clinical and methodological issues in a cost-utility analysis alongside a
randomized clinical trial in patients with rectal cancer undergoing
total mesorectal excision
Brink, Mandy van den
Citation
Brink, M. van den. (2005, June 28). Economic evaluation of preoperative radiotherapy in
rectal cancer : clinical and methodological issues in a cost-utility analysis alongside a randomized clinical trial in patients with rectal cancer undergoing total mesorectal excision. Retrieved from https://hdl.handle.net/1887/4273
Version: Corrected Publisher’s Version
License: Licence agreement concerning inclusion of doctoral thesis in theInstitutional Repository of the University of Leiden
Downloaded from: https://hdl.handle.net/1887/4273
CLI
NI
CAL NATURE AND PROGNOSI
S OF
LOCALLY RECURRENT RECTAL CANCER
AFTER TOTAL MESORECTAL EXCI
SI
ON
W I
TH OR W I
THOUT PREOPERATI
VE
RADI
OTHERAPY
Abstract
Purpose To document the clinical nature and prognosis of locally recurrent rectal cancer after total mesorectal excision (TME) with or without 5×5 Gy pre-operative radiotherapy
(PRT) and to identify patient-, disease- and treatment-related factors associated with differences in prognosis after local recurrence.
Methods For 96 Dutch patients with a local recurrence, who participated in a multicenter randomized clinical trial, data on treatments and follow-up were gathered from surgeons and radiation and medical oncologists. Twenty-three patients (24%) had previously been treated with PRT plus TME and 73 patients (76%) had been treated with TME alone. Eighty-one patients (84%) were followed until death; median follow-up time of the alive patients after local recurrence was 21 months (range 5-48 months).
Results Survival after local recurrence in the PRT+TME group was significantly shorter than in the TME group (median survival 6.1 vs. 15.9 months, HR for death 2.1, p=0.008). Patients with a local recurrence in the PRT+TME group had distant metastases more often (74% vs. 40%, p=0.004), underwent surgical resection of local recurrence less often (17% vs. 35%,p=0.11), and received radiotherapy for local recurrence at a total dose ≥ 45 Gy less often (4% vs. 42%, p=0.001) than patients without PRT. In a multivariate analysis the difference in survival after local recurrence between randomization groups was no longer statistically significant (HR for death of PRT 1.53, p=0.16).
Introduction
In patients with rectal cancer, local recurrences are difficult to treat, may cause severe disabling symptoms and usually have a fatal outcome (1-3). For this reason, the focus has been on the prevention of local recurrences by refining imaging and surgical techniques and by assessing the value of adjuvant preoperative or postoperative treatment modalities (4). Nowadays, Total Mesorectal Excision (TME) is widely accepted as the standard of surgery for patients with primary rectal cancer (5-7) and recently, the additional value of preoperative radiotherapy (PRT) was demonstrated (8).
New treatment techniques have not only reduced local recurrence rates (8-10), but may also have changed the clinical nature and prognosis of locally recurrent rectal cancer. In previous studies, median survival for locally recurrent rectal cancer after conventional surgery with or without PRT was estimated at 7-11 months and 15-16 months respectively (1;2). Suggested explanations for the reduced survival after local recurrence in patients treated with PRT have been that in patients previously treated with PRT local recurrences may be treated less aggressively, since maximal radiotherapy is no longer possible as part of the multimodality treatment for local recurrence, or that the recurrences arising may be more often associated with distant metastases. However, those hypotheses have not been further investigated, possibly owing to the fact that patients with a local recurrence were retrieved over a long time period, from multiple studies, and with varying treatments, thus hampering unequivocal analyses (1;2). This has led to contrasting viewpoints on the prognosis and treatment of local recurrence. For example, Holm et al. (2) concluded that attempted curative treatment for local recurrence is an attractive option, because local failure is not inevitably an early manifestation of disseminated disease, whereas Frykholm et al. (1) concluded that the probability of curative treatment of local recurrence is very low.
The purpose of our study was to document the clinical nature and prognosis of locally recurrent rectal cancer after TME-surgery, with or without short-term PRT. In addition, we tried to identify patient-, disease- and treatment-related factors associated with differences in prognosis.
