Table S3: Specific Dose-Limiting Toxicities and corresponding patient numbers

Abstract Antibody Drug Conjugates (ADCs) are proving to be groundbreaking therapeutics for all breast cancer subtypes. Sacituzumab govitecan (SG), an ADC targeting the TROP2 tumor antigen and harboring a topoisomerase 1 inhibitor payload, improves overall survival in advanced Hormone Receptor positive (HR+) and Triple Negative Breast Cancer (TNBC). However, SG has not been tested in early-stage breast cancer, and response predictors and mechanisms of resistance to SG in any setting remain to be defined. We thus designed the NeoSTAR trial, a Phase 2 study of response-guided neoadjuvant SG for localized TNBC, in which 50 patients received single-agent SG followed by imaging and tumor bed biopsy, with subsequent therapy determined by clinical and pathologic response. We collected fresh pre-treatment and post-SG biopsies and carried out analysis employing single-cell RNAseq, exome sequencing, and spatial analysis. As anticipated, cell composition of TNBCs between subjects was highly heterogenous but revealed non-tumor cell types predictive of pathologic complete response (pCR) to SG alone. As in the metastatic setting, tumor cell TROP2 levels were variable both within and between TNBCs, and were not predictive of response to SG. In contrast, we identified tumor cell-intrinsic pathways predictive of response that were the same as those identified in parallel genome-wide CRISPR screens for SG response pathways in TNBC. Analysis of matched pre/post treatment specimens demonstrated substantial clonal selection and clonal evolution post SG. Of particular interest were small subpopulations of tumor cells with shared properties that were present in multiple tumors in very small numbers pre-treatment but dramatically expanded post-treatment, suggesting the presence of a common resistance phenotype. Ongoing work to be presented focuses on defining the nature of the shared resistant subpopulations, and on comprehensively assessing the genomic correlates of response to SG versus standard chemotherapy. In summary, detailed single-cell and genomic analysis of single-agent ADC therapy in treatment-naïve primary TNBC reveals how intratumoral heterogeneity and subclonal resistance phenotypes shape the landscape of treatment response. These observations provide new insights relevant to the clinical application of this complex class of therapeutics. Citation Format: Laura M. Spring, Bogang Wu, Ting Liu, Jacob Geisberg, Simona Cristea, Veerle Bossuyt, Rachel Occhiogrosso Abelman, Nicole Peiris, James Coates, Siang-Boon Koh, Mengran Zhang, Lianne Ryan, Beverly Moy, Steven J. Isakoff, Sara M. Tolaney, Franziska Michor, Aditya Bardia, Leif W. Ellisen. Intratumoral heterogeneity drives resistance to Antibody Drug Conjugate therapy: Analysis of the NeoSTAR trial of neoadjuvant Sacituzumab govitecan for localized TNBC [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Advances in Breast Cancer Research; 2023 Oct 19-22; San Diego, California. Philadelphia (PA): AACR; Cancer Res 2024;84(3 Suppl_1):Abstract nr PR08.

Purpose: The antibody-drug conjugate (ADC) sacituzumab govitecan (SG) comprises the topoisomerase 1 (TOP1) inhibitor (TOP1i) SN-38, coupled to a monoclonal antibody targeting trophoblast cell surface antigen 2 (TROP-2). Poly(ADP-ribose) polymerase (PARP) inhibition may synergize with TOP1i and SG, but previous studies combining systemic PARP and TOP1 inhibitors failed due to dose-limiting myelosuppression. Here, we assess the proof-of-mechanism and clinical feasibility for SG and talazoparib (TZP) employing an innovative sequential dosing schedule.Patients and Methods: In vitro models tested pharmacodynamic endpoints, and in a phase 1b clinical trial (NCT04039230), 30 patients with metastatic triple-negative breast cancer (mTNBC) received SG and TZP in a concurrent (N = 7) or sequential (N = 23) schedule. Outcome measures included safety, tolerability, preliminary efficacy, and establishment of a recommended phase 2 dose.Results: We hypothesized that tumor-selective delivery of TOP1i via SG would reduce nontumor toxicity and create a temporal window, enabling sequential dosing of SG and PARP inhibition. In vitro, sequential SG followed by TZP delayed TOP1 cleavage complex clearance, increased DNA damage, and promoted apoptosis. In the clinical trial, sequential SG/TZP successfully met primary objectives and demonstrated median progression-free survival (PFS) of 7.6 months without dose-limiting toxicities (DLT), while concurrent dosing yielded 2.3 months PFS and multiple DLTs including severe myelosuppression.Conclusions: While SG dosed concurrently with TZP is not tolerated clinically due to an insufficient therapeutic window, sequential dosing of SG followed by TZP proved a viable strategy. These findings support further clinical development of the combination and suggest that ADC-based therapy may facilitate novel, mechanism-based dosing strategies.

