Search

In this study, the impact of the apolipoprotein B mRNA-editing catalytic subunit-like (APOBEC) enzyme APOBEC3B (A3B) on epidermal growth factor receptor (EGFR)-driven lung cancer was assessed. A3B expression in EGFR mutant (EGFRmut) non-small-cell lung cancer (NSCLC) mouse models constrained tumorigenesis, while A3B expression in tumors treated with EGFR-targeted cancer therapy was associated with treatment resistance. Analyses of human NSCLC models treated with EGFR-targeted therapy showed upregulation of A3B and revealed therapy-induced activation of nuclear factor kappa B (NF-κB) as an inducer of A3B expression. Significantly reduced viability was observed with A3B deficiency, and A3B was required for the enrichment of APOBEC mutation signatures, in targeted therapy-treated human NSCLC preclinical models. Upregulation of A3B was confirmed in patients with NSCLC treated with EGFR-targeted therapy. This study uncovers the multifaceted roles of A3B in NSCLC and identifies A3B as a potential target for more durable responses to targeted cancer therapy.

Immuno-oncology has transformed the treatment of cancer, with several immunotherapies becoming the standard treatment across histologies. Despite these advancements, the majority of patients do not experience durable clinical benefits, highlighting the imperative for ongoing advancement in immuno-oncology. Computational immuno-oncology emerges as a forefront discipline that draws on biomedical data science and intersects with oncology, immunology, and clinical research, with the overarching goal to accelerate the development of effective and safe immuno-oncology treatments from the laboratory to the clinic. In this review, we outline 10 critical challenges and opportunities in computational immuno-oncology, emphasizing the importance of robust computational strategies and interdisciplinary collaborations amid the constantly evolving interplay between clinical needs and technological innovation.

Background and objective: Previous germline studies on renal cell carcinoma (RCC) have usually pooled clear and non–clear cell RCCs and have not adequately accounted for population stratification, which might have led to an inaccurate estimation of genetic risk. Here, we aim to analyze the major germline drivers of RCC risk and clinically relevant but underexplored germline variant types. Methods: We first characterized germline pathogenic variants (PVs), cryptic splice variants, and copy number variants (CNVs) in 1436 unselected RCC patients. To evaluate the enrichment of PVs in RCC, we conducted a case-control study of 1356 RCC patients ancestry matched with 16 512 cancer-free controls using approaches accounting for population stratification and histological subtypes, followed by characterization of secondary somatic events. Key findings and limitations: Clear cell RCC patients (n = 976) exhibited a significant burden of PVs in VHL compared with controls (odds ratio [OR]: 39.1, p = 4.95e-05). Non–clear cell RCC patients (n = 380) carried enrichment of PVs in FH (OR: 77.9, p = 1.55e-08) and MET (OR: 1.98e11, p = 2.07e-05). In a CHEK2-focused analysis with European participants, clear cell RCC (n = 906) harbored nominal enrichment of low-penetrance CHEK2 variants—p.Ile157Thr (OR: 1.84, p = 0.049) and p.Ser428Phe (OR: 5.20, p = 0.045), while non–clear cell RCC (n = 295) exhibited nominal enrichment of CHEK2 loss of function PVs (OR: 3.51, p = 0.033). Patients with germline PVs in FH, MET, and VHL exhibited significantly earlier age of cancer onset than patients without germline PVs (mean: 46.0 vs 60.2 yr, p

Summary: Renal cell carcinoma with sarcomatoid differentiation (sRCC) is associated with poor survival and a heightened response to immune checkpoint inhibitors (ICIs). Two major barriers to improving outcomes for sRCC are the limited understanding of its gene regulatory programs and the low diagnostic yield of tumor biopsies due to spatial heterogeneity. Herein, we characterized the epigenomic landscape of sRCC by profiling 107 epigenomic libraries from tissue and plasma samples from 50 patients with RCC and healthy volunteers. By profiling histone modifications and DNA methylation, we identified highly recurrent epigenomic reprogramming enriched in sRCC. Furthermore, CRISPRa experiments implicated the transcription factor FOSL1 in activating sRCC-associated gene regulatory programs, and FOSL1 expression was associated with the response to ICIs in RCC in two randomized clinical trials. Finally, we established a blood-based diagnostic approach using detectable sRCC epigenomic signatures in patient plasma, providing a framework for discovering epigenomic correlates of tumor histology via liquid biopsy.

