Combined Transcriptome and Circulating Tumor DNA Longitudinal Biomarker Analysis Associates With Clinical Outcomes in Advanced Solid Tumors Treated With Pembrolizumab.

PURPOSE: Immune gene expression signatures are emerging as potential biomarkers for immunotherapy (IO). VIGex is a 12-gene expression classifier developed in both nCounter (Nanostring) and RNA sequencing (RNA-seq) assays and analytically validated across laboratories. VIGex classifies tumor samples into hot, intermediate-cold (I-Cold), and cold subgroups. VIGex-Hot has been associated with better IO treatment outcomes. Here, we investigated the performance of VIGex and other IO biomarkers in an independent data set of patients treated with pembrolizumab in the INSPIRE phase II clinical trial (ClinicalTrials.gov identifier: NCT02644369). MATERIALS AND METHODS: Patients with advanced solid tumors were treated with pembrolizumab 200 mg IV once every 3 weeks. Tumor RNA-seq data from baseline tumor samples were classified by the VIGex algorithm. Circulating tumor DNA (ctDNA) was measured at baseline and start of cycle 3 using the bespoke Signatera assay. VIGex-Hot was compared with VIGex I-Cold + Cold and four groups were defined on the basis of the combination of VIGex subgroups and the change in ctDNA at cycle 3 from baseline (ΔctDNA). RESULTS: Seventy-six patients were enrolled, including 16 ovarian, 12 breast, 12 head and neck cancers, 10 melanoma, and 26 other tumor types. Objective response rate was 24% in VIGex-Hot and 10% in I-Cold/Cold. VIGex-Hot subgroup was associated with higher overall survival (OS) and progression-free survival (PFS) when included in a multivariable model adjusted for tumor type, tumor mutation burden, and PD-L1 immunohistochemistry. The addition of ΔctDNA improved the predictive performance of the baseline VIGex classification for both OS and PFS. CONCLUSION: Our data indicate that the addition of ΔctDNA to baseline VIGex may refine prediction for IO.

A T2-weighted MRI-based radiomic signature for disease-free survival in locally advanced cervical cancer following chemoradiation: An international, multicentre study.

INTRODUCTION: To develop and validate a T2-weighted magnetic resonance imaging (MRI)-based radiomic signature associated with disease-free survival (DFS) in locally advanced cervical cancer. MATERIALS AND METHODS: The study comprised a training dataset of 132 patients (93 Norwegian; 39 The Cancer Imaging Archive (TCIA) and an independent validation Canadian dataset of 199 patients with FIGO stage IB-IVA cervical cancer treated with chemoradiation. Radiomic features were extracted using PyRadiomics. A radiomic signature was developed based on a multivariable radiomic prognostic model for DFS built using the training dataset, with minimal redundancy maximum relevancy feature selection method. Univariate and multivariable Cox regression analyses were then conducted to examine the association of the derived radiomic signature with DFS. RESULTS: A radiomic signature was prognostic for DFS in the training cohort (Norwegian hazard ratio [HR] 5.54, p = 0.002; TCIA HR 3.59, p = 0.04). The radiomic signature remained independently associated with DFS (HR 3.70, p = 0.004) when adjusted for stage and tumor volume. The radiomic signature was also prognostic for DFS in the validation cohort, both on univariate analysis (HR 2.22, p = 0.003), and multivariable analysis adjusted for stage and tumor volume (HR 1.84, p = 0.04). The 4-year DFS rates of patients with radiomic signature score > 0 vs ≤ 0 were 48.2 % vs 87.9 %, and 56.4 % vs 80.8 % for training and validation cohorts respectively. CONCLUSION: An MRI-based radiomic signature can be used as a prognostic biomarker for DFS in patients with locally advanced cervical cancer undergoing chemoradiation.

A Phase II, Open-Label, Randomized Trial of Durvalumab With Olaparib or Cediranib in Patients With Mismatch Repair-Proficient Colorectal or Pancreatic Cancer.