Figure S1, radiographic assessment of key lesions of patient MGH-18; Figure S2, supplementary genomic analysis for patient MGH-18; Figure S3, supporting data for mechanism of TROP2 T256R; Figure S4, TROP2 immunohistochemistry for metastatic lesions of MGH-18; Table S1, detailed cohort characteristics; Table S2, patient treatment histories; Table S3, supplementary mutational data for MGH-18; Table S4, antibodies used.

Antibody drug conjugates (ADC) are emerging as paradigm-changing precision therapeutics. Sacituzumab Govitecan (SG), which combines a Topoisomerase I (TOP1) inhibitor payload (SN38) with hRS7, an antibody targeting the tumor-selective antigen TROP2, is proving to be a successful cancer therapeutic for some highly refractory cancers including triple negative breast cancer (TNBC). However, mechanisms and biomarkers of sensitivity and resistance to SG remain poorly understood. Prior clinical trials have identified positive correlations between TROP2 expression and better outcome for metastatic TNBC patients receiving SG. Additionally, patient-derived specimen analyses by our group identified mutations in both TROP2 and TOP1 in post-progression metastatic lesions, suggesting multifaceted resistance mechanisms. To systematically interrogate ADC sensitivity mechanisms, we utilized genome-wide CRISPR-Cas9 knockout screening to identify novel regulators mediating SG sensitivity in human TNBC cells. Interactome analysis of top SG sensitizing hits showed clustered pathways of interest, including DNA repair/replication, mTORC1/metabolism, and endosome trafficking/sorting. The top druggable hit sensitizing to SG was PARP1, encoding the poly (ADP-ribose) polymerase 1 that is a key enzyme required for single-strand break repair pathways and for DNA replication fork stability. Conversely, the very top SG resistance-inducing hit in the genome was PARG, encoding the glycohydrolase that opposes PARP enzymatic activity. Furthermore, the synergistic lethality between TOP1 and PARP1 was confirmed in experiments combining SG with PARP inhibitor treatment of TNBC cell lines. In addition, to identify SG/ADC specific genes and pathways, we performed a secondary round of CRISPR screens using a custom library comprising top hits discovered in the primary screen, treating cells in parallel with either SG or cytotoxic payload SN38 alone. Pathway analysis of genes selectively modulating sensitivity to SG but not SN38 revealed novel mediators involved in TROP2 turnover and recycling. We then performed validation and mechanistic studies to elucidate how specific genetic permutations regulate antibody delivery and sensitivity to SG. In conclusion, through a systematic approach using CRISPR screening, we identified several novel resistance and sensitizing pathways that are implicated in ADC delivery and target protein turnover. We are currently prioritizing the druggable genes and pathways in preparation for proof-of-concept studies combining ADCs with select targeted therapeutics to achieve greater tumor killing efficacy. Citation Format: Bogang Wu, Sheng Sun, Nayana Thimmiah, Aiko Nagayama, James Coates, Win Thant, David Li, John Doench, Aditya Bardia, Leif Ellisen. Systematic identification of pathways associated with antibody drug conjugate sensitivity in breast cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 1696.

