Intérêt du séquençage de nouvelle génération dans le cancer de la vessie à haut risque traité par BCG

25 mai 2018

Auteurs : C. Jungels, N. Martinez Chanza, S. Albisinni, M. Mercier, N. d’Haene, S. Rorive, T. Roumeguère
Référence : Prog Urol, 2018, 6, 28, 344-350



Bladder cancer is the 9th most diagnosed cancer in the world and even takes the 7th place in men [1]. At diagnosis, almost 80% of urothelial carcinomas are stadified as non-muscular-invasive. In total, 30% of non-muscle invasive bladder cancers (NMIBC) are high-risk tumours. European guidelines define high-risk urothelial carcinoma as T1 infiltration stage, high grade (Grade 3), with the presence of carcinoma in situ (CIS), or for lower grades and stages as concomitant presence of large (>3cm), recurrent and multiple tumours [2]. High-risk tumours associate high recurrence (70% at 5 years) with mortality rates (10% at 5 years) [3, 4]. First line conservative treatment consists of endoscopic resection with adjuvant immunotherapy (intravesical instillations with Bacillus Calmette-Guérin (BCG) or direct cystectomy.

These NMIBC represent a heterogeneous group with wide variations in prognosis and natural history evaluation. Nonetheless, patient care has not been much evaluated lately and we have got only few prognostic factors to adapt treatment and surveillance to our patients.

Several factors (number and time to recurrence, multiplicity and tumour size, presence of CIS) have been proposed to predict recurrence and progression of these tumours using risk calculators like the one presented by the club urologico Espagnol de tratamiento-oncologico (CUETO) or the European organisation for research and treatment of cancer (EORTC) [5, 6].

Extensive catalogues of genetic aberrations in cancers have been generated by high throughput technologies such as next-generation sequencing (NGS).

NGS is a DNA sequencing technology for cancer genomes research. NGS platforms perform sequencing of millions of small fragments of DNA in parallel. Bioinformatics analyses are used to piece together these fragments by mapping the individual reads to the human reference genome.

It may provide a more precise diagnosis and classification of the diseases with a more accurate prognosis. Therefore, using NGS, mainly aiming to identify mutations in tumours, could be the basis of personalised cancer management.

Our study analyzed the results of the sequencing of new generation (NGS) targeted on 50 genes of oncological interest obtained on bladder resection parts in high-risk NMIBC patients treated with BCG. The purpose was to describe this population from a molecular point of view and to try to correlate these results in patients with or without recurrence after BCG.


Patient identification

Patient medical files with high-risk NMIBC [2], which had BCG instillations between January 2014 and December 2016 in Erasme-ULB hospital, have been retrospectively reviewed after ethics committee approbation (n°P2017/355). BCG instillations scheme included 6 induction instillations with control cystoscopies and biopsies and afterwards every 3 months, 1 weekly instillation for 3 weeks to complete a minimum of 18 instillations in one year.

Clinical and histological characteristics have been retrospectively reviewed: age, gender, tumour grade and stage, multiplicity, tumour size, number of BCG instillations, recurrence rate (defined by recurrence confirmed on anatomopathological analysis after endoscopic resection) and progression rate (defined as anatomopathological confirmed muscular-invasive (MI) tumour in follow-up).

Anatomopathological analysis

Anatomopathological analysis has been carried out based on 1973 and 2004 WHO and International urologic pathological society classifications, and the samples have been reviewed by two senior pathologists.

NGS analysis

One representative block of the tumour has been selected by the pathologist: a tumour zone surrounded by a haematoxylin-eosin (HE) section and the percentage of tumour cells estimated in this zone. 10-μm paraffin sections have been realized and the tumour zone has been macro-dissected to extract DNA with a Kit QiAamp DNA FFPE Tissue kit (Qiagen) following the manufacturer's instructions.

