The optimal management of T1G3 bladder cancer

Abbreviations

TUR(B)

transurethral resection (of the bladder)

CIS

carcinoma in situ

SWOG

Southwest Oncology Group.

INTRODUCTION

High-grade bladder cancer infiltrating the lamina propria (G3T1) is an aggressive disease. Recurrence and progression to muscle-infiltrating cancer rates are high. The best treatment for G3T1 TCC of the bladder is still debated. However, at first presentation a bladder-sparing approach is presently supported by most [1]. The incidence of G3T1 bladder cancer is 5%-23% of all non-muscle-invasive TCC of the bladder [2–5].

STAGING

Transurethral resection of the bladder (TURB) is the standard method for clinical staging; any further treatment is based upon grade and stage. However, at cystectomy some 44–46% of patients with clinical stage T1 TCC based on TURB are found to have muscle-infiltrating disease [6,7]. Some authors have proposed substaging of T1 bladder cancer according to invasion above and below the muscolaris mucosae (T1a and T1b), or invasion above, at the level of, and below the muscularis mucosae (T1a, T1b and T1c) [8]. Although a literature review failed to find significant differences in terms of recurrence and progression according to substaging [9], some investigators have suggested that T1a tumours might have a lower progression rate (Table 1) [10–14]. The pathological evaluation of the muscularis mucosae can be difficult. To date, substaging of T1 bladder cancer has no clear prognostic value and its clinical application continues to be debated.

Table 1.  Progression for G3T1 bladder cancer according to substaging

Ref	Progression rate according to infiltration depth, %
	follow-up, months	Above MM	At level of MM	Below MM
MM, muscularis mucosae.
[10]	56	25	25	89
[11]	60	25	–	59
[12]	84	36	–	58
[13]	71	22	–	29
[14]	85	8	34	34

The evaluation of the pathological specimen can differ significantly according to the pathologist. In an extensive study, 1400 TURB specimens were reviewed; only 46% of 88 G3T1 tumours were confirmed as G3T1, while 10% of cases were upstaged to muscle-infiltrating tumours [15]. Similar results were obtained by Witjes et al.[2] in a multicentre study on bladder tumours initially classified as G3T1. The review pathologist found a lower grade in 16/54 (29%), while 61 new G3T1 bladder cancers were diagnosed (25 were previously classified as a lower stage). An experienced pathologist in urothelial cancer is crucial to decrease the risk of staging errors.

TURB

An accurate TURB is necessary for the proper treatment of non-muscle-invasive bladder cancer. The performance and quality of the TURB can influence tumour staging and subsequent therapy. After TUR of G3T1 bladder cancer, residual disease can be found at repeat TUR in 49%[16] of cases when muscle in not present in the TURB specimen, and in 5–13%[17–21] when muscle is present in the specimen. The results of a second TURB for initial G3T1 bladder cancer are shown in Table 2.

Table 2.  The results of a second TUR for G3T1 bladder cancer

Ref	N	Tumour grade	Stage at second TUR, %
			T0	Tis	Ta	T1	T ≥ 2
* patients with infiltrating tumour at restaging excluded from study.
[17]	42	2–3	36	19	17	24	5
[18]	9	3	45	44			11
[19]	39	3	61	10		15	13
[20]	7	3	14
[21]	319	3	40	26	6	28	*

When G3T1 is multifocal, the chance of finding residual disease might be even higher [19]. When residual T1 disease is present at the second TURB and cystectomy is performed, 13% of patients have remaining muscle-infiltrating disease [22]. An appropriate TUR, including muscularis propria must be done, especially when high-grade cancer is suspected. At 2–6 weeks after TURB, a repeat TURB is strongly recommended for G3T1 cancer, especially if no muscle can be found in the specimen.

