Blinded Cohort Research Articles

Population-Based Prostate Cancer Screening With Magnetic Resonance Imaging or Ultrasonography

Screening for prostate cancer using prostate-specific antigen (PSA) testing can lead to problems of underdiagnosis and overdiagnosis. Short, noncontrast magnetic resonance imaging (MRI) or transrectal ultrasonography might overcome these limitations. To compare the performance of PSA testing, MRI, and ultrasonography as screening tests for prostate cancer. This prospective, population-based, blinded cohort study was conducted at 7 primary care practices and 2 imaging centers in the United Kingdom. Men 50 to 69 years of age were invited for prostate cancer screening from October 10, 2018, to May 15, 2019. All participants underwent screening with a PSA test, MRI (T2 weighted and diffusion), and ultrasonography (B-mode and shear wave elastography). The tests were independently interpreted without knowledge of other results. Both imaging tests were reported on a validated 5-point scale of suspicion. If any test result was positive, a systematic 12-core biopsy was performed. Additional image fusion-targeted biopsies were performed if the MRI or ultrasonography results were positive. The main outcome was the proportion of men with positive MRI or ultrasonography (defined as a score of 3-5 or 4-5) or PSA test (defined as PSA ≥3 μg/L) results. Key secondary outcomes were the number of clinically significant and clinically insignificant cancers detected if each test was used exclusively. Clinically significant cancer was defined as any Gleason score of 3+4 or higher. A total of 2034 men were invited to participate; of 411 who attended screening, 408 consented to receive all screening tests. The proportion with positive MRI results (score, 3-5) was higher than the proportion with positive PSA test results (72 [17.7%; 95% CI, 14.3%-21.8%] vs 40 [9.9%; 95% CI, 7.3%-13.2%]; P < .001). The proportion with positive ultrasonography results (score, 3-5) was also higher than the proportion of those with positive PSA test results (96 [23.7%; 95% CI, 19.8%-28.1%]; P < .001). For an imaging threshold of score 4 to 5, the proportion with positive MRI results was similar to the proportion with positive PSA test results (43 [10.6%; 95% CI, 7.9%-14.0%]; P = .71), as was the proportion with positive ultrasonography results (52 [12.8%; 95% CI, 9.9%-16.5%]; P = .15). The PSA test (≥3 ng/mL) detected 7 clinically significant cancers, an MRI score of 3 to 5 detected 14 cancers, an MRI score of 4 to 5 detected 11 cancers, an ultrasonography score of 3 to 5 detected 9 cancer, and an ultrasonography score of 4 to 5 detected 4 cancers. Clinically insignificant cancers were diagnosed by PSA testing in 6 cases, by an MRI score of 3 to 5 in 7 cases, an MRI score of 4 to 5 in 5 cases, an ultrasonography score of 3 to 5 in 13 cases, and an ultrasonography score of 4 to 5 in 7 cases. In this cohort study, when screening the general population for prostate cancer, MRI using a score of 4 or 5 to define a positive test result compared with PSA alone at 3 ng/mL or higher was associated with more men diagnosed with clinically significant cancer, without an increase in the number of men advised to undergo biopsy or overdiagnosed with clinically insignificant cancer. There was no evidence that ultrasonography would have better performance compared with PSA testing alone.

TMT and PRM-Based Quantitative Proteomics Identify Potential Biomarkers for Behçet Syndrome

Background: Development of reliable, noninvasive clinical markers is crucial for early diagnosis of Behcet syndrome (BS). In this study, we aimed to identify potential serum biomarkers and construct a diagnostic panel for BS. Methods: Sera from 26 BS patients and 26 healthy controls (HCs) were collected and candidate biomarkers were explored using an untargeted shotgun proteomics approach based on tandem mass tags (TMTs) and a parallel reaction monitoring (PRM) analysis. ELISAs were further used for quantification of identified markers in 40 BS, 41 HCs, 20 rheumatoid arthritis (RA) and 20 Systemic lupus erythematosus (SLE). A random forest model and logistic regression algorithm were used for predictive model construction and validation. Results: TMT analysis identified 55 serum proteins differentially expressed between HC and BS patients. Using PRM and ELISA, we further identified 2 proteins which were significantly expressed between HC, BS, RA and SLE patients. By random forest modeling, 2 proteins TPM4, FLNA and individual age were selected as a diagnostic panel. A predictive model was built by logistic regression algorithm, which yielded area under the curve (AUC) values of 86.2%. The corresponding nomogram calibration curves showed good consistency. Additionally, this combinational diagnostic panel showed good predictive performance with an AUC of 0.74, specificity of 84.6%, sensitivity of 68.8% and accuracy of 75.8% in a blinded validation cohort. Conclusions: BS patients have distinct differences in serum proteomics profile compared to HCs. A predictive panel of TPM4, FLNA and age maybe used for BS diagnosis after validation in larger multiethnic cohorts. Clinical Trial: This study was registered on Chinese Clinical Trial Registry(ChiCTR2100048988). Funding: This research was funded by the National Natural Science Foundation of China (No. 81770101, 81403041), 1.3.5 project for disciplines of excellence, West China Hospital, Sichuan University (Project No. ZYJC18024; ZYJC18003; ZYGD18015). Declaration of Interest: None to declare. Ethical Approval: The study procedure was approved by the Biomedical Research Ethics Committee, West China Hospital of Sichuan University (No.593, 2019）