Ten nanograms of DNA have been used for sequencing by NGS. The technique of sequencing by NGS has been realized according to the manufacturer's instructions in the anatomy-pathology laboratory of Erasme-ULB hospital on a Ion torrent platform (Life Technologies) with the panel "Ion AmpliSeqâ„¢ Cancer Hotspot Panel" permitting sequencing in several regions of interest (n =207) of 50 genes often mutated in cancers: ABL1, AKT1, ALK, APC, ATM, BRAF, CDH1, CDKN2A, CSF1R, CTNNB1, EGFR, ERBB2, ERBB4, EZH2, FBXW7, FGFR1, FGFR2, FGFR3, FLT3, GNA11, GNAQ, GNAS, HNF1A, HRAS, IDH1, IDH2, JAK2, JAK3, KDR, KIT, KRAS, MET, MLH1, MPL, NMP1, NOTCH1, NRAS, PDGFRA, PIK3CA, PTEN, PTPN11, RB1, RET, SMAD4, SMARCB1, SMO, SRC, STK11, TP53, VHL .

This technique is carried out in 4 steps: preparation of libraries by "Multiplex polymerase chain reaction", then one step of emulsion PCR permitting clonal amplification of DNA fragments of interest purified which are then loaded and sequenced by an Ion torrent chip. Interpretation of the results consists in detecting mutations that are manually checked in order to keep only the exonic and non-silent ones which are reported in COSMIC database (cosmic).


Statistic tests from Pearson khi carré association for values without evaluation in time and Kaplan-Meier method after log-rank verification in the different prognostic groups have been used. We used STATA 12.0 statistical software. Value for statistical significance was defined for P <0.05.


A total of 63 patients with high-risk NMIBC and adjuvant BCG treatment have been identified. Five patients have been excluded for upper urinary tract involvement (which potentially presents different molecular pathways), 2 patients for incomplete data, 3 lost in follow-up and two patients died from a concurrent pathology. NGS was successful in 45 out of the 51 remaining patients (88%). 3 out of the 6 unsuccessful NGS presented CIS without papillary tumour. These 45 remaining patients form the basis of our analysis population.

Median age was 72 years (ranging from 39 to 85) and median follow-up was 24 months (ranging from 4 to 40). Men/women ratio was 4/1. 82.2% of tumours were T1 high grades with or without associated CIS, and at the time of analysis 97.8% of the patients have had at least BCG induction instillations. Table 1 shows the patients and tumour characteristics.

In 33 cases (73% of patients), a mutation has been found. FGFR3, TP53 and PIK3CA were the most frequent. In 12 patients, more than one mutation was found, 3 had double FGFR3/TP53 mutation and 3 had double FGFR3/PIK3CA mutation. Table 2 shows the frequencies of the various mutations found.

Recurrence has occurred in 15 patients (33.3%). Median time to recurrence was 11 months (4-37) and 5 patients out of 15 (33.3%) were high-risk NMIBC. Ten patients had NMIBC recurrence (22.2%) and 5 progressed to muscular-invasive (MI) tumours (11.1%). A total of 10 patients were treated by cystectomy (5 NMIBC, 5 MIBC). 5 recurrences occurred after BCG induction instillations and a total of 9 within the first year after endoscopic resection. Figure 1 shows the distribution of the most frequent mutations detected by NGS in different groups: patients without recurrence, with recurrence, with progression, with recurrence after BCG induction and recurrence within the first year after resection respectively. One patient with a mutated FGFR3 has progressed and presented now metastatic disease after cystectomy. Nine cases of PIK3CA mutation have been noted, 3 of them at the same time as FGFR3 mutation.

Figure 1
Figure 1. 

Distribution of most frequent mutations in different groups.

In 16 patients presenting also CIS, 44% (n =7) had TP53 mutation compared to 24% (n =7/29) in tumours without CIS. No patients with FGFR3 or PIK3CA mutations presented CIS concurrently.

No statistically significant association was found between molecular alteration for FGFR3 (P =0.094) and TP53 (P =0.139) and the risk of recurrence after BCG (Figure 2, Figure 3).