BACKGROUND: The use of immunotherapy in mismatch repair proficient colorectal cancer (pMMR-CRC) or pancreatic adenocarcinoma (PDAC) is associated with limited efficacy. DAPPER (NCT03851614) is a phase 2, basket study randomizing patients with pMMR CRC or PDAC to durvalumab with olaparib (durvalumab + olaparib) or durvalumab with cediranib (durvalumab + cediranib). METHODS: PDAC or pMMR-CRC patients were randomized to either durvalumab+olaparib (arm A), or durvalumab + cediranib (arm B). Co-primary endpoints included pharmacodynamic immune changes in the tumor microenvironment (TME) and safety. Objective response rate, progression-free survival (PFS) and overall survival (OS) were determined. Paired tumor samples were analyzed by multiplexed immunohistochemistry and RNA-sequencing. RESULTS: A total of 31 metastatic pMMR-CRC patients were randomized to arm A (n = 16) or B (n = 15). In 28 evaluable patients, 3 patients had stable disease (SD) (2 patients treated with durvalumab + olaparib and 1 patient treated with durvalumab + cediranib) while 25 had progressive disease (PD). Among patients with PDAC (n = 19), 9 patients were randomized to arm A and 10 patients were randomized to arm B. In 18 evaluable patients, 1 patient had a partial response (unconfirmed) with durvalumab + cediranib, 1 patient had SD with durvalumab + olaparib while 16 had PD. Safety profile was manageable and no grade 4-5 treatment-related adverse events were observed in either arm A or B. No significant changes were observed for CD3+/CD8+ immune infiltration in on-treatment biopsies as compared to baseline for pMMR-CRC and PDAC independent of treatment arms. Increased tumor-infiltrating lymphocytes at baseline, low baseline CD68+ cells and different immune gene expression signatures at baseline were associated with outcomes. CONCLUSIONS: In patients with pMMR-CRC or PDAC, durvalumab + olaparib and durvalumab + cediranib showed limited antitumor activity. Different immune components of the TME were associated with treatment outcomes.

A dynamic microRNA profile that tracks a chemotherapy resistance phenotype in osteosarcoma. Implications for novel therapeutics.

Osteosarcoma is a rare primary bone tumor for which no significant therapeutic advancement has been made since the late 1980s despite ongoing efforts. Overall, the five-year survival rate remains about 65%, and is much lower in patients with tumors unresponsive to methotrexate, doxorubicin, and cisplatin therapy. Genetic studies have not revealed actionable drug targets, but our group, and others, have reported that epigenomic biomarkers, including regulatory RNAs, may be useful prognostic tools for osteosarcoma. We tested if microRNA (miRNA) transcriptional patterns mark the transition from a chemotherapy sensitive to resistant tumor phenotype. Small RNA sequencing was performed using 14 patient matched pre-chemotherapy biopsy and post-chemotherapy resection high-grade osteosarcoma frozen tumor samples. Independently, small RNA sequencing was performed using 14 patient matched biopsy and resection samples from untreated tumors. Separately, miRNA specific Illumina DASL arrays were used to assay an independent cohort of 65 pre-chemotherapy biopsy and 26 patient matched post-chemotherapy resection formalin fixed paraffin embedded (FFPE) tumor samples. mRNA specific Illumina DASL arrays were used to profile 37 pre-chemotherapy biopsy and five post-chemotherapy resection FFPE samples, all of which were also used for Illumina DASL miRNA profiling. The National Cancer Institute Therapeutically Applicable Research to Generate Effective Treatments dataset, including PCR based miRNA profiling and RNA-seq data for 86 and 93 pre-chemotherapy tumor samples, respectively, was also used. Paired differential expression testing revealed a profile of 17 miRNAs with significantly different transcriptional levels following chemotherapy. Genes targeted by the miRNAs were differentially expressed following chemotherapy, suggesting the miRNAs may regulate transcriptional networks. Finally, an in vitro pharmacogenomic screen using miRNAs and their target transcripts predicted response to a set of candidate small molecule therapeutics which potentially reverse the chemotherapy resistance phenotype and synergize with chemotherapy in otherwise treatment resistant tumors. Importantly, these novel therapeutic targets are distinct from targets identified by a similar pharmacogenomic analysis of previously published prognostic miRNA profiles from pre chemotherapy biopsy specimens.

A data science roadmap for open science organizations engaged in early-stage drug discovery.

