Fluorescence-guided transurethral resection of bladder tumours reduces bladder tumour recurrence due to less residual tumour tissue in T  a/T1 patients: a randomized two-centre study


Gregers G. Hermann, Urology Department, Frederiksberg Hospital, Copenhagen University, Ndr Fasanvej 57, DK-2000 Copenhagen, Denmark. e-mail: gregershermann@hotmail.com


Study Type – Therapy (RCT)

Level of Evidence 1b

What’s known on the subject? and What does the study add?

Photodynamic diagnosis (PDD) improves the diagnostic sensitivity of non-invasive bladder cancer as compared to TURB without PDD.

TURB in white light leaves residual tumour in the bladder in up to 49% of the patients. PDD-guided TURB improves the detection of Ta/T1 tumours of the bladder resulting in more complete TURB procedures and thus a reduced tumour recurrence rate.


• To compare the bladder tumour recurrence rate in stage Ta and T1 tumours after conventional transurethral resection of the bladder in white light (WL TURB) and after fluorescence-guided TURB (HAL TURB) using hexaminolaevulinate (HAL: Hexvix®, Photocure, Norway) for photodynamic diagnosis during 12 months of follow-up.

• As secondary objectives, to relate the tumour recurrence rate to fluorescence-detected residual tumour after WL TURB and to assess the false positive rate.


• This was a prospective, comparative, randomized, open-label study carried out in hospital outpatient urology clinics and the operating theatre. A total of 233 patients presenting with suspected superficial bladder tumour were recruited. Both patients with new tumours and patients with recurrent tumours were included.

• The study duration was 2.5 years with follow-up cystoscopic investigations at 4, 8 and 12 months.

• Patients were randomized to cystoscopy and WL TURB (118 patients) or WL TURB followed immediately by HAL TURB (115 patients). Cystoscopy/TURB and bladder biopsies were performed under general anaesthesia. No patients had intravesical chemotherapy immediately after TURB.

• Recurrences were verified histologically.


• The two groups were similar regarding age and previous bladder cancer history.

• In all, 90 patients from the HAL TURB group had bladder tumour. Fluorescence-guided cystoscopy after complete WL TURB identified residual tumour tissue in 44 of 90 patients (49%). In 37 of 83 (45%) residual Ta tumour was found; in three of seven residual T1 was found and in four cases carcinoma in situ.

• True (and false) positive detection rate of photodynamic diagnosis was 64% (25%) and of white light 83% (16%).

• In all, 145 patients were eligible for analysis of tumour recurrence. Twelve patients had their last follow-up after 4 months. The recurrence rate in patients followed for 12 months was 47.3% (35/74) after WL TURB and 30.5% (18/59) after HAL TURB (P= 0.05).

• Kaplan–Meier analyses comprising data from all 145 patients showed that the recurrence-free period was significantly longer in the HAL TURB group than in the WL TURB group (P= 0.02).


• WL TURB often leaves residual tumour in the bladder. HAL TURB improves the detection of Ta/T1 tumours of the bladder resulting in more complete TURB procedures and thus a reduced recurrence rate.


transurethral resection of the bladder


photodynamic diagnosis




TURB in white light


HAL-fluorescence-guided TURB


carcinoma in situ


5-aminolaevulinic acid.


Bladder cancer has a high recurrence rate of up to 45% at first follow-up for multiple tumours [1]. A number of studies have reported that tumour is frequently found at ‘second-look’ transurethral resections of the bladder (TURBs) carried out only a few weeks after the initial resection [2]. Tumour has been found in 30–44% of patients resected up to 2–8 weeks after original surgery [3–5]. These early recurrences may be caused by lack of identification and following removal of small tumours at first investigation or regrowth of residual tumour after incomplete resection [6].

Photodynamic diagnosis (PDD) improves the diagnostic sensitivity of non-invasive bladder cancer compared with TURB without PDD [7–10]. If used during TURB, it may result in a more complete resection of bladder tumours as well as identify more tumours, thus reducing the tumour recurrence rate [11].

The objective of the present study was to assess the impact that improved detection of non-muscle-invasive bladder cancer with hexaminolaevulinate (HAL: Hexvix®, Photocure, Norway) fluorescence during TURB may have on the tumour recurrence rate within 12 months after TURB.