TREATMENT

TURB alone

TURB is the first procedure for bladder carcinoma; when G3T1 tumours are treated by TUR alone, recurrence and progression rates are high. The recurrence rate is 50–80% and progression to muscle-invasive disease occurs in 27–63% of patients (Table 3) [23–31]. Eventually a third of patients with G3T1 bladder cancer treated only with TUR will die from their disease [32]. TURB alone is not a viable treatment for G3T1 bladder cancer, as it is associated with an unacceptably high probability of recurrence and progression.

Table 3.  Recurrence and progression (as %) in patients with G3T1 tumour after TUR alone

Ref	N	Recurrence	Progression	Follow-up, months
[23]	33	79	48	36
[24]	430		31	60
[3]	31	80	33	60/106
[25]	18		50	36
[26]	48	75	27	48
[27]	26		38	108
[28]	34	50	24	40
[29]	32	85	46	79
[30]	30	90	47	62
[31]	52	55	23	23

Radiotherapy

Radiotherapy after TURB of G3T1 bladder cancer in old series is followed by high progression rates (32–54% at 36–60 months); the 5-year disease-free survival is 55%[33]. These results are unacceptable and comparable with the results of TURB alone. According to these data, radiotherapy for high-grade T1 bladder cancer has not been considered a suitable option [34]. However, more advanced radiotherapy techniques have been developed recently. These procedures might improve the outcome of radiotherapy for G3T1 bladder cancer.

In a recent study, 120 patients with G3T1 or T2 bladder cancer were treated with brachytherapy [35]. The 5-year disease-free survival for G3T1 tumours was 100%. In recent reports, the results of radiotherapy for G3T1 bladder cancer appear to be improved, compared with previous studies (Table 4) [36–42]. In any case, these series included few patients and definitive conclusions cannot be drawn.

Table 4.  Progression in G3T1 tumours after radiotherapy

Ref	N	Progression, %	Follow-up, months
Older series
[36]	28	39	36
[37]	11	54	60
[38]	19	47	48
[39]	53	32	60
Recent series
[40]	12	0	40
[41]	19	11	60
[42]	14	20	60

Recently, a multicentre randomized trial was reported in patients with G3T1 bladder cancer [43]; 210 patients were randomized between observation and radiotherapy, or between intravesical therapy and radiotherapy. Radiotherapy did not improve the progression-free interval (hazard ratio 1.07, 95% CI 0.65–1.74; P = 0.785), progression-free survival (1.35; 0.92–1.98; P = 0.133) or overall survival (1.32, 0.86–2.04; P = 0.193). Although promising results were reported in small series with newer radiotherapy methods, in the only randomized trial available radiotherapy did not improve the results. Radiotherapy is not a standard option and the only candidates are patients with G3T1 bladder cancer recurrent after BCG treatment and who absolutely refuse surgery or are at high surgical risk [44].

Combination radio-chemotherapy

Radio-chemotherapy is often considered as a possible treatment option for bladder-sparing protocols in infiltrating bladder cancer. Weiss et al.[45] treated 84 patients with G3T1 cancer with radiotherapy alone or radiotherapy plus platinum-based systemic chemotherapy. At 6 weeks after the end of treatment, 89% of the patients were considered complete responders at re-staging TUR. The respective 5- and 10-year recurrence rates were 35% and 46%, the progression rates were 13% and 29%, and disease-specific survival was 80% and 71%. As the experience with radio-chemotherapy for high-grade superficial bladder is limited to a very few centres, this approach cannot be universally recommended.

Intravesical therapy

BCG immunotherapy is a widely accepted intravesical therapy for high-risk Ta, T1, Tis bladder cancer, with decreased recurrence and progression rates [46,47]. Immunotherapy with BCG after TURB is superior to TURB alone for treating G3T1 bladder cancer. In a study of 80 patients [48] treated with TURB plus BCG or TURB alone, at 61–65 months, BCG treatment was superior to TURB alone in terms of cancer-specific survival (90% vs 70%, P = 0.03), recurrence rate (50% vs 90%, P < 0.01) and progression rate (22% vs 47%, P < 0.001). In a prospective randomized trial, 86 patients with high-risk non-muscle-invasive bladder cancer received TURB plus BCG or TURB alone [49]. The 10-year progression rate was 38% and 63% for patients treated with BCG vs the control group, respectively. The cancer-specific survival was 75% vs 55%. Several authors consider BCG as the first-line treatment for primary G3T1 bladder cancer. At 22–78 months of follow-up, recurrence rates were 20–70% and progression rates 0–33% (Table 5) [13,48,50–71].

