A Rigorous Multi-Laboratory Study of Known PDAC Biomarkers Identifies Increased Sensitivity and Specificity Over CA19-9 Alone.

A blood test that enables surveillance for early-stage pancreatic ductal adenocarcinoma (PDAC) is an urgent need. Independent laboratories have reported PDAC biomarkers that could improve biomarker performance over CA19-9 alone, but the performance of the previously reported biomarkers in combination is not known. Therefore, we conducted a coordinated case/control study across multiple laboratories using common sets of blinded training and validation samples (132 and 295 plasma samples, respectively) from PDAC patients and non-PDAC control subjects representing conditions under which surveillance occurs. We analyzed the training set to identify candidate biomarker combination panels using biomarkers across laboratories, and we applied the fixed panels to the validation set. The panels identified in the training set, CA19-9 with CA199.STRA, LRG1, TIMP-1, TGM2, THSP2, ANG, and MUC16.STRA, achieved consistent performance in the validation set. The panel of CA19-9 with the glycan biomarker CA199.STRA improved sensitivity from 0.44 with 0.98 specificity for CA19-9 alone to 0.71 with 0.98 specificity (p < 0.001, 1000-fold bootstrap). Similarly, CA19-9 combined with the protein biomarker LRG1 and CA199.STRA improved specificity from 0.16 with 0.94 sensitivity for CA19-9 to 0.65 with 0.89 sensitivity (p < 0.001, 1000-fold bootstrap). We further validated significantly improved performance using biomarker panels that did not include CA19-9. This study establishes the effectiveness of a coordinated study of previously discovered biomarkers and identified panels of those biomarkers that significantly increased the sensitivity and specificity of early-stage PDAC detection in a rigorous validation trial.

A blood-based metabolomic signature predictive of risk for pancreatic cancer.

Emerging evidence implicates microbiome involvement in the development of pancreatic cancer (PaCa). Here, we investigate whether increases in circulating microbial-related metabolites associate with PaCa risk by applying metabolomics profiling to 172 sera collected within 5 years prior to PaCa diagnosis and 863 matched non-subject sera from participants in the Prostate, Lung, Colorectal, and Ovarian (PLCO) cohort. We develop a three-marker microbial-related metabolite panel to assess 5-year risk of PaCa. The addition of five non-microbial metabolites further improves 5-year risk prediction of PaCa. The combined metabolite panel complements CA19-9, and individuals with a combined metabolite panel + CA19-9 score in the top 2.5th percentile have absolute 5-year risk estimates of >13%. The risk prediction model based on circulating microbial and non-microbial metabolites provides a potential tool to identify individuals at high risk of PaCa that would benefit from surveillance and/or from potential cancer interception strategies.

Coffee Intake, Caffeine Metabolism Genotype, and Survival Among Men with Prostate Cancer.

BACKGROUND: Coffee intake may lower prostate cancer risk and progression, but postdiagnosis outcomes by caffeine metabolism genotype are not well characterized. OBJECTIVE: To evaluate associations between coffee intake, caffeine metabolism genotype, and survival in a large, multicenter study of men with prostate cancer. DESIGN, SETTING, AND PARTICIPANTS: Data from The PRACTICAL Consortium database for 5727 men with prostate cancer from seven US, Australian, and European studies were included. The cases included had data available for the CYP1A2 -163C>A rs762551 single-nucleotide variant associated with caffeine metabolism, coffee intake, and >6 mo of follow-up. OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS: Multivariable-adjusted Cox proportional hazards models across pooled patient-level data were used to compare the effect of coffee intake (categorized as low [reference], high, or none/very low) in relation to overall survival (OS) and prostate cancer-specific survival (PCSS), with stratified analyses conducted by clinical disease risk and genotype. RESULTS AND LIMITATIONS: High coffee intake appeared to be associated with longer PCSS (hazard ratio [HR] 0.85, 95% confidence interval [CI] 0.68-1.08; p = 0.18) and OS (HR 0.90, 95% CI 0.77-1.07; p = 0.24), although results were not statistically significant. In the group with clinically localized disease, high coffee intake was associated with longer PCSS (HR 0.66, 95% CI 0.44-0.98; p = 0.040), with comparable results for the group with advanced disease (HR 0.92, 95% CI 0.69-1.23; p = 0.6). High coffee intake was associated with longer PCSS among men with the CYP1A2 AA (HR 0.67, 95% CI 0.49-0.93; p = 0.017) but not the AC/CC genotype (p = 0.8); an interaction was detected (p = 0.042). No associations with OS were observed in subgroup analyses (p > 0.05). Limitations include the nominal statistical significance and residual confounding. CONCLUSIONS: Coffee intake was associated with longer PCSS among men with a CYP1A2 -163AA (*1F/*1F) genotype, a finding that will require further replication. PATIENT SUMMARY: It is likely that coffee intake is associated with longer prostate cancer-specific survival in certain groups, but more research is needed to fully understand which men may benefit and why.