Adult patients (over 18 years) with suspected Ta/T1 bladder tumour were consecutively enrolled from two Danish outpatient departments. Suspicion of bladder tumour was based on flexible cystoscopy performed in the outpatient department by a qualified urologist. Patients were excluded if they had porphyria, gross haematuria or a known allergy to HAL.

The trial was conducted in accordance with the International Conference on Harmonization Guidelines for Good Clinical Practice and the Declaration of Helsinki (10 September 2004 version). All patients provided written informed consent.

Eligible patients were randomized just before surgery to one of two treatment groups: TURB in white light (WL TURB) or TURB in white light followed by HAL-fluorescence-guided TURB (HAL TURB, Fig. 1).

Figure 1.

Study design. WL: white light; PDD: photodynamic diagnosis.

HAL was supplied as 85 mg vials of powder for reconstitution in 50 mL of sterile PBS, resulting in an 8 mm solution. The solution was instilled into the bladder by catheter and was to be retained for at least 1 h (mean instillation time 100.5 min, range 12–303 min). A xenon light source (D-light C system®, Karl Storz Gmbh, Tüttlingen, Germany) fitted with a light filter and a three-chip pendulum camera were used to provide blue light for induction of red fluorescence and to provide white light illumination.

Patients in the white light group (WL TURB) underwent cystoscopy and mapping of bladder lesions on a bladder chart, followed by WL TURB. In the fluorescence group (HAL TURB), bladder inspection was first exclusively carried out under white light, suspicious lesions and tumours were mapped and biopsied and all identified lesions were completely resected using WL TURB as if HAL TURB would not be performed. The bladder was then inspected under blue light to identify residual tumour tissue at the resection sites or additional tumours that had been missed in white light inspection. Residual and additional tumours are called residual tumours in this report. Fluorescing lesions were recorded, biopsied and resected. Five qualified urologists performed the surgery.

No patients had intravesical chemotherapy immediately after TURB. Three patients in both treatment arms had previously had a course of mitomycin 30 mg intravesically in the outpatient department. Ten patients in the PDD treatment arm and 11 patients in the white light arm had had BCG therapy.

The histopathology evaluation and staging was performed by local pathologists in accordance with the WHO classification 2004 and TNM 2003 [12,13]. Only patients with macroscopic and histologically confirmed new or recurrent Ta/T1 tumours (with or without dysplasia) were eligible to continue in the study and to be monitored for recurrence. Patients in whom no tumour could be identified or who had a cystectomy or invasive tumour stage T2–T4 were not followed for recurrence. Patients were treated according to Danish national guidelines.

All patients eligible to be followed for recurrence were followed up at 4 and 12 months with standard flexible white light cystoscopy (Fig. 1). Patients with high grade Ta and small single T1 tumours with superficial lamina propria invasion (stage T1a) could receive cystoscopy at 8 months (according to Danish national guidelines). Patients who had a recurrence at any follow-up visit were removed from further assessment, and only patients who were tumour-free were eligible to continue to the next follow-up. A recurrence was defined as macroscopic tumour, and recurrences were histologically verified by local pathologists in accordance with Danish national guidelines. Adverse events were recorded at each follow-up visit.

HAL TURB was estimated to provide a 25% reduction in the recurrence rate. One-year recurrence rate in standard WL TURB was set to 40%, type 1 error risk at 5% and type 2 error risk at 10%. Thus, each study arm should comprise 72 evaluable patients. An additional 90 patients were added to compensate for the excluded patients.

The analysis of the primary endpoint (comparison of recurrence rates) was evaluated using the Cochran–Mantel–Haenszel chi-squared test with a two-sided significance level of 5%. The recurrence-free period was analysed by the Kaplan–Meier method and regression analysis of grouped survival data. Secondary efficacy data (detection) and safety were summarized using descriptive statistics at a patient and lesion level. Homogeneity of data between the two centres was assessed using the Breslow–Day test.