Table 5.  Results of full- and low-dose BCG for treating G3T1 tumours

Ref	Year	N	Recurrence, %	Progression, %	Follow-up, months
* 37% of patients treated with low-dose BCG. na, not available.
Full dose
[50]	1990	24	25	25	22
[51]	1991	62	20	0	46
[52]	1992	16	44	19	59
[53]	1994	17	37	12	36
[54]	1995	26	50	27	54
[55]	1995	49	–	16	60
[56]	1998	78	28	8	56
[57]	1998	109	39	3	78
[58]	1999	25	40	4	63
[59]	2000	44	27	16	43
[13]	2000	49	65	24	71
[60]	2002	43	28	16	53
[61]	2002	41	24	2	40
[62]	2002	69	35	12	45
[48]*	2002	50	50	22	61
[63]	2002	37	43	16	27
[64]	2003	90	70	33	64
[65]	2003	82	34	15	73
[66]	2004	57	42	23	53
[67]	2005	46	39	22	61
Low dose
[68]	1995	21	29	na	60
[69]	1996	95	40	11	46
[70]	1998	35	24	12	45
[71]	1999	51	25	18	85

Different doses and schedules of BCG have been proposed. In patients with G3T1 bladder cancer treated with full-dose and low-dose BCG, apparently equivalent recurrence and progression rates were reported in historical series. In a multicentre Spanish study, Martinez-Pineiro et al.[72] randomized 90 patients with G3T1 bladder cancer to either standard or reduced-dose intravesical BCG. At 61 months the overall recurrence and progression rates did not differ statistically. Side-effects were significantly lower with the reduced dose, although toxicity did not affect the likelihood of study withdrawal.

According to these reports, reduced-dose BCG protocols might be feasible in G3T1 bladder cancer, especially in patients with significant toxicity with the standard regimen. Gruenwald et al.[73] randomized 70 patients with high-risk non-muscle-invasive carcinoma of the bladder to a 12-week course of BCG or a standard 6-week course. The disease-free rate was 70% and 55% for the 12-week course and the 6-week course, respectively, at a mean follow-up of 28 months. A 12-week first course of intravesical BCG might be better for G3T1 bladder cancers.

Maintenance therapy

Lamm et al.[74] reported the results of a multicentre study conducted by the Southwest Oncology Group (SWOG), with 550 patients randomized to maintenance or no maintenance after an induction course of intravesical BCG. The median recurrence-free survival was 36 months in the no-maintenance and 77 months in the maintenance arm (P < 0.001). Based on these and other reported data Lamm et al. concluded that BCG could improve long-term survival, compared to intravesical chemotherapy, only if maintenance was given.

In a previous series in Rome, we evaluated 83 patients treated with BCG for G3T1 bladder cancer [75]; 70 patients completed the maintenance schedule, while 13 had the first 6-week cycle only. The recurrence rate was seven of 13 and 21 (30%) for no-maintenance and maintenance, respectively (P = 0.048). The progression rate was two of 13 and nine (13%), respectively (P = 0.403). Cancer-related death occurred in one of 13 and three (4%) patients, respectively (P = 0.299).

In a recent multicentre Spanish study [76], 398 patients with TaG3 or T1G3 TCC of the bladder and/or carcinoma in situ (CIS) were randomized to a 6-week induction course of 81 mg intravesical BCG with or without maintenance (one dose of BCG every 3 months for 3 years) with a median follow-up of 33.5 months. The recurrence-free survival was 58.9 months in the no-maintenance and 62.1 months in the maintenance arm (P = 0.112). The respective recurrence rate was 33.1% vs 28.4% and the progression rate was 15.3% vs 12.4%. The difference was not significant but a single dose of BCG as maintenance does not provide an adequate cytokine response, and together with the short follow-up, this maintenance schedule would not be expected to improve the results over induction therapy alone. Based on level 1 evidence from the SWOG trial showing a clear benefit for maintenance BCG, this should be considered the standard of care following an initial complete response to TURB and induction BCG.