A polyamine-centric, blood-based metabolite panel predictive of poor response to CAR-T cell therapy in large B cell lymphoma.

Anti-CD19 chimeric antigen receptor (CAR) T cell therapy for relapsed or refractory (r/r) large B cell lymphoma (LBCL) results in durable response in only a subset of patients. MYC overexpression in LBCL tumors is associated with poor response to treatment. We tested whether an MYC-driven polyamine signature, as a liquid biopsy, is predictive of response to anti-CD19 CAR-T therapy in patients with r/r LBCL. Elevated plasma acetylated polyamines were associated with non-durable response. Concordantly, increased expression of spermidine synthase, a key enzyme that regulates levels of acetylated spermidine, was prognostic for survival in r/r LBCL. A broad metabolite screen identified additional markers that resulted in a 6-marker panel (6MetP) consisting of acetylspermidine, diacetylspermidine, and lysophospholipids, which was validated in an independent set from another institution as predictive of non-durable response to CAR-T therapy. A polyamine centric metabolomics liquid biopsy panel has predictive value for response to CAR-T therapy in r/r LBCL.

Mining the Immunopeptidome for Antigenic Peptides in Cancer.

Although harnessing the immune system for cancer therapy has shown success, response to immunotherapy has been limited. The immunopeptidome of cancer cells presents an opportunity to discover novel antigens for immunotherapy applications. These neoantigens bind to MHC class I and class II molecules. Remarkably, the immunopeptidome encompasses protein post-translation modifications (PTMs) that may not be evident from genome or transcriptome profiling. A case in point is citrullination, which has been demonstrated to induce a strong immune response. In this review, we cover how the immunopeptidome, with a special focus on PTMs, can be utilized to identify cancer-specific antigens for immunotherapeutic applications.

A Blood-Based Metabolite Panel for Distinguishing Ovarian Cancer from Benign Pelvic Masses.

PURPOSE: To assess the contributions of circulating metabolites for improving upon the performance of the risk of ovarian malignancy algorithm (ROMA) for risk prediction of ovarian cancer among women with ovarian cysts. EXPERIMENTAL DESIGN: Metabolomic profiling was performed on an initial set of sera from 101 serous and nonserous ovarian cancer cases and 134 individuals with benign pelvic masses (BPM). Using a deep learning model, a panel consisting of seven cancer-related metabolites [diacetylspermine, diacetylspermidine, N-(3-acetamidopropyl)pyrrolidin-2-one, N-acetylneuraminate, N-acetyl-mannosamine, N-acetyl-lactosamine, and hydroxyisobutyric acid] was developed for distinguishing early-stage ovarian cancer from BPM. The performance of the metabolite panel was evaluated in an independent set of sera from 118 ovarian cancer cases and 56 subjects with BPM. The contributions of the panel for improving upon the performance of ROMA were further assessed. RESULTS: A 7-marker metabolite panel (7MetP) developed in the training set yielded an AUC of 0.86 [95% confidence interval (CI): 0.76-0.95] for early-stage ovarian cancer in the independent test set. The 7MetP+ROMA model had an AUC of 0.93 (95% CI: 0.84-0.98) for early-stage ovarian cancer in the test set, which was improved compared with ROMA alone [0.91 (95% CI: 0.84-0.98); likelihood ratio test P: 0.03]. In the entire specimen set, the combined 7MetP+ROMA model yielded a higher positive predictive value (0.68 vs. 0.52; one-sided P < 0.001) with improved specificity (0.89 vs. 0.78; one-sided P < 0.001) for early-stage ovarian cancer compared with ROMA alone. CONCLUSIONS: A blood-based metabolite panel was developed that demonstrates independent predictive ability and complements ROMA for distinguishing early-stage ovarian cancer from benign disease to better inform clinical decision making.