Triple-Negative Breast Cancer (TNBC) is an aggressive breast cancer subset, which lacks expression of estrogen receptors, progesterone receptors, and human epidermal growth factor receptor-2. This subset of breast cancer disproportionately affects Black and African American women and improving TNBC treatment options is vital to reducing breast cancer mortality. The novel antibody-drug conjugate, Sacitizumab Govetican (SG), which targets TROP2 at the cell surface, has shown promising clinical results from the NeoSTAR trial (NCT04230109), a phase II study evaluating neoadjuvant SG therapy in a localized TNBC setting. As part of the NeoSTAR clinical trial, we have collected and processed pre- and post- SG treatment patient samples, with the aim of understanding response to this monotherapy. Herein, we identify biomarkers of response and resistance to SG monotherapy through the use of single cell RNA sequencing of matched pre- and post-treatment patient biopsies combined with exome sequencing these patient samples. Pre-treatment core needle biopsy samples, and if applicable, post-treatment residual disease biopsies were dissociated into single cell suspensions and subjected to single cell RNA sequencing. Additionally, fixed patient tissue samples were processed accordingly for exome sequencing analyses. Overall, we analyzed over 144,000 cells from 37 total scRNA seq libraries with an average of 3800 cells per biopsy sample, demonstrating the feasibility of this method. From these analyses, we observed several cell-type differences between patients who achieved a pathological complete response (pCR) and patients who had residual disease (RD). Specifically, our data shows that tumor infiltrating lymphocytes are a potential prognostic biomarker of response to SG. Furthermore, we detected alterations in among stromal and immune cell subsets, among non-responders, indicating that these cell types maybe indicative of SG resistance. Taken together, we outline biomarkers of response to SG treatment for an improved understanding of resistance mechanisms in the neoadjuvant setting to improve TNBC outcomes among patients. Citation Format: Nicole Peiris, Simona Cristea, Mengran Zhang, Siang Koh, James Coates, Ilze Smidt, Veerle Bossuyt, Laura Spring, Aditya Bardia, Leif Ellisen. Utilizing scRNA sequencing to understand biomarkers of response and resistance to Sacitizumab Govetican in localized TNBC. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 4353.

Supplementary Material, including Supplementary Methods describing the Random Effects Model. Supplementary Results showing example quantitative T1/T2 maps from the rat tumour model (Supplementary Figure S1), the standard curve for making quantitative protein concentration measurements using Coomassie staining (Supplementary Figure S2), the result of a conventional MTRasym APT analysis approach in the rat tumour model (Supplementary Figure S3), and the histological validation that areas of high protein concentration correlate with areas of increased hypoxia, increased vessel area and higher cellularity in the rat tumour model (Supplementary Figure S4). Supplementary Table S1 shows the mean {plus minus} standard deviation of T1 and T2 relaxation times measured in different tissue ROIs from both the rat tumour and mouse tumour models.

Artificial intelligence (AI)-based technologies for healthcare are being developed increasingly day-byday. While they are often used interchangeably, machine learning (ML) encompasses a subset of techniques within the broader field of AI often being more complex and having more abstruse physical interpretations (Figure 1). An increasingly recognised necessity for the robust and successful implementation of ML-based platforms in healthcare is relevant, high-quality data around which such platforms can be trained, tested and trusted (Figure 2(a) and (b)). Centralised healthcare systems such as the United Kingdom’s National Health Service (NHS) are therefore uniquely positioned to exploit the advantages of ML-based tools given their population-wide accessibility, standardisations of data collection and regulatory governance structures. Application of ML to solve and optimise challenges in systems such as the NHS is as far from a mature concept as it is novel; however, the systemic and methodological impact of the use of such technologies in critical care decision making, as well as pursuant legal ramifications, is only now coming to light and is worryingly opaque. Future policy choices relating to ML-driven healthcare technologies will be predicated on those made today as it is unarguably a period of genesis for the field. Shortcomings in governance, legislation, ethical codes, liability structures and policies relating to implementation risk deterring private enterprises from engaging. This also renders the

