AUA 2021: State-of-the-art Lecture: Personalized Medicine in the Management of Prostate Cancer Across the Patient Care Continuum 

( The American Urologic Association annual meeting included a State-of-the-Art Lecture by Dr. Brian Chapin who discussed personalized medicine in the management of prostate cancer across the patient care continuum. Dr. Chapin notes that the way we think about the personalized approach to prostate cancer therapy involves seeing a patient in the clinic and assessing them on a number of factors (clinical features, genetics, genomics, serum markers, receptors, induced responses, and selection pressures) and then developing a personalized approach to their treatment plan:

AUA Brian Chapin-0.jpg 

In setting the stage for personalized care, it is important to understand the available treatment options in the castration-naïve setting, as well as the castration resistant setting:

AUA Brian Chapin-1.jpg 

Dr. Chapin notes that personalized care is compartmentalization by stage, histology and biology. Traditionally we have delineated care by stage of disease in combination with histologic considerations (ie. neuroendocrine, ductal, adenocarcinoma). More recently, there has become an increasingly more sophisticated approach, with regards to assessing the androgen receptor axis (TMPRSS2-ERG, SPOP, AR responsive), loss of tumor suppression (p53, PTEN, Rb), DDR mutations (BRCA, CDK12, FANCONI, CHEK1/2), and mismatch repair (MSI, Lynch syndrome).

Compartmentalization allows for better risk stratification, for example balancing arms in a clinical trial by using stratification factors (ie. M1a/b versus M1c) and balancing groups in retrospective studies by utilizing matching or propensity scoring (ie. NCDB, SEER database). However, this can also generate selection bias in retrospective cohort series, for example in patients with occult node positive prostate cancer, some patients will undergo a completion prostatectomy whereas others will have their radical prostatectomy aborted (thus, unable to compare outcomes). Many of us think of personalized care in prostate cancer by way of DNA, RNA, proteins and receptors, such as in the pre-biopsy, post-biopsy, positive biopsy, post-radical prostatectomy, and metastatic settings:

Brian Chapin-2.jpg 

Dr. Chapin emphasizes that it is critically important to delineate between prognostic versus predictive biomarkers. Prognostic biomarkers are a variable associated with favorable or unfavorable outcomes for patients in the absence of treatment. Predictive biomarkers are a variable used to identify patients or groups of patients most likely to benefit from a specific therapy, for example a patient with a DNA damage repair mutation being a candidate for PARP inhibitor treatment, or a patient with a mismatch repair deficiency being a candidate for anti-PD1 antibody treatment. Prognostic variables can be used incorrectly thus mistakenly influencing management. For example, genomic tests on prostate biopsies have all been based on treated prostate cancer patients, thus these findings may not be applicable to an active surveillance cohort. Findings from these genetic tests are used to make changes in management decisions, resulting in Medicare approval. Dr. Chapin states that it is important to remember that no randomized trials have reported outcomes (although there are several ongoing) assessing if genomic tests improve patient outcomes, and in fact reflexive genomic testing may be detrimental to patients.

Dr. Chapin then discussed several potential predictive markers and prospective trials. The PAM50 gene expression classifier was previously described in the breast cancer literature, but has since been applied to prostatectomy specimens. Zhao et al.1 applied the classifier to 3,782 samples (1,567 retrospective, 2,215 prospective) noting that the PAM50 classifier consistently segregated prostate cancer into three subtypes in both retrospective and prospective cohorts:

  • Luminal A (retrospective 34.3%; prospective 33.3%)
  • Luminal B (retrospective 28.5%; prospective 32.6%)
  • Basal (retrospective 37.1%; prospective 34.1%)

Luminal A, luminal B, and basal curves separate based on PAM50 gene expression, with basal tumors having worse prostate-cancer specific survival:

AUA Brian Chapin-3.jpg 

When assessing luminal B and basal tumors with regards to response to ADT, this study suggests there may be a benefit to treatment of luminal B patients with ADT but no benefit seen in those with basal tumors:

