**5. Clinical significance of molecular subtypes of bladder cancer**

Not only do clinical outcomes differ among the molecular subtypes, but also therapeutic response. The gold-standard for management of MIBCs for disease confined to the pelvis includes radical cystectomy preceded by platinum based neoadjuvant chemotherapy. Although a significant minority of patients treated this way achieves durable response and improved cancer specific survival, a sizeable fraction does not respond. In fact, a meta-analysis has suggested that there is only 5% absolute survival benefit at 5 years for patients treated in this manner [34]. In addition, concerns regarding delayed surgery and risk of serious morbidity have limited the usage rates of NACT for cystectomy patients at 25% or less [35, 36]. Biomarker tests

that predict chemo-response could address these problems by identifying patients most likely to benefit. In this regard, data suggests that NAC confers the greatest benefit in basal tumors [11, 29], while the "p53-like subtype" has been reported to confer chemo-resistance [13]. However, current ISUP working group guidelines of 2020 do not recommend routine subtyping to guide NACT [37].

Data published by the TCGA Consortium stated that about 69% of BCs contain potentially actionable therapeutic targets which associate with specific molecular subtypes [11]. Mutations and amplifications of FGFR3 are seen in 50–80% of superficial bladder cancers and up to 20% of MIBCs [19]. Luminal-papillary tumors demonstrate a high rate of these alterations [11]; however, clinical trials have not yet incorporated molecular classification to determine patient eligibility. Similarly PIK3CA is also a frequently mutant gene and therapies targeting PI3K pathway have also shown preclinical BC trials [38]. Other targets such as ERBB2 and TSC1 are also being investigated as therapeutic targets.

Molecular subtyping may also provide a guide to BC immunotherapy. Antiprogrammed death 1 (PD-1) and anti-programmed death-ligand 1(PD-L1) and cytotoxic T-lymphocyte antigen 4 (CTLA-4) are important second line therapies for MIBC with NAC failure. A few are also approved as first line therapy for cisplatin ineligible cases. Unfortunately, not all patients benefit from immunotherapy [39, 40]. Testing for biomarkers of response involves IHC for PD-1 or PD-L1 or estimation of tumor mutation burden, micro-satellite instability or immunemicroenvironment of the tumor. However, none of these biomarkers have shown overwhelming predictive efficacy over others [41–43]. The luminal infiltrated subtype in the TCGA 2017 subtyping system was found to be enriched in PD-L1, CTLA-4 and other immune signatures. In particular, although this subtype did not respond well to NACT, they showed good response to anti-PD-L1 and anti-PD-1 treatment. The basal subtype has also shown response to immunotherapeutic agents in addition to being sensitive to NACT [11, 29]. Molecular subtypes may thus help define patient selection for immunotherapy.

#### **6. Conclusion**

Compared to traditional classification of BCs, molecular subtypes provide more information regarding tumor biology, prognosis and treatment. In general, BC can be divided into luminal and non-luminal subtypes based on their degree of urothelial differentiation. The luminal subtype is further subdivided into those with papillary features, which are superficial, predominantly non-invasive. Though they carry good prognosis when compared to other treatment-naïve subtypes, they do not respond well to conventional NAC and may benefit from targeted therapies. The luminal infiltrated type has more inflammatory and stromal signatures. They are more invasive than luminal papillary tumors and may respond well to immunotherapies. Basal/squamous tumors express stem cell and squamous differentiation associated gene expression signatures. They are aggressive untreated, but respond well to NACT as well as immunotherapies but are insensitive to radiotherapy. The neural subtype forms a minority of non-luminal tumors with neuronal or neuroendocrine phenotype and usually carry poor prognosis but respond to NACT. There are an increasing number of molecular subtyping systems being constantly updated. While they carry great potential to reform BC prognostication and therapeutics, they are not entirely without limitations. Accessibility is a key issue in the present times. Molecular subtyping is mainly based on "static" research, especially in NMIBC, and enables a one-time detection and analysis of tumor specimens rather than "dynamic" tracking to over the disease course. It has also mainly focused *Molecular Classification of Bladder Cancer DOI: http://dx.doi.org/10.5772/intechopen.97393*

on genome and transcriptome research so far but proteomics and immune status of tumors are also closely related to their development. Therefore, the implementation of multiomics is a key necessity in future studies on molecular subtyping. Intratumor heterogeneity also provides another challenge with patient outcome being dominated by one subtype more than the other. With the rapid development of single-cell high-throughput sequencing, mass spectrometric analysis, immune cell analysis and other technologies, the accuracy of the molecular subtyping prediction system need further improvement. Compared to the existing classification system, molecular subtyping methods offer a more comprehensive analysis, particularly to guide adjuvant chemotherapy, targeted therapy and immunotherapy. In the future, these classifications will become an important complementary approach to traditional pathological classification.
