Optimised photodynamic diagnosis for transurethral resection of the bladder (TURB) in German clinical practice: results of the noninterventional study OPTIC III

Suspected non-muscle invasive bladder cancer (NMIBC) is diagnosed using white light cystoscopy (WLC), followed by biopsy [1]; however, WLC can fail to detect 4–41 % of papillary Ta and T1 tumours, carcinoma in situ (CIS), dysplasia, multifocal growth, and microscopic lesions [2, 3]. Visualisation of bladder lesions including residual tumour tissue, small tumours, and CIS during transurethral resection of the bladder (TURB) can be improved with hexaminolevulinate blue light cystoscopy (HALC; Hexvix^®, Ipsen Pharma GmbH, Ettlingen, Germany) [3‐6]. Results from randomised controlled trials (RCTs) show that WLC plus HALC can detect an additional 7–30 % of cancer lesions versus WLC alone [6‐11].

The quality of TURB and outcomes achieved in clinical studies can vary considerably [6]. Therefore, observational studies reflecting routine clinical practice are useful to assess the benefit of HALC outside of RCTs, since the effect of photodynamic diagnosis may differ between these settings. Prior observational studies conducted in Italy [7], Spain [8] and France [12] showed that in daily clinical practice, HALC can enhance the diagnostic accuracy of WLC, resulting in a lower tumour recurrence rate. Detection rates again varied, potentially reflecting different clinical settings between series.

Because of the observed differences in different countries, and as changes in EU regulations promote post-authorisation studies in different countries in Europe [13], we conducted a study to assess additional detection of cancerous lesions with HALC plus WLC versus WLC alone in patients with NMIBC undergoing TURB in routine clinical practice in various urology departments in Germany in unprecedented detail.

Participant flow is summarised in Fig. 1. Overall, 395 patients received hexaminolevulinate and 379 had valid histology from the diagnostic procedure; 97 % of enrolled patients completed the procedure.

Baseline patient and disease characteristics for the ITT population are summarised in Table 1.

This study showed that in patients with NMIBC undergoing TURB in routine clinical practice, WLC plus HALC detected an additional 34 (6.8 %) cancerous lesions in the PP population versus WLC alone (p < 0.0001). While this proportion is markedly lower than in a previous observational study (23.2 %) [7], findings from both are consistent with results from RCTs (range 7–30 %) [6‐11]. Large ranges in reported results could be due to different documentation procedures and clinical settings across countries. Nevertheless, HALC provides a diagnostic benefit over WLC alone for patients with suspected NMIBC, supporting its use in daily clinical practice. Indeed, the EAU guidelines now recommend use of this imaging modality when available for high-risk patients (when cytology is positive or when high-risk exophytic tumour is expected) [17]. The validity of this approach is confirmed by the current study, in which 8.2 % of additional cancerous lesions were detected in patients at high risk of progression. Initially targeting HALC to high-risk patients seems an appropriate means of introducing the use of this technique within a treatment centre. Thus, while HALC may lead to long-term benefits across all risk groups compared with WLC, high-risk patients seem to benefit most [16].

Importantly, significantly more CIS lesions were detected with WLC plus HALC versus WLC alone. While the effect of HALC was weaker than in a previous observational study, the rate of CIS was also lower (7.2 % here vs. 19.4 % in an observational series) [7]. RCTs also reported a greater (31.9–70.6 %) detection of additional CIS lesions with HALC [3].

In the PP population, 10 % of patients had ≥1 additional positive lesion detected with HALC, and WLC would have missed 2.2 % of NMIBC patients. Therefore, with WLC plus HALC versus WLC alone, more appropriate risk classification and optimised treatment management may be possible.

Comparisons with findings from RCTs should be viewed with caution. In RCTs, carefully selected patient populations managed at specialist centres with potentially more experienced surgeons may produce results that are not easily transferrable to routine practice at larger or more diverse centres. Indeed, TURB is a teaching procedure often performed by less experienced surgeons. Likewise, local pathology laboratories may yield different results, and generally a central pathology laboratory is used in RCTs. Hence, results from studies of routine HALC use are important to better understand the potential benefits of this approach. As outlined above, there is good evidence that HALC may offer benefits in terms of detection and outcome in a range of patients; this benefit may be most significant in those with high-risk tumours (as the data from this study support), and optimising detection of positive lesions will rely on optimising the technique more than via patient selection. Furthermore, differences in local histology/pathology assessments may contribute to varied results between observational studies and clinical trials; conformity in staging and grading is 50–60 % [1]. Importantly, this study included sites experienced and inexperienced with HALC, reflecting routine practice in Germany.