Background: Sacituzumab Govitecan (SG), the first antibody-drug conjugate approved for metastatic TNBC (mTNBC), is comprised of SN-38 (active metabolite of irinotecan), a topoisomerase I (TOP1) inhibitor, coupled via a hydrolyzable linker to monoclonal antibody targeting trophoblast cell surface antigen 2 (Trop-2), an antigen overexpressed in mTNBC. Poly (ADP-ribose) polymerase inhibitors (PARPi) block resolution of TOP1 cleavage complexes (TOP1CCs) induced by TOP1 inhibitors, thus unmasking the inability of remaining pathways to repair DNA damage. However, previous clinical trials combining PARPi with standard TOP1 inhibitors (irinotecan, topotecan) were terminated early due to dose-limiting myelosuppression. We evaluated the combination of SG with PARP inhibitor in both pre-clinical models and phase 1b clinical trial. Methods and Results: In pre-clinical models we demonstrated that the targeted antibody-based delivery of SN-38 increased the ratio of tumor-to-normal cell SN-38, resulting in stabilized TOP1CCs, enhanced DNA damage and increased cytotoxicity with the combination, selectively in tumor cells but not normal cells, despite temporal separation of SG and PARPi exposure. To validate the hypothesis, we conducted a phase 1b investigator-initiated clinical trial combining SG with PARPi (talazoparib) in patients with mTNBC (NCT04039230). Inclusion criteria included female patients ≥ 18 years of age with mTNBC (per ASCO/CAP guidelines) and previous treatment with at least one prior therapeutic regimen for mTNBC. Clinical outcomes were assessed by Objective Response Rate per RECIST v1.1. In the phase 1b clinical trial (SG day 18, every 21 days with talazoparib), the staggered schedule with supportive therapy was relatively well-tolerated without DLTs, as predicted by the pre-clinical models. Furthermore, the staggered schedule demonstrated promising clinical activity. Molecular analysis of paired pre-treatment and on-treatment specimens demonstrated γ-H2AX accumulation, confirming pharmacodynamic inhibition with combination therapy. The dose-escalation portion of clinical trial successfully completed enrollment with a recommended phase-2 dose (R2PD) of sequential SG (10 mg/kg on days 1,8) with talazoparib (1 mg on days 15-21), every 21 days. Conclusion: Staggered dosing of SG and PARPi, leveraging the selective drug delivery mechanism of SG to minimize toxicity while maintaining efficacy, was feasible and demonstrated encouraging evidence of clinical activity with objective responses among patients with mTNBC. The translational study highlights how mechanistic insights and innovative scheduling could be utilized to develop promising drug combinations, including previously rejected combinations, for patients with mTNBC. Citation Format: Aditya Bardia, James T. Coates, Laura Spring, Sheng Sun, Dejan Juric, Nayana Thimmiah, Andrzej Niemierko, Phoebe Ryan, Ann Partridge, Jeffrey Peppercorn, Heather Parsons, Seth Wander, Kelsey Pierce, Victoria Attaya, Donna Fitzgerald, Brenda Lormil, Maria Shellock, Aiko Nagayama, Veerle Bossuyt, Bev Moy, Sara Tolaney, Leif Ellisen. Sacituzumab Govitecan, combination with PARP inhibitor, Talazoparib, in metastatic triple-negative breast cancer (TNBC): Translational investigation [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 2638.

