Introduction:With recent approvals of multiple targeted therapies for triple-negative breast cancer (TNBC), including antibody-drug conjugates and immunotherapy in biomarker-selected populations, it is critical to define the temporal evolution of cell-surface target expression from early-stage to metastatic disease, the co-expression patterns across these markers, and optimal quantification methodologies. Here we report biomarker expression profiles measured by multi-omics and pathology-based platforms in patients with TNBC using a large cohort of matched longitudinal tumor samples. Methods:Patients who underwent neoadjuvant chemotherapy (NAC) for stage I-III TNBC or were diagnosed with any stage TNBC and developed metastatic recurrence were retrospectively identified from an institutional database and prospective research metastatic biopsy protocol. Tumor samples from diagnosis (DX), residual disease (RD) post-NAC (if applicable), and metastasis/recurrence (MR) were collected. Quantification of HER2, TROP2, and PD-L1 expression was performed by immunohistochemistry (IHC), whole-exome sequencing, transcriptome sequencing, and targeted mass spectrometry (MS). For HER2, TROP2, and stromal tumor-infiltrating lymphocytes (sTILs), both manual pathologist assessment and computational pathology quantification were obtained. HER2 status was categorized as HER2-0 or HER2-low by local (L-IHC) and central (C-IHC) review, TROP2 status was defined as low (H-score <100), medium (H-score 100-200) or high (H-score >200), and PD-L1 as low (tumor area positivity, TAP <5%) or high (TAP ≥5%). Pathologist-assessed sTILs were classified as low (<10%), medium (≥10% and <40%) or high (≥40%). Biomarkers were compared between primary (DX/RD) and MR, and between pre- vs post-NAC (DX-RD) samples. Correlations between markers, quantification methods, inferred PAM50 subtype, and clinical variables of interest were evaluated. Results:A total of 359 samples from 110 patients with TNBC with data available from at least one platform were included in the analysis. HER2-low prevalence at DX, RD, and MR was: 51% (50/98), 40% (21/53), and 27% (16/60); TROP2 high/medium was 90% (47/52), 91% (42/46), and 88% (28/32); PD-L1-high was 51%, 50%, and 38% (9/24); and sTILs-high/medium was 88% (59/67), 80% (40/50), and 49% (17/35), respectively. While TROP2-high/medium vs low remained stable over time, HER2 IHC and sTILs significantly decreased from DX/RD to MR samples, both at the cohort-level (HER2, p=0.0081; sTILs, p=4.6x10e-5) and longitudinal patient-level (HER2, p=0.030; sTILs, p=0.0077), with a similar decreasing trend for PD-L1 that did not reach statistical significance. HER2 concordance (0 vs low) between L-IHC and C-IHC was 78% (91/116). ERBB2 , TACSTD2 and CD274 mRNA expression were significantly correlated with IHC protein levels, though only TACSTD2 had limited overlap in distribution of gene expression between high/medium vs low groups. Strong correlation between protein membrane staining intensity from computational pathology, protein expression measured by MS, and pathologist-assed IHC was observed across all biomarkers tested by each method. In comparisons between biomarkers, pathologist-assessed PD-L1 IHC and sTILs were significantly correlated (p=0.0001); 94% (51/54) of PD-L1-high tumors were classified as sTILs high/medium. PAM50 subtype was not significantly correlated with time point or biomarker status, although there was a trend toward more HER2-enriched tumors in HER2-low (20%, 5/25) vs HER2-0 (6%, 3/52) (p=0.086). Across biomarkers and clinical variables, an association between age and sTILs was observed (p=0.038, FDR=0.42) due to a decrease in sTILs high/medium tumors with age, primarily driven by post-treatment (RD/MR) but not DX samples. Conclusions:Multi-platform and multi-omics profiling in this large unique cohort of longitudinal TNBC samples revealed distinct patterns of expression and dynamic changes of key biomarkers of interest for targeted therapies. Given variability with manual IHC scoring, improved methods for quantification of expression may help optimize treatment selection in an individualized manner.

Utilizing generative models to create in-silico data presents an economical alternative to the traditional methods of staining, imaging, and annotating images in the computational pathology workflow. Specifically, appearance transfer diffusion models enable the generation of images without the need for model training, making the process swift and efficient. While originally developed for natural images, these models can transfer foreground objects from a source to a target domain, with less emphasis on the background. In computational pathology, however, every aspect of an image, including the background, can be crucial for understanding the tumor micro-environment. In this study, we adapted an appearance transfer diffusion model to align with the demands of computational pathology by adjusting the AdaIN feature statistics in the denoising process. The effectiveness of this modified method was demonstrated through its application to epithelium segmentation. The results demonstrated superior performance compared to the baseline approach, indicating that the number of manual annotations necessary for model training could be reduced by 75% without sacrificing accuracy. The authors expect that this research will promote the use of zero-shot diffusion models in computational pathology.

