Abstract The evolution to metastatic disease is a major determinant of cancer mortality. Cancer evolution involves a complex interplay between intrinsic genetics and transcriptional alterations and the microenvironment. To define mechanisms underpinning metastatic heterogeneity in late-stage disease, we focus on metastatic castration-resistant prostate cancer and employed single-cell multi-omics and whole-genome sequencing to deeply profile 34 metastatic lesions obtained from 9 patients through rapid autopsy. We find evolutionary convergence of intra-tumour heterogeneity, characterised by recurrent tumour populations acting as critical functional components of the tumour ecosystem, irrespective of clonal and microenvironmental backgrounds. We find little evidence of the microenvironment driving transcriptional heterogeneity, but there are signatures of co-adaptation between the microenvironment and tumour cells. In contrast, clonal evolution primarily foster widespread transcriptional changes that did not result in de novo functional states. Intra-patient functional convergence of tumour ecosystems across metastases indicates system-level selection pressures that drive the heterogeneity landscape of metastatic castration-resistant prostate cancer. Our findings reveal functional evolutionary convergence of metastatic disease into distinct intra-tumour subpopulations, identifying critical determinants for therapeutic targeting.

Breast implant-associated anaplastic large cell lymphoma (BIA-ALCL) is a rare T cell lymphoma in women with textured implants. Little is known about the cell of origin (COO) and whether the tumour immune microenvironment (TIME) is critical for BIA-ALCL survival. Single-cell RNA sequencing revealed BIA-ALCL cells were heterogeneous with unique gene profiles per patient and signature genes BATF3, SERPINS, TNFSFR8, and IL2RA. The COO is a CD4+ or CD8+ memory T cell identified through rearranged TCR and gene expression. The BIA-ALCL TIME included distinct myeloid clusters, dendritic cells (DCs), and monocytes, which may support lymphoma cells. Cytokines IL-13, IL-10, TNF and secretory PDL1 were increased in BIA-ALCL seroma, indicating an immunosuppressive TIME. Interactome analysis revealed a complex network among lymphoma cells, dominated by IL-13, IL-10, and TNF members. Endogenous CD8+ T cells exhibit higher checkpoint expression than benign seromas. No clonal relationship exists between BIA-ALCL and endogenous T cells. Tissue-archetype and gene-expression analysis revealed T-cell exclusion and immunosuppressive TIME in invasive disease. In summary, BIA-ALCL cells are diverse with markedly different TIME across stages. The TIME provides immunosuppressive signals to activated/exhausted T cells and homeostatic signals that promote BIA-ALCL proliferation. These new findings may offer novel therapeutic targets for advanced disease patients.

BACKGROUND:Breast cancer is etiologically heterogeneous, but which risk factors differ in their associations across tumor subtypes remains unclear. We conducted a large, pooled analysis to evaluate independent, dose-response associations between breast cancer risk factors and quantitative tumor features. METHODS:Analyses of 15,731 invasive breast cancers from 24 studies evaluated associations (p-trend) between reproductive and hormonal factors, body mass index (BMI), alcohol, smoking, and family history in relation to quantitative immunohistochemistry measures on tissue microarrays (ER, PR, HER2, KI67, TP53) and tumor grade. Analyses in a subset of 10 population-based studies estimated subtype-specific odds ratios (ORs) comparing cases to controls. A Bayesian False Discovery Probability (BFDP) <0.2 was used to identify associations with strong statistical evidence. RESULTS:Nulliparity and later age at menopause were associated with higher ER-positivity (p-trend=0.021 and 0.001, respectively), with corresponding OR[ER+] (95% CI) = 1.49 (1.16-1.90) for nulliparous vs. parous and 1.07 (1.03-1.11) per 5 years. Current combined menopausal hormone therapy (MHT) use was associated with lower grade (p-trend<0.001), with OR [grade1] = 3.37 (2.69-4.21) for current vs. never users. Higher BMI was associated with lower ER-positivity and higher grade in premenopausal women (p-trend<0.001 and <0.001), with OR[ER+] = 0.80 (0.74-0.87) and OR[grade1] = 0.75 (0.63-0.88) per 5 units, and with higher PR-positivity and higher grade in postmenopausal women (p-trend<0.001 and <0.001), with OR[PR+] = 1.08 (1.03-1.14) and OR[grade 3] = 1.10 (1.04-1.17) per 5 units. CONCLUSION:This pooled analysis of 15,731 cases showed that nulliparity, age at menopause, MHT, and BMI have independent, dose-response associations with ER, PR, and grade, clarifying patterns of etiologic heterogeneity. Associations with HER2, KI67 and TP53, or other risk factors did not meet our threshold for strong evidence.