TERT Promoter Mutation Analysis for Blood-Based Diagnosis and Monitoring of Gliomas.

Liquid biopsy offers a minimally invasive tool to diagnose and monitor the heterogeneous molecular landscape of tumors over time and therapy. Detection of TERT promoter mutations (C228T, C250T) in cfDNA has been successful for some systemic cancers but has yet to be demonstrated in gliomas, despite the high prevalence of these mutations in glioma tissue (>60% of all tumors). Here, we developed a novel digital droplet PCR (ddPCR) assay that incorporates features to improve sensitivity and allows for the simultaneous detection and longitudinal monitoring of two TERT promoter mutations (C228T and C250T) in cfDNA from the plasma of patients with glioma. In baseline performance in tumor tissue, the assay had perfect concordance with an independently performed clinical pathology laboratory assessment of TERT promoter mutations in the same tumor samples [95% confidence interval (CI), 94%-100%]. Extending to matched plasma samples, we detected TERT mutations in both discovery and blinded multi-institution validation cohorts with an overall sensitivity of 62.5% (95% CI, 52%-73%) and a specificity of 90% (95% CI, 80%-96%) compared with the gold-standard tumor tissue-based detection of TERT mutations. Upon longitudinal monitoring in 5 patients, we report that peripheral TERT-mutant allele frequency reflects the clinical course of the disease, with levels decreasing after surgical intervention and therapy and increasing with tumor progression. Our results demonstrate the feasibility of detecting circulating cfDNA TERT promoter mutations in patients with glioma with clinically relevant sensitivity and specificity.

Glottic Keratosis: Significance and Identification of Laryngoscopic Findings.

ObjectiveGlottic keratosis poses a challenge because a decision to biopsy must weigh the likelihood of dysplasia and cancer against the voice outcome after biopsy. We determined the significance of laryngoscopic findings and agreement among clinicians to identify those specific findings.Study DesignRetrospective case-control study.SettingTertiary care university hospital.MethodsAdults with glottic keratosis with preoperative office laryngoscopies were included. Preoperative videostroboscopies were reviewed by a blinded reviewer. Multivariable logistic regression was used to examine the correlation between laryngoscopic appearance of glottic keratosis and presence or absence of high-grade dysplasia or carcinoma on biopsies. Consensus among head and neck cancer surgeons to detect specific laryngoscopic findings was evaluated by presenting representative laryngoscopies to a blinded cohort. Interrater reliability was calculated using Fleiss’s κ.ResultsSixty glottic keratotic lesions met inclusion criteria. On logistic regression, both erythroplakia and aberrant microvasculature like vascular speckling were significantly associated with high-grade dysplasia and carcinoma, P = .002 and P = .03, respectively. Interrater reliability among clinicians to identify erythroplakia and aberrant microvasculature was minimal, κ = 0.35 and κ = 0.29, respectively. Interrater reliability was improved with the use of virtual chromoendoscopy.ConclusionThe presence of erythroplakia and aberrant microvasculature in glottic keratosis is associated with the presence of high-grade dysplasia or carcinoma. Virtual chromoendoscopy can be used to improve reliability for detecting erythroplakia and vascular speckling, and this is a potential area for practice-based learning. Clinicians should identify and consider immediate diagnostic biopsy of suspicious glottic keratosis.