Abstract Current practice for evaluating prognosis of resectable lung adenocarcinoma (LUAD) patients relies on the Tumor, Node, Metastasis staging system. Despite low relapse risk, post-recurrence survival in a stage I population is poor with a median post relapse survival of around 25 months. There is a need for a biomarker that can enrich for stage I patients with relapse risk to identify a population likely to benefit from adjuvant therapy. In this study, we aim to identify a predictive biomarker of relapse in early-stage lung cancer following surgical resection using AI-based computational pathology. The retrospective patient cohort analyzed in this study consisted of a total of 166 patients that underwent surgical resection for a clinical stage I LUAD. Out of these, 54 patients experienced disease recurrence within 5 years, and 112 patients had a confirmed disease-free period of at least 5 years post-surgery. For each patient, at least one digitized slide (average = 2.46) stained with haematoxylin and eosin (H&E) was available for analysis. The project was accepted by the IUCPQ ethics committee (2022-3751, 22138). Manual annotations delineating the tumor core (TC) were drawn by pathologists. We then applied several proprietary image analysis models capable of distinguishing epithelium, stroma, and necrosis within the tumor tissue as well as detecting tumor infiltrating lymphocytes (TILs) and cell nuclei. In this project, we additionally introduce a novel approach rooted in mathematical graph theory to encoding topological aspects of the tissue morphology. A cell-graph constructed from the positions of cell nuclei was used to calculate cell-level graph-theoretic measures which were then aggregated to the slide-level by computing the median. In total, 23 data readouts were obtained of which 14 were based on the standard image analysis pipeline alone, and a further 9 originated from our newly developed graph-based approach. We identified promising biomarkers using the Wilcoxon rank-sum test and the area under the ROC curve (AUROC); robustness of this evaluation was investigating using resampling methods, using 100 bootstrap samples and 100 repeats of 3-fold cross-validation, respectively. This allowed us to evaluate the prognostic value of the biomarkers with respect to 5-year relapse or death. We were able to identify the best prognostic biomarker for relapse to be the median weighted clustering coefficient (CC) across all cells in the TC area, which is derived from the graph-based analyses. The median CC was able to achieve a whole-cohort AUROC of 0.695 (average cross-validated AUROC = 0.675, 95% interval 61.7%-70.4%) as well as a relapse vs relapse-free Wilcoxon test p-value of 4.65x10-5 (bootstrapped average = 0.0002). In addition, we were also able to demonstrate that the biomarker shows some relation to the architecture patterns. Citation Format: Florian J. Song, Alma Andoni, Manal Kordahi, Sara Batelli, Armin Meier, Markus Schick, Emilie Mahieu, Günter Schmidt, Claire E. Myers, Michael Abadier, Abjihit Dasgupta, Christopher Abbosh, Darren Hodgson, Michèle Orain, Fabien C. Lamaze, Yohan Bossé, Philippe Joubert. Identifying poor prognosis stage I lung adenocarcinoma patients through novel morphological biomarker based on computational pathology [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6176.

Abstract Background: Volrustomig (MEDI5752) is designed to fully inhibit PD-1 while preferentially inhibiting CTLA-4 on activated PD-1+ T cells. In a FTiH trial, volrustomig demonstrated meaningful clinical activity in patients with solid tumors (including 1L RCC and NSCLC), acceptable safety profile, sustained PD-1 receptor occupancy, and T cell proliferation at levels greater than seen with clinically tolerable doses of CTLA-4 inhibitors with co-administration of anti-PD(L)-1. We sought to further characterize volrustomig pharmacodynamic effect in the periphery and tumor to demonstrate its biological activity and inform dose optimization. Methods: Pre- and on-treatment blood samples from 86 patients enrolled in dose exploration were collected. Peripheral T cell profiling was assayed by flow cytometry, RNAseq, TCRSeq and serum proteomics. Paired tumor biopsies (N=17), pre- and at week 6 on-treatment, were analyzed for T cell infiltration and function using RNAseq and computational image analysis of immune cell marker density detected by histochemistry and multiplex immune fluorescence assays. Results: Volrustomig achieved dose-dependent peripheral pharmacodynamic changes of biomarkers specific to CTLA-4 blockade at ≥ 225mg. At C1D8 with doses ≥ 500mg, increases in CD4 T cell proliferation (ki67), activation (ICOS), and memory T cells (1516%, 321% and 88%, respectively) were seen, each greater than observed with historical controls of patients treated with tremelimumab 3mg/kg (T3) in combination with durvalumab. Similarly, CD8 proliferation and activation were also higher than T3 at C1D8 with 300% and 161% increase respectively. Additional CD4 and CD8 T cell proliferation and activation was observed at C2D8 supporting the benefit of repeat dosing. Higher peripheral T cell clonal expansion at doses ≥ 500mg was seen and associated with PFS and OS benefit. In patients with paired biopsies, volrustomig achieved robust T cell mediated anti-tumor activity including increases in T cell infiltration (CD8+, 2.3 fold, p=0.04), proliferation (CD8+ Ki67+, 2.1 fold, p=0.04) and, T cell cytotoxicity and activation (GrB+, 2.2 fold, p=0.01; T cell effector and IFNg gene signatures, > 2-fold, p< 0.01). Additionally, volrustomig increased CD8 proliferation in tumor compared to historical control durvalumab alone (>2-fold increase in 70% vs 50% patients). Conclusion: Voltrustomig demonstrated robust peripheral and intra-tumoral T cell activation and proliferation, broadening the pool of antigen experienced T cells, at levels greater than seen with dual checkpoint blockade currently used in the clinic. Citation Format: Ikbel Achour, Michael Kuziora, Florian Song, Steven Eck, Lina Meinecke, Jorge Blando, Jennifer Pearson, Megha Saraiya, Christian Eisen, Felix Segerer, Markus Schick, Nicolas Triltsch, Michael Surace, Gisele Silva Boos, Harald Hessel, Mai Bui, Katrin Pratsch, Alma Andoni, Shelby Denise Gainer, Dan Freeman, Deepa Subramaniam, Ben Tran. Volrustomig a novel PD-1/CTLA-4 bispecific antibody leads to robust increase in peripheral and intra-tumoral pharmacodynamic biomarkers in solid tumors from FTiH study [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 3627.