Follow-up

Herr [77] reported on 48 patients with G3T1 bladder cancer followed for 15 years. These patients were part of a randomized study, with TURB alone or TURB plus intravesical BCG, and 81% of patients received BCG therapy. Progression and cancer-related death rates were 25/48 (52%) and 15/48 (31%), respectively. Overall survival was 69% and 24 patients (50%) had an intact bladder. The rate of tumour progression was 35% after 1–5 years, 16% at 5–10 years, and 12% at 10–15 years. Cancer-specific deaths occurred in 25% of the patients in the first 5 years and in 10% of patients after 5–15 years. After BCG treatment for G3T1 bladder cancer, long-term recurrence and progression can occur, requiring strict surveillance for at least 15 years.

Intravesical chemotherapy

Intravesical immunotherapy with BCG for T1 bladder cancer is superior to intravesical chemotherapy in decreasing recurrence rates. In 334 patients (303 T1) treated with intravesical BCG, epirubicin or adriamycin, the risk of recurrence was significantly higher with intravesical chemotherapy than with BCG. Combined therapy with BCG and epirubicin was associated with a lower recurrence rate [78]. In a recent report, 191 patients with high-risk non-muscle-invasive bladder cancer were treated with BCG, mitomycin or adriamycin. Multivariate analysis showed that BCG treatment was associated with a reduced risk of disease progression at 73 months of follow-up [79].

In two meta-analyses, BCG with a maintenance schedule was associated with reduced recurrence and progression rates compared to mitomycin and other intravesical chemotherapeutic agents [46,80]. In another meta-analysis of 1901 patients with high-risk non-muscle-invasive bladder cancer BCG was more effective than mitomycin in reducing recurrence, but not progression rates [81]. Intravesical chemotherapy did not improve progression and cancer-specific death rates [82].

Several authors reported that one dose of intravesical chemotherapy, 30–60 min after completing TURB, can decrease recurrence rates [83]. In a study on 168 patients, with mainly G2–3T1 bladder cancer, one dose of postoperative intravesical epirubicin did not reduce the recurrence rate for high-grade tumours [84]. There is no scientific proof that one dose of intravesical chemotherapy immediately after TURB can improve the prognosis of G3T1 cancer. Randomized trials comparing intravesical chemotherapy and BCG for G3T1 bladder cancer are not available. While associated with a reduced probability of recurrence for patients with low- and intermediate-risk disease, intravesical chemotherapy should not be considered as first-line treatment for G3T1 tumours.

Alternative and second-line treatments

In one study [85], 43 patients with G3T1 bladder cancer recurring after intravesical BCG were treated with a second TURB and a second course of BCG. At 46 months of follow-up, nine patients (21%) had a cystectomy (five for local progression, four for persistent high-grade T1/Cis). In another study, 34 patients with high-risk non-muscle-invasive bladder cancer recurring after 28 months (range 7–50) were treated with TURB and an additional course of BCG. At a follow-up of 36 months, recurrence and progression rates (requiring cystectomy) were 41% and 6%, respectively [86].

A second course of BCG might be an option for highly motivated selected patients with recurrent G3T1, although strict surveillance and adequate informed consent are mandatory. Interferon has been added to BCG in salvage protocols. In a study by O’Donnell et al.[87] 39 patients with G3T1 bladder cancer refractory to BCG received low-dose BCG plus interferon-α2b. At 24 months, recurrence and progression rates were 47% and 26%, respectively.