Application of Artificial Intelligence to Plasma Metabolomics Profiles to Predict Response to Neoadjuvant Chemotherapy in Triple-Negative Breast Cancer.

There is a need to identify biomarkers predictive of response to neoadjuvant chemotherapy (NACT) in triple-negative breast cancer (TNBC). We previously obtained evidence that a polyamine signature in the blood is associated with TNBC development and progression. In this study, we evaluated whether plasma polyamines and other metabolites may identify TNBC patients who are less likely to respond to NACT. Pre-treatment plasma levels of acetylated polyamines were elevated in TNBC patients that had moderate to extensive tumor burden (RCB-II/III) following NACT compared to those that achieved a complete pathological response (pCR/RCB-0) or had minimal residual disease (RCB-I). We further applied artificial intelligence to comprehensive metabolic profiles to identify additional metabolites associated with treatment response. Using a deep learning model (DLM), a metabolite panel consisting of two polyamines as well as nine additional metabolites was developed for improved prediction of RCB-II/III. The DLM has potential clinical value for identifying TNBC patients who are unlikely to respond to NACT and who may benefit from other treatment modalities.

A Deep Learning Approach to Re-create Raw Full-Field Digital Mammograms for Breast Density and Texture Analysis.

PURPOSE: To develop a computational approach to re-create rarely stored for-processing (raw) digital mammograms from routinely stored for-presentation (processed) mammograms. MATERIALS AND METHODS: In this retrospective study, pairs of raw and processed mammograms collected in 884 women (mean age, 57 years ± 10 [standard deviation]; 3713 mammograms) from October 5, 2017, to August 1, 2018, were examined. Mammograms were split 3088 for training and 625 for testing. A deep learning approach based on a U-Net convolutional network and kernel regression was developed to estimate the raw images. The estimated raw images were compared with the originals by four image error and similarity metrics, breast density calculations, and 29 widely used texture features. RESULTS: In the testing dataset, the estimated raw images had small normalized mean absolute error (0.022 ± 0.015), scaled mean absolute error (0.134 ± 0.078) and mean absolute percentage error (0.115 ± 0.059), and a high structural similarity index (0.986 ± 0.007) for the breast portion compared with the original raw images. The estimated and original raw images had a strong correlation in breast density percentage (Pearson r = 0.946) and a strong agreement in breast density grade (Cohen κ = 0.875). The estimated images had satisfactory correlations with the originals in 23 texture features (Pearson r ≥ 0.503 or Spearman ρ ≥ 0.705) and were well complemented by processed images for the other six features. CONCLUSION: This deep learning approach performed well in re-creating raw mammograms with strong agreement in four image evaluation metrics, breast density, and the majority of 29 widely used texture features.Keywords: Mammography, Breast, Supervised Learning, Convolutional Neural Network (CNN), Deep learning algorithms, Machine Learning AlgorithmsSee also the commentary by Chan in this issue.Supplemental material is available for this article.©RSNA, 2021.

Protein citrullination as a source of cancer neoantigens.