The Structural Genomics Consortium is an international open science research organization with a focus on accelerating early-stage drug discovery, namely hit discovery and optimization. We, as many others, believe that artificial intelligence (AI) is poised to be a main accelerator in the field. The question is then how to best benefit from recent advances in AI and how to generate, format and disseminate data to enable future breakthroughs in AI-guided drug discovery. We present here the recommendations of a working group composed of experts from both the public and private sectors. Robust data management requires precise ontologies and standardized vocabulary while a centralized database architecture across laboratories facilitates data integration into high-value datasets. Lab automation and opening electronic lab notebooks to data mining push the boundaries of data sharing and data modeling. Important considerations for building robust machine-learning models include transparent and reproducible data processing, choosing the most relevant data representation, defining the right training and test sets, and estimating prediction uncertainty. Beyond data-sharing, cloud-based computing can be harnessed to build and disseminate machine-learning models. Important vectors of acceleration for hit and chemical probe discovery will be (1) the real-time integration of experimental data generation and modeling workflows within design-make-test-analyze (DMTA) cycles openly, and at scale and (2) the adoption of a mindset where data scientists and experimentalists work as a unified team, and where data science is incorporated into the experimental design.

Radiomic-Based Approaches in the Multi-metastatic Setting: A Quantitative Review.

Background: Radiomics traditionally focuses on analyzing a single lesion within a patient to extract tumor characteristics, yet this process may overlook inter-lesion heterogeneity, particularly in the multi-metastatic setting. There is currently no established method for combining radiomic features in such settings, leading to diverse approaches with varying strengths and limitations. Our quantitative review aims to illuminate these methodologies, assess their replicability, and guide future research toward establishing best practices, offering insights into the challenges of multi-lesion radiomic analysis across diverse datasets. Methods: We conducted a comprehensive literature search to identify methods for integrating data from multiple lesions in radiomic analyses. We replicated these methods using either the author's code or by reconstructing them based on the information provided in the papers. Subsequently, we applied these identified methods to three distinct datasets, each depicting a different metastatic scenario. Results: We compared ten mathematical methods for combining radiomic features across three distinct datasets, encompassing a total of 16,850 lesions in 3,930 patients. Performance of these methods was evaluated using the Cox proportional hazards model and benchmarked against univariable analysis of total tumor volume. We observed variable performance in methods across datasets. However, no single method consistently outperformed others across all datasets. Notably, while some methods surpassed total tumor volume analysis in certain datasets, others did not. Averaging methods showed higher median performance in patients with colorectal liver metastases, and in soft tissue sarcoma, concatenation of radiomic features from different lesions exhibited the highest median performance among tested methods. Conclusions: Radiomic features can be effectively selected or combined to estimate patient-level outcomes in multi-metastatic patients, though the approach varies by metastatic setting. Our study fills a critical gap in radiomics research by examining the challenges of radiomic-based analysis in this setting. Through a comprehensive review and rigorous testing of different methods across diverse datasets representing unique metastatic scenarios, we provide valuable insights into effective radiomic analysis strategies.

Radiomic-based prediction of lesion-specific systemic treatment response in metastatic disease.

Despite sharing the same histologic classification, individual tumors in multi metastatic patients may present with different characteristics and varying sensitivities to anticancer therapies. In this study, we investigate the utility of radiomic biomarkers for prediction of lesion-specific treatment resistance in multi metastatic leiomyosarcoma patients. Using a dataset of n=202 lung metastases (LM) from n=80 patients with 1648 pre-treatment computed tomography (CT) radiomics features and LM progression determined from follow-up CT, we developed a radiomic model to predict the progression of each lesion. Repeat experiments assessed the relative predictive performance across LM volume groups. Lesion-specific radiomic models indicate up to a 4.5-fold increase in predictive capacity compared with a no-skill classifier, with an area under the precision-recall curve of 0.70 for the most precise model (FDR = 0.05). Precision varied by administered drug and LM volume. The effect of LM volume was controlled by removing radiomic features at a volume-correlation coefficient threshold of 0.20. Predicting lesion-specific responses using radiomic features represents a novel strategy by which to assess treatment response that acknowledges biological diversity within metastatic subclones, which could facilitate management strategies involving selective ablation of resistant clones in the setting of systemic therapy.

Generalizability of deep learning in organ-at-risk segmentation: A transfer learning study in cervical brachytherapy.