1229. Volatile Biomarkers of Influenza Infection in the Breath

BackgroundAnnual influenza epidemics cause significant morbidity and mortality. New, emerging strains threaten to cause catastrophic pandemics. Assay of exhaled breath for volatile organic compounds (VOCs) via gas chromatography-mass spectroscopy (GC-MS) is an emerging diagnostic modality ideally suited to fill the gap in influenza diagnostics.MethodsPatients with influenza like illness (ILI) presenting to the Troop Medical Clinic on JBSA Fort Sam Houston, TX, from 3/2017 to 3/2019 submitted a 2-minute breath sample in addition to a nasopharyngeal swab collected for polymerase chain reaction (PCR) assay for influenza virus. ILI was defined as temperature > 100.40F AND respiratory symptoms like cough, sputum production, chest pain and/or sore throat. Breath VOCs were assayed with GC-MS and data were analyzed in order to identify the significant breath VOC biomarkers that discriminated between ILI patients with and without a PCR assay positive for influenza with greater than random accuracy.ResultsDemographic, clinical, PCR and breath data were available for 237 episodes of ILI. PCR was positive for influenza for 32 episodes (30 influenza A and 2 B). The median age of participants was 21 (IQR 19, 23) and 69% were male. There were no differences in age, gender, education, race, or smoking, between the influenza positive and negative groups. Likewise, there was no difference in days of limited activity or missed work, or symptoms at presentation between the groups. The algorithm achieved near maximal predictive accuracy of 78% with four biomarkers (74% sensitivity and 70% specificity). Based on their mass spectra, these biomarker VOCs were tentatively identified as 2-amino-1-propanol, 2-butanamine, n-nitro, 3-methyl-hexanal, and heptane, which are consistent with products of oxidative stress.Figure. Accuracy, Sensitivity, and Specificity of Influenza Breath Test. Receiver operating characteristic (ROC) of the breath test (sensitivity versus 1-specificity). The accuracy of the breath test was 78%. With a cutoff point at the “shoulder” of the ROC curve, the test had 74% sensitivity and 70% specificity. ConclusionOur findings bolster available benchtop and clinical data suggesting that breath testing may be a useful diagnostic modality for influenza infection. The next step will be to study the predictive algorithm developed in this protocol in a blinded validation cohort. If the predictive algorithm performs well in a validation study, adaptation for its use in a portable, tabletop GC would be warranted to allow for a rapid, accurate, universal point-of-care influenza diagnostic test.DisclosuresMichael Phillips, MD, Menssana Research, Inc (Grant/Research Support)

Highly accurate diagnosis of papillary thyroid carcinomas based on personalized pathways coupled with machine learning.

Thyroid nodules are neoplasms commonly found among adults, with papillary thyroid carcinoma (PTC) being the most prevalent malignancy. However, current diagnostic methods often subject patients to unnecessary surgical burden. In this study, we developed and validated an automated, highly accurate multi-study-derived diagnostic model for PTCs using personalized biological pathways coupled with a sophisticated machine learning algorithm. Surprisingly, the algorithm achieved near-perfect performance in discriminating PTCs from non-tumoral thyroid samples with an overall cross-study-validated area under the receiver operating characteristic curve (AUROC) of 0.999 (95% confidence interval [CI]: 0.995-1) and a Brier score of 0.013 on three independent development cohorts. In addition, the algorithm showed excellent generalizability and transferability on two large-scale external blind PTC cohorts consisting of The Cancer Genome Atlas (TCGA), which is the largest genomic PTC cohort studied to date, and the post-Chernobyl cohort, which includes PTCs reported after exposure to radiation from the Chernobyl accident. When applied to the TCGA cohort, the model yielded an AUROC of 0.969 (95% CI: 0.950-0.987) and a Brier score of 0.109. On the post-Chernobyl cohort, it yielded an AUROC of 0.962 (95% CI: 0.918-1) and a Brier score of 0.073. This algorithm also is robust against other various types of clinical scenarios, discriminating malignant from benign lesions as well as clinically aggressive thyroid cancer with poor prognosis from indolent ones. Furthermore, we discovered novel pathway alterations and prognostic signatures for PTC, which can provide directions for follow-up studies.

