|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 22, No 20 (October 15), 2004: pp. 4103-4108 © 2004 American Society of Clinical Oncology. DOI: 10.1200/JCO.2004.04.127 Is Eligibility for a Chemotherapy Protocol a Good Prognostic Factor for Invasive Bladder Cancer After Radical Cystectomy?From the Department of Urology, University of Bern, Bern, Switzerland Address reprint requests to Stephan Madersbacher, MD, FEBU, Associate Professor, Department of Urology, Inselspital, CH-3010 Bern, Switzerland; e-mail: urs.studer{at}insel.ch
PURPOSE: To assess whether eligibility to an adjuvant chemotherapy protocol in itself represents a good prognostic factor after radical cystectomy for bladder cancer. PATIENTS AND METHODS: Between April 1984 and May 1989, our institution entered 35 patients with invasive bladder cancer into the Swiss Group for Clinical and Epidemiological Cancer Research (SAKK) study 09/84. They were randomly assigned to either observation or three postoperative courses of cisplatin monotherapy after cystectomy. This study had a negative result. The outcome of these 35 patients (protocol group) was compared with an age- and tumor-stage–matched cohort (matched group; n = 35) who also underwent cystectomy during the same period, but were not entered into the SAKK study, as well as the remaining 57 patients treated during the study period for the same indication (remaining group). RESULTS: Median overall survival decreased from 76.3 months in the protocol group to 52.1 months in the matched group and to 20.3 months in the remaining group. The respective times of median recurrence-free survival were 67.2, 16.0, and 9.4 months. Tumor progression occurred in 46% of the protocol group compared with 69% in the matched group and 65% in the remaining group (P < .05). Cancer-related death was noted in 40% of the protocol group, 57% in the matched group, and 56% in the remaining group. CONCLUSION: These data suggest that being willing and fit enough for a chemotherapy protocol is a good prognostic factor for invasive bladder cancer. This eligibility bias emphasizes the need for prospective, randomized trials, and indicates that single-group studies using historical or matched controls have to be interpreted with caution.
Application of the conclusions of clinical trials to general practice is inherently limited by the bias of patient selection criteria.1-6 These systematic selection biases may result from the specific inclusion and exclusion criteria required for research by scientific as well as individual considerations.1-6 The comparison of new treatment modalities with so-called matched controls and particularly historic controlled series may be misleading as well.7 Such comparisons confuse the results by introducing errors resulting from case selection bias, stage migration, differences in follow-up, and the evolution of supportive care. These confounding, recall, and detection biases are particularly problematic for the results of oncologic trials because the respective surgical or medical therapies can be associated with considerable treatment-related morbidity. In 1984, the Swiss Group for Clinical and Epidemiological Cancer Research (SAKK) study 09/84 was initiated to study the efficacy of adjuvant cisplatin monotherapy after radical cystectomy for bladder cancer in a randomized clinical trial (RCT).8 The result of this RCT was negative.8 Had this RCT been a single-group phase II study, the excellent 70% 5-year survival rate for organ-confined tumors and the 40% 5-year survival rate for non–organ-confined tumors would have been impressive.8 We therefore hypothesized that the favorable outcome in the SAKK study 09/84 was due to the study selection criteria.8 We compared the outcome of the 35 patients with invasive bladder cancer who were entered into the SAKK study 09/84 by our institution (irrespective of the study arm) to an age- and tumor-stage–matched group of patients (n = 35) who also underwent radical cystectomy during the study period at our institution, but were not eligible for or refused to be randomly assigned into the SAKK 09/84 trial, as well as to the remaining 57 patients treated by radical cystectomy at our institution during the study period. All patients were preoperatively staged N0M0, and the surgical techniques of pelvic lymphadenectomy and radical cystectomy were standardized throughout the study period.