Figure 2
Figure 2. 

Kapan-Meier survival curve in months without recurrence in function of FGFR3 status (0=wild type; 1=mutation).

Figure 3
Figure 3. 

Kapan-Meier survival curve in months without recurrence in function of TP53 status (0=wild type; 1=mutation).


Patient care in high risk NMIBC is a major challenge regarding limitations to classify prognostic risks still generally based on histopathological data subject to interpretation variability [3, 5, 6, 7, 8, 9] and a lack of biomarkers capable of predicting response to BCG treatment [6, 10].

We found high frequencies of TP53 and FGFR3 mutations similar to the results of the cancer genome atlas (TCGA). As a reminder, TCGA project provide broad insight into the underlying genetic aberrations existing across multiple cancer types and the potential for determining the clinical significance of critical genetic aberrations. Mutation of FGFR3 gene was the most frequent found in this study (33%) and was more common in patients without recurrence or progression but without reaching statistical significance, in comparison to recurring or progressing patients. Literature describes FGFR3 mutations in mostly 50% of bladder cancers, being more frequent in low grade and stage tumours [9, 11, 12, 13, 14, 15, 16, 17] with favourable progression risk [11, 13, 16, 18]. Implication of FGFR3 in recurrence risk is still controversial with divergent results in literature [16, 18, 19, 20, 21]. Only one study showed better specific survival in NMIBC presenting FGFR3 mutation [14]. The study of van Rhijn and al. [9] reached statistically better prognostic values combining molecular factors (FGFR3 and MIB-1) with EORTC score, which is one of the recommended standard features [4]. Deregulation of FGFR3 by mutation, over-expression or both at the same time is frequent in bladder cancer, namely in 81% of NMIBC and 54% of MIBC [13]. This explains the interest in developing treatments targeting FGFR3 by different possible mechanisms: Anti-FGFR treatments like tyrosine kinase inhibitors or antibodies blocking FGFR3 [12]. Several trials with anti-FGFR agents alone or in combination therapies in NMIBC and MIBC are in progress.

Beside FGFR3 , alterations of TP53 are frequently described in urothelial tumours. These are more often linked to high grade and stage tumours [12, 15, 22, 23], with worse prognostic outcomes and higher progression risk [12, 15, 23, 24, 25]. TP53 alterations are often present at an early stage in tumours presenting CIS [12, 26, 27]. Predictive character of TP53 in response to BCG treatment or chemotherapy is still controversial [22]. Recent trials study the possibility to reintroduce TP53 activity in tumours with the mutation to promote apoptosis in cancer cells and diminish abnormal cell proliferation, both effects of TP53 [22]. TP53 mutation was the 2nd most found in the present study with 14/45 cases (31%). Even if the frequency was higher in early recurrent patients (after BCG induction or within 12 months), no statistical link has been reached.

FGFR3 and TP53 mutations are rarely associated without being exclusive [12, 15]. The absence of this exclusivity is confirmed in this study with three patients presenting both mutations. High grade NMIBC present a heterogeneous group where mutational pathways may cross. The association of FGFR3 mutation to TP53 seems to have a protective effect [16]. None of the patients presenting FGFR3/TP53 had occurred recurrence or progression at the moment of analysis.

PIK3CA mutations are frequently linked to FGFR3 mutations and have been described more often in low grade and stage tumours, even if this association is less established [17, 27]. Specific PIK3CA inhibitors occur clinical evaluation [28].

Mutations in RAS genes (HRAS, KRAS, NRAS ) are present in 13% of bladder cancers without having clear association to specific grade, stage or established prognostic value [17, 27]. These mutations are mutually exclusive to FGFR3 mutations [17, 29], a fact we have found out in the current study with none of the 4 RAS mutations found in presence of FGFR3 mutation. This suggests different pathways [29].