BACKGROUND: Citrulline post-translational modification of proteins is mediated by protein arginine deiminase (PADI) family members and has been associated with autoimmune diseases. The role of PADI-citrullinome in immune response in cancer has not been evaluated. We hypothesized that PADI-mediated citrullinome is a source of neoantigens in cancer that induces immune response. METHODS: Protein expression of PADI family members was evaluated in 196 cancer cell lines by means of indepth proteomic profiling. Gene expression was assessed using messenger RNA data sets from The Cancer Genome Atlas. Immunohistochemical analysis of PADI2 and peptidyl-citrulline was performed using breast cancer tissue sections. Citrullinated 12-34-mer peptides in the putative Major Histocompatibility Complex-II (MHC-II) binding range were profiled in breast cancer cell lines to investigate the relationship between protein citrullination and antigen presentation. We further evaluated immunoglobulin-bound citrullinome by mass spectrometry using 156 patients with breast cancer and 113 cancer-free controls. RESULTS: Proteomic and gene expression analyses revealed PADI2 to be highly expressed in several cancer types including breast cancer. Immunohistochemical analysis of 422 breast tumor tissues revealed increased expression of PADI2 in ER- tumors (p<0.0001); PADI2 protein expression was positively correlated (p<0.0001) with peptidyl-citrulline staining. PADI2 expression exhibited strong positive correlations with a B cell immune signature and with MHC-II-bound citrullinated peptides. Increased circulating citrullinated antigen-antibody complexes occurred among newly diagnosed breast cancer cases relative to controls (p=0.0012). CONCLUSIONS: An immune response associated with citrullinome is a rich source of neoantigens in breast cancer with a potential for diagnostic and therapeutic applications.

Pan-Cancer Early Detection: Hype or Hope?

Universal cancer screening based on circulating DNA, proteins, metabolites, or other combinations has the potential to revolutionize early cancer detection, especially for cancers with no available screening modalities. Two recent publications in Science and Annals of Oncology highlight the potential benefits and limitations of single-test, multiple cancer screens.

ß-Catenin maintains lung epithelial progenitors after lung specification.

The entire lung epithelium arises from SRY box 9 (SOX9)-expressing progenitors that form the respiratory tree and differentiate into airway and alveolar cells. Despite progress in understanding their initial specification within the embryonic foregut, how these progenitors are subsequently maintained is less clear. Using inducible, progenitor-specific genetic mosaic mouse models, we showed that ß-catenin (CTNNB1) maintains lung progenitors by promoting a hierarchical lung progenitor gene signature, suppressing gastrointestinal (GI) genes, and regulating NK2 homeobox 1 (NKX2.1) and SRY box 2 (SOX2) in a developmental stage-dependent manner. At the early, but not later, stage post-lung specification, CTNNB1 cell-autonomously maintained normal NKX2.1 expression levels and suppressed ectopic SOX2 expression. Genetic epistasis analyses revealed that CTNNB1 is required for fibroblast growth factor (Fgf)/Kirsten rat sarcoma viral oncogene homolog (Kras)-mediated promotion of the progenitors. In silico screening of Eurexpress and translating ribosome affinity purification (TRAP)-RNAseq identified a progenitor gene signature, a subset of which depends on CTNNB1. Wnt signaling also maintained NKX2.1 expression and suppressed GI genes in cultured human lung progenitors derived from embryonic stem cells.

Baseline and longitudinal plasma caveolin-1 level as a biomarker in active surveillance for early-stage prostate cancer.

OBJECTIVES: To evaluate the role of caveolin-1 (Cav-1) as a predictor of disease reclassification (DR) in men with early prostate cancer undergoing active surveillance (AS). PATIENTS AND METHODS: We analysed archived plasma samples prospectively collected from patients with early prostate cancer in a single-institution AS study. Of 825 patients enrolled, 542 had ≥1 year of follow-up. Baseline and longitudinal plasma Cav-1 levels were measured using an enzyme-linked immunosorbent assay. Tumour volume or Gleason grade increases were criteria for DR. Logistic regression analyses were used to assess associations between clinicopathological characteristics and reclassification risk. RESULTS: In 542 patients, 480 (88.6%) had stage cT1c disease, 542 (100.0%) had a median prostate-specific antigen level of 4.1 ng/mL, and 531 (98.0%) had a median Cancer of the Prostate Risk Assessment score of 1. In all, 473 (87.3%) had a Gleason score of 3+3. After a median of 3.1 years of follow-up, disease was reclassified in 163 patients (30.1%). The mean baseline Cav-1 level was 2.2 ± 8.5 ng/mL and the median 0.2 ng/mL (range, 0-85.5 ng/mL). In univariate analysis, baseline Cav-1 was a significant predictor for risk of DR (odds ratio [OR] 1.82, 95% confidence interval [CI] 1.24-2.65; P = 0.002). In multivariate analysis, with adjustments for age, tumour length, group risk stratification and number of positive cores, reclassification risk associated with Cav-1 remained significant (OR 1.91, 95% CI 1.28-2.84; P = 0.001). CONCLUSION: Baseline plasma Cav-1 level was an independent predictor of disease classification. New methods for refining AS and intervention may result.