Background Pathogenic germline variants in certain genes are associated with somatic tumour mutation signatures. The use of somatic tumour mutation data has the potential to improve the identification of true pathogenic variants but remains underexplored. We investigated the integration of tumour homologous recombination (HR) deficiency status as a predictor of pathogenicity for germline BRCA1 and BRCA2 variants, building on the established link between HR deficiency and germline pathogenic variants in these genes. Methods We analysed breast tumour whole-genome sequence and matching germline data from 350 patients across four datasets: Familial Breast Cancer (N = 77), The Cancer Genome Atlas (TCGA-BRCA, N = 96), the MAGIC study (N = 136), and Q-IMPROvE (N = 41). A total of 15,156 germline variants (including structural variations) in BRCA1, BRCA2, and other cancer genes (ATM, BARD1, BRIP1, CHEK2, PALB2, PTEN, RAD51C, RAD51D, TP53) underwent variant curation. Patients were categorised based on germline classification as BRCA1 positive (N = 27), BRCA2 positive (N = 21), and BRCA1/2 negative (N = 232), excluding those with BRCA1/2 variants of uncertain significance (N = 8) and pathogenic or only uncertain variants in other cancer genes (N = 62). Somatic HR status (deficient or proficient) was predicted using three algorithms: HRDetect, CHORD, and HRDsum. HR-deficient and HR-proficient status were significant predictors of germline BRCA1/2 pathogenic variant status (positive and negative directions). Findings The CHORD algorithm, which estimates BRCA1 and BRCA2 subtype specifically, added precision contributing evidence towards pathogenicity for the corresponding gene, reaching pathogenic moderate strength for the relevant gene-subtype. Finally, we assessed CHORD HR predictions for variants of uncertain significance in BRCA1 and BRCA2, and reported their tumour HR status for potential use as additional evidence in variant curation. Interpretation Analysis across multiple tumour whole-genome sequencing datasets has shown that HR status prediction algorithms can separate profiles for BRCA1 and BRCA2 pathogenic variants and provide further evidence at increased weight to aid in the classification of germline BRCA1 and BRCA2 variants. Tumour sequencing offers a promising strategy for reducing the uncertainty in germline variant interpretation. Funding This work was funded by the National Breast Cancer Foundation.

BACKGROUND:Communicating genetic test results within families can facilitate cascade testing that informs cancer prevention and surveillance. Researchers at the University of Washington, USA, developed a public website called ConnectMyVariant to support family communication of genetic test results. Australian families share similar challenges communicating genetic information. AIMS:To explore the acceptability of ConnectMyVariant resources, identify content requiring adaptation, and need for further resources. METHODS:Semi-structured interviews were conducted across three stages to review the ConnectMyVariant resources, assess relevance and acceptability for Australians, and identify need for adaptation. Stages 1 and 2 involved individuals with an identified hereditary cancer variant and their relatives. Stage 3 involved genetic counsellors. Resource content was revised between stages based on interview findings. Content analysis was used to analyse the interviews. RESULTS:Stage 1 interviews (n = 31) highlighted the value of the resources but identified changes needed to make them suitable for an Australian context. Adaptations included adding ethics and consent around risk communication, reducing communicator burden, simplifying and softening language, incorporating local support services and genetic testing processes, including implications on life insurance. Stage 2 interviews (n = 6) confirmed the adaptations were acceptable. In stage 3 interviews (n = 7), genetic counsellors recognised the need for a communication resource and recommended minor changes to reflect local genetic testing processes, clinical language, and service preferences. CONCLUSION:This study demonstrates there is value in a communication resource for Australian families, and highlights the importance of undertaking cultural adaptation of resources. The authors are actively disseminating resources across community and clinical settings.