Several gene mutations found in our study like CTNNB1, CDKN2A, BRAF, NOTCH, FBXW7, MET or KIT have been described in urothelial carcinoma with unknown clinical impact.

Carcinoma in situ is a less explored disease with few available prognostic factors. In our study, NGS has had results in only 3 out of 6 cases with solely CIS. A possible explanation is that collecting enough DNA material from the tumour site can be difficult in these cases. If mutations of TP53 are frequently seen in tumours with CIS, FGFR3 mutations are rare in these tumours [16, 27]. In our study, no mutation of FGFR3 was seen when CIS was present.

A recent study failed to confirm TP53 or FGFR3 as prognostic factors for recurrence after BCG instillations but was able to statistically bind ARID1A , a chromatin-modifying gene to recurrence [30]. This mutation was not tested in the current study but is poorly described in the literature so far.

This study is one of the first to focus on NGS in the specific group of patients with high-risk NMIBC treated by BCG instillations. Our study has limitations due to its monocentric character with a small number of patients, a short period of follow-up and has limited power in detecting genomic significant associations between the mutations found and recurrence or progression of the disease. In 73% of the cases, at least one mutation of the 50 genes searched in our panel has been found which arouses our interest in performing molecular analysis in this population to find prognostic and predictive biomarkers with personalized treatments, which may be developed.


Next generation sequencing in non-muscular invasive bladder cancer could be an additional aid in our therapeutic decision making in the near future. In an area where personalized medicine is rapidly growing, we need larger studies to define molecular characteristics in tumours, and try to find predictive biomarkers and specific targeted treatments.

Disclosure of interest

The authors declare that they have no competing interest.

Table 1 - Patients and tumour characteristics.
Median age  72 years  (39-85 years) 
Gender  36 men (80%)  9 women (20%) 
Median follow-up  24 months  (4-40 months) 
Previous urothelial carcinoma  Patients 
Yes  12  27% 
No  33  73% 
Ta LG:  1 (hist. pT1 HG)  2.2% 
Ta HG:  8.9% 
T1 HG:  24  53.3% 
T1 LG+CIS:  2.2% 
T1 HG+CIS:  12  26.7% 
CIS:  6.7% 
Number of BCG instillations    (5-30) 
<6  2.2% 
10  22.2% 
7-12  12  26.7% 
>12  22  48.9% 
Number of recurrence  15  33.3% 
Number of progression  11.1% 

Légende :
HG: high grade; LG: low grade; CIS: carcinoma in situ.

Table 2 - NGS characteristics.
0 mutation detected  12  26.7 
FGFR3  15  33.3 
TP53  14  31.1 
PI3KCA  20.0 
CTNNB1  4.4 
BRAF  2.2 
NOTCH  2.2 
CDKN2A  2.2 
FBXW7  2.2 
KIT  2.2 
MET  2.2 
≥2 mutations  12  26.7 
FGFR3+TP53  6.7 
FGFR3+PI3KCA  6.7 