A Multi-Institutional External Validation of a Deep-Learning Based Platform for Prediction of Outcomes following SBRT Treatment for Early-Stage Non-Small Cell Lung Cancer

Despite favorable tumor control of NSCLC with SBRT, some patients remain at high risk for treatment failure. Our group previously developed a deep learning (DL) platform that integrated clinical, radiographic, and dosimetric data to predict outcomes in patients with early-stage NSCLC at a single institution. This study aimed to externally validate the generalizability of DL based predictions on a blinded cohort of patients at a geographically distinct institution and assess model performance versus traditional statistical models. A DL algorithm was trained on a cohort of 527 patients with cT1-cT2 NSCLC treated with SBRT at a single-institution. The DL algorithm incorporated feed-forward neural network architecture to model clinical co-variates, 2D convolutional neural network (CNN) and long-short term memory recurrent neural network architectures to model radiographic data, and 1D CNN to model dosimetric data. Outcomes of interest were all-cause mortality as well as local, regional, distant, and overall treatment failures at 2-years. The DL model was tested with a blinded cohort of 62 patients treated at an external institution. Discriminatory ability of the DL model on the validation set was assessed using receiver operating characteristic curves (ROC) and compared to predictions generated from a traditional Cox Proportional Hazard model. Statistically significant differences in ROC curves were tested using the Delong method. The training and blinded validation cohorts showed no statistically significant differences in disease presentation, workup, or outcomes. When tested on the blinded external validation cohort, the DL platform maintained strong discriminatory ability in predicting all-cause mortality (AUC 0.73) as well as local (AUC 0.79), regional (AUC 0.69), distant (AUC 0.72), and overall (AUC 0.85) treatment failures. Discriminatory ability of the DL platform significantly outperformed the predictions generated from a traditional Cox Proportional Hazard model for all outcomes of interest (Table 1). A DL platform that effectively integrates clinical, imaging, and dosimetric data successfully models outcomes in early-stage NSCLC patients treated with SBRT on a blinded external validation cohort. The DL platform shows strong generalizability across multiple institutions and consistently outperforms predictions generated from a traditional Cox Proportional Hazard model. Integration of similar DL models that effectively analyze multiple data streams may improve clinical outcome prediction within clinical practice and guide personalization of therapy.Abstract 1103; Table2-Year OutcomesAUC DL Model95% Confidence IntervalAUC Cox Model95% Confidence IntervalDelong TestAny Treatment Failure.85.70-1.00.65.52-.78p<.001Local Failure.79.62-.96.58.47-.69p<.001Distant Failure.72.56-.88.60.47-.73p = .008Regional Failure.69.52-.86.61.51-.71p = .041All-Cause Mortality.73.61-.85.60.48-.72p<.001 Open table in a new tab

An ultra-high-density protein microarray for high throughput single-tier serological detection of Lyme disease

Current serological immunoassays have inherent limitations for certain infectious diseases such as Lyme disease, a bacterial infection caused by Borrelia burgdorferi in North America. Here we report a novel method of manufacturing high-density multiplexed protein microarrays with the capacity to detect low levels of antibodies accurately from small blood volumes in a fully automated system. A panel of multiple serological markers for Lyme disease are measured using a protein microarray system, Lyme Immunochip, in a single step but interpreted adhering to the standard two-tiered testing algorithm (enzyme immunoassay followed by Western blot). Furthermore, an enhanced IgM assay was supplemented to improve the test’s detection sensitivity for early Lyme disease. With a training cohort (n = 40) and a blinded validation cohort (n = 90) acquired from CDC, the Lyme Immunochip identified a higher proportion of Lyme disease patients than the two-tiered testing (82.4% vs 70.6% in the training set, 66.7% vs 60.0% in the validation set, respectively). Additionally, the Immunochip improved sensitivity to 100% while having a lower specificity of 95.2% using a set of investigational antigens which are being further evaluated with a large cohort of blinded samples from the CDC and Columbia University. This universal microarray platform provides an unprecedented opportunity to resolve a broad range of issues with diagnostic tests, including multiplexing, workflow simplicity, and reduced turnaround time and cost.

Rhinovirus Detection in the Nasopharynx of Children Undergoing Cardiac Surgery Is Not Associated With Longer PICU Length of Stay: Results of the Impact of Rhinovirus Infection After Cardiac Surgery in Kids (RISK) Study.