The false positive diagnosis rate (i.e. fluorescing areas that were subsequently found to have negative histology) on a patient level was calculated as the number of patients with one or more false positive results divided by the total number of patients examined with fluorescence cystoscopy. The false detection rate on a lesion level for HAL was calculated as the total number of false positive lesions divided by the total number of lesions that were suspected with blue light.

Analyses for detection were performed using the ‘intention to treat’ population whereas recurrence analyses were performed using the per protocol population comprising patients with follow-up. Study monitoring, data management and statistical analyses were performed by Trial Form Support AB (formerly Clinical Data Care), Lund, Sweden. Approval was granted by the Ethical Committee of Copenhagen and Odense Counties (Dnr 02-049/04) and the Danish Medicine Agency (Dnr 2612-2607).



A total of 233 patients were enrolled from two centres in Denmark. Of these, 115 (49%) were randomized to the fluorescence cystoscopy group and 118 (51%) to the white light group (Fig. 2). Thirteen patients were excluded from the HAL group and one from the white light group (Fig. 2). Fourteen patients were lost to follow-up.

Figure 2.

Patient data flow diagram. [Change analyzed to analysed, twice]

In all, 34 patients from the white light group and 26 patients from the HAL group were excluded from analysis of recurrence as they had no macroscopic Ta/T1 bladder tumour suitable for follow-up. Of these, 32 patients had no bladder tumour pathology; 12 patients had invasive tumour stage T2–T4 and 10 patients suffered from stage T1 tumour requiring immediate cystectomy; five patients had only carcinoma in situ (CIS) at baseline and in one patient a bladder perforation during TURB prevented the fluorescence procedure (Fig. 2).

Therefore the population for analysis of recurrence included 145 patients, 68 in the HAL group and 77 in the white light group. As Table 1 shows, the population consisted mainly of patients with tumours at low and medium risk for recurrence [14]. There was no significant difference between patients’ characteristics, adverse events, detection rate or recurrence rates between the two centres.

Table 1.  Characteristics of patients and bladder tumours in the per protocol populations for analysis of recurrence
 WL TURB (N= 77) n (%)HAL TURB (N= 68) n (%) P (test)chi-square test
  1. N/A, not applicable.

Tumour number   
Single48 (62)44 (65)0.84
 2–729 (38)24 (35)
Tumour size   
 <3 cm57 (74)54 (79)0.45
 ≥3 cm20 (26)14 (21)
Stage Ta low grade69 (90)57 (84) 
Stage Ta high grade5 (6)8 (12)0.59
Stage T1 low grade00
Stage T1 high grade3 (4)2 (2)
N/A01 (1)
Low grade69 (90)57 (84)0.68
High grade8 (10)10 (15)
N/A01 (1)
Papillary/solid tumour74/3 (96/4)65/3 (96/4)0.85
Pedunculated/broadbased tumour36/41 (47/53)35/33 (52/48)0.57

Analysis of fluorescence-guided tumour detection was performed using patients in the HAL group who had any form of bladder pathology. The tumour detection endpoint was a within-patient comparison. A fluorescence-guided procedure was performed for 102 patients. Twelve of these did not contribute with a tumour diagnosis as 11 patients had no bladder tumour or dysplasia at baseline and one patient had a bladder perforation at baseline preventing the fluorescence procedure. Thus, 90 patients were included in the analysis of tumour detection.

Mean patient age (range) for the population for tumour detection analyses was 69 (41–92) years in the WL TURB arm and 71 (35–96) years in the HAL TURB arm. The female to male ratio was 1 : 3. No patient had perioperative intravesical chemotherapy immediately after TURB.


HAL TURB identified tumour or dysplasia in 44 of 90 patients (49%; 95% CI 38–60%) after WL TURB (Fig. 3). Of these, 37 of 83 patients with at least one Ta tumour (45%; 95% CI 30–52%) had residual Ta tumour and three of eight patients with at least one T1 tumour (43%; 95% CI 9–76%) had residual T1 tumour. Some patients had tumours of different stages and more than one residual tumour. HAL-guided cystoscopy identified CIS in four patients where white light cystoscopy was normal. HAL cystoscopy identified flat low grade dysplasia in 12 patients, which was diagnosed in only four patients in white light.

Figure 3.