PURPOSE: Deep learning can automate delineation in radiation therapy, reducing time and variability. Yet, its efficacy varies across different institutions, scanners, or settings, emphasizing the need for adaptable and robust models in clinical environments. Our study demonstrates the effectiveness of the transfer learning (TL) approach in enhancing the generalizability of deep learning models for auto-segmentation of organs-at-risk (OARs) in cervical brachytherapy. METHODS: A pre-trained model was developed using 120 scans with ring and tandem applicator on a 3T magnetic resonance (MR) scanner (RT3). Four OARs were segmented and evaluated. Segmentation performance was evaluated by Volumetric Dice Similarity Coefficient (vDSC), 95 % Hausdorff Distance (HD95), surface DSC, and Added Path Length (APL). The model was fine-tuned on three out-of-distribution target groups. Pre- and post-TL outcomes, and influence of number of fine-tuning scans, were compared. A model trained with one group (Single) and a model trained with all four groups (Mixed) were evaluated on both seen and unseen data distributions. RESULTS: TL enhanced segmentation accuracy across target groups, matching the pre-trained model's performance. The first five fine-tuning scans led to the most noticeable improvements, with performance plateauing with more data. TL outperformed training-from-scratch given the same training data. The Mixed model performed similarly to the Single model on RT3 scans but demonstrated superior performance on unseen data. CONCLUSIONS: TL can improve a model's generalizability for OAR segmentation in MR-guided cervical brachytherapy, requiring less fine-tuning data and reduced training time. These results provide a foundation for developing adaptable models to accommodate clinical settings.

RADCURE: An open-source head and neck cancer CT dataset for clinical radiation therapy insights.

PURPOSE: This manuscript presents RADCURE, one of the most extensive head and neck cancer (HNC) imaging datasets accessible to the public. Initially collected for clinical radiation therapy (RT) treatment planning, this dataset has been retrospectively reconstructed for use in imaging research. ACQUISITION AND VALIDATION METHODS: RADCURE encompasses data from 3346 patients, featuring computed tomography (CT) RT simulation images with corresponding target and organ-at-risk contours. These CT scans were collected using systems from three different manufacturers. Standard clinical imaging protocols were followed, and contours were manually generated and reviewed at weekly RT quality assurance rounds. RADCURE imaging and structure set data was extracted from our institution's radiation treatment planning and oncology information systems using a custom-built data mining and processing system. Furthermore, images were linked to our clinical anthology of outcomes data for each patient and includes demographic, clinical and treatment information based on the 7th edition TNM staging system (Tumor-Node-Metastasis Classification System of Malignant Tumors). The median patient age is 63, with the final dataset including 80% males. Half of the cohort is diagnosed with oropharyngeal cancer, while laryngeal, nasopharyngeal, and hypopharyngeal cancers account for 25%, 12%, and 5% of cases, respectively. The median duration of follow-up is five years, with 60% of the cohort surviving until the last follow-up point. DATA FORMAT AND USAGE NOTES: The dataset provides images and contours in DICOM CT and RT-STRUCT formats, respectively. We have standardized the nomenclature for individual contours-such as the gross primary tumor, gross nodal volumes, and 19 organs-at-risk-to enhance the RT-STRUCT files' utility. Accompanying demographic, clinical, and treatment data are supplied in a comma-separated values (CSV) file format. This comprehensive dataset is publicly accessible via The Cancer Imaging Archive. POTENTIAL APPLICATIONS: RADCURE's amalgamation of imaging, clinical, demographic, and treatment data renders it an invaluable resource for a broad spectrum of radiomics image analysis research endeavors. Researchers can utilize this dataset to advance routine clinical procedures using machine learning or artificial intelligence, to identify new non-invasive biomarkers, or to forge prognostic models.

Clinically impactful metabolic subtypes of pancreatic ductal adenocarcinoma (PDAC).