To determine whether children with asymptomatic carriage of rhinovirus in the nasopharynx before elective cardiac surgery have an increased risk of prolonged PICU length of stay. Prospective, single-center, blinded observational cohort study. PICU in a tertiary hospital in The Netherlands. Children under 12 years old undergoing elective cardiac surgery were enrolled in the study after informed consent of the parents/guardians. The parents/guardians filled out a questionnaire regarding respiratory symptoms. On the day of the operation, a nasopharyngeal swab was obtained. Clinical data were collected during PICU admission, and PICU/hospital length of stay were reported. If a patient was still intubated 3 days after operation, an additional nasopharyngeal swab was collected. Nasopharyngeal swabs were tested for rhinovirus and other respiratory viruses with polymerase chain reaction. Of the 163 included children, 74 (45%) tested rhinovirus positive. Rhinovirus-positive patients did not have a prolonged PICU length of stay (median 2 d each; p = 0.257). Rhinovirus-positive patients had a significantly shorter median hospital length of stay compared with rhinovirus-negative patients (8 vs 9 d, respectively; p = 0.006). Overall, 97 of the patients (60%) tested positive for one or more respiratory virus. Virus-positive patients had significantly shorter PICU and hospital length of stay, ventilatory support, and nonmechanical ventilation. Virus-negative patients had respiratory symptoms suspected for a respiratory infection more often. In 31% of the children, the parents reported mild upper respiratory complaints a day prior to the cardiac surgery, this was associated with postextubation stridor, but no other clinical outcome measures. Preoperative rhinovirus polymerase chain reaction positivity is not associated with prolonged PICU length of stay. Our findings do not support the use of routine polymerase chain reaction testing for respiratory viruses in asymptomatic children admitted for elective cardiac surgery.

Utility of a multigene testing for preoperative evaluation of indeterminate thyroid nodules: A prospective blinded single center study in China.

BackgroundThyroid nodules are highly prevalent, with fine‐needle aspiration (FNA) commonly used as the standard preoperative tool for their diagnosis. However, the method classifies some of the samples as indeterminate, leading to unnecessary surgery. In this study, we evaluated the value of next‐generation sequencing (NGS) for cancer diagnosis in indeterminate thyroid nodules.Materials and methodsWe performed a prospective, blinded cohort study on 189 patients, with 196 Bethesda III/IV nodules. Specifically, we analyzed DNA mutations and RNA fusions across the FNA samples using NGS, then reviewed follow‐up reports from 84 nodules following definitive surgery, to determine the assay performance.ResultsEnough DNA and RNA were obtained in 188 nodules, revealing mutations or fusions in 34.6% of them. The most frequently mutated genes were RAS, followed by BRAF V600E. Based on surgical pathology, 39% (33/84) and 4.8% (4/84) of the nodules were malignant and intermediate, respectively. According to the risk stratification criteria, 28 cases were categorized High‐Risk group, all of the resected nodules (n = 20) were malignant. Twenty‐four thyroid nodules were in the Low‐Risk group, 28.6% (4/14) surgically removed nodules were malignant. In the Benign‐Like category, 18.0% (9/50) were malignant. Five out of 13 nodules with benign mutations were resected, including SPOP, EZH1, and ZNF148, all of them were benign. If genetic alterations annotated with High‐Risk or Low‐Risk was considered as positive, and negative if Benign‐Like. Multigene testing revealed sensitivity, specificity, positive predictive values (PPV), and negative predictive value (NPV) of 73%, 80%, 71%, and 82%, respectively. In addition, if four intermediate nodules were counted as malignant, the PPV and NPV were 71% and 74%.ConclusionOur results allow for further stratification of Bethesda III/IV thyroid nodules based on the risk of their malignancy. SPOP, EZH1, and ZNF148 mutations may be used as benign markers.

Blood biomarkers indicate that the preclinical stages of Alzheimer's disease present overlapping molecular features

It is still debated whether non-specific preclinical symptoms of Alzheimer’s disease (AD) can have diagnostic relevance. We followed the evolution from cognitively normal to AD by NMR-based metabolomics of blood sera. Multivariate statistical analysis of the NMR profiles yielded models that discriminated subjective memory decline (SMD), mild cognitive impairment (MCI) and AD. We validated a panel of six statistically significant metabolites that predicted SMD, MCI and AD in a blind cohort with sensitivity values ranging from 88 to 95% and receiver operating characteristic values from 0.88 to 0.99. However, lower values of specificity, accuracy and precision were observed for the models involving SMD and MCI, which is in line with the pathological heterogeneity indicated by clinical data. This excludes a “linear” molecular evolution of the pathology, pointing to the presence of overlapping “gray-zones” due to the reciprocal interference of the intermediate stages. Yet, the clear difference observed in the metabolic pathways of each model suggests that pathway dysregulations could be investigated for diagnostic purposes.