Background Small cell lung cancer (SCLC) is an aggressive and largely immune-cold cancer type, for which chemotherapy combined with Immuno-oncology (IO) therapies is providing benefit only in a subgroup of patients. SCLC is a highly heterogeneous cancer with at least four major subtypes.1 Among them, the 'inflamed' subtype is characterized by an inflamed immune gene signature and high expression of MHC class I (MHC-I) antigen presentation and shows the greatest benefit from the addition of IO treatment to chemotherapy,2 suggesting that MHC-I could serve as a biomarker for IO therapies. Here, we aimed to assess the spatial characteristics of immune cells in MHC-I high SCLC cases to investigate and support its role as a potential biomarker for IO therapies. Methods We combined a computational pathology approach with multiplex immunofluorescence (mIF) to profile the SCLC tumor microenvironment (TME). To this end, 126 SCLC formalin-fixed, paraffin-embedded tissue samples were stained with two mIF panels consisting of six markers each: (A) PanCK, CD8, CD68, PD-1, PD-L1, and Ki67; (B) CD20, NKp46, CD1c, CD66b, ICOS, and FOXP3. Based on these panels, we investigated the location and phenotype of each cell in the tumor center and within the stroma and tumor parenchyma. Additional slides from the same tissue blocks were immunohistochemically stained with MHC-I and scored by pathologists. Starting from the observation that high MHC-I expression was associated with higher densities of CD8+ T-cells,3 we further explored the TME characteristics of MHC-I SCLC cases. Results Beyond higher densities of CD8+ cytotoxic T-cells, we observed higher densities of FOXP3+ regulatory T-cells, and ICOS+ T-cells in the tumor center of MHC-I high cases. Considering the role of MHC-I in antigen presentation and T-cell activation, we investigated the proportion of CD8;PD-1;Ki67+ T-cells out of all CD8+ cells. Of note, we observed a compelling association of a high proportion of CD8;PD-1;Ki67+ T-cells with high MHC-I. This effect was particularly prominent in the tumor parenchyma and absent in the stroma, revealing an association with functionally relevant presentation of tumor antigens by MHC-I on SCLC tumor cells. Interestingly, we did not observe alterations in other immune cell populations like myeloid dendritic cells, macrophages, and granulocytes. Conclusions We utilized computational pathology to comprehensively profile the composition and spatial arrangement of the TME in inflamed SCLC cases defined by high MHC-I expression. Our findings provide the functional rationale for MHC-I as a biomarker for a potentially increased response to IO therapies.4 References Gay CM, Stewart CA, Park EM, Diao L, Groves SM, Heeke S, Nabet BY, Fujimoto J, Solis LM, Lu W, Xi Y, Cardnell RJ, Wang Q, Fabbri G, Cargill KR, Vokes NI, Ramkumar K, Zhang B, Della Corte CM, Robson P, Swisher SG, Roth JA, Glisson BS, Shames DS, Wistuba II, Wang J, Quaranta V, Minna J, Heymach JV, Byers LA. Patterns of transcription factor programs and immune pathway activation define four major subtypes of SCLC with distinct therapeutic vulnerabilities. Cancer Cell. 2021 Mar 8;39(3):346–360.e7. doi: 10.1016/j.ccell.2020.12.014. Epub 2021 Jan 21. PMID: 33482121; PMCID: PMC8143037. Mahadevan NR, Knelson EH, Wolff JO, Vajdi A, Saigí M, Campisi M, Hong D, Thai TC, Piel B, Han S, Reinhold BB, Duke-Cohan JS, Poitras MJ, Taus LJ, Lizotte PH, Portell A, Quadros V, Santucci AD, Murayama T, Cañadas I, Kitajima S, Akitsu A, Fridrikh M, Watanabe H, Reardon B, Gokhale PC, Paweletz CP, Awad MM, Van Allen EM, Lako A, Wang XT, Chen B, Hong F, Sholl LM, Tolstorukov MY, Pfaff K, Jänne PA, Gjini E, Edwards R, Rodig S, Reinherz EL, Oser MG, Barbie DA. Intrinsic Immunogenicity of Small Cell Lung Carcinoma Revealed by Its Cellular Plasticity. Cancer Discov. 2021 Aug;11(8):1952–1969. doi: 10.1158/2159–8290.CD-20–0913. Epub 2021 Mar 11. PMID: 33707236; PMCID: PMC8338750. Vuko M, Xie M, Gavaldon MA, Segerer F, Spitzmueller A, Hessel H, Testori M, Zimmermann J, Surace M, Heininen-Brown M, Canales JR, Saran S, Angell H, Schmidt G, Sade H, Barrett C, Schick M, Fabbri G. 155 MHC class I antigen presentation is associated with an inflamed SCLC tumor microenvironment characterized by a higher density of cytotoxic T-cells in closer proximity to tumor cells. Journal for ImmunoTherapy of Cancer 2022;10:doi: 10.1136/jitc-2022-SITC2022.0155 Rudin CM, Balli D, Lai WV, Richards AL, Nguyen E, Egger JV, Choudhury NJ, Sen T, Chow A, Poirier JT, Geese WJ, Hellmann MD, Forslund A. Clinical benefit from immunotherapy in patients with small cell lung cancer is associated with tumor capacity for antigen presentation. J Thorac Oncol. 2023 May 18:S1556–0864(23)00554–3. doi: 10.1016/j.jtho.2023.05.008. Epub ahead of print. PMID: 37210008. Ethics Approval All samples from which data in this report were generated, were obtained from an internal repository. All protocols, amendments, and participant informed consent documents were approved by the appropriate institutional review boards.

