Ovarian recurrence risk assessment using machine learning, clinical information, and serum protein levels to predict survival in high grade ovarian cancer.

In ovarian cancer, there is no current method to accurately predict recurrence after a complete response to chemotherapy. Here, we develop a machine learning risk score using serum proteomics for the prediction of early recurrence of ovarian cancer after initial treatment. The developed risk score was validated in an independent cohort with serum collected prospectively during the remission period. In the discovery cohort, patients scored as low-risk had a median time to recurrence (TTR) that was not reached at 10 years compared to 10.5 months (HR 4.66, p < 0.001) in high-risk patients. In the validation cohort, low-risk patients had a median TTR which was not reached compared to 4.7 months in high-risk patients (HR 4.67, p = 0.009). In advanced-stage patients with a CA125 < 10, low-risk patients had a median TTR of 68 months compared to 6 months in high-risk patients (HR 2.91, p = 0.02). The developed risk score was capable of distinguishing the duration of remission in ovarian cancer patients. This score may help guide maintenance therapy and develop innovative treatments in patients at risk at high-risk of recurrence.

High serum levels of inflammatory markers are associated with early recurrence in patients with high-grade serous ovarian cancer after platinum therapy.

OBJECTIVE: To determine if inflammatory biomarkers can predict the long-term outcome of platinum therapy in patients with high-grade serous ovarian cancer. METHODS: Women diagnosed with high-grade serous epithelial ovarian cancer (n = 70) at a single institution were enrolled in a prospective serum collection study between 2005 and 2020. Seventeen markers of inflammation and oxidative stress were measured in serum samples on a chemistry analyzer. Association was tested for serum levels with progression-free survival (PFS), time to recurrence (TTR), overall survival (OS), and time to death (TTD) using Cox proportional hazards and Kaplan-Meier curves. Patient survival was censored at 10 years. RESULTS: Higher serum levels of LDH were associated with worse PFS (HR 2.57, p = 0.028). High serum levels of BAP (HR 0.38, p = 0.025), GSP (HR 0.40, p = 0.040), HDL-c (HR 0.27, p = 0.002), and MG (HR 0.36, p = 0.017) were associated with improved PFS. Higher expression of LDH was associated with worse OS (HR 2.16, p = 0.023). Higher levels of CK.nac (HR 0.39, p = 0.033) and HDL-c (HR 0.35, p = 0.029) were associated with improved OS. Similar outcomes were found with TTR and TTD analyses. CONCLUSION: General inflammatory biomarkers may serve as a guide for prognosis and treatment benefit. Future studies needed to further define their role in predicting prognosis or how these markers may affect response to therapy.

Development of a Single Molecule Counting Assay to Differentiate Chromophobe Renal Cancer and Oncocytoma in Clinics.

Malignant chromophobe renal cancer (chRCC) and benign oncocytoma (RO) are two renal tumor types difficult to differentiate using histology and immunohistochemistry-based methods because of their similarity in appearance. We previously developed a transcriptomics-based classification pipeline with "Chromophobe-Oncocytoma Gene Signature" (COGS) on a single-molecule counting platform. Renal cancer patients (n = 32, chRCC = 17, RO = 15) were recruited from Augusta University Medical Center (AUMC). Formalin-fixed paraffin-embedded (FFPE) blocks from their excised tumors were collected. We created a custom single-molecule counting code set for COGS to assay RNA from FFPE blocks. Utilizing hematoxylin-eosin stain, pathologists were able to correctly classify these tumor types (91.8%). Our unsupervised learning with UMAP (Uniform manifold approximation and projection, accuracy = 0.97) and hierarchical clustering (accuracy = 1.0) identified two clusters congruent with their histology. We next developed and compared four supervised models (random forest, support vector machine, generalized linear model with L2 regularization, and supervised UMAP). Supervised UMAP has shown to classify all the cases correctly (sensitivity = 1, specificity = 1, accuracy = 1) followed by random forest models (sensitivity = 0.84, specificity = 1, accuracy = 1). This pipeline can be used as a clinical tool by pathologists to differentiate chRCC from RO.

Serum Levels of Inflammatory Proteins Are Associated With Peripheral Neuropathy in a Cross-Sectional Type-1 Diabetes Cohort.