SAKK 09/84 Study Results of SAKK 09/84 study have been published previously.8 In brief, patients entered into the SAKK 09/84 multicenter study were those preoperatively staged N0M0 who required radical cystectomy. Reasons for exclusion were other previous or concomitant malignancies, a WHO performance status of grade 2 or more, age of 75 years or older, or a postoperative creatinine clearance level of less than 60 mL/min. Between April 1984 and May 1989, 77 eligible patients were stratified after radical cystectomy and pelvic lymph node dissection on the basis of nodal status (stage pN0 v pN1-2) and randomly assigned to observation or postoperative cisplatin chemotherapy (three courses of cisplatin 90 mg/m2 given for 3 consecutive days at monthly intervals).8 The Department of Urology, University of Bern (Bern, Switzerland) contributed 48% of the patients in this trial.
Surgical Technique Pelvic lymphadenectomy. In brief, standard pelvic lymphadenectomy involves meticulous removal of all lymphatic and connective tissue within the following boundaries: laterally, the genitofemoral nerve; distally, the femoral canal; proximally, the crossing of the ureter with the common iliac artery; inferiorly, the side wall of the obturator muscle and floor of the obturator fossa down to the internal-iliac vessels at the level of superior and inferior vesical arteries; and medially, the side wall of the bladder. Radical cystectomy. After the dorsolateral bladder pedicles containing superior and inferior vesical vessels are divided along the hypogastric arteries, the endopelvic fascia is incised along the dorsolateral side of the prostate and Santorini's plexus is ligated. The ureters are divided where they cross iliac vessels. This allows en bloc removal of the distal ureters and paraureteral lymphatic vessels, together with the cystectomy specimen. The dorsomedial pedicle is resected along the pararectal and presacral plane on the tumor-bearing side, on the opposite side along the dorsolateral wall of the seminal vesicle until the base of the prostate is reached. Santorini's plexus is then divided over the ventral aspect of the prostate and the membranous urethra is transected as close as possible to the apex of the prostate by dissecting it from the donut-shaped apex in an ascending manner. In women, the procedure usually involves en bloc resection of the uterus together with most of the anterior wall of the vagina and the entire urethra unless an orthotopic bladder substitution is performed.
Pathology
Follow-Up
Statistical Analysis
Patients Between April 1984 and May 1989, a total of 127 patients underwent pelvic lymphadenectomy and radical cystectomy for invasive bladder cancer at our institution (Table 1). All of them were preoperatively staged N0M0 and had no preoperative chemotherapy or radiation therapy. Of these 127 patients, 35 (28%) entered into the SAKK 09/84 study. The remaining 92 (72%) were either not eligible or not willing to participate. To study a potential eligibility bias influencing the oncologic outcome, patients were divided into the following three groups: patients treated in our institution and included to the SAKK 09/84 study (n = 35; protocol group), irrespective of the study arm (no treatment or cisplatin monotherapy); an age- and tumor-stage–matched control group to the SAKK 09/84 study group (n = 35; matched group; matching was performed one to one regarding tumor stage and ± 2 years regarding age); and 57 patients (remaining group) who had surgery during the study period but were not included in either of the two above-mentioned groups (Table 1). Patient characteristics including tumor stage and comorbidities of the entire population and of the three study groups are listed in Table 1. The protocol group was slightly younger than the other two groups, but not significantly so. With the exception of chronic obstructive pulmonary disease, all other recorded comorbidities such as diabetes, hypertension, and coronary artery disease were less frequent in the protocol group. Serum creatinine 14 days postoperatively was also lower in the protocol group compared with the matched group and the remaining group; the latter was statistically significant (Table 1). Tumor stages including lymph node status were evenly distributed between the protocol group and the matched group (Table 1). Patients with positive lymph nodes and pT4 tumors were less frequent in the protocol group and the matched group compared with the remaining-group (Table 1).