AUA Brian Chapin-4.jpg 

AUA Brian Chapin-5.jpg 

Another example is the post-operative radiation therapy outcomes score (PORTOS), which is a genetic prediction score for post-op radiation. PORTOS is made up of 24 genes selected from a training set of 196 men and validate in a separate cohort of 330 men, with a clinical endpoint of metastasis over 10 years of follow-up.2 In this study, patients with a high PORTOS score had a benefit in cumulative incidence of distance metastasis with radiotherapy (HR 0.15, 95% CI 0.04-0.60), whereas patients with a low PORTOS score (HR 0.92, 95% CI 0.56-1.51) did not have a benefit with radiotherapy:

AUA Brian Chapin-6.jpg 

Biomarker examples have also been described in the advanced prostate cancer setting, specifically assessing androgen indifferent or aggressive variant prostate cancer. In a phase 1-2 trial assessing cabazitaxel plus carboplatin for men with mCRPC, at a median follow-up of 31.0 months, combination cabazitaxel plus carboplatin improved the median progression-free survival from 4.5 months (95% CI 3.5-5.7) to 7.3 months (95% CI 5.5-8.2; HR 0.69, 95% CI 0.50-0.95) with cabazitaxel alone.3 Dr. Chapin notes that what is particularly interesting is that for patients with a positive aggressive variant prostate cancer molecular signature, there was an improvement in overall survival with the addition of platinum based chemotherapy, whereas there was no benefit in those that were aggressive variant prostate cancer molecular signature negative. Dr. Chapin highlighted that there are several trials ongoing in the localized setting assessing a personalized approach, including the Genomic Umbrella Neoadjuvant Study (GUNS):

AUA Brian Chapin-7.jpg 

Decipher is a 22 gene classifier that provides risk stratification based on radical prostatectomy specimen analysis, which is prognostic for metastasis. However, there is no data on predictive ability, which is undergoing prospective evaluation in the NRG-GU009 PREDICT-RT trial:

AUA Brian Chapin-8.jpg 

Importantly, there are several barriers to overcome in the era of personalized medicine, including (i) assessing if findings are transferrable across stages of the disease; (ii) tumor heterogeneity, whether intertumoral, intratumoral, or comparing a metastases to the primary tumor; (iii) in order to move from prognostic to predictive markers, prospective trials are required; (iv) it is important to determine drivers in the setting of co-occurences (ie. DDR, +/- p53, +/- Rb1, +/- PTEN); and (v) assess selection pressures over time (ie. the predominant clone).

Dr. Chapin concluded his presentation with the following take-home messages from his presentation of personalizing medicine in the management of prostate cancer:

  • Progress is being made in the personalized care of prostate cancer patients
  • It is important to delineate between predictive versus prognostic markers
  • We need validation of markers to predict a therapeutic benefit
  • Prospective trials with generation of biobanks are needed moving forward
  • Skeptical optimism is appropriate until validation is completed

Presented by: Brian Chapin, MD, Associate Professor of Urology, MD Anderson Cancer Center, Houston, TX

Written by: Zachary Klaassen, MD, MSc – Urologic Oncologist, Assistant Professor of Urology, Georgia Cancer Center, Augusta University/Medical College of Georgia, @zklaassen_md on Twitter during the 2021 American Urological Association, (AUA) Annual Meeting, Fri, Sep 10, 2021 – Mon, Sep 13, 2021.


  1. Zhao SG, Chang SL, Erho N, et al. Associations of Luminal and Basal Subtyping of Prostate Cancer with Prognosis and Response to Androgen Deprivation Therapy. JAMA Oncol. 2017 Dec 1;3(12):1663-1672.
  2. Zhao SG, Chang SL, Spratt DE, et al. Development and validation of a 24-gene predictor of response to postoperative radiotherapy in prostate cancer: A matched, retrospective analysis. Lancet Oncol. 2016 Nov;17(11):1612-1620.
  3. Corn PG, Heath EI, Zurita A, et al. Cabazitaxel plus carboplatin for the treatment of men with metastatic castration-resistant prostate cancers: A randomized, open-label, phase 1-2 trial. Lancet Oncol. 2019 Oct;20(10):1432-1443.