Schema of study numbers for each analysis to describe different cohort numbers due to pathology review and missing data. MOC, mucinous ovarian carcinoma; MBOT, mucinous borderline ovarian tumor; LGI, lower gastrointestinal; UGI, upper gastrointestinal; SISH, silver in situ hybridization.

Associations between gene expression and stage group and OS in MOC.

Kaplan–Meier curves of OS in (A) main tumor groups (n = 582)—MBOT, MOC, LGI, and UGI; (B) patients with MOC by FIGO stage (n = 184); (C) patients with MOC by pattern of invasion in all stages (n = 178); and (D) patients with stage I MOC (n = 134), by pattern of invasion.

Supplementary Figures S1-S13

Heat map of unsupervised clustering analysis. Contains all samples with a concordant pathology diagnosis (n = 497), with MOC grouped by FIGO stage. Labels show main clusters and diagnoses. Gene-expression values are normalized and logarithm base 2 transformed. dx, diagnosis.

Background: The mechanisms by which parity and breastfeeding can mediate protection against breast cancer (BC), particularly triple-negative breast cancer (TNBC), remain unclear. CD8+ T cells with a tissue-resident (TRM)-like phenotype have been shown to be present in primary TNBCs with high quantities of immune infiltrate and are associated with better clinical outcomes. We have previously shown TRM-like cells are also present in healthy, non-cancer-affected normal human breast tissue. We hypothesized that the inflammatory processes induced post pregnancy could result in increased CD8+ T cell quantity and this could facilitate protection against BC in the normal breast. Methods:We compared immune cell quantities in healthy, cancer-unaffected normal breast tissue collected from parous versus nulliparous women using flow cytometry and from single cell RNA sequencing datasets. Using immunocompetent and deficient murine models, we characterised mammary tissue infiltrating lymphocyte populations at early (Day 10) and late involution (Day 28) timepoints, without and with lactation respectively. We next correlated immune cell profiles and rate of tumor growth by implanting mammary tumours at the different stages of involution. We further evaluated the quantity of immune infiltration from a dataset of 736 women who had developed BC with parity and breastfeeding history. Results:We observed significantly higher quantities of CD45+ immune cells and CD8+ TRM-like cells in the normal cancer unaffected breast tissue of parous women compared to nulliparous women using both flow cytometry (n=90) and single cell RNA sequencing datasets (n=263). In murine models, significantly more CD8+ TRM-like cells accumulated in mammary tissue at the later D28 vs earlier D10 involution timepoint compared with virgin control mice. The increase in TRM-like cells in mammary tissue correlated with significantly decreased tumor growth compared to virgin controls, whereas the early involution timepoint had similar tumor growth rates to virgin mice. The tumors implanted at D28 compared with the D10 involution also had significantly higher intratumoral T cell content. To specifically investigate if these tumor protective effects were mediated by T cells, we used RAG2-/-yc-/- immunodeficient mice inoculated with AT3-OVA tumors into the mammary gland at the early D10 and late D28 involution timepoints after injection of OT-I cells (transgenic CD8+ T cells that specifically recognise OVA antigen) prior to mating. We observed significantly reduced mammary tumor growth at the later D28 vs earlier D10 timepoint, corresponding to increased conversion of naive OT-I cells to a resident memory phenotype (CD69+CD103+). We next examined a cohort of patients who were diagnosed with BC and their intratumoral immune content according to subtype and parity/breastfeeding history (n=736). Consistent with the murine data, we found significantly increased tumoral immune infiltration in parous women who had developed postpartum TNBC compared with nulliparous women. Conclusions:Our data, from both human normal breast samples and murine models of parity, provide evidence that lactation and involution modulate local breast immunity that results in increased quantity of resident CD8+ T cells in cancer unaffected normal breast tissue. In mice, these cells could restrain mammary tumor outgrowth, suggesting a protective mechanism by which lactation and involution can attenuate long term risk of BC development. Parity was also associated with increased tumor immune infiltrate in women who developed early stage TNBC, which has important implications for prognosis. Citation Format: Balaji Virassamy, Franco Caramia, Peter Savas, Michael A Harris, Jia-Wern Pan, Jianan Wang, Roberto Salgado, Heather Thorne, Soo-Hwang Teo, Jane Visvader, Paul J Neeson, Phillip K Darcy, Laura K Mackay, Sherene Loi. T cell mediated breast cancer protection following lactation and involution [abstract]. In: Proceedings of the San Antonio Breast Cancer Symposium 2024; 2024 Dec 10-13; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2025;31(12 Suppl):Abstract nr P1-01-02.