Chronic low-grade inflammation is involved in the pathogenesis of type-1 diabetes (T1D) and its complications. In this cross-section study design, we investigated association between serum levels of soluble cytokine receptors with presence of peripheral neuropathy in 694 type-1 diabetes patients. Sex, age, blood pressure, smoking, alcohol intake, HbA1c and lipid profile, presence of DPN (peripheral and autonomic), retinopathy and nephropathy was obtained from patient's chart. Measurement of soluble cytokine receptors, markers of systemic and vascular inflammation was done using multiplex immunoassays. Serum levels were elevated in in DPN patients, independent of gender, age and duration of diabetes. Crude odds ratios were significantly associated with presence of DPN for 15/22 proteins. The Odds ratio (OR) remained unchanged for sTNFRI (1.72, p=0.00001), sTNFRII (1.45, p=0.0027), sIL2Rα (1.40, p=0.0023), IGFBP6 (1.51, p=0.0032) and CRP (1.47, p=0.0046) after adjusting for confounding variables, HbA1C, hypertension and dyslipidemia. Further we showed risk of DPN is associated with increase in serum levels of sTNFRI (OR=11.2, p<10), sIL2Rα (8.69, p<10-15), sNTFRII (4.8, p<10-8) and MMP2 (4.5, p<10-5). We combined the serum concentration using ridge regression, into a composite score, which can stratify the DPN patients into low, medium and high-risk groups. Our results here show activation of inflammatory pathway in DPN patients, and could be a potential clinical tool to identify T1D patients for therapeutic intervention of anti-inflammatory therapies.

Retrospective Validation of a 168-Gene Expression Signature for Glioma Classification on a Single Molecule Counting Platform.

Gene expression profiling has been shown to be comparable to other molecular methods for glioma classification. We sought to validate a gene-expression based glioma classification method. Formalin-fixed paraffin embedded tissue and flash frozen tissue collected at the Augusta University (AU) Pathology Department between 2000-2019 were identified and 2 mm cores were taken. The RNA was extracted from these cores after deparaffinization and bead homogenization. One hundred sixty-eight genes were evaluated in the RNA samples on the nCounter instrument. Forty-eight gliomas were classified using a supervised learning algorithm trained by using data from The Cancer Genome Atlas. An ensemble of 1000 linear support vector models classified 30 glioma samples into TP1 with classification confidence of 0.99. Glioma patients in TP1 group have a poorer survival (HR (95% CI) = 4.5 (1.3-15.4), p = 0.005) with median survival time of 12.1 months, compared to non-TP1 groups. Network analysis revealed that cell cycle genes play an important role in distinguishing TP1 from non-TP1 cases and that these genes may play an important role in glioma survival. This could be a good clinical pipeline for molecular classification of gliomas.

Senescence-Associated Secretory Phenotype Determines Survival and Therapeutic Response in Cervical Cancer.

Molecular biomarkers that can predict survival and therapeutic outcome are still lacking for cervical cancer. Here we measured a panel of 19 serum proteins in sera from 565 patients with stage II or III cervical cancer and identified 10 proteins that have an impact on disease specific survival (DSS) (Hazzard's ratio; HR = 1.51-2.1). Surprisingly, all ten proteins are implicated in senescence-associated secreted phenotype (SASP), a hallmark of cellular senescence. Machine learning using Ridge regression of these SASP proteins can robustly stratify patients with high SASP, which is associated with poor survival, and patients with low SASP associated with good survival (HR = 3.09-4.52). Furthermore, brachytherapy, an effective therapy for cervical cancer, greatly improves survival in SASP-high patients (HR = 3.3, p < 5 × 10-5) but has little impact on survival of SASP-low patients (HR = 1.5, p = 0.31). These results demonstrate that cellular senescence is a major determining factor for survival and therapeutic response in cervical cancer and suggest that senescence reduction therapy may be an efficacious strategy to improve the therapeutic outcome of cervical cancer.

A 73-gene proliferative transcriptomic signature predicts uterine serous carcinoma patient survival and response to primary therapy.

OBJECTIVES: To develop a transcriptomic signature capable of predicting overall survival (OS) for uterine serous carcinoma (USC). METHODS: RNAseq data for 58 USC patients were obtained from TCGA. Expression of 73 candidate genes was measured for 67 Augusta University (AU) samples using NanoString technology. RESULTS: Analysis of the TCGA RNAseq data identified 73 genes that individually predict prognosis for USC patients and an elastic net model with all 73 genes (USC73) distinguishes a good OS group with low USC73 score from a poor OS group with high USC73 score (5-year OS = 83.3% and 13.3% respectively, HR = 40.1; p = 3 × 10-8). This finding was validated in the independent AU cohort (HR = 4.3; p = 0.0004). The poor prognosis group with high USC73 score consists of 37.9% and 32.8% of patients in the TCGA and AU cohort respectively. USC73 score and pathologic stage independently contribute to OS and together provide the best prognostic value. Early stage, low USC73 patients have the best prognosis (5-year OS = 85.1% in the combined dataset), while advanced stage, high USC73 patients have the worst prognosis (5-year OS = 6.4%, HR = 30.5, p = 1.2 × 10-12). Consistent with the observed poor survival, primary cell cultures from high USC73 patients had higher proliferation rate and cell cycle progression; and high USC73 patients had lower rates of complete response to standard therapy. CONCLUSIONS: The USC73 transcriptomic signature and stage independently predict OS of USC patients and the best prediction is achieved using USC73 and stage. USC73 may also serve as a therapeutic biomarker to guide patient care.