The diversity of structural variants (SVs) in melanoma and how they impact oncogenesis are incompletely known. We performed harmonized analysis of SVs across melanoma histologic and genomic subtypes, and we identified distinct global properties between subtypes. These included the frequency and size of SVs and SV classes, their relation to chromothripsis events, and the impact on cancer-related genes of SVs that alter topologically associated domain (TAD) boundaries. Following our prior identification of double-stranded break repair deficiency in a subset of triple-wild-type cutaneous melanoma, we identified MRE11 and NBN loss-of-function SVs in melanomas with this mutational signature. Experimental knockouts of MRE11 and NBN, followed by olaparib cell viability assays in melanoma cells, indicated that dysregulation of each of these genes may cause sensitivity to PARP inhibitors in cutaneous melanomas. Broadly, harmonized analysis of melanoma SVs revealed distinct global genomic properties and molecular drivers, which may have biological and therapeutic impact., Introduction Cutaneous melanoma is among the most highly mutated cancers because of the impact of UV mutagenesis, which leads to many C>T transitions across the genome (1, 2). For this [...]

Abstract Background Breast cancer patients from the indigenous Arab population present much earlier than patients from Western countries and have traditionally been underrepresented in cancer genomics studies. The contribution of polygenic and Mendelian risk toward the earlier onset of breast cancer in the population remains elusive. Methods We performed low-pass whole genome sequencing (lpWGS) and whole-exome sequencing (WES) from 220 female breast cancer patients unselected for positive family history from the indigenous Arab population. Using publicly available resources, we imputed population-specific variants and calculated breast cancer burden-sensitive polygenic risk scores (PRS). Variant pathogenicity was also evaluated on exome variants with high coverage. Results Variants imputed from lpWGS showed high concordance with paired exome (median dosage correlation: 0.9459, Interquartile range: 0.9410–0.9490). After adjusting the PRS to the Arab population, we found significant associations between PRS performance in risk prediction and first-degree relative breast cancer history prediction (Spearman rho=0.43, p = 0.03), where breast cancer patients in the top PRS decile are 5.53 (95% CI 1.76–17.97, p = 0.003) times more likely also to have a first-degree relative diagnosed with breast cancer compared to those in the middle deciles. In addition, we found evidence for the genetic liability threshold model of breast cancer where among patients with a family history of breast cancer, pathogenic rare variant carriers had significantly lower PRS than non-carriers (p = 0.0205, Mann-Whitney U test) while for non-carriers every standard deviation increase in PRS corresponded to 4.52 years (95% CI 8.88–0.17, p = 0.042) earlier age of presentation. Conclusions Overall, our study provides a framework to assess polygenic risk in an understudied population using lpWGS and identifies common variant risk as a factor independent of pathogenic variant carrier status for earlier age of onset of breast cancer among indigenous Arab breast cancer patients.

Abstract Background Longitudinal data on key cancer outcomes for clinical research, such as response to treatment and disease progression, are not captured in standard cancer registry reporting. Manual extraction of such outcomes from unstructured electronic health records is a slow, resource-intensive process. Natural language processing (NLP) methods can accelerate outcome annotation, but they require substantial labeled data. Transfer learning based on language modeling, particularly using the Transformer architecture, has achieved improvements in NLP performance. However, there has been no systematic evaluation of NLP model training strategies on the extraction of cancer outcomes from unstructured text. Results We evaluated the performance of nine NLP models at the two tasks of identifying cancer response and cancer progression within imaging reports at a single academic center among patients with non-small cell lung cancer. We trained the classification models under different conditions, including training sample size, classification architecture, and language model pre-training. The training involved a labeled dataset of 14,218 imaging reports for 1112 patients with lung cancer. A subset of models was based on a pre-trained language model, DFCI-ImagingBERT, created by further pre-training a BERT-based model using an unlabeled dataset of 662,579 reports from 27,483 patients with cancer from our center. A classifier based on our DFCI-ImagingBERT, trained on more than 200 patients, achieved the best results in most experiments; however, these results were marginally better than simpler “bag of words” or convolutional neural network models. Conclusion When developing AI models to extract outcomes from imaging reports for clinical cancer research, if computational resources are plentiful but labeled training data are limited, large language models can be used for zero- or few-shot learning to achieve reasonable performance. When computational resources are more limited but labeled training data are readily available, even simple machine learning architectures can achieve good performance for such tasks.