Abstract Sacituzumab govitecan (SG), the first antibody–drug conjugate (ADC) approved for triple-negative breast cancer, incorporates the anti-TROP2 antibody hRS7 conjugated to a topoisomerase-1 (TOP1) inhibitor payload. We sought to identify mechanisms of SG resistance through RNA and whole-exome sequencing of pretreatment and postprogression specimens. One patient exhibiting de novo progression lacked TROP2 expression, in contrast to robust TROP2 expression and focal genomic amplification of TACSTD2/TROP2 observed in a patient with a deep, prolonged response to SG. Analysis of acquired genomic resistance in this case revealed one phylogenetic branch harboring a canonical TOP1E418K resistance mutation and subsequent frameshift TOP1 mutation, whereas a distinct branch exhibited a novel TACSTD2/TROP2T256R missense mutation. Reconstitution experiments demonstrated that TROP2T256R confers SG resistance via defective plasma membrane localization and reduced cell-surface binding by hRS7. These findings highlight parallel genomic alterations in both antibody and payload targets associated with resistance to SG. Significance: These findings underscore TROP2 as a response determinant and reveal acquired SG resistance mechanisms involving the direct antibody and drug payload targets in distinct metastatic subclones of an individual patient. This study highlights the specificity of SG and illustrates how such mechanisms will inform therapeutic strategies to overcome ADC resistance. This article is highlighted in the In This Issue feature, p. 2355

Platinum chemotherapies are highly effective cytotoxic agents but often induce resistance when used as monotherapies. Combinatorial strategies limit this risk and provide effective treatment options for many cancers. Here, we repurpose atovaquone (ATQ), a well-tolerated & FDA-approved anti-malarial agent by demonstrating that it potentiates cancer cell death of a subset of platinums. We show that ATQ in combination with carboplatin or cisplatin induces striking and repeatable concentration- and time-dependent cell death sensitization in vitro across a variety of cancer cell lines. ATQ induces mitochondrial reactive oxygen species (mROS), depleting intracellular glutathione (GSH) pools in a concentration-dependent manner. The superoxide dismutase mimetic MnTBAP rescues ATQ-induced mROS production and pre-loading cells with the GSH prodrug N-acetyl cysteine (NAC) abrogates the sensitization. Together, these findings implicate ATQ-induced oxidative stress as key mediator of the sensitizing effect. At physiologically achievable concentrations, ATQ and carboplatin furthermore synergistically delay the growth of three-dimensional avascular spheroids. Clinically, ATQ is a safe and specific inhibitor of the electron transport chain (ETC) and is concurrently being repurposed as a candidate tumor hypoxia modifier. Together, these findings suggest that ATQ is deserving of further study as a candidate platinum sensitizing agent.

Machine learning (ML) provides a broad framework for addressing high-dimensional prediction problems in classification and regression. While ML is often applied for imaging problems in medical physics, there are many efforts to apply these principles to biological data toward questions of radiation biology. Here, we provide a review of radiogenomics modeling frameworks and efforts toward genomically guided radiotherapy. We first discuss medical oncology efforts to develop precision biomarkers. We next discuss similar efforts to create clinical assays for normal tissue or tumor radiosensitivity. We then discuss modeling frameworks for radiosensitivity and the evolution of ML to create predictive models for radiogenomics.

The radiotherapy process is a series of events during which discrepancies between the planned treatment and actual treatment delivered can occur.This necessitates a comprehensive quality assurance (QA) programme, including regular quality control (QC) checks and audits.As more advanced technology is introduced in the clinical setting, QA activities must continually evolve to provide a safe framework for implementation of technical radiotherapy.With image guided and adaptive strategies being increasingly employed to ensure accurate delivery of treatment in scenarios such as dose escalation and hypofractionation; techniques must be implemented in a safe and effective manner.QA in the clinical trial arena has played a leading role in striving for accuracy and consistency of radiotherapy treatment delivery through monitoring protocol compliance in a multi-centre setting.Clinical trials can also evaluate the feasibility and effectiveness of a new technology.A comprehensive trial QA programme not only accredits centres for recruitment to a trial but also benefits the general standard of radiotherapy delivered.This presentation will aim to demonstrate how we can extend the clinical trial QA experience to routine practice to ensure quality of image guidance through discussing examples of clinical trial benchmarking and credentialing processes and their perceived impacts on clinical practice.