The IL-6–gp130–STAT3 signaling axis is a major regulator of inflammation. Activating mutations in the gene encoding gp130 and germline gain-of-function mutations in STAT3 (STAT3 GOF ) are associated with multi-organ autoimmunity, severe morbidity, and adverse prognosis. To dissect crucial cellular subsets and disease biology involved in activated gp130 signaling, the gp130-JAK-STAT3 axis was constitutively activated using a transgene, L-gp130 , specifically targeted to T cells. Activating gp130 signaling in T cells in vivo resulted in fatal, early onset, multi-organ autoimmunity in mice that resembled human STAT3 GOF disease. Female mice had more rapid disease progression than male mice. On a cellular level, gp130 signaling induced the activation and effector cell differentiation of T cells, promoted the expansion of T helper type 17 (T H 17) cells, and impaired the activity of regulatory T cells. Transcriptomic profiling of CD4 + and CD8 + T cells from these mice revealed commonly dysregulated genes and a gene signature that, when applied to human transcriptomic data, improved the segregation of patients with transcriptionally diverse STAT3 GOF mutations from healthy controls. The findings demonstrate that increased gp130-STAT3 signaling leads to T H 17-driven autoimmunity that phenotypically resembles human STAT3 GOF disease.

Discovery of explainable biomarkers is a complex process which is typically driven by a-priori hypothesis and expert annotations. This contribution introduces an almost entirely annotation and hypothesis free workflow to discover predictive biomarkers derived from cell phenotypes. It relies on self-supervised learning, clustering and survival analysis of cell centric image patches. The workflow is successfully evaluated on mIF images of 2L+ mNSCLC samples from a clinical study (NCT01693562). Two potential biomarkers are identified that closely align with the known relevant biology.

Abstract Background: Anti-PD-L1 therapy has demonstrated clinical activity in patients with metastatic non-small cell lung cancer (mNSCLC). However, only subgroups of patients respond and their identification via PD-L1 as a biomarker remains imperfect. PD-L1 expression is commonly assessed by pathologist tumor cell (TC) scoring of immunohistochemically (IHC) stained tissue. We developed a system for digitally scoring PD-L1 in IHC (PD-L1 QCS), which demonstrated robust scoring across studies [1]. Here, we present a comparison of PD-L1 QCS against manual scoring of PD-L1 (SP263 assay, Ventana) in the MYSTIC clinical trial. Methods: PD-L1 QCS on digitized whole slide images (WSI) comprises two deep learning models, enabling segmentation of single TCs followed by PD-L1 expression quantification via their optical density (OD). Positive cells are classified based on an OD threshold, allowing robust digital calculation of the TC percentage [1]. The analysis included 502 WSI from the MYSTIC trial (NCT02453282), representing 256 patients treated with anti-PD-L1 therapy and 246 treated with chemotherapy as standard-of-care (SoC) [2]. First, an optimal cut-point was determined by optimization against outcome, classifying patients with ≥0.575% TC as biomarker positive (BM+). Next, the approach was compared against manual TC scoring at 1%, 25% and 50% cut-off. Results: In durvalumab treated patients, median overall survival (mOS) in the PD-L1 QCS BM+ subgroup (prevalence 54.3%) was 12.1 months longer than in the BM- subgroup (19.9m vs. 7.8m, HR=0.45, CI [0.33, 0.60]). Analogous comparison of subgroups based on manually scored TC proportion at 1% (prev. 75.0%), 25% (prev. 42.6%) or 50% (prev. 29.7%) cut-points yielded a mOS difference of 7.8m (HR=0.52, CI [0.38, 0.72]), 8.3m (HR=0.61, CI [0.45, 0.82]) and 11.0m (HR=0.55, CI [0.40, 0.77]) respectively. Comparing durvalumab treatment against SoC within the PD-L1 QCS BM+ subgroup yielded a HR of 0.62 (CI [0.46, 0.82], log rank p=0.0008, 7.4m mOS delta). In comparison, a HR of 0.69 (CI [0.46,1.02], p=0.0642, 8.4m mOS delta) was obtained for manual TC scoring at 50%. Conclusion: We compared a computational pathology approach for continuous PD-L1 scoring for the selection of mNSCLC patients for anti-PD-L1 treatment against established manual scoring. Our results suggest that PD-L1 QCS has the potential to identify a larger patient subgroup that retains benefit from anti-PD-L1 treatment and more precisely identifies non-responders. References: 1. Lesniak, Jan, et al. "Quantitative computational assessment of PD-L1 enables robust patient selection for biomarker-informed anti-PD-L1 treatment of NSCLC patients." J. Immunother. Cancer, Vol. 10., 2022. 2. Rizvi NA, et al. "Durvalumab with or without tremelimumab vs standard chemotherapy in first-line treatment of mNSCLC: the MYSTIC phase 3 randomized clinical trial." JAMA Oncology. 2020;6.5:661-674. Citation Format: Jan Martin Lesniak, Markus Schick, Thomas Kunzke, Federico Pollastri, Juan Pedro Vigueras-Guillén, Harald Hessel, Susanne Haneder, Pallavi Sontakke, Karma DaCosta, Regina Alleze, Hadassah Sade, J Carl Barrett, Günter Schmidt, Ross Stewart. Enhanced patient selection for anti-PD-L1 treatment in metastatic NSCLC with quantitative continuous scoring of PD-L1 [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 2492.