Overall survival in MOC by pattern of invasion.

Parity and breastfeeding reduce the risk of breast cancer, particularly triple-negative breast cancer (TNBC)1,2, yet the immunological mechanisms underlying this protection remain unclear. Here we show that parity is associated with increased numbers of CD8+ T cells, including cells with a tissue-resident-memory-like phenotype within human normal breast tissue. In mouse models, pregnancy followed by lactation and involution drove the accumulation of CD8+ T cells in the mammary gland, coinciding with reduced tumour growth and increased intratumoural immune cell infiltration, effects that were abrogated by CD8+ T cell depletion. Importantly, this CD8+ T-cell-dependent tumour control was observed only after a complete cycle of lactation and involution. Consistent with this, primary triple-negative breast cancers from parous women exhibited greater T cell infiltration and improved clinical outcomes. Together, these findings, spanning preclinical models and over 1,000 patient samples, provide insights into how reproductive history shapes breast immunity, positioning CD8+ T cells as key mediators of parity-associated protection and informing strategies for both the prevention and the treatment of breast cancer.

Supplementary Tables S1-S11

Neuroendocrine prostate cancer (NEPC) tumours are classified by pathology into several distinct subtypes. Gene expression profiling has revealed transcriptional heterogeneity across NEPC, but this is rarely considered in the context of variation between pathologies. Diagnosis typically relies on immunohistochemical markers (CHGA, SYP, NCAM1) and genomic alterations in RB1, PTEN and TP53. We hypothesized that NEPC pathologies have unique transcriptional features. Single‐cell RNA sequencing of 18 632 tumour cells from nine patient‐derived xenograft models representing five pathologies (small‐cell and large‐cell neuroendocrine carcinomas, focal neuroendocrine differentiation (Focal NED), low‐grade neuroendocrine and amphicrine) demonstrated pathway‐specific enrichment. Focal NED and amphicrine tumours exhibited cellular subpopulations enriched for KRAS, IL2‐STAT5 and TNF signalling pathways, absent in small‐ and large‐cell carcinomas, which were instead enriched for Myc and E2F pathways. Furthermore, focal NED cells exhibited minimal clonal divergence from adjacent adenocarcinoma cells, while small cell carcinoma cells were clonally distinct. These data underscore significant transcriptional variation among NEPC pathologies, highlighting focal NED's unique biological context and its clinical implications.

Patient characteristics and analytical cohorts.