Background: Pancreatic ductal adenocarcinoma (PDAC) is a lethal disease characterized by a diverse tumor microenvironment. The heterogeneous cellular composition of PDAC makes it challenging to study molecular features of tumor cells using extracts from bulk tumor. The metabolic features in tumor cells from clinical samples are poorly understood, and their impact on clinical outcomes are unknown. Our objective was to identify the metabolic features in the tumor compartment that are most clinically impactful. Methods: A computational deconvolution approach using the DeMixT algorithm was applied to bulk RNASeq data from The Cancer Genome Atlas to determine the proportion of each gene's expression that was attributable to the tumor compartment. A machine learning algorithm designed to identify features most closely associated with survival outcomes was used to identify the most clinically impactful metabolic genes. Results: Two metabolic subtypes (M1 and M2) were identified, based on the pattern of expression of the 26 most important metabolic genes. The M2 phenotype had a significantly worse survival, which was replicated in three external PDAC cohorts. This PDAC subtype was characterized by net glycogen catabolism, accelerated glycolysis, and increased proliferation and cellular migration. Single cell data demonstrated substantial intercellular heterogeneity in the metabolic features that typified this aggressive phenotype. Conclusion: By focusing on features within the tumor compartment, two novel and clinically impactful metabolic subtypes of PDAC were identified. Our study emphasizes the challenges of defining tumor phenotypes in the face of the significant intratumoral heterogeneity that typifies PDAC. Further studies are required to understand the microenvironmental factors that drive the appearance of the metabolic features characteristic of the aggressive M2 PDAC phenotype.

Differential DNA damage repair and PARP inhibitor vulnerability of the mammary epithelial lineages.

It is widely assumed that all normal somatic cells can equally perform homologous recombination (HR) and non-homologous end joining in the DNA damage response (DDR). Here, we show that the DDR in normal mammary gland inherently depends on the epithelial cell lineage identity. Bioinformatics, post-irradiation DNA damage repair kinetics, and clonogenic assays demonstrated luminal lineage exhibiting a more pronounced DDR and HR repair compared to the basal lineage. Consequently, basal progenitors were far more sensitive to poly(ADP-ribose) polymerase inhibitors (PARPis) in both mouse and human mammary epithelium. Furthermore, PARPi sensitivity of murine and human breast cancer cell lines as well as patient-derived xenografts correlated with their molecular resemblance to the mammary progenitor lineages. Thus, mammary epithelial cells are intrinsically divergent in their DNA damage repair capacity and PARPi vulnerability, potentially influencing the clinical utility of this targeted therapy.

Radiomic-Based Prediction of Lesion-Specific Systemic Treatment Response in Metastatic Disease.

Despite sharing the same histologic classification, individual tumors in multi metastatic patients may present with different characteristics and varying sensitivities to anticancer therapies. In this study, we investigate the utility of radiomic biomarkers for prediction of lesion-specific treatment resistance in multi metastatic leiomyosarcoma patients. Using a dataset of n=202 lung metastases (LM) from n=80 patients with 1648 pre-treatment computed tomography (CT) radiomics features and LM progression determined from follow-up CT, we developed a radiomic model to predict the progression of each lesion. Repeat experiments assessed the relative predictive performance across LM volume groups. Lesion-specific radiomic models indicate up to a 5-fold increase in predictive capacity compared with a no-skill classifier, with an area under the precision-recall curve of 0.79 for the most precise model (FDR = 0.01). Precision varied by administered drug and LM volume. The effect of LM volume was controlled by removing radiomic features at a volume-correlation coefficient threshold of 0.20. Predicting lesion-specific responses using radiomic features represents a novel strategy by which to assess treatment response that acknowledges biological diversity within metastatic subclones, which could facilitate management strategies involving selective ablation of resistant clones in the setting of systemic therapy.

Spatial transformation of multi-omics data unlocks novel insights into cancer biology.

The application of next-generation sequencing (NGS) has transformed cancer research. As costs have decreased, NGS has increasingly been applied to generate multiple layers of molecular data from the same samples, covering genomics, transcriptomics, and methylomics. Integrating these types of multi-omics data in a combined analysis is now becoming a common issue with no obvious solution, often handled on an ad hoc basis, with multi-omics data arriving in a tabular format and analyzed using computationally intensive statistical methods. These methods particularly ignore the spatial orientation of the genome and often apply stringent p-value corrections that likely result in the loss of true positive associations. Here, we present GENIUS (GEnome traNsformatIon and spatial representation of mUltiomicS data), a framework for integrating multi-omics data using deep learning models developed for advanced image analysis. The GENIUS framework is able to transform multi-omics data into images with genes displayed as spatially connected pixels and successfully extract relevant information with respect to the desired output. We demonstrate the utility of GENIUS by applying the framework to multi-omics datasets from the Cancer Genome Atlas. Our results are focused on predicting the development of metastatic cancer from primary tumors, and demonstrate how through model inference, we are able to extract the genes which are driving the model prediction and are likely associated with metastatic disease progression. We anticipate our framework to be a starting point and strong proof of concept for multi-omics data transformation and analysis without the need for statistical correction.