Methods
Study population
patients were clinically evaluated for local recurrence every 3 months during the first year after surgery and annually thereafter. Local recurrence was defined as radiologic and/or histologic evidence of a tumor within the lesser pelvis or the perineal wound after a macroscopically complete resection. For the purpose of this paper, non-Dutch patients (n=331), ineligible patients (n=50) and patients without a local resection or with a macroscopically incomplete local resection (n=46), were excluded, leaving 1434 patients for the analyses of local recurrence rates. Local recurrence location was categorized according to an adapted version of the subdivision proposed by Philipsen et al. (11), into recurrences located at the level of the anastomosis or perineum, and pelvic recurrences. Distant recurrences were defined as radiologic and/or histologic evidence of a tumor in any other area. A radiation oncologist checked all reported local and distant recurrences.
Design
In August 2002, at a median follow-up time of 43 months since TME-surgery, 96 Dutch patients had a local recurrence. For those patients, we asked surgeons from 48 hospitals in the Netherlands for copies of all medical correspondence since local recurrence diagnosis, to obtain detailed data on treatments for local recurrence and follow-up. In August 2003, additional follow-up information was collected for patients still alive in August 2002. In case of non-response or incomplete data, reminder phone calls were made. In case of missing or unknown data on details of treatments, radiation or medical oncologists were also asked for copies of all medical correspondence. Types of treatment for local recurrence were classified according to whether the local recurrence was surgically resected or not, and by types of multimodality treatment approaches. Treatments for local recurrence were also categorized according to the administered radiotherapy dose, i.e. no radiotherapy or radiotherapy by total doses of < 45 Gy, and radiotherapy by a total dose ≥ 45 Gy, to study whether PRT limits the possibilities for maximal radiotherapy treatment of local recurrence.
Statistical analyses
Results
Response and follow-up
At a median follow-up of 43 months, the actuarial rate of local recurrence was 4.1% in the PRT+TME group (n=719) and 11.5% in the TME group (n=715) (Figure 1, p<0.0001). The hazard ratio for local recurrence after PRT+TME was 0.33 (95% CI 0.21–0.53). In the PRT+TME group, 74% and 91% of observed local recurrences had occurred within 2 and 3 years respectively. In the TME group, 78% and 93% of observed local recurrences had occurred within 2 and 3 years respectively (p=0.98 for the difference between randomization groups).
Figure 1. Local recurrence rates since TME-surgery by randomization group.
Of the 96 patients with a local recurrence, 81 patients (84%) were followed until death. For the remaining 15 patients, median time between date of local recurrence and date of last contact was 21 months (range 5-48 months). For 94 patients (98%) we received additional information on treatments after local recurrence from surgeons, radiation, or medical oncologists. Surgeons returned information on 88 out of 96 patients (92%), radiation oncologists on 49 out of 66 patients (74%), and medical oncologists on 28 out of 37 patients (76%).
Patient characteristics
Table 1 lists the characteristics of patients with locally recurrent rectal cancer by randomization group. Age at local recurrence diagnosis, tumor-node-metastasis (TNM) system stage of the primary tumor, type of resection of the primary tumor, confirmation by
No. at risk PRT+TME TME 719 715 641 631 567 547 377 391 189 204 56 62 Months since TME-surgery
CT or MRI, time to local recurrence, recurrence location and distant metastases diagnosed before and at the time of local recurrence diagnosis, were all not significantly different between randomization groups.
Table 1. Characteristics of patients with a local recurrence by randomization group
Variable PRT+TME
(n=23) (n=73) TME P-value
Demographics
Males (%) 78 59 0.09
Median age at local recurrence diagnosis (years) 62 67 0.26
Primary tumor TNM stage* (%) I+II III IV 35 48 17 25 66 9 0.29
Distance from the anal verge in cm (%) >10 5–10 ≤ 5 23 18 59 18 47 35 0.05 Abdominoperineal resection† (%) 48 36 0.29
Positive circumferential margin (%) 52 34 0.12
Local recurrence (LR)
Confirmation
Radiologically by CT or MRI (%)
Histologically or cytologically (%) 74 52 71 73 0.80 0.07 Median time to LR in months (range) 16 (4-58) 14 (3-59) 0.30 LR at the anastomosis or perineum only (%) 30 40 0.42
Distant metastases (DM)
At time of LR diagnosis‡ (%) Hazard ratio for DM§
Median DM-free interval in months§ Before and within 1 month after LR (%)
44 4.2 0.9 74 40 1 20 40 0.75 <0.0001 -- 0.004 * Due to the small number of TNM I tumors (n=1 in the PRT+TME group and n=4 in the TME alone group), TNM I and II
were grouped.