Ultrasensitive high-resolution profiling of early seroconversion in patients with COVID-19.

Sensitive assays are essential for the accurate identification of individuals infected with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Here, we report a multiplexed assay for the fluorescence-based detection of seroconversion in infected individuals from less than 1 µl of blood, and as early as the day of the first positive nucleic acid test after symptom onset. The assay uses dye-encoded antigen-coated beads to quantify the levels of immunoglobulin G (IgG), IgM and IgA antibodies against four SARS-CoV-2 antigens. A logistic regression model trained using samples collected during the pandemic and samples collected from healthy individuals and patients with respiratory infections before the first outbreak of coronavirus disease 2019 (COVID-19) was 99% accurate in the detection of seroconversion in a blinded validation cohort of samples collected before the pandemic and from patients with COVID-19 five or more days after a positive nasopharyngeal test by PCR with reverse transcription. The high-throughput serological profiling of patients with COVID-19 allows for the interrogation of interactions between antibody isotypes and viral proteins, and should help us to understand the heterogeneity of clinical presentations.

Lipoarabinomannan antigenic epitope differences in tuberculosis disease subtypes

An accurate urine test for diverse populations with active tuberculosis could be transformative for preventing TB deaths. Urinary liporabinomannan (LAM) testing has been previously restricted to HIV co-infected TB patients. In this study we evaluate urinary LAM in HIV negative, pediatric and adult, pulmonary and extrapulmonary tuberculosis patients. We measured 430 microbiologically confirmed pretreatment tuberculosis patients and controls from Peru, Guinea Bissau, Venezuela, Uganda and the United States using three monoclonal antibodies, MoAb1, CS35, and A194, which recognize distinct LAM epitopes, a one-sided immunoassay, and blinded cohorts. We evaluated sources of assay variability and comorbidities (HIV and diabetes). All antibodies successfully discriminated TB positive from TB negative patients. ROAUC from the average of three antibodies’ responses was 0.90; 95% CI 0.87–0.93, 90% sensitivity, 73.5% specificity (80 pg/mL). MoAb1, recognizing the 5-methylthio-d-xylofuranose(MTX)-mannose(Man) cap epitope, performed the best, was less influenced by glycosuria and identified culture positive pediatric (N = 19) and extrapulmonary (N = 24) patients with high accuracy (ROAUC 0.87, 95% CI 0.77–0.98, 0.90 sensitivity 0.80 specificity at 80 pg/mL; ROAUC = 0.96, 95% CI 0.92–0.99, 96% sensitivity, 80% specificity at 82 pg/mL, respectively). The MoAb1 antibody, recognizing the MTX-Man cap epitope, is a novel analyte for active TB detection in pediatric and extrapulmonary disease.

Abstract 2316: Integrated genomic and epigenomic cell-free DNA (cfDNA) analysis for the detection of early-stage colorectal cancer