Decoding biological age from face photographs using deep learning.

Because humans age at different rates, a person's physical appearance may yield insights into their biological age and physiological health more reliably than their chronological age. In medicine, however, appearance is incorporated into medical judgments in a subjective and non-standardized fashion. In this study, we developed and validated FaceAge, a deep learning system to estimate biological age from easily obtainable and low-cost face photographs. FaceAge was trained on data from 58,851 healthy individuals, and clinical utility was evaluated on data from 6,196 patients with cancer diagnoses from two institutions in the United States and The Netherlands. To assess the prognostic relevance of FaceAge estimation, we performed Kaplan Meier survival analysis. To test a relevant clinical application of FaceAge, we assessed the performance of FaceAge in end-of-life patients with metastatic cancer who received palliative treatment by incorporating FaceAge into clinical prediction models. We found that, on average, cancer patients look older than their chronological age, and looking older is correlated with worse overall survival. FaceAge demonstrated significant independent prognostic performance in a range of cancer types and stages. We found that FaceAge can improve physicians' survival predictions in incurable patients receiving palliative treatments, highlighting the clinical utility of the algorithm to support end-of-life decision-making. FaceAge was also significantly associated with molecular mechanisms of senescence through gene analysis, while age was not. These findings may extend to diseases beyond cancer, motivating using deep learning algorithms to translate a patient's visual appearance into objective, quantitative, and clinically useful measures.

A pan-cancer clinical platform to predict immunotherapy outcomes and prioritize immuno-oncology combinations in early-phase trials.

BACKGROUND: Immunotherapy is effective, but current biomarkers for patient selection have proven modest sensitivity. Here, we developed VIGex, an optimized gene signature based on the expression level of 12 genes involved in immune response with RNA sequencing. METHODS: We implemented VIGex using the nCounter platform (Nanostring) on a large clinical cohort encompassing 909 tumor samples across 45 tumor types. VIGex was developed as a continuous variable, with cutoffs selected to detect three main categories (hot, intermediate-cold and cold) based on the different inflammatory status of the tumor microenvironment. FINDINGS: Hot tumors had the highest VIGex scores and exhibited an increased abundance of tumor-infiltrating lymphocytes as compared with the intermediate-cold and cold. VIGex scores varied depending on tumor origin and anatomic site of metastases, with liver metastases showing an immunosuppressive tumor microenvironment. The predictive power of VIGex-Hot was observed in a cohort of 98 refractory solid tumor from patients treated in early-phase immunotherapy trials and its clinical performance was confirmed through an extensive metanalysis across 13 clinically annotated gene expression datasets from 877 patients treated with immunotherapy agents. Last, we generated a pan-cancer biomarker platform that integrates VIGex categories with the expression levels of immunotherapy targets under development in early-phase clinical trials. CONCLUSIONS: Our results support the clinical utility of VIGex as a tool to aid clinicians for patient selection and personalized immunotherapy interventions. FUNDING: BBVA Foundation; 202-2021 Division of Medical Oncology and Hematology Fellowship award; Princess Margaret Cancer Center.

Heterogeneity in leukemia cells that escape drug-induced senescence-like state.

Erythropoietin (EPO) suppresses drug-induced apoptosis in EPO-receptor-positive leukemia cells and allows cells to persist after drug treatment by promoting cellular senescence. Importantly a small proportion of senescent cells can re-enter the cell cycle and resume proliferation after drug treatment, resulting in disease recurrence/persistence. Using a single-cell assay to track individual cells that exit a drug-induced senescence-like state, we show that cells exhibit asynchronous exit from a senescent-like state, and display different rates of proliferation. Escaped cells retain sensitivity to drug treatment, but display inter-clonal variability. We also find heterogeneity in gene expression with some of the escaped clones retaining senescence-associated gene expression. Senescent leukemia cells exhibit changes in gene expression that affect metabolism and senescence-associated secretory phenotype (SASP)-related genes. Herein, we generate a senescence gene signature and show that this signature is a prognostic marker of worse overall survival in AML and multiple other cancers. A portion of senescent leukemia cells depend on lysosome activity; chloroquine, an inhibitor of lysosome activity, promotes senolysis of some senescent leukemia cells. Our study indicates that the serious risks associated with the use of erythropoietin-stimulating agents (ESAs) in anemic cancer patients may be attributed to their ability to promote drug-tolerant cancer cells through the senescence program.