† Hartmann procedures (n=1 in the PRT+TME group and n=4 in the TME alone group) were classified as anterior resections.
‡ Distant metastases known before and at the time of local recurrence diagnosis.
§ In patients without distant metastases before and at the time of local recurrence diagnosis.
PRT+TME group, median time to distant metastases after local recurrence was 0.9 months (95% CI 0.3-1.5), compared to 20.1 months (95% CI 15.0-25.2) in the TME group. Therefore, for all other analyses, distant metasases diagnosed within one month of local recurrence diagnosis were also considered as occuring simultaneous with local recurrence, resulting in distant metastases percentages, diagnosed before and within one month of local recurrence diagnosis, of 74% and 40% in the PRT+TME and TME group respectively (p=0.004).
Figure 2. Distant metastases after local recurrence diagnosis by randomization group in patients without distant metastases before and at the time of local recurrence diagnosis.
Treatments for local recurrence
Table 2 lists the treatments for local recurrence. Surgical resection of local recurrence took place more often in the TME group (35%) than in the PRT+TME group (17%), although not significantly so (p=0.11). However, patients in the PRT+TME group were significantly more often treated with hyperthermia (p=0.02) and chemotherapy (p=0.03). Patients in the TME group received radiotherapy treatment for local recurrence at a dose ≥ 45 Gy (42%) significantly more often than patients in the PRT+TME group (4%) (p=0.001).
Despite distant metastases, 11% and 7% of patients underwent a surgical resection of local recurrence and were irradiated at a dose ≥ 45 Gy respectively, compared to respectively 48% and 56% of patients without distant metastases (all p<0.0001). Patients with pelvic recurrences underwent significantly less often surgical resection of local recurrence than patients with recurrences located at the anastomosis or perineum only (22% versus 46%, p=0.02). The difference in radiotherapy treatment at a dose ≥ 45 Gy for local recurrence was
not statistically significant between patients with and without pelvic recurrences (29% versus 40%, p=0.27).
Table 2. Treatments for locally recurrent rectal cancer by randomization group PRT+TME (n=23) TME (n=71)† P-value Types of treatments* Surgical resection (%) with IORT (%) without RT and CT (%)
with pre-or postoperative RT without CT (%) with pre-or postoperative RT and CT (%)
17 4 0 0 4 35 14 0 7 7 0.11 0.21 -- 0.19 0.65 without IORT (%) without RT and CT (%)
with pre-or postoperative RT without CT (%) with pre-or postoperative RT and CT (%)
13 13 0 0 21 7 13 1 0.39 0.37 0.07 0.57
Hyperthermia combined with RT (%) 13 2 0.02
RT only (%) 9 20 0.22
CT only (%) 31 11 0.03
Combined RT and CT (%) 4 10 0.41
None or supportive care‡ (%) 26 22 0.73
Radiotherapy dose ≥ 45 Gy (%) 4 42 0.001
* Abbrevations: IORT=intraoperative radiotherapy, RT=radiotherapy, CT=chemotherapy. † Patients of whom no treatment data were available (n=2) were excluded from analyses. ‡ Three patients in the TME alone group refused further treatment.
Survival
Figure 3. Survival after local recurrence diagnosis by randomization group.
The results of the univariate analyses of survival after local recurrence are shown in tables 3 and 4. PRT+TME, abdominoperineal resection of the primary tumor, positive circumferential resection margin of the primary tumor, pelvic recurrences, distant metastases diagnosed before and within one month of local recurrence diagnosis, no surgical resection of local recurrence and radiotherapy treatment for local recurrence at a dose < 45 Gy were all related to a shorter survival after local recurrence diagnosis.
