2.2.2. Case 19260: Right Prostate Tumour Core Biopsy

Patient 19260 was also treatment-naïve at the time of his prostatectomy and sampling for WGS. He biochemically relapsed 16 months postoperatively at which time he had salvage radiotherapy with a good PSA response.

*TMPRSS2-ERG* fusion positive with a low PGA (2.4%), this case presented with a pathogenic somatic missense mutation in *BRAF* (c1406G > T; p.Gly469Val). Known to confer increased kinase activity [48], this mutation may sensitise the patient to BRAF +/− MEK inhibitor therapy. Of interest in CRPC [30], with a report of response to targeted therapy in a *BRAF* mutant patient [52], clinical trials of MEK inhibitors are currently underway (NCT02881242). Though not relevant to this patient's upfront treatment, it could prove useful in the event of relapse.

#### 2.2.3. Case 19145: Left Prostate Tumour Core Biopsy

This *TMPRSS2-ERG* positive tumour had a high PGA (10.2%), but lacked any known deleterious somatic mutation. SCNAs/SVs of note include heterozygous losses in *PTEN, FANCA, CDK12, TP53, NCOR1* and *NCOR2*, an inter-chromosomal translocation with breakpoints overlapping *RAD51B* and *PTEN* and a large heterozygous deletion overlapping with *TP53* and *NCOR1*.

Responses to PARP inhibition have been seen in patients with *FANCA* alterations [23,53] and preclinical data suggest that *PTEN* loss sensitises cancers to PARP inhibitors, with reported cases of exceptional responses to olaparib [54,55]. However, resistance to single agent PARP inhibition has been described in Pten/p53 deficient mouse models, though a synergistic response was seen upon PARP inhibition in combination with PI3K inhibition. [56]. *NCOR1* and *NCOR2* are transcriptional corepressors that negatively regulate androgen receptor (AR) signalling and androgen-induced cell proliferation [57–59]; losses in these genes increase with disease progression and are associated with anti-androgen and ADT resistance [60,61]. *TP53* loss may also predict inferior responses to novel androgen signalling inhibitors (ASIs), such as enzalutamide and abiraterone, in CRPC [62]. CDK12 loss may predict sensitivity to immune checkpoint inhibiting therapies [63].

Many of the observed alterations in this case have therapeutic potential but are still the subject of early phase clinical trials. The presence of the *NCOR1*/*2* losses, however, may indicate a vulnerability in this patient for early development of CRPC. His four week course of ADT preoperatively may have induced treatment resistant clones even at this early stage. These losses together with *TP53* loss and high PGA indicate this patient may develop early resistance to ADT and, given his high-risk disease at presentation, he would be an ideal candidate for escalation of his initial treatment with chemotherapy.

#### 2.2.4. Case 19651: Bilateral Prostate and Internal Iliac Node Tumour Core Biopsies

Reporting a family history of PCa, via his father, and breast cancer in his mother and sister, it was not surprising that this patient carries a pathogenic germline *BRCA2* stop-gain mutation (rs80359031; c.7988A > T; p.Glu2663Val) confirmed to predispose carriers to BRCA-associated cancers.

The somatic heterogeneity across the four tumour samples is striking (Figures 1 and 4A). Of the 78 overlapping SNVs (out of 24,195) present across all four samples, none had notable therapeutic relevance. Phylogenetic reconstruction of this cancer's evolution reveals distinct differences between the left primary and the other three samples (Figure 4B). Notably, the left prostatic primary acquired a somatic pathogenic *BRCA2* stop-gain mutation ((c.6308C > G; p.Ser2103Ter), variant allele frequency (VAF; 26%). Additionally, genes associated with several different growth signalling pathways, including MAPK/ERK, TGF-β, PI3K and WNT, are impacted by SCNAs in the left primary but there are few events in the other samples. No relevant SVs within the left lymph node were noted on WGM. As expected with the combined germline and somatic *BRCA2* mutations, there was a high rate of large deletions in the left primary [10], including a 3Mb deletion overlapping multiple tumour suppressor genes (TSGs) including *BTG* and *DCN*.

Inter- and intra-patient heterogeneity have been well-described in PCa [13] and most recently in multi-focal primary tumours [64], with significant therapeutic implications. The germline mutation not only informs screening for secondary cancers and testing in relatives, *BRCA2* mutations may also be associated with a worse prognosis [65–69] and confer sensitivity to platinum-based chemotherapy [70] and PARP-inhibitors [23,53]. However, there is increasing evidence that responses are markedly improved with biallelic loss and many of the PARP inhibitor clinical trials have refined their inclusion criteria to include only patients with biallelic alterations. Acquiring a somatic *BRCA2* mutation in a single primary tumour could result in a differential response to targeted therapies that would not be predicted based on the typical single site sampling performed in clinical practice.

Aside from the germline *BRCA2* mutation, there is no unifying therapeutically relevant event across all four samples. Having had short-term ADT preoperatively, losses in *NCOR1* and *NCOR2* as well as other SCNAs associated with CRPC within the left primary raise the possibility that early ADT resistance is developing after minimal treatment.

In practice, knowledge of this patient's genomic landscape at baseline may have prompted his treating clinician to escalate his treatment with combination systemic therapy such as the rucaparib arm of the STAMPEDE trial. The loss of *NCOR1* and *NCOR2* and the poorer prognosis conferred by his *TP53* and *BRCA2* status represent potential indications for early chemohormonal therapy (ADT with docetaxel chemotherapy) despite him having low-volume, node only metastases [3,4,46,71]. *BRCA2* alterations may also sensitize this patient to radiotherapy due to impaired DDR. Therefore, had his genomic data been available early, an upfront strategy with radiotherapy to his primary in combination with ADT and docetaxel may have been used. At progression, he may be considered for a clinical trial with a PARP inhibitor, potentially in combination with another agent given his somatic *BRCA2* discordance. A metastatic biopsy at a site of progression could prove useful in determining whether new sites of disease harbour the somatic *BRCA2* alteration.