Overall and Recurrence-Free Survival The median overall survival of the entire study population was 38.9 months (range, 0.3 to 176.4 months; mean, 62.6 ± 55.8 months [standard deviation]). Median overall survival decreased from 76.3 months (range, 1.1 to 176.4 months) in the protocol group to 52.1 months (range, 1.2 to 166.7 months) in the matched group and to 20.3 months (range, 0.3 to 147.7 months) in the remaining 57 patients (Table 2). The difference between the protocol and the matched group, as assessed by a pair-wise log-rank test, was not statistically significant (P = .6; Fig 1). The difference between the protocol group and the remaining 57 patients was of borderline statistical significance (P = .05; Fig 1). The estimated 5-year probability for overall survival was 52% for the protocol group, 48% for the matched group, and 30% for the remaining group (Table 2).
Median recurrence-free survival was significantly longer in the protocol group (median, 67.2 months; range, 1.1 to 176.4 months) compared with the matched group (median, 16.0 months; range, 1.2 to 166.7 months; P = .02;) and the remaining group (median, 9.4 months; range, 1.3 to 147.7 months; P = .01; Table 2, Fig 2). Recurrence-free survival between the matched group and the remaining 57 patients was similar (P = .83; Table 2; Fig 2). The estimated 5-year probability for recurrence-free survival was 60% for the protocol group, 33% for the matched group, and 32% for the remaining group.
Progression, No Evidence of Disease, and Cause of Death Of the 127 patients, 75% have died. This emphasizes the maturity of our data (Table 3). Overall, clinical progression was observed in 61% of the patients. Sixty-six patients (52%) died as a result of their cancer, five (4%) died as a result of complications (sepsis, n = 3; cardiac arrest, n = 1; pulmonary embolism, n = 1), and 24 (19%) died as a result of other causes. In the protocol group the progression rate was 46% compared with 69% in the matched group and 65% in the remaining group. The difference between the protocol and the other two groups was statistically significant. Transitional cell cancer–related death occurred in 40% of patients in the protocol group compared with 57% in the matched group and 56% in the remaining patients (Table 3).
Awareness of selection biases and recruitment rates is important for interpreting findings from individual studies. Variables related to the inclusion of patients may be related to demographic, comorbid, diagnostic, or other clinical variables that in turn may affect treatment outcome.5 In the larger context of public health issues, patient selection bias is one factor that may limit the ability to extrapolate findings from research trials to the clinical setting. Pelvic lymphadenectomy and radical cystectomy are currently considered the standard of care for patients with muscle-invasive bladder cancer.9, 11, 12 The excellent 70% 5-year survival rate of patients undergoing radical cystectomy for undifferentiated bladder cancer confined to the bladder wall and the surprising 40% 5-year survival rate in patients with growth of urothelial cancer into the perivesical fat or the prostate in the SAKK 09/84 trial led us to the hypothesis that this favorable outcome might be due to the study selection process.8 As reported here, we were indeed able to confirm this hypothesis, given that the overall survival and in particular the recurrence-free survival was longer for patients included to the SAKK study than for those not eligible or not willing to participate (Figs 1 and 2). This indicates that promising results from nonrandomized clinical trials must be interpreted with caution and that the comparison to matched controls may lead to erroneous conclusions. One strength of this study is its single-center nature, which is particularly important for surgical series. During the study period the surgical techniques of pelvic lymphadenectomy and radical cystectomy were standardized and remained unchanged. The second point of strength is the maturity of the data; 75% of the 127 patients have died and the follow-up period was long. The long follow-up and maturity of data stand in contrast to the majority of phase II oncologic trials that frequently report promising data. One limitation of the current study is the relatively small sample size. However, of relevance is not so much the sample size but rather the fact that the differences observed were highly significant. It is therefore unlikely that larger sample sizes would have resulted in other conclusions. The second limitation is the retrospective nature of the study. Although the role of a potential eligibility bias in this setting can only be determined in a retrospective study (such as in the current one), we were not able to determine exactly why patients were not included or why they refused to participate in the SAKK 09/84 trial. Although this is the first demonstration of a selection bias in the field of urologic medical oncology, such bias has been shown for other medical specialties. Rahman et al,13 for instance, reviewed 18 successive doxorubicin-containing protocols (a total of 1,581patients) for the treatment of metastatic breast cancer and studied the impact of the eligibility to a high-dose chemotherapy (HDCT) protocol on the outcome. The complete response rate, for example, was 27% for HDCT candidates versus 7% for the noncandidates.13 The authors concluded that the encouraging results of single-arm trials of HDCT could be due partially to the selection of patients, emphasizing the need for randomized trials.13 Similar conclusions were drawn by Garcia-Carbonero14 in analyzing high-risk breast cancer patients. In the field of cardiovascular medicine, Bjorn et al3 emphasized that patients in conventional cardiovascular trials not using consecutive recruitment are selected and are not representative of the disease under investigation. Patients in cardiovascular studies are younger, are more often male, and have a lower overall risk.3 Robinson et al15 evaluated this issue in a multicenter schizophrenia treatment study. Six thousand twelve diagnostically appropriate subjects were screened for the study; of these, 1,320 met the eligibility criteria and only 528 (9% of the screened sample) entered onto the study. Although the authors concluded that in this setting selection biases would rarely alter the overall study outcomes to a clinically relevant degree, selection biases did limit the ability to make inferences about the results for select small subgroups of the study population.15 The mechanisms leading to the prolonged overall and recurrence-free survival of the protocol group compared with the other two groups remain to be elucidated. Given that the SAKK-09/84 study was negative, this difference cannot be attributed to the (marginal) beneficial effect of an adjuvant cisplatin monotherapy after cystectomy. Given that the protocol and matched groups were comparable with respect to age and tumor stage, other factors have to be involved. There is no doubt that clinical trials tend to have a positive patient selection, including motivation. Patients willing to receive adjuvant therapy after radical cystectomy want the maximum to be done for their health and probably also consulted a physician for help at the first symptoms. Other patients caring less about their health might have a lead-time bias between first symptoms and treatment because they see their physician only when symptoms become stressing. Such patients also are often not interested in adjuvant treatment with uncertain benefit. Furthermore, a placebo effect cannot be excluded even in an oncologic setting with hard end points such as progression or survival. Patients participating in trials tend to get closer attention and have more visits, which might also bias the results. In addition to the selection bias, the rather low number of patients entering clinical trials is of concern. Even in our setting (an academic center and with extensive experience participating in clinical trials), only 35 of 127 (27.6%) patients were included in the SAKK 09/84-trial. An extreme example in this respect is one of the landmark studies on metastatic renal-cell cancer.16 It took 80 urologic institutions 7 years to recruit 246 patients with a relatively common disease (metastatic renal-cell cancer).16 The recruitment rate was only one patient every 2 years per institution.17 This slow recruitment suggests that occult criteria were unintentionally applied by doctors when patients were invited to participate in the study, or were unintentionally applied by patients when agreeing to enter the study.17 Recruitment of such a small proportion of eligible patients evidently raises questions about the applicability of the study results to the overall population of patients with metastatic renal cancer.17 In conclusion, our data suggest that being willing and fit enough to enter a chemotherapy protocol seems to be a good prognostic factor for muscle-invasive bladder cancer. The promising data from usually short-term, nonrandomized, phase II clinical trials, therefore, have to be interpreted with great caution.18 In addition, so-called matched controls do not reflect the overall patient population and as such may be misleading. Whenever possible, only results from well-designed, prospective, randomized trials should be used when making recommendations for standard care. It should be of public interest for governments to generously sponsor and support such trials and the participating patients.
The authors indicated no potential conflicts of interest.
Presented at the 27th Congress of the European Association of Urology, Birmingham, UK, February 23-26, 2002. Authors' disclosures of potential conflicts of interest are found at the end of this article.