Background Men with neuroendocrine prostate cancer (NEPC) have a poor prognosis. NEPC is commonly diagnosed by immunohistochemical markers (CHGA, SYP and NCAM1) and genomic features (mutations in RB1, PTEN, TP53). But by pathology, NEPC tumours are variable, leading to a classification of NE subtypes such as small cell and large cell neuroendocrine carcinomas, focal neuroendocrine differentiation (Focal NED), and Amphicrine. We postulated the diversity observed in NEPC pathologies might arise from differences in transcriptional profiles and the aim of this study is to utilize single-cell RNA sequencing to define the transcriptional differences between NEPC subtype pathologies. Methods Gene expression profiles were obtained for 18,632 individual tumour cells from 9 patient-derived xenograft (PDX) models representing five distinct neuroendocrine pathologies of prostate cancer. Integration and clustering of cell-level data demarked transcriptionally distinct sub-populations of cells. Differential gene expression, gene set enrichment and transcriptional factor regulon analysis identified expression signatures unique to specific neuroendocrine pathologies. Copy-number estimated from expression data revealed the clonal structure of PDXs with mixed adenocarcinoma and neuroendocrine pathologies. Results Significant differences were observed in the transcriptional profiles of NEPC pathology subtypes. Focal NED cells maintain AR signaling, similar to the amphicrine subtype but different from small and large cell carcinomas. Cellular sub-populations enriched for expression of KRAS, IL2-STAT5 and TNF-signaling genes were found in focal NED and amphicrine pathologies, but not in small or large cell carcinomas. In contrast, sub-populations enriched for the YAP, Myc and E2F pathways were detected in small cell, large cell and amphicrine tumours, but not in focal NED cells. Each pathology showed unique patterns of master regulator activity as well, further implicating focal NED as a transcriptionally distinct entity. Based on copy number alterations within PDXs of mixed pathology, focal NED cells showed little clonal divergence from neighboring adenocarcinoma cells, whereas cells with small cell neuroendocrine pathology were clonally distinct. Conclusions Neuroendocrine prostate cancer subtypes can be identified by pathology and our study shows that transcriptional features identified by single-cell RNA-sequencing also distinguish neuroendocrine subtypes pathologies from each other. In particular, our data redefine focal neuroendocrine differentiation as a pathology expressing androgen receptors (AR), exhibiting its distinctive composition of transcriptionally unique sub-populations. These findings advocate for differences in the treatment of NEPC tumors, particularly those displaying focal NED. ### Competing Interest Statement The authors have declared no competing interest.

While there is a great clinical need to understand the biology of metastatic cancer in order to treat it more effectively, research is hampered by limited sample availability. Research autopsy programmes can crucially advance the field through synchronous, extensive, and high-volume sample collection. However, it remains an underused strategy in translational research. Via an extensive questionnaire, we collected information on the study design, enrolment strategy, study conduct, sample and data management, and challenges and opportunities of research autopsy programmes in oncology worldwide. Fourteen programmes participated in this study. Eight programmes operated 24 h/7 days, resulting in a lower median postmortem interval (time between death and start of the autopsy, 4 h) compared with those operating during working hours (9 h). Most programmes (n = 10) succeeded in collecting all samples within a median of 12 h after death. A large number of tumour sites were sampled during each autopsy (median 15.5 per patient). The median number of samples collected per patient was 58, including different processing methods for tumour samples but also non-tumour tissues and liquid biopsies. Unique biological insights derived from these samples included metastatic progression, treatment resistance, disease heterogeneity, tumour dormancy, interactions with the tumour micro-environment, and tumour representation in liquid biopsies. Tumour patient-derived xenograft (PDX) or organoid (PDO) models were additionally established, allowing for drug discovery and treatment sensitivity assays. Apart from the opportunities and achievements, we also present the challenges related with postmortem sample collections and strategies to overcome them, based on the shared experience of these 14 programmes. Through this work, we hope to increase the transparency of postmortem tissue donation, to encourage and aid the creation of new programmes, and to foster collaborations on these unique sample collections. (c) 2024 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.