Better survival is observed in cervical cancer patients positive for specific anti-glycan antibodies and receiving brachytherapy.

OBJECTIVE: To measure anti-glycan antibodies (AGA) in cervical cancer (CC) patient sera and assess their effect on therapeutic outcome. PATIENTS AND METHODS: Serum AGA was measured in 276 stage II and 292 stage III Peruvian CC patients using a high content and throughput Luminex multiplex glycan array (LMGA) containing 177 glycans. Association with disease-specific survival (DSS) and progression free survival (PFS) were analyzed using Cox regression. RESULTS: AGAs were detected against 50 (28.3%) of the 177 glycans assayed. Of the 568 patients, 84.5% received external beam radiation therapy (EBRT) plus brachytherapy (BT), while 15.5% only received EBRT. For stage-matched patients (Stage III), receiving EBRT alone was significantly associated with worse survival (HR 6.4, p < 0.001). Stage III patients have significantly worse survival than Stage II patients after matching for treatment (HR = 2.8 in EBRT+BT treatment group). Furthermore, better PFS and DSS were observed in patients positive for AGA against multiple glycans belonging to the blood group H, Lewis, Ganglio, Isoglobo, lacto and sialylated tetrarose antigens (best HR = 0.49, best p = 0.0008). CONCLUSIONS: Better PFS and DSS are observed in cervical cancer patients that are positive for specific antiglycan antibodies and received brachytherapy.

Proteins of TNF-α and IL6 Pathways Are Elevated in Serum of Type-1 Diabetes Patients with Microalbuminuria.

Soluble cytokine receptors may play an important role in development of microalbuminuria (MA) in type-1 diabetes (T1D). In this study, we measured 12 soluble receptors and ligands from TNF-α/IL6/IL2 pathways in T1D patients with MA (n = 89) and T1D patients without MA (n = 483) participating in the PAGODA study. Twelve proteins in the sera from T1D patients with and without MA were measured using multiplex Luminex assays. Ten serum proteins (sTNFR1, sTNFR2, sIL2Rα, MMP2, sgp130, sVCAM1, sIL6R, SAA, CRP, and sICAM1) were significantly elevated in T1D patients with MA. After adjusting for age, duration of diabetes, and sex in logistic regression, association remained significant for seven proteins. MA is associated with increasing concentrations of all 10 proteins, with the strongest associations observed for sTNFR1 (OR = 108.3, P < 10-32) and sTNFR2 (OR = 65.5, P < 10-37), followed by sIL2Rα (OR = 12.9, P < 10-13), MMP2 (OR = 5.5, P < 10-6), sgp130 (OR = 5.2, P < 10-3), sIL6R (OR = 4.6, P < 10-4), and sVCAM1 (OR = 3.3, P < 10-4). We developed a risk score system based on the combined odds ratios associated with each quintile for each protein. The risk scores cluster MA patients into three subsets, each associated with distinct risk for MA attributable to proteins in the TNF-α/IL6 pathway (mean OR = 1, 13.5, and 126.3 for the three subsets, respectively). Our results suggest that the TNF-α/IL6 pathway is overactive in approximately 40% of the MA patients and moderately elevated in the middle 40% of the MA patients. Our results suggest the existence of distinct subsets of MA patients identifiable by their serum protein profiles.

Elevated Serum Levels of Soluble TNF Receptors and Adhesion Molecules Are Associated with Diabetic Retinopathy in Patients with Type-1 Diabetes.

