*2.2. Primary Prostate Samples with Synchronous Lymph Node Metastases: 19011, 19260, 19145 and 19651*

These cases each presented with elevated PSA levels and prostate adenocarcinomas confirmed on biopsy. Only 19651 had evidence of nodal metastases on conventional imaging preoperatively. However, at radical prostatectomy with lymph node dissection, all had pathologically involved nodes. Genomic alterations of potential relevance are summarized in Figure 3 (19011, 19260, 19145) and Figure 4 (19651). Though nodal involvement at presentation is associated with a high PCa mortality rate [42], the optimal management strategy for these men has not been established. Retrospective data suggest that adjuvant ADT with radiotherapy compared to ADT or observation is beneficial for men with lymph node metastases identified at radical prostatectomy [43]. Based on their ISUP grade group or Gleason score and tumour (T) stage, 19011, 19145 and 19651 also meet eligibility criteria for the STAMPEDE trial arms C and G that have shown benefit for adding docetaxel or abiraterone respectively to ADT [3,44,45]. Though some studies of men with high-risk localized PCa treated with neoadjuvant and/or adjuvant docetaxel demonstrate improved outcomes, these improvements occur in a small proportion, with significant toxicity to many [4,46,47]. These studies have all been based on clinical risk factors, thus, there is an urgent need for biomarkers that better select men likely to benefit, thereby avoiding over- and undertreatment.

**Figure 3.** Summary of genomic alterations in primary prostate samples with synchronous lymph node metastases (Cases 19011, 19260 and 19145). Relevant somatic variants by type listed with SCNAs: blue indicates loss. Circos plots depict mutational load in each tumour sample. The outermost (first) track: autosome (chromosomes 1 to Y) ideograms with centromeres shown in red and the pter-qter orientation in a clockwise direction (length in Mbp); second track: somatic copy-number gains (red) and losses (blue); third track: somatic SNV allele frequencies (not corrected for tumour purity) coloured according to their mutation changes per Alexandrov et al. [39]; fourth and fifth tracks: allele frequencies (not corrected for tumour purity) of small deletions (red) and insertions (blue); innermost circle: acquired genomic rearrangements, including deletions (blue), tandem duplications (red), inversions (orange), insertions (black) and interchromosomal translocations (grey). MET: metastases.

**Figure 4.** Case 19651. (**A**) Summary of relevant genomic alterations (germline and somatic); Circos plots as per Figure 3. GL: germline mutations; MET: metastases; (**B**) Phylogenetic reconstruction of cancer evolution predicted based on somatic SNV and copy number data. Each circle is a predicted tumour subclone, from the leftmost ancestral clone, with pie charts representing cancer cell fractions (proportion of the four samples harbouring the corresponding clone). Sizes of the circles are proportionate to the number of additional small somatic mutations acquired. Number of SCNAs acquired are indicated.

#### 2.2.1. Case 19011: Left Prostate Tumour Core Biopsy

This *TMPRSS2-ERG* positive case had a somatic missense mutation in *MED12* (c.3670C > G; p.Leu1224Val), that is potentially pathogenic in many cancers, including prostate [48], via its upregulation of Wnt/β-catenin signalling [49]. A non-coding somatic SNV upstream of *CTNNB1* that is within the binding region of 188 TFs, may indicate misregulation of this gene involved in Wnt/β-catenin signalling [50]. No therapeutically relevant SCNA or SV was identified. PGA was low (2.7%).

Although no targetable alteration is seen in this case, the lack of mutations and low PGA may still be valuable in guiding decision-making. This patient is unlikely to respond to targeted therapies, like PARP inhibitors, but also to non-targeted agents that rely upon high mutational loads, such as immune checkpoint inhibition. The lack of poor prognostic markers, such as *TP53* loss, could mean this low volume, locally advanced PCa may respond well to aggressive local therapy without escalation to systemic therapy (e.g., addition of docetaxel). A low PGA is associated with a lower risk of BCR following definitive local therapy [51]. Despite meeting criteria for perioperative therapy trials, his genomic profile suggests aggressive local therapy will be sufficient. However, should he relapse however, the alterations in Wnt pathway-associated regions could confer resistance to AR targeting agents [32].
