ASCO GU 2019: Biologic Basis for Sequencing Novel Treatments for Metastatic Prostate Cancer

• Choice of Androgen receptor (AR) pathway inhibitor vs. docetaxel for castration-sensitive prostate cancer (CSPC)
• Optimal sequencing of approved drugs for castration-resistant prostate cancer (CRPC)
• Patient selection for targeted therapy or immunotherapy
• Early detection of treatment resistance
• Design of rational combination therapies

• Overall increased genomic testing has led to the identification of numerous molecular alterations in advanced prostate cancer – which has led to more biomarker-driven trials but also off label use of current drugs

• International Stand Up to Cancer (SU2C) Dream Team work – have been focusing on 150 metastatic CRPC biopsies
  • Found DNA repair alterations in 22% of specimens – including homologous recombination genes and mismatch repair genes
  • These have prognostic, therapeutic and family implications
• The mutations in the homologous recombination genes have had significant implications:
  • Poor prognosis for germline carriers (PROREPAIR-B trial)
  • Response to PARP inhibitors
  • Response to platinum chemotherapy
  • A possible response to Radium-223?
  • A possible response to AR-targeted therapies (depends on the study)
  • A possible response to checkpoint inhibitors?
  • A possible response to PSMA radionucleotide therapy?

But,

• Are all HR mutations the same? Do somatic or germline mutations have different outcomes?
  • In PROREPAIR 2, germline BRCA2 mutations had worse prognosis than other mutations
• Could similar biologic and clinical results be produced by co-targeting PARP and AR?
  • Phase 2 abiraterone + olaparib, unselected for HR mutations, the PFS was improved by 5 months (13.8 vs. 8.2 months)
  • DNA repair crosstalk with AR pathway?
  • But, who are the best responders to this combination?
• Mismatch repair/microsatellite instability (MMR and MSI)
  • MMR loss results in hypermutations while MSI results in neoantigens and PDL-1 expression
  • 3-5% of mCRPC patients have these – maybe more, depending on testing
  • Pembrolizumab has been pan-approved for all histologies for patients with MSI-high tumors
  • However, as seen in Abida et al. (JAMA Onc 2018), only 6/11 patients with MSI-high tumors responded to Pembo – why don’t they all respond? What is the optimal approach to capturing this population?

• 1) Monoclonal – where all metastases are derived from one clone from the original multifocal tumor. Therefore, they likely share some mutations with the founding lesion/primary tumor. This model is easier to treat broadly, as a single mutation may be shared by all lesions.
• 2) Polyclonal – where the original founding tumor/primary has multiple clones with different founder mutations. As these each metastasizes and the mets cross-talk, there is a much broader array of mutations. It becomes harder to

• When would we NOT choose an AR-targeted therapy?
• Germline HSD3B1 (1245C) – steroid biosynthesis enzyme – is associated with decreased response to primary ADT – may be useful as a marker for therapy intensification?
• AR ligand binding domain mutation – useful to select an AR-targeted agent vs taxane in mCRPC?
• AR splice variants and amplification - useful to select an AR-targeted agent vs taxane in mCRPC?

Some of her suggested strategies are as follows:

• Not effective: sequencing abiraterone/enzalutamide in CRPC
• Not effective: adding abiraterone to enzalutamide at the time of enza resistance
• Combination strategies and targeting cross-talk pathways
  • Combining abi/enza up front
  • Targeting AR + chemotherapy, AR + PARP targeting, etc.
• Newer AR targeting agents with different mechanisms
• Bipolar androgen / high dose T (rapid cycling on androgen)
• Earlier targeting before resistance develops – but we need to determine which patients benefit from this strategy!

• Small cell / neuroendocrine tumors – due to histologic transformation
• Loss of AR expression / AR signaling (ie low PSA)
• Clinical features such as visceral metastases, lytic bone metastases, low PSA, short ADT response
• Molecular features such as combined TP53, RB1 and PTEN loss? (Hamid et al. EU 2018)
• Can we use any of these to better identify patients for platinum chemotherapy vs. AR pathway inhibitor? Identify new therapies? Co-target AR and other susceptible foci?

• There is a growing need for biomarkers to inform the optimal choice and sequencing of systemic therapies
• Certain “targetable” genomic alterations are enriched in CRPC – ie BRCA2, AKT, CDK12, MMR
• Resistance pathways are primarily AR-driven – has led to more potent AR-targeted approaches
• Detection of non-AR driven disease – identifies patients with poor prognosis
• Challenges for precision medicine include tumor evolution, heterogeneity, pathway cross-talk, non-genomic factors, many clinical factors