HSP90 inhibition enhances cancer immunotherapy by upregulating interferon response genes.

T-cell-based immunotherapies are promising treatments for cancer patients. Although durable responses can be achieved in some patients, many patients fail to respond to these therapies, underscoring the need for improvement with combination therapies. From a screen of 850 bioactive compounds, we identify HSP90 inhibitors as candidates for combination with immunotherapy. We show that inhibition of HSP90 with ganetespib enhances T-cell-mediated killing of patient-derived human melanoma cells by their autologous T cells in vitro and potentiates responses to anti-CTLA4 and anti-PD1 therapy in vivo. Mechanistic studies reveal that HSP90 inhibition results in upregulation of interferon response genes, which are essential for the enhanced killing of ganetespib treated melanoma cells by T cells. Taken together, these findings provide evidence that HSP90 inhibition can potentiate T-cell-mediated anti-tumor immune responses, and rationale to explore the combination of immunotherapy and HSP90 inhibitors.Many patients fail to respond to T cell based immunotherapies. Here, the authors, through a high-throughput screening, identify HSP90 inhibitors as a class of preferred drugs for treatment combination with immunotherapy.

Mass spectrometry based proteomics for absolute quantification of proteins from tumor cells.

In-depth quantitative profiling of the proteome and sub-proteomes of tumor cells has relevance to tumor classification, the development of novel therapeutics, and of prognostic and predictive markers and to disease monitoring. In particular the tumor cell surface represents a highly relevant compartment for the development of targeted therapeutics and immunotherapy. We have developed a proteomic platform to profile tumor cells that encompasses enrichment of surface membrane proteins, intact protein fractionation and label-free mass spectrometry based absolute quantification. Here we describe the methodology for capture, identification and quantification of cell surface proteins using biotinylation for labeling of the cell surface, avidin for capture of biotinylated proteins and ion mobility mass spectrometry for protein identification and quantification.

Mining the pre-diagnostic antibody repertoire of TgMMTV-neu mice to identify autoantibodies useful for the early detection of human breast cancer.

BACKGROUND: The use of autoantibodies for the early detection of breast cancer has generated much interest as antibodies can be readily assayed in serum when antigen levels are low. Ideally, diagnostic autoantibodies would be identified in individuals who harbored pre-invasive disease/high risk lesions leading to malignancy. Prospectively collected human serum samples from these individuals are rare and not often available for biomarker discovery. We questioned whether transgenic animals could be used to identify cancer-associated autoantibodies present at the earliest stages of the malignant transformation of breast cancer. METHODS: We collected sera from transgenic mice (TgMMTV-neu) from the time of birth to death by spontaneous mammary tumors. Using sera from a time point prior to the development of tumor, i.e. "pre-diagnostic", we probed cDNA libraries derived from syngeneic tumors to identify proteins recognized by IgG antibodies. Once antigens were identified, selected proteins were evaluated via protein arrays, for autoantibody responses using plasma from women obtained prior to the development of breast cancer and matched controls. The ability of the antigens to discriminate cases from controls was assessed using receiver-operating-characteristic curve analyses and estimates of the area under the curve. RESULTS: We identified 6 autoantibodies that were present in mice prior to the development of mammary cancer: Pdhx, Otud6b, Stk39, Zpf238, Lgals8, and Vps35. In rodent validation cohorts, detecting both IgM and IgG antibody responses against a subset of the identified proteins could discriminate pre-diagnostic sera from non-transgenic control sera with an AUC of 0.924. IgG and IgM autoantibodies, specific for a subset of the identified antigens, could discriminate the samples of women who eventually developed breast cancer from case-matched controls who did not develop disease. The discriminatory potential of the pre-diagnostic autoantibodies was enhanced if plasma samples were collected greater than 5 months prior to a breast cancer diagnosis (AUC 0.68; CI 0.565-0.787, p=0.0025). CONCLUSION: Genetically engineered mouse models of cancer may provide a facile discovery tool for identifying autoantibodies useful for human cancer diagnostics.