Antoni S., Ferlay J., Soerjomataram I., Znaor A., Jemal A., Bray F. Bladder cancer incidence and mortality: a global overview and recent trends Eur Urol 2017 ;  71 (1) : 96-108 [cross-ref]
Babjuk M., Böhle A., Burger M., Capoun O., Cohen D., Compérat E.M., et al. EAU Guidelines on non-muscle-invasive urothelial carcinoma of the bladder: update 2016 Eur Urol 2017 ;  71 (3) : 447-461 [cross-ref]
Chamie K., Litwin M.S., Bassett J.C., Daskivich T.J., Lai J., Hanley J.M., et al. Recurrence of high-risk bladder cancer: a population-based analysis: burden of high-risk bladder cancer Cancer 2013 ;  119 (17) : 3219-3227 [cross-ref]
Sylvester R.J., van der Meijden A.P., Oosterlinck W., Witjes J.A., Bouffioux C., Denis L., et al. Predicting recurrence and progression in individual patients with stage Ta T1 bladder cancer using EORTC risk tables: a combined analysis of 2596 patients from seven EORTC trials Eur Urol 2006 ;  49 (3) : 466-477 [cross-ref]
Fernandez-Gomez J., Madero R., Solsona E., Unda M., Martinez-Piñeiro L., Gonzalez M., et al. Predicting nonmuscle invasive bladder cancer recurrence and progression in patients treated with Bacillus Calmette-Guerin: the CUETO scoring model J Urol 2009 ;  182 (5) : 2195-2203 [cross-ref]
Cambier S., Sylvester R.J., Collette L., Gontero P., Brausi M.A., van Andel G., et al. EORTC nomograms and risk groups for predicting recurrence, progression, and disease-specific and overall survival in non-muscle-invasive stage Ta-T1 urothelial bladder cancer patients treated with 1-3 years of maintenance Bacillus Calmette-Guérin Eur Urol 2016 ;  69 (1) : 60-69 [cross-ref]
Murphy W.M., Takezawa K., Maruniak N.A. Interobserver discrepancy using the 1998 World Health Organization/International Society of Urologic Pathology Classification of urothelial neoplasms: practical choices for patient care J Urol 2002 ;  168 (3) : 968-972 [cross-ref]
Witjes J.A., Moonen P.M.J., van der Heijden A.G. Review pathology in a diagnostic bladder cancer trial: effect of patient risk category Urology 2006 ;  67 (4) : 751-755 [inter-ref]
van Rhijn B.W., Zuiverloon T.C., Vis A.N., Radvanyi F., van Leenders G.J., Ooms B.C., et al. Molecular grade (FGFR3/MIB-1) and EORTC risk scores are predictive in primary non-muscle-invasive bladder cancer Eur Urol 2010 ;  58 (3) : 433-441 [cross-ref]
Saint F., Salomon L., Quintela R., Cicco A., Hoznek A., Abbou C.C., et al. Do prognostic parameters of remission versus relapse after Bacillus Calmette-Guérin (BCG) immunotherapy exist? Analysis of a quarter century of literature Eur Urol 2003 ;  43 (4) : 351-360
di Martino E., Tomlinson D.C., Knowles M.A. A decade of FGF receptor research in bladder cancer: past, present, and future challenges Adv Urol 2012 ;  2012 : 1-10 [cross-ref]
Neuzillet Y., Paoletti X., Ouerhani S., Mongiat-Artus P., Soliman H., de The H., et al. Meta-analysis of the relationship between FGFR3 and TP53 mutations in bladder cancer. Koul H, éditeur PLoS One 2012 ;  7 (12) : e4899310.1371/journal.pone.0048993
Tomlinson D., Baldo O., Harnden P., Knowles M. FGFR3 protein expression and its relationship to mutation status and prognostic variables in bladder cancer J Pathol 2007 ;  213 (1) : 91-98 [cross-ref]
van Oers J.M., Zwarthoff E.C., Rehman I., Azzouzi A.R., Cussenot O., Meuth M., et al. FGFR3 mutations indicate better survival in invasive upper urinary tract and bladder tumours Eur Urol 2009 ;  55 (3) : 650-658 [cross-ref]
van Rhijn B.W., van der Kwast T.H., Vis A.N., Kirkels W.J., Boevé E.R., Jöbsis A.C., et al. FGFR3 and P53 characterize alternative genetic pathways in the pathogenesis of urothelial cell carcinoma Cancer Res 2004 ;  64 (6) : 1911-1914 [cross-ref]
