Boost
The addition of an extra radiation dose to the tumor area was not associated with a decrease of local treatment failure rate, prevalence of complications requiring colostomy or of the overall need for colostomy. The administration of a radiation boost was not associated with a statistically significant improvement of local-tumor-free interval or colostomy-free interval.
Comparison of interstitial with external boost radiation demonstrated not only a trend for lower incidence of local treatment failure (19 vs. 36%, P = 0.10), but also a trend for a higher rate of complications requiring colostomy (31 vs. 14%, P = .079), resulting in a similar overall need for colostomy (46 vs. 38%, P > 0.1). In this relatively small number of patients, the use of interstitial RT resulted in an almost statistically significant (P = .066) improvement of local-tumor-free interval. It has to be noted, however, that considerably more patients treated by this modality had early stage disease (92 vs. 55% stage I/II). No difference in colostomy-free interval was observed.
Radiotherapy versus Chemoradiotherapy
Local recurrence was more frequently noted after chemoRT than after RT without concurrent chemotherapy. The need for colostomy was higher in the chemoRT group. However, these differences did not reach apparent statistical significance (P = 0.14, P = 0.12) and it has to be noted that the chemoRT group contained a considerably higher proportion of patients with more advanced disease (48 vs. 28% stage III disease). The addition of chemotherapy to radiation was not associated with an increase in incidence of treatment-related complications requiring colostomy creation, neither with a difference in local-tumor-free or colostomy-free interval in this relatively small number of patients.
Multivariate Analysis
Factors tested in multivariate analysis for local-tumor-free interval, which had a P value of less than 0.2 in univariate analysis, were: stage I/II versus stage III (in univariate analysis P = .040), T1/2 versus T3/ 4 score (P = .0029), infiltration of adjacent organs (P = .053), total radiation dose of < vs.
50 Gy (P = 0.17) and use of interstitial radiation boost (P = 0.47). Multivariate analysis revealed that only T-score was an independent and statistically significant factor to predict local tumor control (P = .009; hazard ratio 3.00, 95% confidence interval 1.316.88). Regarding the colostomy-free interval the parameters stage I/II versus stage III (in univariate analysis P = 0.12), T1/2 versus T3/4 score (P = .0011) and infiltration of adjacent organs (P = .0007) were included in a multivariate analysis. In predicting the prolonged colostomy-free interval, T-score was the only independent and statistically significant variable (P = .002; hazard ratio 3.14, 95% confidence interval 1.506.55).
| DISCUSSION |
|---|
|
|
|---|
In clinical practice, the concern for increased toxicity makes many physicians omit concurrent chemotherapy for elderly patients with small tumors.9,10 Clinical studies on primary RT for anal cancer demonstrated a 5-year rate for overall survival of 5272% with no additional benefit of concurrent chemotherapy, while the estimated cancer-specific and disease-free survival after 5 years were 6084% and 5472%, respectively.6,7,1120 Similar outcome was observed in the present study. Although studies have demonstrated the efficacy of RT, there remains uncertainty as to the most appropriate RT dosage, fractionation, radiation boost, overall treatment time (including any gap before the boost phase of treatment) and chemotherapy regimen.2,13,15 Moreover, whether inguinal lymph nodes should be routinely irradiated during primary treatment remains controversial.2,14 Of importance is that inter-study comparisons of primary RT results are notoriously difficult, due to variations in RT treatment, the use of various or no concurrent chemotherapy schemes and patient selection, especially regarding stage. Some studies included patients with RT after excisional biopsy, which should actually be considered as adjuvant rather than primary RT.
Although surgery is no longer the primary treatment modality for anal squamous cell carcinoma, currently it still has an important role. Firstly, initial examination and biopsy under anesthesia is often required for adequate assessment of the tumor. Secondly, the surgical management of any inguinal node disease should be considered. Thirdly, temporary stoma formation can be done electively in patients who might cope poorly with acute side effects of pelvic RT, who have an obstructing tumor or an anal ulcer before treatment, who are at risk of developing a rectovaginal fistula and in those who have lost anal sphincter function because of a very advanced tumor. Additionally, colostomy formation may be necessary for RT-related complications. Finally, surgery is an option for salvage of recurrent or residual primary tumor after non-surgical treatment.2
The local treatment failure rate in the present series is consistent with the local tumor control failure rates of 1955% that have been reported in other series of anal cancer patients treated by primary RT (Table 2
).68,1525 Different definitions for discrimination between locally persistent and recurrent disease have been used. Generally, locally persistent disease has been noted in 446% of cases, while local recurrences, which may be seen up to years after initial treatment, have been observed in 630% (Table 2
).68,12,15,1825 The wide variation in residual tumor and local recurrence rates are not only due to different interpretation of these terms, but more as a result of patient selection, treatment variations and differences in follow-up period. The use of concomitant chemotherapy significantly increased the locoregional control rate in comparative studies.6,7,15,17 The impaired local tumor control in our chemoRT group is probably due to the considerably higher number of patients with more advanced disease that had been treated by this modality. In our series with a relatively small number of patients in the subgroups, a definite statistically significant decreased local treatment failure rate could only be found for early stage disease and low T-scores. Risk factors for local treatment failure that have been reported in various univariate analyses include: age (<66 years,26 >70 years17) male gender,7,26 advanced tumor size or extent,11,12,15,16 skin ulceration,7 high histological tumor grade,15 positive nodal status,3,7 no tumor response at interval evaluation,12,15 boost with external beam instead of interstitial RT,26 longer time period to complete RT,3,15 lower total radiation dose (<35 Gy,11 <50 Gy17,24) and, as earlier mentioned, omission of concurrent chemotherapy.6,7,15,17 In multivariate analyses, advanced age,13,17 male gender,7 advanced tumor size or extent,12,13,15,17 skin ulceration,7 nodal involvement,7,13 omission of concurrent chemotherapy,7,15,17 poor response in the initial treatment period,12,15 and lower total RT dose 17 have been noted as independent adverse predictors of local tumor control.