In a small non-randomized series, 19 patients with BCG refractory G3T1 bladder cancer were studied [88]; 10 received a second course of BCG and nine received intravesical gemcitabine. At 17 months, three of nine patients treated with gemcitabine were disease-free, compared to only one of 10 treated with BCG. The disease-specific survival at 27 months was nine of nine for gemcitabine and eight of 10 for BCG. ‘Salvage’ intravesical second-line treatments are still investigational and they should be used only in highly selected patients.

Cystectomy

When G3T1 bladder cancer does not respond to BCG, the progression rates can be as high as 81%[89]. Radical cystectomy is the standard approach for patients with muscle-infiltrating bladder cancer and some authors advocate radical cystectomy as the first option for high-grade T1 bladder cancer [90]. Orthotopic bladder substitutions have improved the acceptance by patients of radical cystectomy.

Herr and Sogani [91] reported on 90 patients with high-risk non-muscle-invasive bladder cancer treated with cystectomy; the survival was 49% at 96 months. In another study, 32 patients with T1 (72% G3) were treated with cystectomy and the 5-year progression rate was five of 17 [92]. Cystectomy for high-grade T1 bladder cancer achieves good results, with 5- and 10-year survival rates of 67–95% and 48–62%, respectively (Table 6) [92–98].

Table 6.  Results for radical cystectomy in G3T1 tumours

Ref	Year	N	Survival, %
			5-years	10 years
* mainly G3.
[92]	1988	32	67	57
[95]	1998	45	92
[96]*	2001	208	74	51
[97]	2003	77	76	58
[98]	2008	98	75	48

BCG OR CYSTECTOMY?

From a comparison of historical series, no clear advantage emerges in favour of aggressive treatment of primary G3T1 bladder cancer. Several studies compared intravesical BCG and cystectomy in the same institution. A retrospective study compared the long-term results of patients with G3T1 treated with intravesical BCG (if refusing surgery) or immediate cystectomy [99]. Of 121 patients, 92 received immunotherapy and 29 had an immediate cystectomy. At a median follow-up of 6.9 years, in the BCG group 31% of the patients remained disease-free and 36% developed local recurrence. In 29% of patients, deferred cystectomy was required. The cancer-specific survival in patients with an organ-preserving approach or immediate cystectomy was 80% and 69%, respectively. There was no significant difference in overall survival between patients undergoing early vs deferred cystectomy.

In another German paper [98] studying 223 patients with G3T1 bladder cancer, 125 had repeated TURB followed by intravesical BCG, while 98 had a radical cystectomy. The 5- and 10-year survival rates were 82% and 65% in the TURB group vs 75% and 48% in the cystectomy group, respectively; the difference was not statistically significant. However, with Cox regression analysis, patients with all three major risk factors (multifocality, tumour diameter > or <3 cm and associated CIS) were likely to benefit from immediate cystectomy.

No randomized trials comparing BCG vs cystectomy in G3T1 bladder cancer are available. To date, evaluation of retrospective studies supports the concept that immediate cystectomy is not superior to intravesical BCG and eventual deferred cystectomy for the treatment of primary tumours. A strict follow-up is crucial to identify patients with early (3–6 months) recurrence of G3 tumours or with progression after BCG treatment, and who require immediate cystectomy.

PROGNOSTIC FACTORS

Up to 27% of G3T1 tumours will progress despite intravesical immunotherapy [53]. It would be helpful to recognize patients in whom BCG therapy will fail. Clinical features of the tumour and biological markers have been investigated for their predictive value.

Clinical characteristics

Solsona et al.[79] treated 80 patients who had G3T1 bladder cancer with intravesical BCG or chemotherapy. An incomplete response at the 3-month cystoscopy (no visible tumour, negative random biopsies) had a significant predictive value for progression in 29 non-responding patients. In another Spanish paper, 159 patients with G3T1 bladder cancer were treated with BCG [98]. An earlier recurrence was related to a higher progression risk during the 70-month follow-up. The progression rates at 70 months were 42%, 29% and 4% according to recurrence time (at 3, 6 or >6 months, respectively). Again, 92 patients with high-risk non-muscle-invasive bladder cancer were treated with BCG [99]. The progression rate was significantly higher for patients with non-muscle-invasive recurrence within 6 months.

