The percentage of participants with a >= 50% decline in PSA values will be reported with 95% exact confidence interval. For each participant, percentage decline in PSA values are calculated as 100% times the difference between PSA values taken at baseline and 12 weeks divided by PSA values at baseline. Percentage of participants determined as 100% times the number of participants with >= 50% decline divided by overall number of participants.

Evaluable patients will include those patients for whom data from genomic studies and IHC tests are available from the pre-treatment biopsy samples and who have completed at least 12 weeks of therapy or who have confirmed disease progression prior to 12 weeks of therapy, and for whom baseline and 12 week PSA data were available. 36 subjects shown in participant flow. 2 subjects were excluded from this PSA analysis: 1 had missing PSA data, 1 had falling PSA in setting of rapid disease progression

Evaluate the association between PSA response at 12 weeks after initiating therapy, and TP53 copy number alterations and mutations at baseline. Simple Logistic regression was planned but due to sample size we used Fisher's Exact test. Percentage of patients in each arm with TP53 copy number alterations and mutations will be reported and odds ratio with two-sided 95% confidence interval will be calculated.

For the primary endpoint, evaluable patients will include those patients for whom data from genomic studies and IHC tests are available from the pre-treatment biopsy samples and who have completed at least 12 weeks of therapy or who have confirmed disease progression prior to 12 weeks of therapy. Targeted DNA sequencing was performed on 26 of 36 biopsies that contained sufficient material.

Evaluate the association between PSA response at 12 weeks after initiating therapy, and PTEN copy number alterations and mutations at baseline. Simple Logistic regression was planned but due to sample size we used Fisher's Exact test. Percentage of patients in each arm with PTEN copy number alterations and mutations will be reported and odds ratio with two-sided 95% confidence interval will be calculated.

For the primary endpoint, evaluable patients will include those patients for whom data from genomic studies and IHC tests are available from the pre-treatment biopsy samples and who have completed at least 12 weeks of therapy or who have confirmed disease progression prior to 12 weeks of therapy. Targeted DNA sequencing was performed on 26 of 36 biopsies that contained sufficient material.

Evaluate the association between PSA response at 12 weeks after initiating therapy, and RB1 copy number alterations and mutations at baseline. Simple Logistic regression was planned but due to sample size we used Fisher's Exact test. Percentage of patients in each arm with RB1 copy number alterations and mutations will be reported and odds ratio with two-sided 95% confidence interval will be calculated.

For the primary endpoint, evaluable patients will include those patients for whom data from genomic studies and IHC tests are available from the pre-treatment biopsy samples and who have completed at least 12 weeks of therapy or who have confirmed disease progression prior to 12 weeks of therapy. Targeted DNA sequencing was performed on 26 of 36 biopsies that contained sufficient material.

Median AR mRNA expression between responders and non-responders.

Evaluable patients will include those patients for whom data from genomic studies and IHC tests are available from the pre-treatment biopsy samples and who have completed at least 12 weeks of therapy or who have confirmed disease progression prior to 12 weeks of therapy, and for whom baseline and 12 week PSA data were available. RNA sequencing was performed on 25 of 36 biopsies that contained sufficient material.

Median AR-V7 expression between responders and non-responders.

Evaluable patients will include those patients for whom data from genomic studies and IHC tests are available from the pre-treatment biopsy samples and who have completed at least 12 weeks of therapy or who have confirmed disease progression prior to 12 weeks of therapy, and for whom baseline and 12 week PSA data were available. RNA sequencing was performed on 25 of 36 biopsies that contained sufficient material.

Evaluate the association between PSA response at 12 weeks after initiating therapy, and AR copy number alterations and mutations at baseline. Simple Logistic regression was planned but due to sample size we used Fisher's Exact test. Percentage of patients in each arm with AR copy number alterations and mutations will be reported and odds ratio with two-sided 95% confidence interval will be calculated.

For the primary endpoint, evaluable patients will include those patients for whom data from genomic studies and IHC tests are available from the pre-treatment biopsy samples and who have completed at least 12 weeks of therapy or who have confirmed disease progression prior to 12 weeks of therapy. AR immunohistochemistry was available in 22 patients with sufficient samples.

The number of participants, responders and non-responders, that were found to have protein expression of AR.