Table 3. Univariate analyses of survival after local recurrence: patient- and disease related variables (n=96)
Variable Hazard ratio for death 95% CI P-value
Demographics Gender Males Females 1.00 1.08 0.68-1.70 0.75
Age at local recurrence diagnosis* 1.01 0.99-1.03 0.63
Primary tumor Randomization TME PRT+TME 1.00 2.09 1.21-3.57 0.008 TNM stage I+II III IV 1.00 1.57 1.99 0.93-2.67 0.89-4.45 0.15 0.09 0.09
Variable Hazard ratio for death 95% CI P-value Distance from the anal verge in cm
> 10 5–10 ≤ 5 1.00 0.83 1.26 0.45-1.53 0.68-2.32 0.26 0.54 0.46 Type of resection
Low anterior resection Abdominoperineal resection 1.00 1.51 0.96-2.37 0.08 Circumferential margin Negative Positive 1.00 2.17 1.37-3.43 0.001 Local recurrence (LR) Time to LR* 0.99 0.97-1.01 0.44 Recurrence location
At the anastomosis or perineum only Pelvic 1.00 2.00 1.23-3.25 0.005 Distant metastases (DM) ≤ 1 month of LR diagnosis No Yes 1.00 1.86 1.15-2.99 0.01
* Shown is the hazard ratio for death per increase in life years (age) or months (time to LR).
Table 4. Univariate analyses of survival after local recurrence: treatment-related variables (n=94)*
Variable Hazard ratio for death 95% CI P-value
Treatment for LR Surgical resection of LR Yes No 1.00 3.24 1.87-5.61 <0.0001 Hyperthermia Yes No 1.00 0.98 0.24-4.03 0.97 Radiotherapy (RT) only Yes No 1.00 0.73 0.41-1.29 0.28 Chemotherapy (CT) only Yes No 1.00 0.81 0.44-1.49 0.50
Combined RT and CT without surgical resection Yes
No
1.00
1.36 0.65-2.84 0.42
Radiotherapy dose for LR ≥ 45 Gy
None, or < 45 Gy
1.00
All variables that were univariately related (p<0.10) to survival were included in a multivariate Cox regression analysis to investigate whether those variables could explain the observed difference in survival between randomization groups (table 5). In this analysis, the difference in survival after local recurrence between randomization groups was no longer statistically significant (HR for death of PRT was 1.53, p=0.16). No surgical resection of local recurrence and radiotherapy at a dose < 45 Gy were associated with shorter survival after local recurrence (p=0.06 and p=0.001 respectively). Note, however, that the results of this analysis may not be interpreted as evidence of the effect of treatment, since patients with a more favorable prognosis are likely to be selected for treatment.
Table 5. Multivariate analysis of survival after local recurrence (n=94)*
Variables
Hazard ratio for death 95% CI P-value Randomization
TME PRT+TME
1.00
1.53 0.85-2.77 0.16
Type of primary resection Abdominoperineal resection Low anterior resection
1.00
1.31 0.75-2.29 0.34
Primary circumferential margin Negative
Positive
1.00
1.41 0.84-2.36 0.19
Local recurrence (LR) location
At the anastomosis or perineum only Pelvic
1.00
1.56 0.89-2.73 0.12
Distant metastases ≤ 1 month of LR diagnosis No Yes 1.00 0.72 0.39-1.32 0.29 Surgical resection of LR Yes No 1.00 1.88 0.96-3.68 0.06
Radiotherapy dose for LR ≥ 45 Gy
None, or < 45 Gy
1.00
3.20 1.58-6.47 0.001 * Variables were included in the multivariate analysis if they were univariately related (p<0.10) to survival after local
recurrence. Two patients were excluded from the multivariate analysis, because data on treatments for local recurrence were missing.
Discussion
PRT reduces local recurrence rates and thus may increase overall survival (8;12), patients who do develop a local recurrence after previous PRT have a significantly shorter survival than local recurrence patients that have not received PRT for the primary tumor. Median survival after local recurrence has been reduced from 16 months after previous TME-surgery without PRT, to only 6 months for patients with PRT. There are several possible explanations for this finding.
The first explanation concerns the nature of recurrences. PRT may be less effective for patients with unfavorable primary tumor characteristics (e.g. positive resection margins, higher primary tumor load), who are more at risk of developing distant metastases (8;13;14). Our results indeed showed that the majority of patients (74%) with a local recurrence after PRT+TME had simultaneous distant metastases, compared to 40% of patients in the TME group. In addition, local recurrences that develop after previous PRT may have different characteristics locally that negatively affect survival. In univariate analyses, the difference in survival after local recurrence between patients with and without PRT was not only associated with differences in distant metastases, but also with differences in circumferential resection margins and recurrence location. Further histopathological and molecular research of primary tumor and recurrence characteristics is needed to gain more insight into the pathway by which local recurrences develop after PRT.