Number of patients having 1, 2, 3 or more residual lesions identified by HAL TURB after TURB in white light. A patient may have more than one kind of lesion. [Remove border]


Nine patients in the HAL TURB group and three patients in the WL TURB group had their last follow-up at 4 months. Their follow-up data are included in the Kaplan–Meier analyses (Fig. 4) but do not contribute to analyses presented in Table 2.

Figure 4.

Kaplan–Meier estimates of 12-month recurrence-free survival stratified by treatment (WL TURB, n = 77, or HAL TURB (PDD), n = 68); P = 0.02.

Table 2.  Tumour recurrence rate after WL TURB and HAL TURB within 4 and 12 months after TURB
Patients at risk (n)No. of recurrencesRecurrence rate (95% CI)Patients at risk (n)No. of recurrencesRecurrence rate (95% CI)
Recurrences at 4 months742331.1%591016.9%
Recurrences at 12 months511223.5%49 816.3%
Recurrences in total743547.3%591830.5%

The recurrence-free period was significantly longer in the HAL TURB group than in the WL TURB group (P= 0.02; Fig. 4). The analyses of recurrence-free survival within 12 months showed a relative reduction in recurrence rate at 35.5% in the HAL TURB group (Fig. 4). In all, 18 of 59 patients (30.5%) from the HAL TURB group and 35 of 74 WL TURB patients (47.3%) recurred within 12 months (P= 0.05; Table 2). The recurrence rate during the first 4 months was 16.9% after HAL TURB and 31.1% after WL TURB. The recurrence rates during the interval from 4 to 12 months were 16.3% after HAL TURB and 23.5% after WL TURB. Thus, the recurrence rate was lower after HAL TURB throughout the 12 months of follow-up.


A false positive diagnosis was made in 25 of 102 patients (25%) examined with HAL fluorescence cystoscopy. A 19 in 117 (16%) false positive diagnosis rate was found in the white light group but the two rates cannot be compared as suspicious areas detected under white light were removed before HAL fluorescence was used. Biopsies taken from false positive fluorescent areas revealed chronic inflammation in two patients, chronic inflammation and cystitis cystica in one patient and normal urothelium in 22 patients. On a lesion level the false positive rates were 55% and 32% (WL/HAL).

A true positive diagnosis was made in 65 of 102 patients (64%) examined with HAL fluorescence cystoscopy after WL TURB, and in 97 of 117 patients (83%) true positive detection was observed using only white light cystoscopy. Adverse events were generally mild.


This is the first study to examine the frequency of HAL-fluorescence-detected residual tumours immediately after standard WL TURB and the efficacy of immediate removal of residual tumour tissue or overlooked tumours on tumour recurrence.

Our findings are in line with previously reported findings showing improved recurrence-free survival with 5-aminolaevulinic acid (5-ALA) fluorescence-guided cystoscopy and TURB [15–17]. Denzinger et al. [17] reported a 35% reduction in 1-year tumour recurrence rate from 16.2% to 10.4%. Babjuk et al. [15] observed a 44% reduction in 1-year recurrence rate from 61% to 34% with 5-ALA guided TURB. Overall, the anticipated reduction in recurrence rate in non-invasive bladder cancer adding HAL fluorescence during TURB may be around 25–35% but the recurrence-free survival rate can differ by up to a factor of 6 between centres [18].

Another recently completed clinical study using HAL reported a 22% reduction in the 9-month recurrence rate from 46% to 36% with fluorescence-guided cystoscopy [19]. In that study, patients were first inspected with white light and then with HAL fluorescence cystoscopy to identify additional tumours, before TURB was carried out in white light and fluorescence guided. A different treatment protocol was used in the present study. Patients were first inspected and resected with white light before undergoing inspection and further resection with blue light HAL fluorescence cystoscopy. It was therefore possible not only to identify previously overlooked lesions, but also to identify areas of residual tumour where the resection under white light had not been complete. Another difference in the treatment protocols was a 12-month follow-up in our study vs 9 months in the cited study.