For the primary endpoint, evaluable patients will include those patients for whom data from genomic studies and IHC tests are available from the pre-treatment biopsy samples and who have completed at least 12 weeks of therapy or who have confirmed disease progression prior to 12 weeks of therapy. AR protein expression was analyzed using immunohistochemical analysis from 21 patients with sufficient samples.

Median Normalized Enrichment Score (NES) AR activity levels of responders and non-responders. Gene Set Enrichment Analysis (GSEA) is used to interpret gene expression data. GSEA enrichment score (ES) reflects the degree to which a gene set (GS) is overrepresented at the top or bottom of a ranked list of genes. ES is calculated by walking down the list, increasing a running-sum statistic when a gene is in the GS and decreasing when it's not. Magnitude of increment depends on correlation of the gene with the phenotype. ES is the max deviation from zero encountered in walking the list. Positive ES indicates GS enrichment at the top of the list; negative indicates GS enrichment at the bottom. GSEA calculates NES as actual ES divided by mean (ESs against all permutations of the dataset). Low AR activity has been linked to stemness and lineage plasticity that are recognized as a cause of acquired resistance to AR-targeting therapies.

Evaluable patients will include those patients for whom data from genomic studies and IHC tests are available from the pre-treatment biopsy samples and who have completed at least 12 weeks of therapy or who have confirmed disease progression prior to 12 weeks of therapy, and for whom baseline and 12 week PSA data were available. RNA sequencing was performed on 25 of 36 biopsies that contained sufficient material.

Number of patients who achieved a 50% or greater PSA decline any time after 12 weeks post-baseline.

The percentage of participants with an objective response will be reported with 95% exact confidence interval. Objective radiographic response is evaluated using the new international criteria proposed by the revised Response Evaluation Criteria in Solid Tumors (RECIST) guideline (version 1.1). Changes in the largest diameter (unidimensional measurement) of the tumor lesions and the shortest diameter in the case of malignant lymph nodes are used in the RECIST criteria. For each participant, objective response is calculated as 100% times the difference between the sum of measurable target lesions at baseline and the smallest sum of measurable target lesions achieved after the initiation of therapy, divided by the sum of target lesions at baseline.

Only those subjects who have measurable disease present at baseline, have received at least one cycle of therapy, and have had their disease re-evaluated will be considered evaluable for objective response.

Correlations between baseline molecular features and pathways and PFS will be assessed using cox regression model. In addition, Kaplan-Meier plots will be used to graphically illustrate the survival distributions across the strata of categorical molecular predictors.

Correlations between baseline molecular features and pathways and DSS will be assessed using cox regression model. In addition, Kaplan-Meier plots will be used to graphically illustrate the survival distributions across the strata of categorical molecular predictors.

Correlations between baseline molecular features and pathways and OS will be assessed using cox regression model. In addition, Kaplan-Meier plots will be used to graphically illustrate the survival distributions across the strata of categorical molecular predictors.

Correlations between baseline molecular features and pathways and time to PSA progression will be assessed.

Random Forests classification will be used to identify molecular features and pathways present in patients with disease progression or who discontinue Enzalutamide treatment.

We did not perform a random forest classifier. Rather, we used gene set enrichment analysis to understand transcriptional differences between responders and non-responders.

Linear regression model will be used to assess the association for changes in CTC counts from baseline and maximal prostate-specific antigen (PSA) observed while on study.

Data were not collected

Degree of prostate-specific antigen (PSA) change from baseline to 12 weeks. PSA at week 12 minus PSA at baseline divided by PSA at baseline and multiplied by 100%. Decline shown as negative percent and incline shown as positive percent.

Evaluable patients will include those patients for whom data from genomic studies and IHC tests are available from the pre-treatment biopsy samples and who have completed at least 12 weeks of therapy or who have confirmed disease progression prior to 12 weeks of therapy, and for whom baseline and 12 week PSA data were available. 36 subjects shown in participant flow. 2 subjects were excluded from this PSA analysis: 1 had missing PSA data, 1 had falling PSA in setting of rapid disease progression

Correlations between baseline molecular features and pathways and maximal PSA decline observed will be assessed. Linear regression model will be used to assess the association for changes in circulating tumor cells (CTC) counts from baseline and maximal PSA observed while on study.

Data were not collected

The association between molecular predictors and survival outcomes (e.g., progression-free survival [PFS], disease-free survival [DSS] and overall survival [OS]) and time on treatment will be assessed using cox regression model.