Immuno-oncology holds promise for transforming patient care having achieved durable clinical response rates across a variety of advanced and metastatic cancers. Despite these achievements, only a minority of patients respond to immunotherapy, underscoring the importance of elucidating molecular mechanisms responsible for response and resistance to inform the development and selection of treatments. Breakthroughs in molecular sequencing technologies have led to the generation of an immense amount of genomic and transcriptomic sequencing data that can be mined to uncover complex tumor-immune interactions using computational tools. In this review, we discuss existing and emerging computational methods that contextualize the composition and functional state of the tumor microenvironment, infer the reactivity and clonal dynamics from reconstructed immune cell receptor repertoires, and predict the antigenic landscape for immune cell recognition. We further describe the advantage of multi-omics analyses for capturing multidimensional relationships and artificial intelligence techniques for integrating omics data with histopathological and radiological images to encapsulate patterns of treatment response and tumor-immune biology. Finally, we discuss key challenges impeding their widespread use and clinical application and conclude with future perspectives. We are hopeful that this review will both serve as a guide for prospective researchers seeking to use existing tools for scientific discoveries and inspire the optimization or development of novel tools to enhance precision, ultimately expediting advancements in immunotherapy that improve patient survival and quality of life.

Background Recent trials suggest that programmed cell death 1 (PD-1)-directed immunotherapy may be beneficial for some patients with anal squamous cell carcinoma and biomarkers predictive of response are greatly needed.Methods This multicenter phase II clinical trial (NCT02919969) enrolled patients with metastatic or locally advanced incurable anal squamous cell carcinoma (n=32). Patients received pembrolizumab 200 mg every 3 weeks. The primary endpoint of the trial was objective response rate (ORR). Exploratory objectives included analysis of potential predictive biomarkers including assessment of tumor-associated immune cell populations with multichannel immunofluorescence and analysis of circulating tumor tissue modified viral-human papillomavirus DNA (TTMV-HPV DNA) using serially collected blood samples. To characterize the clinical features of long-term responders, we combined data from our prospective trial with a retrospective cohort of patients with anal cancer treated with anti-PD-1 immunotherapy (n=18).Results In the phase II study, the ORR to pembrolizumab monotherapy was 9.4% and the median progression-free survival was 2.2 months. Despite the high level of HPV positivity observed with circulating TTMV-HPV DNA testing, the majority of patients had low levels of tumor-associated CD8+PD-1+ T cells on pretreatment biopsy. Patients who benefited from pembrolizumab had decreasing TTMV-HPV DNA scores and a complete responder’s TTMV-HPV DNA became undetectable. Long-term pembrolizumab responses were observed in one patient from the trial (5.3 years) and three patients (2.5, 6, and 8 years) from the retrospective cohort. Long-term responders had HPV-positive tumors, lacked liver metastases, and achieved a radiological complete response.Conclusions Pembrolizumab has durable efficacy in a rare subset of anal cancers. However, despite persistence of HPV infection, indicated by circulating HPV DNA, most advanced anal cancers have low numbers of tumor-associated CD8+PD-1+ T cells and are resistant to pembrolizumab.