Abstract Background: The NeoCOAST study (NCT03794544) investigated the efficacy of neoadjuvant immuno-oncology combinations in patients with resectable non-small-cell lung cancer (NSCLC), using major pathological response (MPR) as the primary endpoint (Cascone T, et al. Cancer Discov 2023;13:2394-411). Patients received a single cycle of the following treatments: durvalumab monotherapy (durva; anti-PD-L1), durva + oleclumab (ole; anti-CD73), or durva + monalizumab (mona; anti-NKG2A). Here we analyze the tumor microenvironment to investigate the mechanism of action of ole. Methods: A cohort of 47 patients (17 durva, 13 durva + mona, 17 durva + ole) with tumor tissue samples from pre-treatment (baseline, n=31) and surgery (post-treatment, n=24; n=8 paired pre- and post-treatment samples) were evaluated. Tumor sections were stained with (1) multiplex immunofluorescence (Panel 1: CD8-PD-L1-PD1-CD68-Ki67-panCK, Panel 2: NKp46-CD20), (2) CD73 immunohistochemistry (IHC), and (3) NKG2A IHC. Stains were analyzed using deep learning algorithms, manual pathology, and digital pathology scoring, respectively. Results: All arms showed similar infiltration of tumors by T cells (CD8), B cells (CD20), NK cells (NKp46), and macrophages (CD68) at the pre-treatment time point (Kruskal-Wallis, p=0.5-0.9 n.s.). Percent CD73+ tumor cells (TCs) at baseline was correlated with increased B cell, T cell, NK cell, and macrophage abundance (Spearman’s Rho=0.4-0.6, p<0.05). Percent CD73+ TC and B cell abundance was higher in patients with MPR vs those without MPR in the durva + ole arm, but not in the durva monotherapy or durva + mona arms. Increased CD8 T cell and NK cell abundance was not associated with MPR in any arm. Moreover, comparison of baseline to post-treatment tumor samples revealed a greater increase in CD8 T cells in the durva + ole arm (FC=2.1, p=0.02), particularly in proliferating activated T cells (CD8+Ki67+PD-1+; FC=8.0, p=0.003), compared to durva alone (FC=1.8, p=0.4 and FC=2.9, p=0.04, respectively). Conclusion: Increased abundance of CD73+ TCs is correlated with high B cells in the tumor microenvironment and associated with patients who experienced MPR in the durva + ole arm. Treatment with neoadjuvant durva + ole results in a greater increase in proliferating CD8 T cells than durva alone. Citation Format: Ina Bisha, Tze Heng Tan, Manuela Weitkunat, Alma Andoni, Iris Dino, Philip Martin, Megha Saraiya, Karma DaCosta, Pallavi Sontakke, Markus Schick, Michael Surace, Jorge Blando, Italia Grenga, Rakesh Kumar, Lara McGrath. Multiplex immunofluorescence profiling of the tumor microenvironment and CD73: Activity of neoadjuvant oleclumab in patients with non-small-cell lung cancer in NeoCOAST [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 2023.

