Basel, Switzerland ( At the Advanced Prostate Cancer Consensus Conference (APCCC) 2019 meeting, Dr. Johann De Bono continued the Molecular Biomarkers and Novel Imaging in Advanced Prostate Cancer (APC) session with a high-level discussion of the clinical utility of molecular markers. Dr. De Bono notes that there are many biomarker classes, but the current talk focuses on predictive biomarkers. In the pre-diagnosis setting, this includes risk biomarkers for prevention or germline DNA repair defects, as well as diagnostic biomarkers, such as multi-parametric MRI, urinary tumor DNA, or cfDNA. In the pre-treatment phase, there are many prognostic and predictive biomarkers, and in the post-treatment phase, there are pharmacodynamic, response, and surrogate biomarkers.

Dr. De Bono notes that there are several perspectives to keep in mind. First and foremost, we need to keep perspective with regards to validation and qualification of biomarkers. Analytic validation matters in that there may be false positives and false negatives. Additionally, clinical qualification matters in that prospective trials are necessary to adequately evaluate these markers, and qualifying predictive biomarkers also requires anticancer drugs targeting that specific subset of patients.

Second, is the perspective of multiplicity. Biomarkers frequently serve multiple purposes, for example as a prognostic and predictive biomarker. One assay may not be enough (ie. next-generation sequencing) and may require additional testing, for example NGS and IHC for MMR defects and ATM detection. Furthermore, there may be multiple alterations in one tumor and it is important to have a predictive biomarker hierarchy. For example, MMR mutations cause mutations such as subclonal ATM or BRCA2 mutations.

Third, is the perspective of next-generation sequencing. It is important to note if the alteration leads to a germline SNP or SNV, whether it is a mutation or truncating/frameshift, whether the amino acid sequence is affected, and whether this leads to loss of function. For NGS, there are also issues with correcting for tumor purity and tumor ploidy (ie. triploid). Just because there is an alteration detected, does not mean there is a loss of function. It is important to note if there is an alteration requirement for complete protein loss, haploinsufficiency or a dominant negative effect.

There are several predictive biomarkers in advanced prostate cancer in 2019, including AR biomarkers, PI3K/AKT and PTEN, transformation biomarkers, and DNA repair defects:

  • AR biomarkers: Are usually not present at diagnosis and are detected in the blood. AR mutations are largely in the AR ligand binding domain and may indicate continued AR dependence actionable by AR targeting. AR amplification is an AR driven sub-clone and may indicate continued AR dependence actionable by AR targeting. AR rearrangements may impact AR splicing and signaling, and AR splice variants (mainly AR-V7) indicates tumor burden and may indicate disease drive by a constitutively active AR. AR driven transcripts indicate continued AR signaling and a low PSA despite disease burden may indicate low AR signaling. According to Dr. De Bono, definitive studies are needed with novel agents with antitumor activity against continued AR signaling.
  • PI3K/AKT and PTEN inhibitors: PTEN loss is associated with less benefit from abiraterone and a poor prognosis, and PI3K/AKT aberrations are also emerging as biomarkers for endocrine resistance. In phase II studies, abiraterone and AKT inhibitors are superior to abiraterone in PTEN loss disease, but we are awaiting results of phase III trials.
  • Transformation biomarkers: Dr. De Bono notes that we need to settle on a best descriptor for what is termed neuroendocrine, small round blue cells, luminal-to-basal, AR independent clones, RB1/TP53 loss. Identifying transformation may matter in that cancer tries to find ways to evade treatment, usually at the sub-clonal level. Lineage plasticity occurs following AR targeting pressure (ie. from abiraterone), can lose “AR signaling dependency”, and can result in neuroendocrine phenotype and PSMA negativity.
  • DNA repair defects: Are found in a small percentage (<5%) of mCRPC patients and are often but not always associated with high TILs. Furthermore, Dr. De Bono notes that these defects often but not always respond to PD-1/PD-L1 checkpoint inhibitors. Orthogonal assays are needed to detect all DNA repair defects, given that the MSI PCR assay works poorly on FFPE, and NGS and IHC can miss many deleterious aberrations

Dr. De Bono concluded with several take-home messages from his clinical utility of molecular markers talk:

  • Assay analytic validation and clinical qualification is an urgent need
  • Molecular stratification for mCRPC is going to become a standard of care
  • More data is needed to prove utility of AR and PTEN/PI3K/AKT assays
  • PSMA targeting agents need to pursue patient selection
  • Many new therapeutic strategies are emerging and further major advances in prostate cancer care are on the horizon in the next 5 years

Presented by: Professor Johann S. De Bono, PhD, MSc, Head of the Division of Clinical Studies, The Institute of Cancer Research, The Royal Marsden NHS Foundation Trust, Sutton, United Kingdom

Written by: Zachary Klaassen, MD, MSc – Assistant Professor of Urology, Georgia Cancer Center, Augusta University/Medical College of Georgia, Twitter: @zklaassen_md at the 2019  Advanced Prostate Cancer Consensus Conference (APCCC) #APCCC19, Aug 29 – 31, 2019 in Basel, Switzerland