AIMS: To examine the association of the serum levels of TNF receptors, adhesion molecules, and inflammatory mediators with diabetic retinopathy (DR) in T1D patients. METHODS: Using the multiplex immunoassay, we measured serum levels of eight proteins in 678 T1D subjects aged 20-75 years. Comparisons were made between 482 T1D patients with no complications and 196 T1D patients with DR. RESULTS: The levels of sTNFR-I, sTNFR-II, CRP, SAA, sgp130, sIL6R, sVCAM1, and sICAM1 were significantly higher in the T1D patients with DR as compared to T1D patients with no complications. Multivariate logistic regression analysis revealed significant association for five proteins after adjustment for age, sex, and disease duration (sTNFR-I: OR = 1.57, sgp130: OR = 1.43, sVCAM1: OR = 1.27, sICAM1: OR = 1.42, and CRP: OR = 1.15). Conditional logistic regression on matched paired data revealed that subjects in the top quartile for sTNFR-I (OR = 2.13), sTNFR-II (OR = 1.66), sgp130 (OR = 1.82), sIL6R (OR = 1.75), sVCAM1 (OR = 1.98), sICAM1 (OR = 2.23), CRP (OR = 2.40) and SAA (OR = 2.03), had the highest odds of having DR. CONCLUSIONS: The circulating markers of inflammation, endothelial injury, and TNF signaling are significantly associated with DR in patients with T1D. TNFR-I and TNFR-II receptors are highly correlated, but DR associated more strongly with TNFR-I in these patients.

Serum protein profile at remission can accurately assess therapeutic outcomes and survival for serous ovarian cancer.

BACKGROUND: Biomarkers play critical roles in early detection, diagnosis and monitoring of therapeutic outcome and recurrence of cancer. Previous biomarker research on ovarian cancer (OC) has mostly focused on the discovery and validation of diagnostic biomarkers. The primary purpose of this study is to identify serum biomarkers for prognosis and therapeutic outcomes of ovarian cancer. EXPERIMENTAL DESIGN: Forty serum proteins were analyzed in 70 serum samples from healthy controls (HC) and 101 serum samples from serous OC patients at three different disease phases: post diagnosis (PD), remission (RM) and recurrence (RC). The utility of serum proteins as OC biomarkers was evaluated using a variety of statistical methods including survival analysis. RESULTS: Ten serum proteins (PDGF-AB/BB, PDGF-AA, CRP, sFas, CA125, SAA, sTNFRII, sIL-6R, IGFBP6 and MDC) have individually good area-under-the-curve (AUC) values (AUC = 0.69-0.86) and more than 10 three-marker combinations have excellent AUC values (0.91-0.93) in distinguishing active cancer samples (PD & RC) from HC. The mean serum protein levels for RM samples are usually intermediate between HC and OC patients with active cancer (PD & RC). Most importantly, five proteins (sICAM1, RANTES, sgp130, sTNFR-II and sVCAM1) measured at remission can classify, individually and in combination, serous OC patients into two subsets with significantly different overall survival (best HR = 17, p<10(-3)). CONCLUSION: We identified five serum proteins which, when measured at remission, can accurately predict the overall survival of serous OC patients, suggesting that they may be useful for monitoring the therapeutic outcomes for ovarian cancer.

The expression of inflammatory genes is upregulated in peripheral blood of patients with type 1 diabetes.

OBJECTIVE: Our previous gene expression microarray studies identified a number of genes differentially expressed in patients with type 1 diabetes (T1D) and islet autoantibody-positive subjects. This study was designed to validate these gene expression changes in T1D patients and to identify gene expression changes in diabetes complications. RESEARCH DESIGH AND METHODS: We performed high-throughput real-time RT-PCR to validate gene expression changes in peripheral blood mononuclear cells (PBMCs) from a large sample set of 928 T1D patients and 922 control subjects. RESULTS: Of the 18 genes analyzed here, eight genes (S100A8, S100A9, MNDA, SELL, TGFB1, PSMB3, CD74, and IL12A) had higher expression and three genes (GNLY, PSMA4, and SMAD7) had lower expression in T1D patients compared with control subjects, indicating that genes involved in inflammation, immune regulation, and antigen processing and presentation are significantly altered in PBMCs from T1D patients. Furthermore, one adhesion molecule (SELL) and three inflammatory genes mainly expressed by myeloid cells (S100A8, S100A9, and MNDA) were significantly higher in T1D patients with complications (odds ratio [OR] 1.3-2.6, adjusted P value = 0.005-10(-8)), especially those patients with neuropathy (OR 4.8-7.9, adjusted P value <0.005). CONCLUSIONS: These findings suggest that inflammatory mediators secreted mainly by myeloid cells are implicated in T1D and its complications.

Genetically dependent ERBB3 expression modulates antigen presenting cell function and type 1 diabetes risk.