The second explanation concerns the types of treatments for local recurrences. PRT may limit the possibilities for aggressive treatment of local recurrence, e.g. high dose radiotherapy treatments for local recurrence cannot be given to patients who have previously been irradiated, because of the maximally tolerated dose. Our results confirm that patients of the PRT group received radiotherapy treatment for local recurrence at a dose ≥ 45 Gy less often. Also, patients with a local recurrence after PRT underwent surgical resection of local recurrence less often. In a multivariate analysis, those treatment variables were significantly associated with survival after local recurrence diagnosis. However, patients with a more favorable prognosis are clinically selected (e.g. by distant metastases, recurrence location) for treatment. Therefore, we cannot determine the separate effects of clinical patient characteristics and treatments for local recurrence on the difference in survival between randomization groups.
A limitation of our study could be that PRT may postpone recurrences and patients with an unfavorable prognosis may present themselves with shorter follow-up. For several reasons, we do not believe that longer follow-up would significantly affect our results. First of all, median time between TME-surgery and local recurrence diagnosis was not significantly different between randomization groups and was also not significantly related to survival after local recurrence diagnosis. Secondly, we selected local recurrence patients from the TME-study at a median follow-up time of 43 months and the local recurrence curves in both randomization groups were flattening out, suggesting that most of the recurrences have become overt.
chemotherapy, either adjuvantly or neo-adjuvantly, could have improved the prognosis of patients in the PRT+TME group (25). We believe it is thereby important to distinguish the different purposes of chemotherapy. The main aim of concurrent chemoradiation is radiosensitization, that is to improve local control and attack any micrometastases. Patients who develop a local recurrence despite chemoradiation may have very unfavorable primary tumor characteristics, and we would expect the results for those patients to be at best equal to, but probably worse than the results in our patients. The main aim of postoperative adjuvant chemotherapy is to prevent distant recurrences in selected groups of patients. For some patients in our study the addition of postoperative chemotherapy could have prevented or delayed the occurrence of distant metastases, and as a result we would expect a more favorable prognosis after local recurrence. Note however that a decision on the addition of chemotherapy should be based on the results of chemotherapy in all patients, including the large majority of patients who do not develop a local recurrence. So far, studies that investigate the results of post-operative chemotherapy combined with pre-operative short-term radiotherapy have not yet reported their results.
In the past, many studies have focused on the development of aggressive treatment modalities for locally recurrent rectal cancer to improve survival, and reduce re-recurrence rates. The main rationale for this focus was that 50%-80% of patients had no manifest distant metastases at the time of local recurrence diagnosis, and that 21%-50% of patients died without distant metastases, suggesting that there were patients who could be cured, and that local recurrence was not inevitably an early manifestation of widespread disease (1-3;15;17-20). The adoption of PRT+TME as the new standard of treatment for patients with rectal cancer calls for reconsideration of this viewpoint. Our results showed that at least 74% of patients with a local recurrence after previous PRT+TME have simultaneous distant metastases. For many patients, aggressive treatment approaches that may also be associated with significant morbidity and mortality (16;18-24) may no longer be appropriate.
In conclusion, the introduction of PRT has reduced local recurrence rates and may thus increase overall survival, but has also changed the clinical nature and prognosis of those patients who do develop locally recurrent rectal cancer. Median survival after local recurrence has decreased from 16 months without previous PRT, to 6 months with previous PRT. The majority of patients who present with a local recurrence after previous PRT have simultaneous distant metastases. Patients with a local recurrence after previous PRT should be meticulously scanned for distant metastases before aggressive treatments for local recurrence are given.