Enhanced detection during TURB with the use of HALC may improve diagnosis of NMIBC in clinical practice, improve outcomes, and reduce recurrence rates [4]. The finding that HALC detected a significantly higher rate of true-positive lesions in patients with high risk of progression (according to EORTC scores) than WLC may indicate that HALC use could improve detection of these patients in routine clinical practice, potentially reducing morbidity and mortality [8, 11, 17, 18]. While false-positive rates were relatively high with WLC alone and when HALC was added, these were within the ranges reported in a previous observational study that highlighted the wide variation in the rate of false positives at different centres (2.6–28.6 % with WLC alone and 6.1–39.3 % with the addition of HALC) [8]. The present study highlights the importance of thoroughly assessing patients in true need of a TURB in advance to avoid biopsy in merely inflamed tissue, as a rather large fraction of patients showed inflammation only (24.1 %).

Although this study did not assess recurrence rate or survival, it is clear that improved visualisation techniques allow for more appropriate disease staging and risk classification and thus guide more appropriate subsequent treatment decisions, as well as allow for better resection of the tumour. It is reasonable, therefore, to expect that use of HALC would result in longer recurrence-free survival compared to WLC alone. The benefits of improved detection with HALC have already been demonstrated in several studies, with significantly reduced recurrence rates at 1 year. This could potentially reduce the requirement for patients to undergo frequent TURBs, which not only have a significant negative impact on patient quality of life, but also incur a considerable cost [3]. A recent study from Sweden modelling the cost consequences of HALC found that the technique had a minimal cost impact if introduced across all risk groups, and reduced TURBs, cystectomies, bed days, and operating room time. Notably, the use of HALC translated into cost savings from year 2–5 in this model [16].

Despite the large number of patients with primary tumours (68.6 %) versus previous studies (43.3 %) [3], this study represents a typical population of patients with suspected diagnosis of NMIBC within in-patient surgical urological institutions in Germany. Similar to other observational studies, there were no reports of spontaneous AEs related to hexaminolevulinate [7, 8].

No subgroup analyses by tumour type were conducted in the present study; however, a meta-analysis that did conduct such analyses found the benefit of HALC was in patients with Ta, T1, CIS, primary and recurrent cancer, and was independent of the level of risk [3]. Additional detection rates with HALC (compared with WLC alone) ranged from 9.7 to 40.2 % for Ta tumours, with 239 of 1621 (14.7 %) of additional Ta tumours only being detected with HALC. For T1 tumours, detection of additional tumours with HALC ranged from 3.6 to 54.5 % of the total T1 tumours detected. In total, 40 of 372 (10.8 %) of additional T1 tumours were detected only by HALC. As in the present study, the benefit of HALC was most pronounced in CIS; detection of additional CIS lesions ranged from 31.9 to 70.6 % of the total CIS lesions detected; and 215 of 527 (40.8 %) of additional CIS lesions were detected with HALC alone [3].

Despite meticulous online documentation with detailed bladder maps and minimised time delay between procedures and documentation, there remained a number of study limitations. Limitations were consistent with other observational studies and include: incomplete information versus RCTs, and lack of blinding between WLC and HALC, which could bias true-positive and false-positive observations. However, achieving blinding in routine clinical settings is impossible and also hard to accomplish in a diagnostic study. Of note, no tertiary or academic centres—both of which institutions treat a significant number of patients—were included in the present study, and not all regions of Germany were equally reflected; either or both of these factors may have introduced bias.

Study strengths included observation of a complete and typical spectrum of patients with a realistic high number of primary tumours and staging distribution, and prevention of patient selection bias through recruitment of consecutive patients with the required diagnosis.

In conclusion, this study demonstrates that HALC significantly improves the detection of NMIBC versus WLC alone in daily clinical practice in Germany.