van Rhijn B.W., van der Kwast T.H., Vis A.N., Kirkels W.J., Boevé E.R., Jöbsis A.C., et al. Molecular grading of urothelial cell carcinoma with fibroblast growth factor receptor 3 and MIB-1 is superior to pathologic grade for the prediction of clinical outcome J Clin Oncol 2003 ;  21 (10) : 1912-1921 [cross-ref]
Kompier L.C., Lurkin I., van der Aa M.N.M., van Rhijn B.W.G., van der Kwast T.H., Zwarthoff E.C., et al. FGFR3, HRAS, KRAS, NRAS and PIK3CA mutations mutations in bladder cancer and their potential as biomarkers for surveillance and therapy PLoS One 2010 ;  5 (11) : e1382110.1371/journal.pone.0013821
Burger M., van der Aa M.N.M., van Oers J.M.M., et al. Prediction of progression of non-muscle-invasive bladder cancer by WHO 1973 and 2004 grading and by FGFR3 mutation status: a prospective study Eur Urol 2008 ;  54 (4) : 835-844 [cross-ref]
Hernández S., López-Knowles E., Lloreta J., Kogevinas M., Amorós A., Tardón A., et al. Prospective study of FGFR3 mutations as a prognostic factor in nonmuscle invasive urothelial bladder carcinomas J Clin Oncol 2006 ;  24 (22) : 3664-3671
van Rhijn B.W., Liu L., Vis A.N., Bostrom P.J., Zuiverloon T.C., Fleshner N.E., et al. Prognostic value of molecular markers, sub-stage and European Organisation for the Research and Treatment of Cancer risk scores in primary T1 bladder cancer BJU Int 2012 ;  110 (8) : 1169-1176 [cross-ref]
Lamont F.R., Tomlinson D.C., Cooper P.A., Shnyder S.D., Chester J.D., Knowles M.A. Small molecule FGF receptor inhibitors block FGFR-dependent urothelial carcinoma growth in vitro and in vivo Br J Cancer 2011 ;  104 (1) : 75-82 [cross-ref]
Slaton J.W., Benedict W.F., Dinney C.P. P53 in bladder cancer: mechanism of action, prognostic value, and target for therapy Urology 2001 ;  57 (5) : 852-859 [inter-ref]
Knowles M.A. The genetics of transitional cell carcinoma: progress and potential clinical application BJU Int 1999 ;  84 (4) : 412-427
van Rhijn B.W., Catto J.W., Goebell P.J., Knüchel R., Shariat S.F., van der Poel H.G., et al. Molecular markers for urothelial bladder cancer prognosis: toward implementation in clinical practice Urol Oncol 2014 ;  32 (7) : 1078-1087 [cross-ref]
Malats N., Bustos A., Nascimento C.M., Fernandez F., Rivas M., Puente D., et al. P53 as a prognostic marker for bladder cancer: a meta-analysis and review Lancet Oncol 2005 ;  6 (9) : 678-686 [inter-ref]
Spruck C.H., Ohneseit P.F., Gonzalez-Zulueta M., Esrig D., Miyao N., et al. Two molecular pathways to transitional cell carcinoma of the bladder Cancer Res 1994 ;  54 (3) : 784-788
Goebell P.J., Knowles M.A. Bladder cancer or bladder cancers? Genetically distinct malignant conditions of the urothelium Urol Oncol 2010 ;  28 (4) : 409-428 [cross-ref]
Ross R.L., McPherson H.R., Kettlewell L., Shnyder S.D., Hurst C.D., Alder O., et al. PIK3CA dependence and sensitivity to therapeutic targeting in urothelial carcinoma BMC Cancer 2016 ;  16 : 55310.1186/s12885-016-02570
Jebar A.H., Hurst C.D., Tomlinson D.C., Johnston C., Taylor C.F., Knowles M.A. FGFR3 and Ras gene mutations are mutually exclusive genetic events in urothelial cell carcinoma Oncogene 2005 ;  24 (33) : 5218-5225 [cross-ref]
Pietzak E.J., Bagrodia A., Cha E.K., Drill E.N., Iyer G., Isharwal S., et al. Next-generation sequencing of nonmuscle invasive bladder cancer reveals potential biomarkers and rational therapeutic targets Eur Urol 2017 ;  72 (6) : 952-959 [cross-ref]

© 2018 
Elsevier Masson SAS. Tous droits réservés.