In a previous report from our group, 82 patients with G3T1 bladder cancer were treated with BCG. At 73 months, the number of lesions, previous lower grade bladder cancer, and associated CIS were unrelated to recurrence or progression [1]. In another institution, a multivariate analysis was reported for 51 patients with G3T1 bladder tumour who had received intravesical BCG. Tumour size (P = 0.027) and associated CIS (P = 0.024) were significantly related to progression [70], On the contrary, CIS did not relate to BCG failure in 44 patients with G3T1 cancer in another study [58]. In a report of 102 patients with high-risk non-muscle-invasive bladder cancer, tumour size and multifocality were unrelated to the response to BCG, while patients with recurrent disease and CIS were related to treatment results [100]. Lebret et al.[101] found that tumour size, but not multifocality, predicted recurrence and progression in 35 patients with G3T1 bladder cancer treated with BCG. In another study, multiple bladder tumours were a negative prognostic factor [85]. According to Cheng et al.[102] tumours of >1 cm have a higher risk of progression. In a recent paper describing 223 patients with G3T1 bladder cancer treated with intravesical BCG or cystectomy, the authors identified three major risk factors, i.e. multifocality, tumour diameter > or <3 cm and associated CIS. Patients with all three factors were at high risk of BCG failure [96].

BIOLOGICAL MARKERS

Molecular and genetic aspects, such as altered expression of p53, have been implicated in the aggressive behaviour of bladder tumours [103,104]. Retinoblastoma (Rb) gene inactivation has also been investigated as a possible predictive factor for progression. p53 is a protein involved in cell-cycle regulation and p53 mutations can be found in many different tumours, including TCC. The role of p53 as predictive factor of BCG response in bladder cancer is controversial. According to Saint et al.[105], p53 overexpression (representing the altered state) is an independent predictive factor of recurrence for high-risk non-muscle-invasive bladder cancer. In 59 patients with T1 bladder cancer treated with BCG [106], p53 was highly correlated with progression risk (P = 0.003). In another study of 119 patients with G3T1 bladder cancer, p53 did not independently predict a patient’s prognosis [107]. In a literature review, p53 suppressor gene expression did not relate to intravesical BCG results in high-grade superficial bladder cancer [108].

In patients with T1 bladder cancer, abnormal expression of pRb and/or p53 is significantly correlated with tumour progression [109]. Tumours with either absent or over-expression of pRb (both corresponding to the altered state) have a shorter recurrence-free survival rate than those with heterogeneous (normal) levels of pRb [110].

Reports of G3T1 bladder cancer have conflicting data on the prognostic value of clinical and biological characteristics of the tumour. However, especially if more than one factor is present, multifocality, CIS and tumour size should be considered to choose the proper treatment. p53 expression might have some predictive value for the risk of progression in bladder cancer, but it has poor prognostic value for high-grade non-muscle-invasive bladder cancer treated with BCG.

CONCLUSIONS

High-grade G3T1 bladder cancer is a potentially lethal disease. TURB alone is not an adequate treatment option. Intravesical therapy with full- or reduced-dose BCG plus maintenance is the best bladder-sparing therapy for primary G3T1 bladder cancer. As recurrence and progression can occur up to 15–20 years after BCG treatment for G3T1 bladder cancer, these patients need strict lifelong surveillance. In BCG-refractory tumours, second-line treatments (second BCG course, intravesical chemotherapy, low-dose BCG plus interferon-α2b, gemcitabine, radiotherapy) might be considered in highly selected patients, under strict control, or in patients refusing or unfit for open surgery. Cystectomy can be proposed as first-line treatment, but patients must be aware that a bladder-sparing protocol with delayed cystectomy (if needed) would not compromise long-term survival. No clear prognostic factors exist to identify patients at high risk for disease progression, apart from early recurrence after BCG, that should be treated with immediate cystectomy.

CONFLICT OF INTEREST

None declared.