Abstract Background We explored potential predictive biomarkers of immunotherapy response in patients with extensive-stage small-cell lung cancer (ES-SCLC) treated with durvalumab (D) + tremelimumab (T) + etoposide-platinum (EP), D + EP, or EP in the randomized phase 3 CASPIAN trial. Methods 805 treatment-naïve patients with ES-SCLC were randomized (1:1:1) to receive D + T + EP, D + EP, or EP. The primary endpoint was overall survival (OS). Patients were required to provide an archived tumor tissue block (or ≥ 15 newly cut unstained slides) at screening, if these samples existed. After assessment for programmed cell death ligand-1 expression and tissue tumor mutational burden, residual tissue was used for additional molecular profiling including by RNA sequencing and immunohistochemistry. Results In 182 patients with transcriptional molecular subtyping, OS with D ± T + EP was numerically highest in the SCLC-inflamed subtype (n = 10, median 24.0 months). Patients derived benefit from immunotherapy across subtypes; thus, additional biomarkers were investigated. OS benefit with D ± T + EP versus EP was greater with high versus low CD8A expression/CD8 cell density by immunohistochemistry, but with no additional benefit with D + T + EP versus D + EP. OS benefit with D + T + EP versus D + EP was associated with high expression of CD4 (median 25.9 vs. 11.4 months) and antigen-presenting and processing machinery (25.9 vs. 14.6 months) and MHC I and II (23.6 vs. 17.3 months) gene signatures, and with higher MHC I expression by immunohistochemistry. Conclusions These findings demonstrate the tumor microenvironment is important in mediating better outcomes with D ± T + EP in ES-SCLC, with canonical immune markers associated with hypothesized immunotherapy mechanisms of action defining patient subsets that respond to D ± T. Trial registration ClinicalTrials.gov, NCT03043872.

Abstract While targeted cancer therapies often rely on subjective and semi-quantitative visual assessment of protein biomarkers by pathologists through immunohistochemically stained tissue, the transformative force of computational pathology is reshaping healthcare by unleashing unprecedented diagnostic accuracy and unlocking personalized treatments. In recent years, we have established a large integrated computational pathology unit to foster collaboration between interdisciplinary teams of computer scientists, pathologists, molecular biologists, and data scientists. This integration of cutting-edge technologies enabled us to develop a computational pathology approach called Quantitative Continuous Scoring (QCS). QCS deploys the power of Deep Learning (DL) to provide objective and continuous expression data of biomarkers in digitized IHC whole slide images (WSI), particularly of proteins expressed at low levels. While manual scoring of IHC WSIs is limited by subjectivity and semi-quantitative assessment of protein expression, QCS overcomes these limitations with an unprecedented accuracy. QCS utilizes two DL-based algorithms, which we developed fully supervised by using pathologist input as the reference standard. These algorithms identify invasive tumour areas and segment each tumour cell across the WSI into cell nuclei, cytoplasm and membrane. Based on an accurate subcellular segmentation, we can compute biomarker expression, on a continuous scale, as mean Optical Density (OD) in each subcellular compartment based on the Hue-Saturation-Density (HSD) model. Therefore, this approach enables precise detection of the low biomarker expression range with single-cell resolution. Importantly, it also allows the computation of the spatial distribution of tumour cells across the WSI. We have successfully used QCS to drive the selection of antibody clones for IHC assays and to delineate the mode of action and PK/PD mechanisms. Of note, the combination of assessing continuous target expression and capturing the spatial distribution of tumor cells has provided surrogate markers to predict potential bystander activity of antibody drug conjugates (ADCs). This approach outperformed traditional pathologist scoring in identifying patient populations having maximum treatment benefit through retrospective analysis of multiple clinical trials. At present, all computational pathology approaches are developed based on conventional IHC assays that have been optimized for manual scoring. Importantly, we draw a vision in which the assay serves as a critical catalyst for unleashing the full potential of computational pathology by providing high-quality, standardized data inputs. We suggest an approach utilizing orthogonal methods as a reference standard to develop highly sensitive IHC assays, capable of detecting even subtle molecular and cellular changes with precision and exhibiting exceptional specificity for accurately identifying and distinguishing target biomarkers from background noise. In summary, we here describe and discuss a computational pathology-based approach for precise biomarker quantification and superior patient selection with broad applicability and the potential to transform the very fabric of how we diagnose and treat cancer. Citation Format: Mark Gustavson, Markus Schick, Ansh Kapil, Anatoliy Shumilov, Carl Barrett, Hadassah Sade. Computational pathology: Revolutionizing diagnostics and clearing the way for precision medicine [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO4-26-01.

The creation of in-silico datasets can expand the utility of existing annotations to new domains with different staining patterns in computational pathology. As such, it has the potential to significantly lower the cost associated with building large and pixel precise datasets needed to train supervised deep learning models. We propose a novel approach for the generation of in-silico immunohistochemistry (IHC) images by disentangling morphology specific IHC stains into separate image channels in immunofluorescence (IF) images. The proposed approach qualitatively and quantitatively outperforms baseline methods as proven by training nucleus segmentation models on the created in-silico datasets.