|
While the literature focuses on the therapeutic impact of RT in cases of anal cancer, the RT-related complications are generally only rarely referred to. RT treatment is not only associated with acute side effects, like skin reactions and diarrhea, but in approximately 4050% of cases with much more important late complications, affecting mainly the urogenital, anorectal and musculoskeletal system.6,10 Severe late complications are seen in 936% of cases,7,12,15,16,18,20 most frequently in association with large anatomic tumor extent,20 higher total RT dose,12,20 large external beam radiation doses per fraction,16 uninterrupted RT course,16 and larger volume, higher dose and longer interstitial implant boost RT.12,18 Generally, the incidence of severe late side effects is not significantly increased after chemoRT.6,7,20 Primary (chemo)RT also seems to be feasible in elderly patients (
75 years), with complication rates being similar to those observed in younger patients.9,30 With the use of intensity-modulated RT, morbidity may be reduced in the future.31 Whether the need for colostomy will also decrease is uncertain, since complications that require colostomy are usually nevertheless located in the target area. In patients with local tumor control, the need for at least a temporary colostomy after complications of RT, including anal ulcer, anal canal stenosis, rectovaginal or rectovesical fistula and fecal incontinence, ranges from 4 to 15%.6,12,15,16,1820,23,25 In our series, this rate was slightly higher, possibly because of more liberal indication for creation of a colostomy in case of anal ulcer. Complete continence or normal anal function is not always obtained after RT.3,4,13,15,25 The reliability of the published data on post-radiation fecal incontinence is poor because patients frequently fail to admit fecal incontinence, and because prospective studies are lacking to assess fecal incontinence as a specific end point using adequate, validated and reproducible methodology.32 Therefore, post-radiation fecal incontinence and the need for colostomy creation may be generally underestimated. The published rates of late new-onset fecal incontinence after pelvic surgery and pelvic radiation are between 3 and 53%. Moreover, as in the present series, colostomies fashioned before RT to avoid complications are frequently not reversed.6,17
As also found in our series, in patients with tumors invading surrounding tissues (T4) the rate of local tumor control with good anal function after RT with curative intent is particularly low.15 Therefore, pre-operative chemoRT followed by APR seems to be reasonable treatment option in case of T4 tumors or poor pre-treatment anal function.15,33 In patients with large tumors, promising results regarding local tumor control, decrease of distant recurrences and maintenance of anal sphincter function have been reported with induction chemotherapy preceding concomitant chemoRT.34 Results of initiated randomized trials are to be awaited in the coming years to confirm the benefit of this induction chemotherapy.
In various series, sphincter conservation was possible in 4091% of the cases after primary RT, with highest probability for smaller tumours.68,12,17,18,25 In our series this rate was 64%. However, some of those patients who avoided creation of a colostomy were not free of local tumor.
In conclusion, RT is considered as primary treatment of anal cancer to avoid the mutilating APR of the distal rectum and anus with formation of a permanent end-colostomy. Although this attempt succeeds in the majority of cases, in approximately one-third of the patients treated by curative intended primary RT creation of a colostomy is necessary for treatment failure and RT-related complications. In the present study, patients with early stage disease, low T-scores and absence of infiltration of adjacent organs were at lower risk. In T4 tumors pre-operative RT with APR is a reasonable alternative option, because of the poor results of non-surgical treatment in these cases. Before initiating primary RT, anal cancer patients should be well informed regarding the considerable risk of a colostomy to avoid unpleasant events after completion of RT.
Received for publication June 6, 2006. Accepted for publication June 6, 2006.
| REFERENCES |
|---|
|
|
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |