In some cases of radical cystectomy for bladder cancer, no residual tumor is found in the cystectomy specimen (the pT0 classification). The aim of this study was to evaluate the outcome of such patients in a large cystectomy series.
In some cases of radical cystectomy for bladder cancer, no residual tumor is found in the cystectomy specimen (the pT0 classification). The aim of this study was to evaluate the outcome of such patients in a large cystectomy series.
All 900 patients with radical cystectomy and pelvic lymphadenectomy for TCC of the bladder in the period January 1986 to September 2003 who received no neoadjuvant therapy were included. Cystectomy specimens from 181 (20.1%) patients were graded as pT0. Complete follow-up was obtained in all cases. Tumor-specific survival (pT0 vs. pT+) was calculated with the Kaplan–Meier method and compared with the log-rank test.
The rate of lymph node metastases in the pT0 group was 6.6%. pT0 status was found with Ta/is/1 in 36.8%, T2a in 41.8%, and T2b in 10.9%. The 169 patients with pT0pN0 tumors had 10-year tumor-specific survival rates of 91.0. There was no statistically significant survival benefit for pT0pN0 tumors compared with pT+pN0 tumors for maximal tumor classifications of pTa/pTis/pT1 and pT2b, but patients with a pT0T2apN0 tumor had a statistically significantly better tumor-specific survival than those with a pT2apN0 tumor (P = 0.002). No patient with a pT0pN0 tumor had a local recurrence. The rate of incidental second primary malignancies in a specimen was 15.5%.
A pT0pN0 cystectomy specimen indicates a curative therapy, but there is a substantial risk of tumor recurrence. In the group of tumors with a maximal classification of pT2a, the pT0 tumors constitute a subgroup with a significantly higher likelihood of survival. Cancer 2005. © 2005 American Cancer Society.
Radical cystectomy is the accepted standard therapy for locally confined, muscle-invasive, or highly aggressive superficial primary malignant tumors of the bladder.1 Usually a transurethral resection of the bladder tumor (TUR-BT) is performed first to confirm the histological diagnosis. In some cases, the pathologist will not find any residual tumor in the cystectomy specimen; this defines the lesion as pT0. There are few and contradictory data on the clinical effect of such a situation. Two courses have to be considered: Complete resection of the tumor with the transurethral resection alone and tumor elimination with neoadjuvant chemotherapy and/or radiotherapy preceding radical cystectomy. In the latter case, the initial tumor staging is based on the data from the TUR-BT specimen and clinical imaging, both of which can be flawed by a distinct staging error.2 The true tumor stage can be accurately determined in the surgery-only group but not in the neoadjuvant-therapy group. By using pT0 cases that belong to both groups, most studies presented so far have been analyzing inhomogeneous populations.
Our study used a large, single-center, surgery-only series of patients who had transitional cell carcinoma (TCC) of the bladder only. Our aim was to analyze the stage-dependent effect of a pT0 cystectomy specimen on survival in a homogenous population.
This analysis was based on the complete data on patients in whom cystectomy was performed at our institution in the period January 1986 through August 2003 (n = 1131). Complete follow-up was obtained on all patients especially with respect to tumor recurrence and/or progression and tumor-related and tumor-unrelated death.
Inclusion criteria for this study were: primary malignancy of the urinary bladder in which transitional cell carcinoma was the only histological component, radical cystectomy and bilateral pelvic lymphadenectomy (at least in the obturator fossa and along the internal and external iliac artery),3, 4 and urethrectomy in all cases that had a positive urethral margin on frozen section.
Exclusion criteria were: neoadjuvant radiotherapy and/or chemotherapy and intravesical therapy in the interval between the last TUR-BT and cystectomy.
All histopathologic reports were classified according to the American Joint Committee on Cancer (AJCC) 1997 classification.5 All TUR-BT specimens that had uncertain findings were reevaluated preoperatively by a second uro-pathologist. The examination of the cystectomy specimen comprised thorough analysis of the TUR-BT lesion and its environment, of all suspect areas of the bladder wall, and of exemplary samples of unsuspicious areas. Every case in which the first analysis showed no residual tumor was reexamined with additional slices taken from the tumor site and its surroundings.
The pT classification was defined as the tumor stage of the cystectomy specimen, the clinical T classification as the maximal preoperative tumor classification from the TUR-BT specimens, and the maximal pT classification as the maximal tumor classification for all specimens (TUR-BT, open biopsy, and cystectomy).
Patients were compared according to the stage of the cystectomy specimen—tumor-free (pT0) versus residual tumor (pT+)—and according to the maximal pT classification. Data on tumor diameter could be obtained in only 60% of the cases, so we did not include this measure in our analyses.
Overall survival, tumor-specific survival, and progression-free survival were calculated with the Kaplan–Meier method.6 Differences in survival rates between pT0 and pT+ groups were evaluated with the log-rank test (significance, P < 0.05).7 The composition of the groups was compared with the chi-square test. For the recurrence-free survival and tumor-specific survival, a Cox model was calculated with several variables (stage, lymph node status, pT0 cystectomy specimen, local recurrence, distant metastasis, and age at the time of surgery). Local recurrence and distant metastasis were considered time-dependent.
Of the 1131 patients, 900 (79.6%) met the criteria for inclusion. Of these, 181 (20.1%) had a tumor-free cystectomy specimen. Because all patients with a pT0 cystectomy specimen had a maximal pT classification ≤pT2b, the analysis of the pT+ group was confined to the analogous maximal tumor classifications (≤pT2b: n = 373). The rate of T0 cystectomy specimens among these 554 patients with a maximal pT classification ≤pT2b was 32.7%.
Of 900 patients who had no neoadjuvant therapy and whose cystectomies showed primary TCC of the bladder, 181 (20.1%) had histologically proven pT0 cystectomy specimens. Twelve (6.6%) patients with tumor-free cystectomy specimens had at least one positive lymph node: two with clinical stage T1 and five each with clinical T2a and T2b tumors. Among lymph node-negative patients, the maximal pT classification of tumor-free cystectomy specimens was 23.8% (5/21) pTa, 34.8% (8/23) pTis, 38.4% (63/164) pT1, 41.8% (82/196) pT2a, and 10.9% (11/101) pT2b (Table 1).
Among patients whose TUR-BT specimens had a clinical classification of Ta, 17.2% proved to have a tumor-free cystectomy specimen, but another 17.2% were understaged and had muscle-invasive disease. Of the patients with clinical Tis classification, 38.1% had pT0 cystectomy specimens, but no patient was understaged. Specimens of the patients with a clinical T1 classification were tumor-free in 28.4% and muscle-invasive in 27.9%. There were no significant differences between patients with tumor-bearing and tumor-free cystectomy specimens with regard to age, sex, number of bladder-tumor events before cystectomy, follow-up, tumor grading, incidence of simultaneous prostate cancer, or form of urinary diversion (Table 2).
|Age (median)||62.0 yrs||63.0 yrs|
|Follow-up (median)||64.0 mos||60.0 mos|
|Male||80.7 %||84.8 %|
|Female||19.3 %||15.2 %||0.268|
|Neobladder||86.7 %||81.1 %|
|Other forms of urinary diversion||13.3 %||18.9 %||0.122|
|Max. pTa/pTis/pT1||43.1 %||35.9 %|
|Max. pT2a/pT2b||56.9 %||64.1 %||0.125|
|pN0||93.4 %||90.1 %|
|pN+||6.6 %||9.9 %||0.263|
|GI-II||30.4 %||22.9 %|
|GIII||69.6 %||77.1 %||0.074|
|Primary tumor||72.9 %||72.8 %|
|Recurrent tumor||27.1 %||27.2 %||0.944|
|No prostate cancer||82.2 %||80.2 %|
|Prostate cancer||17.8 %||19.8 %||0.702|
The overall, tumor-specific, and recurrence-free survival rates of all 181 patients with tumor-free cystectomy specimens were 79.2%, 89.7%, and 88.3% at 5 years, respectively, and 65.9%, 86.5%, and 85.2% at 10 years, respectively. Of the 169 patients who had negative lymph nodes, those rates were 83.8%, 94.4%, and 91.7% at 5 years and 69.4%, 91.0%, and 88.4% at 10 years, respectively.
In patients with pT0 pN0 tumors, the tumor-specific survival rate was 95.7–100.0% at 5 years and 92.0–100.0% at 10 years if the clinical T classification did not exceed T2a; in patients with T2b pT0 pN0 tumors, that rate was only 70.7% at 5 and 10 years. That difference was statistically significant in the log-rank test (P = 0.004), but the differences were not significant for Ta, Tis, T1, and T2a tumors (Fig. 1).
When tumor-specific survival was analyzed according to the maximal clinical tumor stage in lymph node-negative patients, no significant differences were seen in superficial tumors (maximum pT classification, pTa/is/1) (Fig. 2), but there was a significant difference in tumors with maximal pT2a classification: the 10-year tumor-specific survival was 92.0% in the group with tumor-free cystectomy specimens, but only 70.9% in the group with residual tumors (P = 0.002) (Fig. 3).
The number of patients with T2b pT0 pN0 tumors was small (11 patients). The log-rank test did not reveal a statistical difference in tumor-specific survival when these patients were compared with the pT2b pN0 group (P = 0.936) (Fig. 4). Patients with tumor-free cystectomy specimens had a significantly better recurrence-free survival when the lymph nodes were tumor-free than when cystectomy specimens were tumor-free but there were lymph node metastases (P = 0.001).
The number of TUR-BTs immediately before cystectomy had a significant effect on the probability of tumor-free cystectomy: 28.2% of patients with pT0 status had two or more TUR-BTs, but only 18.6% of pT+ patients (P = 0.019). Lymph node-negative patients had a distinctly different pattern of site of first recurrence from patients with identical maximal tumor stage with pT0 versus pT+ cystectomy specimen: Local recurrences as first site of recurrence occurred in 5.7% of pT+ patients, but no case was detected in the pT0 group. Of the patients with pT0 tumors, 4.7% developed distant metastases, but 14.0% of those with pT+ tumors.
The rate of de novo TCC of the upper urinary tract or the urethra was 2.96% in the pT0 patients and 2.7% in the pT+ patients. In 28 (15.5%) patients with pT0 TCC cystectomy specimens, a second primary malignancy was detected in the specimen. One patient had a carcinoma in situ of the cervix, one had a Hodgkin lymphoma in the pelvic lymph nodes, and 26 patients had adenocarcinoma of the prostate diagnosed. The prostate cancer was organ-confined in 20 cases and infiltrating the periprostatic fat in 5 cases; 1 patient had an infiltration of both seminal vesicles. The Gleason score was 4 in 23%, 5 in 27%, 6 in 27%, 7 in 19%, and 8 in 4%. The mean preoperative serum prostate-specific antigen concentration was 3.74 ng/mL (range, 0.8–11.6 ng/mL).
Theoretically, a pT0 situation may have various causes8, 9: 1) complete R0-resection by the previous TUR-BTs; 2) successful neoadjuvant radio- and/or chemotherapy; 3) residual tumor in the cystectomy specimen that is too small to be detected by the pathologist; or 4) cystectomy undertaken because of misdiagnosis of benign lesions in a TUR-BT specimen.
The rate of pT0 cystectomy specimens ranged from 6.2% in a larger cystectomy series1 to 20.1% in the present series. In a subgroup of 12 patients with a tumor progression of an initial superficial bladder tumor after failure of conservative treatment, Yiou10 found a pT0 rate of 41%. Most series that consist of mixed populations (with and without neoadjuvant therapy) show a rate of tumor-free cystectomy specimens of 15%.11–13 That rate is influenced by two factors: the early indication for cystectomy and the aggressiveness of the surgeon who performs the TURs. Our approach is to offer early cystectomy in all cases of primary T1 G3 or muscle-invasive bladder cancer and in all cases of superficial tumor that show a highly aggressive growth. But we try to perform complete TUR of every bladder tumor independently of invasion of the detrusor muscle as long as transurethral management seems appropriate. With this approach, the rate of tumor-free cystectomy specimens must be higher than in series that includes deferred cystectomy and/or diagnostic TUR-BT to demonstrate muscle invasion.
Some reported cases of tumor-free cystectomy specimens must be attributed to the effects of neoadjuvant therapy. The evaluation of such cases is difficult because the true tumor stage before neoadjuvant therapy is based only on the results of the TUR-BT specimen and imaging techniques. Understaging or overstaging has to be expected in up to 40% of cases.2, 14, 15 In a contemporary multicenter study of neoadjuvant chemotherapy, the rate of pT0 cystectomy was 38% in the chemotherapy arm but only 15% in the surgery-only arm.16 This study was unable to find a difference in tumor-specific survival between patients with and without neoadjuvant chemotherapy when no residual tumor was seen in the cystectomy specimen. In the complete study, the possible staging error of 40% is higher than the observed difference according to tumor stage. Similar data was obtained in a Swedish cystectomy series12 with neoadjuvant radiation therapy (20 Gy). Of 276 cases, the clinical and pathologic stages were in accordance in 44%; the clinical stage was less than the pathological stage in 23%, and greater in 33%. The rate of pT0 cystectomy specimens was 15.6%, so the relevance of neoadjuvant therapy could not be evaluated.
In prostate cancer a pT0 radical prostatectomy specimen heralds cure: no patient developed a postoperative PSA-recurrence in a series from a nationwide German prostate cancer registry.17 However, in bladder cancer patients a pT0 cystectomy specimen has been seen as having only a survival advantage.18, 19 In a survival nomogram that was calculated from more than 10,000 cystectomy datasets, a pT0 cystectomy specimen was associated with a distinct survival benefit compared with all other tumor stages.20 The tumor-specific survival rate in our series does not differ from that presented by Stein and Skinner21 on the basis of 66 cases: 5 years, 92% (Ulm: 91.7%), 10 years, 86% (Ulm: 88.4%).
Detailed analyses of series with pT0 cystectomy specimens are rare. In 1994, Thrasher et al.11, 22 published their series of 433 cystectomy patients that included 66 (15.2%) pT0 specimens; 22 of these 66 patients had had neoadjuvant radiation therapy. Comparing the Kaplan–Meier survival projections for tumor-specific survival according to the maximal tumor classification (Ta/is, T1, and T2a), they found no benefit in the pT0 group. They declared that their results refuted the axiom that a stage pT0 cystectomy specimen confers a survival advantage in patients who underwent radical cystectomy for TCC of the bladder. Our data confirm this finding in pTa/is/1 tumors, but show a statistically significant benefit in maximum pT2a-classified tumors. In the Thrasher et al. study, the group consisted of 23 pT0 and 71 pT+ cases. The rate of neoadjuvant radiotherapy was 43% (10/23) in the pT0 group and 27% (19/71) in the pT+ group. Therefore, the number of cases with accurate preoperative staging was small, and it was not possible to draw any conclusions about patients who did not have neoadjuvant therapy.
Thrasher et al.11, 22 analyzed and compared overall survival rates but not tumor-specific survival rates. This leads to the question, What survival rate is the best measure to look at: overall or tumor-specific? There is a tendency to offer early cystectomy to younger patients or to patients without comorbidity rather than to older and comorbid patients. In the latter group, the risk of dying apart from the presence of a tumor is distinctly higher. Therefore, some bias has to be considered in analyzing overall survival rate. In our study, we preferred the tumor-specific survival rate because it has a higher potential for excluding this bias.
The major advantage of our study is the homogeneity of its population. We had a large, single-center series of patients who were treated by a small number of surgeons using the same standards for indications and the same surgical techniques. All cases included bilateral pelvic lymph node resection. The cases were limited to patients with TCC as the only histologic tumor component. All patients who had neoadjuvant therapy—even intravesical application of bacillus Calmette–Guerin or mitomycin in the interval between the last TUR-BT and cystectomy—were excluded. The remaining 181 patients with a pT0 cystectomy specimen constituted the largest homogenous series published so far.
We included no case with a clinical tumor extension over T2b; this was in accordance with the well-known experience that bladder cancer with invasion to the perivesical fat cannot be cured with a transurethral approach alone. In patients with a clinical T classification of T2b or less, the rate of tumor-free cystectomy specimens was as high as 32.7%.
In all of the pT0 cases the clinical tumor stage was based on the TUR-BT specimens alone. Assessment of the depth of infiltration is difficult: artifacts caused by squeezing the material, coagulation, fixation, and staining present problems, as does the loss of anatomic landmarks in repeated TUR-BTs.14, 23
Muscle invasion can be diagnosed reliably, but sometimes it is impossible to differentiate T2a from T2b tumors. Definite diagnosis requires additional information on tumor-free second resections or deep muscle layers, on perivesical fat in the TUR-BT specimen, and on the residual detrusor muscle at the site of TUR in the cystectomy specimen. Our study integrated all that information to determine the depth of invasion as accurately as possible. We classified a case as clinical stage T2a only if there was a tumor-free deep muscle layer in the TUR-BT or a tumor-free muscle layer at the site of TUR-BT of at least half the depth of the total detrusor muscle in the cystectomy specimen. All other cases with muscle invasion and without infiltration of the perivesical fat were classified as clinical T2b. Although some staging error is inevitable, this differentiation yields the best assessment of the true tumor stage. It was confirmed by the different rates of lymph node metastases—5.7% in pT0 T2a cases and 31.3% in pT0 T2b cases—and the statistically significantly different tumor-specific survival rates in the two stages.
Our series showed no survival benefit in patients who had maximal pTa/is/1 tumors and tumor-free cystectomy specimens. That result does not differ from that obtained by Thrasher et al.11 Radical cystectomy provides an almost perfect tumor-specific survival rate in patients with superficial bladder cancer even if there is residual tumor in the cystectomy specimens. A very large number of patients is required to gain the statistical power necessary to show a statistically significant survival benefit in this group because the difference in survival rate is relatively small. A similar statistical problem is seen in patients with maximal pT2b tumors: tumor-free cystectomy specimens are rare events, accounting for less than 15% of the cases. Our 11 cases with pT0 status and tumor-free lymph nodes had a benefit in tumor-specific survival that, because of the small number of cases, cannot reach statistical significance. However, a significant tumor-specific survival benefit could be demonstrated in patients with maximal pT2a tumors. That result contradicts the observations of Thrasher et al.11 In contrast to their population, the number of patients in our study in the pT0 and the pT+ group are similar and large enough to provide statistical power, and the survival benefit of 21% at 5 and 10 years cannot be attributed to any kind of bias in this homogenous setting.
Those findings may reflect some beneficial results in the organ-sparing therapy of early muscle-invasive TCC.24–26 Different reasons for such a survival benefit have to be taken into account. Among all the patients who had maximal pT2a tumors, those whose tumors were completely resected with TUR-BT alone constitute a favorable subgroup: their tumors were more often small, monofocal, and without concomitant carcinoma in situ than those of patients whose cystectomy specimens had residual tumor. The larger the tumor burden, the less probable is complete TUR and the higher is the surgeon's tendency to undertake diagnostic and not complete TUR-BT. A possible explantation for this phenomenon is the effect of tumor size as an independent prognostic factor in patients with T2 bladder cancer. This was stated by Cheng et al.,27, 28 who found a better prognosis in tumors less than 3 cm in diameter. They found a cancer-specific 10-year survival of 88% in patients with organ-confined and small (less than 3 cm) cancers. That was confirmed by Wijkstrom et al.12 in a series of 276 patients. In a series of preoperatively irradiated patients, Pollack et al.29 found a tumor size of less than 5 cm to be an independent predictor of patient survival. Narayan et al.30 also found the largest tumor diameter to be an independent predictor of patient survival.
It is well known that some patients who undergo radical cystectomy for organ-confined bladder cancer will develop tumor recurrence. That has been attributed to the presence of occult micrometastases at the time of surgery in up to 50% of all patients.31, 32 The tumor-specific survival benefit obtained in the maximal pT2a tumors when pT0 tumors are compared with pT+ tumors may be a hint that micrometastases do not develop immediately with infiltration of the detrusor muscle, but increase with the growth of the tumor burden; the rate was 8% in patients with tumor-free cystectomy specimens, but it was 29% in patients with residual tumor in our series of maximal pT2a tumors. In contrast with those results, the rate of manifest lymph node metastases is not very different between the pT0 and the pT+ group (6.6% vs. 9.9%) in the total group of maximal pT2b tumors.
Another remarkable fact is that no patient with a pT0 pN0 situation had local tumor recurrence, whereas the rate was 5.7% in the pT+ pN0 group. There are several possible explanations. The pT+ pN0 group had a distinctly higher number of patients with pT2b tumors. Those tumors have a higher probability of undetected extension or micrometastases to the perivesical fat that may become the origin of local recurrence. Another reason may be intraoperative tumor spillage. If the bladder is tumor-free, there is no risk of spilling tumor cells, but there is a slight, but not insignificant, risk in all patients with residual tumor in the cystectomy specimens.
From the excellent long-term survival rates after radical cystectomy, it cannot be concluded that TUR-BT would have had similar results. It is well known that many bladder cancers are of polyclonal origin, and there would have been tumor recurrence in some cases. In T1 G3 tumors, the rate of recurrence after organ-sparing therapy is expected to be about 45–60% within 5 years.33 The data presented show that 95% of patients with early muscle-invasive TCC of the bladder can be cured as long as no lymph node is involved. The cure rate decreases significantly with progression of the disease, not only with progression from T2a to T2b but also with growth of the tumor burden and with development of micrometastases. That result must be seen as a strong argument in favor of early cystectomy. Only cystectomy in the stage of monofocal, small, early muscle-invasive bladder cancer can provide an optimal cure rate. Attempts at organ-sparing therapy in patients who are not eligible for cystectomy may be successful in the short term but show an increased risk of recurrence and progression in the long term.33 The substantial understaging in our study even in clinical Ta tumors (17.2% had muscle-invasive tumors in the cystectomy specimens) and the rate of lymph node metastases of 6.6% stress this point.
There are no specific recommendations for follow-up of patients with tumor-free cystectomy specimens.34, 35 The observation that patients with pT0 pN0 cystectomy specimens do not develop local recurrences has not been described before. Our findings might justify a modification of the follow-up regimen for the pT0-subgroup.
The finding of a second primary malignancy in a cystectomy specimen is not rare, in light of the data recently published by Revelo et al.,36 who found unsuspected prostate cancer in 41% of cystoprostatectomy specimens. The rate of clinically significant prostate cancers reached 48% in this study.
It can be concluded that, in contrast with some earlier publications, a tumor-free cystectomy specimen without neoadjuvant therapy constitutes an entity with a beneficial effect on tumor-specific survival that reaches the level of significance in maximal pT2a tumors. The survival benefit must be seen as a strong argument for early cystectomy, but the prognostic significance needs to be validated in a large prospective study.