Medulloblastoma is among the most common malignant brain tumors in children. Recent studies have identified at least four subgroups of the disease that differ in terms of molecular characteristics and patient outcomes. Despite this heterogeneity, most patients with medulloblastoma receive similar therapies, including surgery, radiation, and intensive chemotherapy. Although these treatments prolong survival, many patients still die from the disease and survivors suffer severe long-term side effects from therapy. We hypothesize that each patient with medulloblastoma is sensitive to different therapies and that tailoring therapy based on the molecular and cellular characteristics of patients' tumors will improve outcomes. To test this, we assembled a panel of orthotopic patient-derived xenografts (PDX) and subjected them to DNA sequencing, gene expression profiling, and high-throughput drug screening. Analysis of DNA sequencing revealed that most medulloblastomas do not have actionable mutations that point to effective therapies. In contrast, gene expression and drug response data provided valuable information about potential therapies for every tumor. For example, drug screening demonstrated that actinomycin D, which is used for treatment of sarcoma but rarely for medulloblastoma, was active against PDXs representing Group 3 medulloblastoma, the most aggressive form of the disease. Functional analysis of tumor cells was successfully used in a clinical setting to identify more treatment options than sequencing alone. These studies suggest that it should be possible to move away from a one-size-fits-all approach and begin to treat each patient with therapies that are effective against their specific tumor. SIGNIFICANCE: These findings show that high-throughput drug screening identifies therapies for medulloblastoma that cannot be predicted by genomic or transcriptomic analysis.

Summary: The success of precision oncology—which aims to match the right therapies to the right patients based on molecular status—is predicated on a robust pipeline of molecular targets against which therapies can be developed. Recent advances in genomics and functional genetics have enabled the unbiased discovery of novel molecular targets at scale. We summarize the promise and challenges in integrating genomic and functional genetic landscapes of cancer to establish the next generation of cancer targets.

An assimilation bias occurs when people's evaluative judgment is positively influenced by a previously observed signal. We study this effect by examining investors' appraisal of M&A deals announced 1 day after other firms in the same 1-digit SIC as the merging parties release earnings surprises. Consistent with assimilation effects, acquirers' M&A announcement stock return initially correlates with the previous day's earnings surprises. This effect reverses after 1 week. Assimilation generates other distortions as more positive surprises are related to increases in bid competition, takeover premiums, and withdrawn M&As. Evidence from IPOs corroborates the presence of assimilation effects in financial markets. [ABSTRACT FROM AUTHOR]

We contrast the investment strategies of hedge funds and mutual funds around mergers and acquisitions (M&A). We find that hedge funds, on average, increase their holdings of soon‐to‐be takeover targets by 7.5% during the quarter before M&A announcements. Conversely, mutual funds, on average, reduce their equity holdings in impending targets by 3.0% over the same time period. The reduction in M&A holdings by mutual funds is less pronounced for more actively managed funds. Our results suggest that hedge funds enjoy superior access to private information or possess superior ability to process public information related to M&A transactions. [ABSTRACT FROM AUTHOR]

(Cell 173, 371–385.e1–e9; April 5, 2018) It has come to our attention that we made two errors in preparation of this manuscript. First, in the STAR Methods, under the subheading of “Hypermutators and Immune Infiltrates” within the “Quantification and Statistical Analysis” section, we inadvertently referred to Figures S7A–S7C for data corresponding to sample stratification by hypermutator status alone in the last sentence. It should have referred to Figure S6A–S6C. Second, the lists of highly frequent missense mutations for COAD (colon adenocarcinoma) and READ (rectum adenocarcinoma) displayed in Figure S7 were incorrect because when we ordered the mutations in the initial analysis, we mistakenly combined the two cancer types COAD and READ for analysis, despite the fact that they were listed as two separate cancer types in the x-axis of the figure. After re-ordering the mutations by frequency for COAD and READ independently, information on highly frequent missense mutations for each of these cancer types is different and updated now in the revised Figure S7. These errors don't change the major conclusions of the paper and have been corrected online. We apologize for any confusion they may have caused. [Figure-presented]

To accelerate cancer research that correlates biomarkers with clinical endpoints, methods are needed to ascertain outcomes from electronic health records at scale. Here, the authors train natural language processing to extract outcomes for participants in a precision oncology study.