Type 1 diabetes (T1D) is an autoimmune disease resulting from the complex interaction between multiple susceptibility genes, environmental factors and the immune system. Over 40 T1D susceptibility regions have been suggested by recent genome-wide association studies; however, the specific genes and their role in the disease remain elusive. The objective of this study is to identify the susceptibility gene(s) in the 12q13 region and investigate the functional link to the disease pathogenesis. A total of 19 SNPs in the 12q13 region were analyzed by the TaqMan assay for 1,434 T1D patients and 1,865 controls. Thirteen of the SNPs are associated with T1D (best p = 4x10(-11)), thus providing confirmatory evidence for at least one susceptibility gene in this region. To identify candidate genes, expression of six genes in the region was analyzed by real-time RT-PCR for PBMCs from 192 T1D patients and 192 controls. SNP genotypes in the 12q13 region are the main factors that determine ERBB3 mRNA levels in PBMCs. The protective genotypes for T1D are associated with higher ERBB3 mRNA level (p<10(-10)). Furthermore, ERBB3 protein is expressed on the surface of CD11c(+) cells (dendritic cells and monocytes) in peripheral blood after stimulation with LPS, polyI:C or CpG. Subjects with protective genotypes have significantly higher percentages of ERBB3(+) monocytes and dendritic cells (p = 1.1x10(-9)); and the percentages of ERBB3(+) cells positively correlate with the ability of APC to stimulate T cell proliferation (R(2) = 0.90, p<0.0001). Our results indicate that ERBB3 plays a critical role in determining APC function and potentially T1D pathogenesis.

IFIH1 polymorphisms are significantly associated with type 1 diabetes and IFIH1 gene expression in peripheral blood mononuclear cells.

Genome-wide association (GWA) studies revealed a number of single nucleotide polymorphisms (SNPs) significantly associated with type 1 diabetes (T1D). In an attempt to confirm some of these candidate associations, we genotyped 2046 Caucasian patients and 2417 normal controls from the United States for SNPs in five genomic regions. While no evidence was obtained for four genomic regions (rs2929366/NM_144715 on chromosome 3, rs9127/Q7Z4C4 on chromosome 5, rs1445898/CAPSL on chromosome 5 and rs2302188/NM_033543 on chromosome 19), we provide strong evidence for association between T1D and multiple SNPs in the IFIH1 linkage disequilibrium (LD) block on chromosome 2q. Among the 10 SNPs genotyped for the 2q region, four SNPs located within the IFIH1 gene or at the 5' region of IFIH1 showed significant association with T1D in the Georgia population [odds ratio (OR) = 1.7-1.9] with the best P-value found at SNP rs1990760 (P = 8 x 10(-8) and OR = 1.9). Several SNPs outside of the IFIH1 gene also showed significant but weaker associations. Furthermore, IFIH1 gene expression levels in peripheral blood mononuclear cells are significantly correlated with IFIH1 genotypes, and higher IFIH1 levels are found in individuals with the susceptible genotypes (P = 0.005). Thus, both genetic association and gene expression data suggest that IFIH1 is the most plausible candidate gene implicated in T1D in this LD block.

The risk factors for deep and superficial chest surgical-site infections after coronary artery bypass graft surgery are different.

OBJECTIVE: We sought to determine risk factors for deep and superficial chest wound infections after coronary artery bypass graft surgery to develop predictive models. METHODS: We retrospectively analyzed data collected on 1980 consecutive patients undergoing coronary artery bypass surgery at our institution between January 1, 1996, and June 30, 1999, by using the Society of Thoracic Surgery database. Independent risk factors for surgical-site infection were identified with multivariate logistic regression. RESULTS: There were 37 (1.9%) deep chest and 46 (2.3%) superficial chest surgical-site infections. Obese diabetic patients had a 7.7-fold increased risk of deep chest infections after controlling for intra-aortic balloon pump use (odds ratio, 3.1) and postoperative transfusion (odds ratio, 2.3). Independent risk factors for superficial surgical-site infections included obesity (odds ratio, 3.1), diabetes in persons 65 years of age or older (odds ratio, 2.7), and current smoking (odds ratio, 2.5). Use of antiplatelet drugs was associated with a lower risk of superficial infections (odds ratio, 0.4). Predicted operative mortality as a marker of severity of illness was not clearly predictive of deep or superficial surgical-site infection. Mortality in the year after the operation was increased in patients with deep chest infections compared with that seen in uninfected control subjects (8/37 [21.6%] vs 114/1612 [7.1%], P =.004) but not in patients with superficial chest infections (7/47 [15.2%] vs 114/1612 [7.1%], P =.075). CONCLUSIONS: Risk factors for deep and superficial chest surgical-site infections after coronary artery bypass surgery differ, suggesting different mechanisms of pathogenesis. Appropriate risk stratification models specific to these important outcomes must be developed.