Acknowledgments
References
(1) Frykholm GJ, Pahlman L, Glimelius B: Treatment of local recurrences of rectal carcinoma. Radiother Oncol 34(3):185-194, 1995
(2) Holm T, Cedermark B, Rutqvist LE: Local recurrence of rectal adenocarcinoma after 'curative' surgery with and without preoperative radiotherapy. Br J Surg 81(3):452-455, 1994
(3) Wiggers T, de Vries MR, Veeze-Kuypers B: Surgery for local recurrence of rectal carcinoma. Dis Colon Rectum 39(3):323-328, 1996
(4) Wiggers T: Staging of rectal cancer. Br J Surg 90(8):895-896, 2003
(5) Enker WE, Thaler HT, Cranor ML, et al: Total mesorectal excision in the operative treatment of carcinoma of the rectum. J Am Coll Surg 181(4):335-346, 1995
(6) Heald RJ: Total mesorectal excision is optimal surgery for rectal cancer: a Scandinavian consensus. Br J Surg 82(10):1297-1299, 1995
(7) MacFarlane JK, Ryall RD, Heald RJ: Mesorectal excision for rectal cancer. Lancet 341(8843):457-460, 1993
(8) Kapiteijn E, Marijnen CA, Nagtegaal ID, et al: Preoperative radiotherapy combined with total mesorectal excision for resectable rectal cancer. N Engl J Med 345(9):638-646, 2001
(9) Enker WE: Potency, cure, and local control in the operative treatment of rectal cancer: Arch Surg 127(12):1396-1401, 1992
(10) Heald RJ, Karanjia ND: Results of radical surgery for rectal cancer. World J Surg 16(5):848-857, 1992 (11) Philipsen SJ, Heilweil M, Quan SH, et al: Patterns of pelvic recurrence following definitive resections of
rectal cancer. Cancer 53(6):1354-1362, 1984
(12) van den Brink M, van den Hout WB, Stiggelbout AM, et al: Cost-utility analysis of pre-operative radiotherapy in patients with rectal cancer undergoing total mesorectal excision. J Clin Oncol (in press) (13) Marijnen CA, Nagtegaal ID, Kapiteijn E, et al: Radiotherapy does not compensate for positive resection margins in rectal cancer patients: report of a multicenter randomized trial. Int J Radiat Oncol Biol Phys 55(5):1311-1320, 2003
(14) Nagtegaal ID, Marijnen CA, Kranenbarg EK, et al: Circumferential margin involvement is still an important predictor of local recurrence in rectal carcinoma: not one millimeter but two millimeters is the limit. Am J Surg Pathol 26(3):350-357, 2002
(15) Wong CS, Cummings BJ, Brierley JD, et al: Treatment of locally recurrent rectal carcinoma--results and prognostic factors. Int J Radiat Oncol Biol Phys 40(2):427-435, 1998
(16) Suzuki K, Dozois RR, Devine RM, et al: Curative reoperations for locally recurrent rectal cancer. Dis Colon Rectum 39(7):730-736, 1996
(17) Valentini V, Morganti AG, De Franco A, et al: Chemoradiation with or without intraoperative radiation therapy in patients with locally recurrent rectal carcinoma: prognostic factors and long term outcome. Cancer 86(12):2612-2624, 1999
(18) Law WL, Chu KW: Resection of local recurrence of rectal cancer: results. World J Surg 24(4):486-490, 2000
(19) Alektiar KM, Zelefsky MJ, Paty PB, et al: High-dose-rate intraoperative brachytherapy for recurrent colorectal cancer. Int J Radiat Oncol Biol Phys 48(1):219-226, 2000
(20) Lopez-Kostner F, Fazio VW, Vignali A, et al: Locally recurrent rectal cancer: predictors and success of salvage surgery. Dis Colon Rectum 44(2):173-178, 2001
(21) Salo JC, Paty PB, Guillem J, et al: Surgical salvage of recurrent rectal carcinoma after curative resection: a 10-year experience. Ann Surg Oncol 6(2):171-177, 1999
(22) Garcia-Aguilar J, Cromwell JW, Marra C, et al: Treatment of locally recurrent rectal cancer. Dis Colon Rectum 44(12):1743-1748, 2001
(23) Mannaerts GH, Rutten HJ, Martijn H, et al: Comparison of intraoperative radiation therapy-containing multimodality treatment with historical treatment modalities for locally recurrent rectal cancer. Dis Colon Rectum 44(12):1749-1758, 2001
(24) Wanebo HJ, Koness RJ, Vezeridis MP, et al: Pelvic resection of recurrent rectal cancer. Ann Surg 220(4):586-595, 1994