Proteomic profiling of the tumor microenvironment: recent insights and the search for biomarkers.

Although gain of oncogene functions and loss of tumor suppressor functions are driving forces in tumor development, the tumor microenvironment, comprising the extracellular matrix, surrounding stroma, signaling molecules and infiltrating immune and other cell populations, is now also recognized as crucial to tumor development and metastasis. Many interactions at the tumor cell-environment interface occur at the protein level. Proteomic approaches are contributing to the definition of the protein constituents of the microenvironment and their sources, modifications, interactions and turnover, as well as providing information on how these features relate to tumor development and progression. Recently, proteomic studies have revealed how cancer cells modulate the microenvironment through their secreted proteins and how they can alter their protein constituents to adapt to the microenvironment. Moreover, the release of proteins from the microenvironment into the circulatory system has relevance for the development of blood-based cancer diagnostics. Here, we review how proteomic approaches are being applied to studies of the tumor microenvironment to decipher tumor-stroma interactions and to elucidate the role of host cells in the tumor microenvironment.

Impact of protein stability, cellular localization, and abundance on proteomic detection of tumor-derived proteins in plasma.

Tumor-derived, circulating proteins are potentially useful as biomarkers for detection of cancer, for monitoring of disease progression, regression and recurrence, and for assessment of therapeutic response. Here we interrogated how a protein's stability, cellular localization, and abundance affect its observability in blood by mass-spectrometry-based proteomics techniques. We performed proteomic profiling on tumors and plasma from two different xenograft mouse models. A statistical analysis of this data revealed protein properties indicative of the detection level in plasma. Though 20% of the proteins identified in plasma were tumor-derived, only 5% of the proteins observed in the tumor tissue were found in plasma. Both intracellular and extracellular tumor proteins were observed in plasma; however, after normalizing for tumor abundance, extracellular proteins were seven times more likely to be detected. Although proteins that were more abundant in the tumor were also more likely to be observed in plasma, the relationship was nonlinear: Doubling the spectral count increased detection rate by only 50%. Many secreted proteins, even those with relatively low spectral count, were observed in plasma, but few low abundance intracellular proteins were observed. Proteins predicted to be stable by dipeptide composition were significantly more likely to be identified in plasma than less stable proteins. The number of tryptic peptides in a protein was not significantly related to the chance of a protein being observed in plasma. Quantitative comparison of large versus small tumors revealed that the abundance of proteins in plasma as measured by spectral count was associated with the tumor size, but the relationship was not one-to-one; a 3-fold decrease in tumor size resulted in a 16-fold decrease in protein abundance in plasma. This study provides quantitative support for a tumor-derived marker prioritization strategy that favors secreted and stable proteins over all but the most abundant intracellular proteins.

Progress in mining the human proteome for disease applications.

Currently available technologies allow in-depth analysis of multiple facets of the proteome that have clinical relevance and that complement current genomics-based approaches. Although some progress has been made in our knowledge of the human proteome in health and in disease, there is an urgent need to chart a coherent road map with clearly defined milestones to guide proteomics efforts. Areas of emphasis include: (1) building resources, (2) filling gaps in our understanding of biological variation, and (3) systematically characterizing proteome alterations that occur in well-defined disease states, all of which require an organized and collaborative effort.

A systems approach to the proteomic identification of novel cancer biomarkers.

The proteomics field has experienced rapid growth with technologies achieving ever increasing accuracy, sensitivity, and throughput, and with availability of computational tools to address particular applications. Given that the proteome represents the most functional component encoded for in the genome, a systems approach to disease investigations and biomarker identification benefits substantially from integration of proteome level studies. Here we present proteomic approaches that have allowed systematic searches for potential cancer markers by integrating cancer cell profiling with additional sources of data, as illustrated with recent studies of ovarian cancer.