Abstract Abnormal pH is a common feature of malignant tumors and has been associated clinically with suboptimal outcomes. Amide proton transfer magnetic resonance imaging (APT MRI) holds promise as a means to noninvasively measure tumor pH, yet multiple factors collectively make quantification of tumor pH from APT MRI data challenging. The purpose of this study was to improve our understanding of the biophysical sources of altered APT MRI signals in tumors. Combining in vivo APT MRI measurements with ex vivo histological measurements of protein concentration in a rat model of brain metastasis, we determined that the proportion of APT MRI signal originating from changes in protein concentration was approximately 66%, with the remaining 34% originating from changes in tumor pH. In a mouse model of hypopharyngeal squamous cell carcinoma (FaDu), APT MRI showed that a reduction in tumor hypoxia was associated with a shift in tumor pH. The results of this study extend our understanding of APT MRI data and may enable the use of APT MRI to infer the pH of individual patients' tumors as either a biomarker for therapy stratification or as a measure of therapeutic response in clinical settings. Significance: These findings advance our understanding of amide proton transfer magnetic resonance imaging (APT MRI) of tumors and may improve the interpretation of APT MRI in clinical settings.

54Advances in patient-specific biological information and biotechnology have contributed to a new era of computational biomedicine. Different -omics (genomics, transcriptomics, proteomics, metabolomics) have emerged as valuable resource of data for modeling outcomes and complementing current clinical and imaging information. Here, we provide an overview of these of -omics with specific focus on radiogenomics. We highlight the current status and potential of this data in outcome modeling.

Tumour hypoxia is a well-recognised barrier to anti-cancer therapy and represents one of the best validated targets in oncology. Previous attempts to tackle hypoxia have focussed primarily on increasing tumour oxygen supply; however, clinical studies using this approach have yielded only modest clinical benefit, with often significant toxicity and practical limitations. Therefore, there are currently no anti-hypoxia treatments in widespread clinical use. As an emerging alternative strategy, we discuss the relevance of inhibiting tumour oxygen metabolism to alleviate hypoxia and highlight recently initiated clinical trials using this approach.

Advances in patient-specific information and biotechnology have contributed to a new era of computational medicine. Radiogenomics has emerged as a new field that investigates the role of genetics in treatment response to radiation therapy. Radiation oncology is currently attempting to embrace these recent advances and add to its rich history by maintaining its prominent role as a quantitative leader in oncologic response modeling. Here, we provide an overview of radiogenomics starting with genotyping, data aggregation, and application of different modeling approaches based on modifying traditional radiobiological methods or application of advanced machine learning techniques. We highlight the current status and potential for this new field to reshape the landscape of outcome modeling in radiotherapy and drive future advances in computational oncology.

Radiation therapy is a first-line treatment option for localized prostate cancer and radiation-induced normal tissue damage are often the main limiting factor for modern radiotherapy regimens. Conversely, under-dosing of target volumes in an attempt to spare adjacent healthy tissues limits the likelihood of achieving local, long-term control. Thus, the ability to generate personalized data-driven risk profiles for radiotherapy outcomes would provide valuable prognostic information to help guide both clinicians and patients alike. Big data applied to radiation oncology promises to deliver better understanding of outcomes by harvesting and integrating heterogeneous data types, including patient-specific clinical parameters, treatment-related dose-volume metrics, and biological risk factors. When taken together, such variables make up the basis for a multi-dimensional space (the "RadoncSpace") in which the presented modeling techniques search in order to identify significant predictors. Herein, we review outcome modeling and big data-mining techniques for both tumor control and radiotherapy-induced normal tissue effects. We apply many of the presented modeling approaches onto a cohort of hypofractionated prostate cancer patients taking into account different data types and a large heterogeneous mix of physical and biological parameters. Cross-validation techniques are also reviewed for the refinement of the proposed framework architecture and checking individual model performance. We conclude by considering advanced modeling techniques that borrow concepts from big data analytics, such as machine learning and artificial intelligence, before discussing the potential future impact of systems radiobiology approaches.