Many targeted cancer therapies rely on biomarkers assessed by scoring of immunohistochemically (IHC)-stained tissue, which is subjective, semiquantitative, and does not account for expression heterogeneity. We describe an image analysis-based method for quantitative continuous scoring (QCS) of digital whole-slide images acquired from baseline human epidermal growth factor receptor 2 (HER2) IHC-stained breast cancer tissue. Candidate signatures for patient stratification using QCS of HER2 expression on subcellular compartments were identified, addressing the spatial distribution of tumor cells and tumor-infiltrating lymphocytes. Using data from trastuzumab deruxtecan-treated patients with HER2-positive and HER2-negative breast cancer from a phase 1 study (NCT02564900; DS8201-A-J101; N = 151), QCS-based patient stratification showed longer progression-free survival (14.8 vs 8.6 months) with higher prevalence of patient selection (76.4 vs 56.9%) and a better cross-validated log-rank p value (0.026 vs 0.26) than manual scoring based on the American Society of Clinical Oncology / College of American Pathologists guidelines. QCS-based features enriched the HER2-negative subgroup by correctly predicting 20 of 26 responders.

Aberrant activity of the SUMOylation pathway has been associated with MYC overexpression and poor prognosis in aggressive B-cell lymphoma (BCL) and other malignancies. Recently developed small-molecule inhibitors of SUMOylation (SUMOi) target the heterodimeric E1 SUMO activation complex (SAE1/UBA2). Here, we report that activated MYC signaling is an actionable molecular vulnerability in vitro and in a preclinical murine in vivo model of MYC-driven BCL. While SUMOi conferred direct effects on MYC-driven lymphoma cells, SUMO inhibition also resulted in substantial remodeling of various subsets of the innate and specific immunity in vivo. Specifically, SUMOi increased the number of memory B cells as well as cytotoxic and memory T cells, subsets that are attributed a key role within a coordinated anti-tumor immune response. In summary, our data constitute pharmacologic SUMOi as a powerful therapy in a subset of BCL causing massive remodeling of the normal B-cell and T-cell compartment.

The DNA damage response (DDR) acts as a barrier to malignant transformation and is often impaired during tumorigenesis. Exploiting the impaired DDR can be a promising therapeutic strategy; however, the mechanisms of inactivation and corresponding biomarkers are incompletely understood. Starting from an unbiased screening approach, we identified the SMC5-SMC6 Complex Localization Factor 2 (SLF2) as a regulator of the DDR and biomarker for a B-cell lymphoma (BCL) patient subgroup with an adverse prognosis. SLF2-deficiency leads to loss of DDR factors including Claspin (CLSPN) and consequently impairs CHK1 activation. In line with this mechanism, genetic deletion of Slf2 drives lymphomagenesis in vivo. Tumor cells lacking SLF2 are characterized by a high level of DNA damage, which leads to alterations of the post-translational SUMOylation pathway as a safeguard. The resulting co-dependency confers synthetic lethality to a clinically applicable SUMOylation inhibitor (SUMOi), and inhibitors of the DDR pathway act highly synergistic with SUMOi. Together, our results identify SLF2 as a DDR regulator and reveal co-targeting of the DDR and SUMOylation as a promising strategy for treating aggressive lymphoma.

Proteasome inhibition is a highly effective treatment for multiple myeloma (MM). However, virtually all patients develop proteasome inhibitor resistance, which is associated with a poor prognosis. Hyperactive small ubiquitin-like modifier (SUMO) signaling is involved in both cancer pathogenesis and cancer progression. A state of increased SUMOylation has been associated with aggressive cancer biology. We found that relapsed/refractory MM is characterized by a SUMO-high state, and high expression of the SUMO E1-activating enzyme (SAE1/UBA2) is associated with poor overall survival. Consistently, continuous treatment of MM cell lines with carfilzomib (CFZ) enhanced SUMO pathway activity. Treatment of MM cell lines with the SUMO E1-activating enzyme inhibitor subasumstat (TAK-981) showed synergy with CFZ in both CFZ-sensitive and CFZ-resistant MM cell lines, irrespective of the TP53 state. Combination therapy was effective in primary MM cells and in 2 murine MM xenograft models. Mechanistically, combination treatment with subasumstat and CFZ enhanced genotoxic and proteotoxic stress, and induced apoptosis was associated with activity of the prolyl isomerase PIN1. In summary, our findings reveal activated SUMOylation as a therapeutic target in MM and point to combined SUMO/proteasome inhibition as a novel and potent strategy for the treatment of proteasome inhibitor-resistant MM.