The generally high rates reported for early tumour recurrence may in fact be partly attributable to inadequate resection. Early re-resection of Ta/T1 tumours 2–6 weeks after white light TURB has identified the presence of residual tumour in around one-third of patients [20,21]. In both of these studies, more complete resection (re-resection) resulted in reduced recurrence rates of 14% and 18%. Although the initial resection in these studies and the present study was carried out by experienced urologists, it seems clear that some tumour resections are incomplete leaving tumour in the resection area and some tumours can be overlooked using white light inspection. Fluorescence cystoscopy improves the detection and resection of tumours and may reduce the need for early re-resection of some tumours.

Our findings showed that the recurrence-free survival at 1 year was higher in the HAL fluorescence cystoscopy group, suggesting that the effects may endure for more than 1 year. Other studies report reduced recurrence rates up to 8 years after 5-ALA fluorescence-guided TURB [4,17]. However, patients at low risk of recurrence may benefit more than patients with a higher risk of recurrence [17].

Denzinger et al. [17,22] followed 46 patients with primary T1 tumour for 7.3 years and found that the progression to muscle-invasive disease was not influenced by fluorescence cystoscopy, even though the detection rate was improved compared with white light cystoscopy. The follow-up period in the present study was too short to evaluate the effect that reduced recurrence may have on progression.

In the present study, additional Ta and T1 tumours were detected with fluorescence cystoscopy. Detection of dysplasia was also increased with HAL fluorescence cystoscopy. Although this was not the primary focus of the study, it is a significant finding, as other studies have reported that 15–19% of patients with low grade dysplasia subsequently progress to CIS or invasive disease [23,24]. Other studies have also reported improved detection of low grade dysplasia with fluorescence cystoscopy (15% vs 9%[6]; 9% vs 5%[25]) which is in line with the findings in the present study. Coagulation of dysplasia identified with fluorescence cystoscopy may also have contributed to the reduction in recurrence rate.

The false positive diagnosis rate on a patient level was 25% for fluorescence cystoscopy and 16% for white light cystoscopy but on a lesion level the figures were 55% and 32%. The latter are higher than reported elsewhere [6,26] which may be a result of our different study design, performing WL TURB just before the fluorescence procedure as false fluorescence often appears at the edge of a tumour resection area.

A more aggressive approach to identify all possible pathological sites in the bladder for testing the accuracy of PDD compared with white light cystoscopy could also explain the differences. This is supported by the false positive diagnosis rate on a patient level which is similar to those in the cited studies. The high false positive rate may also be due to a start-up phase of learning to differentiate between areas of false fluorescence and tumour or dysplasia.

As PDD improves the sensitivity of detecting non-muscle-invasive bladder neoplasia the false negative tumour detection rate is simultaneously reduced. To our knowledge only one study has evaluated the false negative detection rate of PDD. Re-resection of the bladder was performed 6 weeks after either TURB in white light in 51 patients or fluorescence-guided TURB in 51 patients with Ta/T1 bladder tumour [11]. Residual tumour was identified in 39% and 16%, respectively. These data show that also fluorescence-guided TURB procedures may be incomplete. Cystoscope optics, cameras and light sources have improved substantially during the last 10 years. It is likely that the frequency of residual tumour in the cited study would have been lower using equipment of 2010 standard.

The present study, in low and medium risk bladder cancer patients, indicates that the reduction in tumour recurrence rate was due solely to improved detection and resection with fluorescence-guided cystoscopy. No patients were given immediate bladder instillation of chemotherapy after TURB. Immediate instillation of mitomycin to avoid tumour cell re-implantation may further reduce the tumour recurrence rate, but no studies have yet evaluated whether immediate intravesical chemotherapy combined with fluorescence-guided TURB is superior to either of the modalities as single supplements to WL TURB.

Relative to WL TURB, fluorescence-guided TURB using HAL improves the detection and resection of Ta/T1 tumours of the bladder. Improved detection and resection with fluorescence cystoscopy reduces residual tumour and tumour appearance during 12 months of follow-up.


The authors wish to thank our research nurses Helle Pedersen, Tine Leth and Charlotte Englund for excellent administration assistance. Photocure ASA is sponsor of the study. The Juchum and the Boemske Foundations have provided the study with financial support.


Gregers G. Hermann is a member of advisory board for Photocure ASA.