1. Lee JY, Marks JE, Simpson JR: Age as a criterion for eligibility in a lung cancer clinical trial. Am J Clin Oncol 5: 449-452, 1982[Medline] 2. Wittes J, Lakatos E, Black D, et al: Selecting screening criteria for clinical trials: An example from the systolic hypertension in the elderly program. Control Clin Trials 19: 121-132, 1999 3. Bjorn M, Brendstrup C, Karlsen S, et al: Consecutive screening and enrollment in clinical trials: The way to representative patient samples? J Card Failure 4: 225-230, 1998[CrossRef][Medline] 4. Canellos GP: Selection bias in trials of transplantation for metastatic breast cancer: Have we picked the apple before it was ripe? J Clin Oncol 15: 3169-3170, 1997[Medline] 5. Winger MJ, Macdonald DR, Schold SC Jr, et al: Selection bias in clinical trials of anaplastic glioma. Ann Neurol 26: 531-534, 1989[CrossRef][Medline] 6. Taylor JM, Withers HR: Influence of selection bias on dose-time analysis: A theoretical investigation. Radiother Oncol 40: 259-270, 1996[CrossRef][Medline] 7. Bhansali MS, Patil PK, Badwe RA, et al: Historical control bias: Adjuvant chemotherapy in esophageal cancer. Dis Esophagus 10: 51-54, 1997[Medline] 8. Studer UE, Bacchi M, Biederman C, et al: Adjuvant cisplatin chemotherapy following cystectomy for bladder cancer: Results of a prospective randomized tiral. J Urol 152: 81-84, 1994[Medline]
9. Madersbacher S, Hochreiter W, Burkhard F, et al: Radical cystectomy for bladder cancer today: A homogeneous series without neoadjuvant therapy. J Clin Oncol 21: 690-696, 2003 10. Fleming ID, Cooper JS, Henson DE, et al (eds.): AJCC Cancer Staging Manual (ed 5). Philadelphia, PA, Lippincott-Raven, 1997, pp. 241-243
11. Stein JP, Lieskovsky G, Cote R, et al: Radical cystectomy in the treatment of invasive bladder cancer: Long-term results in 1,054 patients. J Clin Oncol 19: 666-675, 2001 12. Sternberg CN, Parmar MKB: Neoadjuvant chemotherapy is not (yet) standard treatment for muscle-invasive bladder cancer. J Clin Oncol 19: 21-26, 2001 (suppl) 13. Rahman ZU, Frye DK, Buzdar AU, et al: Impact of selection process on response rate and long-term survival of potential high-dose chemotherapy candidates treated with standard-dose doxorubicin-containing chemotherapy in patients with metastatic breast cancer. J Clin Oncol 15: 3171-3177, 1997[Abstract] 14. Garcia-Carbonero R, Hidalgo M, Paz-Ares L, et al: Patient selection in high-dose chemotherapy trials: Relevance in high-risk breast cancer. J Clin Oncol 15: 3178-3184, 1997[Abstract] 15. Robinson D, Woerner MG, Pollack S, et al: Subject selection biases in clinical trials: Data from a multicenter schizophrenia treatment study. J Clin Psychopharmacol 16: 170-176, 1996[CrossRef][Medline]
16. Flanigan RC, Salmon SE, Blumenstein BA, et al: Nephrectomy followed by interferon alpha-2b compared with interferon alpha-2b alone for metastatic renal-cell cancer. N Engl J Med 345: 1655-1659, 2001
17. Tannock IF: Removing the primary tumour after the cancer has spread. N Engl J Med 345: 1699-1700, 2001 18. Liu PY, LeBlanc M, Desai M: False positive rates of randomized phase II designs. Control Clin Trials 20: 343-352, 1999[CrossRef][Medline] Submitted April 26, 2004; accepted August 4, 2004.
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2004 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
.05 was considered as significant. 