Safety, Immunologic, and Clinical Activity of Durvalumab in Combination with Olaparib or Cediranib in Advanced Leiomyosarcoma: Results of the DAPPER Clinical Trial.

PURPOSE: Non-inflamed (cold) tumors such as leiomyosarcoma do not benefit from immune checkpoint blockade (ICB) monotherapy. Combining ICB with angiogenesis or PARP inhibitors may increase tumor immunogenicity by altering the immune cell composition of the tumor microenvironment (TME). The DAPPER phase II study evaluated the safety, immunologic, and clinical activity of ICB-based combinations in pretreated patients with leiomyosarcoma. PATIENTS AND METHODS: Patients were randomized to receive durvalumab 1,500 mg IV every 4 weeks with either olaparib 300 mg twice a day orally (Arm A) or cediranib 20 mg every day orally 5 days/week (Arm B) until unacceptable toxicity or disease progression. Paired tumor biopsies, serial radiologic assessments and stool collections were performed. Primary endpoints were safety and immune cell changes in the TME. Objective responses and survival were correlated with transcriptomic, radiomic, and microbiome parameters. RESULTS: Among 30 heavily pretreated patients (15 on each arm), grade ≥ 3 toxicity occurred in 3 (20%) and 2 (13%) on Arms A and B, respectively. On Arm A, 1 patient achieved partial response (PR) with increase in CD8 T cells and macrophages in the TME during treatment, while 4 had stable disease (SD) ≥ 6 months. No patients on Arm B achieved PR or SD ≥ 6 months. Transcriptome analysis showed that baseline M1-macrophage and B-cell activity were associated with overall survival. CONCLUSIONS: Durvalumab plus olaparib increased immune cell infiltration of TME with clinical benefit in some patients with leiomyosarcoma. Baseline M1-macrophage and B-cell activity may identify patients with leiomyosarcoma with favorable outcomes on immunotherapy and should be further evaluated.

Multi-institutional Prognostic Modeling in Head and Neck Cancer: Evaluating Impact and Generalizability of Deep Learning and Radiomics.

Artificial intelligence (AI) and machine learning (ML) are becoming critical in developing and deploying personalized medicine and targeted clinical trials. Recent advances in ML have enabled the integration of wider ranges of data including both medical records and imaging (radiomics). However, the development of prognostic models is complex as no modeling strategy is universally superior to others and validation of developed models requires large and diverse datasets to demonstrate that prognostic models developed (regardless of method) from one dataset are applicable to other datasets both internally and externally. Using a retrospective dataset of 2,552 patients from a single institution and a strict evaluation framework that included external validation on three external patient cohorts (873 patients), we crowdsourced the development of ML models to predict overall survival in head and neck cancer (HNC) using electronic medical records (EMR) and pretreatment radiological images. To assess the relative contributions of radiomics in predicting HNC prognosis, we compared 12 different models using imaging and/or EMR data. The model with the highest accuracy used multitask learning on clinical data and tumor volume, achieving high prognostic accuracy for 2-year and lifetime survival prediction, outperforming models relying on clinical data only, engineered radiomics, or complex deep neural network architecture. However, when we attempted to extend the best performing models from this large training dataset to other institutions, we observed significant reductions in the performance of the model in those datasets, highlighting the importance of detailed population-based reporting for AI/ML model utility and stronger validation frameworks. We have developed highly prognostic models for overall survival in HNC using EMRs and pretreatment radiological images based on a large, retrospective dataset of 2,552 patients from our institution.Diverse ML approaches were used by independent investigators. The model with the highest accuracy used multitask learning on clinical data and tumor volume.External validation of the top three performing models on three datasets (873 patients) with significant differences in the distributions of clinical and demographic variables demonstrated significant decreases in model performance. Significance: ML combined with simple prognostic factors outperformed multiple advanced CT radiomics and deep learning methods. ML models provided diverse solutions for prognosis of patients with HNC but their prognostic value is affected by differences in patient populations and require extensive validation.