Background Exploratory data from single-arm PhI/II clinical trials provide a unique opportunity for predictive biomarker discovery, but have the limitation to estimate the prognostic value in absence of a control arm. This poses a significant risk to subsequent biomarker-driven PhIII trials. A Digital Twin is the computational description a real patient based on their clinical data,1 including baseline histopathology imaging. In this work, we demonstrate the generation of a virtual randomized PhIII clinical trial (vPhIII) via Digital Twins and their application for predictive biomarker discovery. Methods Eligible baseline tissue samples of NSCLC patients enrolled in the PhI/II trial 'CP1108' (NCT01693562) and in the randomized PhIII trial 'MYSTIC' (NCT02453282)4 5 were considered for computational pathology analysis. Relying on Quantitative Continuous Scoring for PD-L1 (PD-L1 QCS), the PD-L1 protein expression was digitally quantified and multidimensional whole slide image (WSI) features for PD-L1 expression were obtained.2 3 For each PhI/II patient, its Digital Twin is generated by identifying the most similar matching patient in the PhIII control arm using Euclidean distance and Delaunay triangulation of dimensionality-reduced patient features. This group of Digital Twins comprises the actual observed overall survival (OS) information from the respective study, and therefore could serve as vPhIII. Each PD-L1 QCS feature is evaluated for its median OS time (mOS) benefit in the QCS-positive vPhIII sub-group. The feature providing longest mOS benefit is selected and evaluated for its predictive value validated in the real PhIII cohort. Results N=121 Digital Twins were generated as vPhIII cohort (figure 1a), while average mOS benefit analysis for each feature indicated that the 20% quantile of PD-L1 tumor cell expression provides optimal stratification (figure 1b). The mOS comparison of the real PhIII (figure 2a) with the vPhIII (figure 2b) showed the OS benefit from durvalumab treatment has been underestimated by the Digital Twins model. Although the selected QCS feature did not indicate significant treatment benefit in the vPhIII (figure 3a), a retrospective analysis of MYSTIC hints towards beneficial patient stratification (figure 3b). Conclusions Digital Twins based on imaging data are a promising approach to generate virtual randomized and biomarker stratified PhIII trials based on single-arm PhI/II and historic Standard-of-Care data. This proof-of-concept study demonstrates technical feasibility of this innovative methodology. Although further validation is required, the Digital Twin approach may open new ways towards maximizing the success probability of drug development programs, and faster implementation of Precision Oncology in clinical routine. Acknowledgements The authors thank Harald Hessel, Susanne Haneder, Pallavi Sontakke, Karma DaCosta and Regina Alleze (Pathology and Pathology Informatics at AstraZeneca Computational Pathology) for curation and annotation of whole slide image data used in this study. Trial Registration NCT01693562, NCT02453282 References Chengyue Wu, Guillermo Lorenzo, et al. Integrating mechanism-based modeling with biomedical imaging to build practical digital twins for clinical oncology. Biophys Rev (Melville). 2022 Jun;3(2):021304 Schmidt G, Brieu N, Spitzmueller A, Kapil A. A SCORING METHOD FOR AN ANTI-HER2 ANTIBODY-DRUG CONJUGATE THERAPY. Patent application WO2022054009 (A2) Jan Lesniak, Markus Schick, Ross Stewart, et al. Quantitative computational assessment of PD-L1 enables robust patient selection for biomarker-informed anti-PD-L1 treatment of NSCLC patients. Journal for ImmunoTherapy of Cancer 2022;10:doi: 10.1136/jitc-2022-SITC2022.0583 Scott J. Antonia, Neil H. Segal, et al. Clinical Activity, Tolerability, and Long-Term Follow-Up of Durvalumab in Patients With Advanced NSCLC. J Thorac Oncol. 2019 Oct;14(10):1794−1806 Pages 1794–1806,Rizvi NA, Cho BC, Reinmuth N, et al. Durvalumab With or Without Tremelimumab vs Standard Chemotherapy in First-line Treatment of Metastatic Non-Small Cell Lung Cancer: The MYSTIC Phase 3 Randomized Clinical Trial. JAMA Oncol. 2020 May 1;6(5):661−674 Ethics Approval Clinical studies NCT01693562 and NCT02453282, from which data in this report were obtained, were carried out in accordance with the Declaration of Helsinki and GoodClinical Practice guidelines. The study protocols, amendments, and participant informed consent documents were approved by the appropriate institutional review boards.

Evasion from drug-induced apoptosis is a crucial mechanism of cancer treatment resistance. The pro-apoptotic protein NOXA marks an aggressive pancreatic ductal adenocarcinoma (PDAC) subtype. To identify drugs that unleash the death-inducing potential of NOXA, we performed an unbiased drug screening experiment. In NOXA-deficient isogenic cellular models we identified an inhibitor of the transcription factor heterodimer CBFβ/RUNX1. By genetic gain and loss of function experiments we validated that the mode of action depends on RUNX1 and NOXA. Of note, RUNX1 expression is significantly higher in PDACs compared to normal pancreas. We show that pharmacological RUNX1 inhibition significantly blocks tumor growth in vivo and in primary patient-derived PDAC organoids. Through genome wide analysis, we detected that RUNX1-loss reshapes the epigenetic landscape, which gains H3K27ac enrichment at the NOXA promoter. Our study demonstrates a previously unknown mechanism of NOXA-dependent cell death, which can be triggered pharmaceutically. Therefore, our data show a novel way to target a therapy resistant PDAC, an unmet clinical need. Significance Recent evidence demonstrated the existence of molecular subtypes in pancreatic ductal adenocarcinoma (PDAC), which resist all current therapies. The paucity of therapeutic options, including a complete lack of targeted therapies, underscore the urgent and unmet medical need for the identification of targets and novel treatment strategies for PDAC. Our study unravels a function of the transcription factor RUNX1 in apoptosis regulation in PDAC. We show that pharmacological RUNX1 inhibition in PDAC is feasible and leads to NOXA-dependent apoptosis. The development of targeted therapies that influence the transcriptional landscape of PDAC might have great benefits for patients who are resistant to conventional therapies. RUNX1 Inhibition as a new therapeutic intervention offers an attractive strategy for future therapies.