Abstract Introduction: A novel non-invasive blood-based cell-free DNA (cfDNA) analysis incorporating genomic and epigenomic assessment has demonstrated high sensitivity and specificity in patients with newly diagnosed colorectal cancer (CRC) (Kim, et al. CCR, 2019). Using an improved epigenomic analysis, we tested a new cohort of individuals with either a negative colonoscopy for advanced neoplasia (CRC or advanced adenoma) or newly diagnosed early-stage CRC. Methods: Whole blood samples were collected from 162 patients with a known diagnosis of CRC (pre-operative; 20 Stage I; 98 Stage II; 39 Stage II; 5 Stage IV), 38 self-declared cancer-free donors, and 205 individuals who were screen-negative for advanced neoplasia by colonoscopy. 5-8mL of plasma was isolated and total cfDNA was extracted and partitioned based on methylation level. Sequencing libraries were prepared and enriched using an integrated genomic and epigenomic cfDNA panel (Guardant Health, Redwood City, CA, USA). Sequencing results were used to analyze genomic, methylation, and fragmentomic signals. By using a thermodynamic model to approximate the physical binding process of methylation partitioning, modeling cfDNA fragmentation with increased resolution, and training on colonoscopy screen-negative samples, we improved the performance of the cfDNA epigenomic analysis. A training cohort of 117 colonoscopy screen-negative controls, 38 self-declared cancer-free controls, and 49 CRC patients (10 Stage I; 21 Stage II; 13 Stage III; 5 Stage IV) was used to train a linear model to combine these multimodal signals. The same cohort was used to establish the classification threshold prior to generating results for any of the validation samples. Results: The assay performance was tested on a blinded held-out cohort of 113 CRC patients (10 Stage I; 77 Stage II; 26 Stage III) and 88 colonoscopy screen-negative controls. Median age was 66 years (39-86) for CRC patients (52% female) and 57 years (20-84) in the colonoscopy screen-negative cohort (73% female). Overall sensitivity for CRC detection was 90.3% (90% Stage I; 88% Stage II; 96% Stage III) and specificity was 96.6%. We define a per sample ‘signal-to-noise-ratio' (SNR) score as the number of standard deviations (SD) from the mean model score observed in the colonoscopy screen-negative cohort. In these terms, our cfDNA detection threshold is 3.8 SD above the mean. The median SNR observed per stage was 6.8 SD (1.2 - 8.6) for Stage I, 6.9 SD (-0.18 - 8.6) for Stage II, and 7.3 SD (2.4 - 8.6) for Stage III. Conclusion: These results provide further support that an integrated genomic and epigenomic cfDNA assay consistently provides sufficient sensitivity and specificity for clinical detection of early-stage CRC. A prospective CRC screening study in a larger cohort of participants is needed to further validate assay performance. Citation Format: Oscar Westesson, Haley Axelrod, Jason Dean, Yupeng He, Paul Sample, Elena Zotenko, Ruth McCole, Mohit Goel, Charbel Eid, Jessica Kurata, Yu Kong, Anna Hartwig, Matthew Snyder, Will Greenleaf, Victoria M. Raymond, Darya Chudova, Ariel Jaimovich, AmirAli Talasaz. Integrated genomic and epigenomic cell-free DNA (cfDNA) analysis for the detection of early-stage colorectal cancer [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 2316.

Portable Automated Surveillance of Surgical Site Infections Using Natural Language Processing: Development and Validation.

We present the development and validation of a portable NLP approach for automated surveillance of SSIs. The surveillance of SSIs is labor-intensive limiting the generalizability and scalability of surgical quality surveillance programs. We abstracted patient clinical text notes after surgical procedures from 2 independent healthcare systems using different electronic healthcare records. An SSI detected as part of the American College of Surgeons' National Surgical Quality Improvement Program was used as the reference standard. We developed a rules-based NLP system (Easy Clinical Information Extractor [CIE]-SSI) for operative event-level detection of SSIs using an training cohort (4574 operative events) from 1 healthcare system and then conducted internal validation on a blind cohort from the same healthcare system (1850 operative events) and external validation on a blind cohort from the second healthcare system (15,360 operative events). EasyCIE-SSI performance was measured using sensitivity, specificity, and area under the receiver-operating-curve (AUC). The prevalence of SSI was 4% and 5% in the internal and external validation corpora. In internal validation, EasyCIE-SSI had a sensitivity, specificity, AUC of 94%, 88%, 0.912 for the detection of SSI, respectively. In external validation, EasyCIE-SSI had sensitivity, specificity, AUC of 79%, 92%, 0.852 for the detection of SSI, respectively. The sensitivity of EasyCIE-SSI decreased in clean, skin/subcutaneous, and outpatient procedures in the external validation compared to internal validation. Automated surveillance of SSIs can be achieved using NLP of clinical notes with high sensitivity and specificity.

Optimal level of bispectral index for conscious sedation in awake fiberoptic nasotracheal intubation

BackgroundBispectral index (BIS) has been in practice as an objective tool to measure the depth of conscious sedation. The data on adequate levels for awake fiberoptic nasotracheal intubation (AFNI) is however scarce. We aimed to obtain an optimal level of BIS required for achieving the adequate conscious sedation in AFNI procedure. MethodsIn a prospective, observational, outcome assessor blinded cohort trial, 94 consecutive patients with anticipated difficult intubation and undergoing AFNI for any elective surgery were enrolled. The topical anesthesia and sedation were induced with lidocaine and dexmedetomidine, while keeping the patient awake. The sedation levels were targeted to BIS ≤90 and Ramsay sedation score (RSS)≥2 to attempt intubation. Propofol bolus were administered if patients get agitated. The primary outcome was “Stable BIS”, the lowest BIS at which intubation could be feasibly performed. A receiver operator characteristic curve, Youden index, and correlation analysis were used. ResultsThe optimal criterion for BIS was obtained as ≤86 (AUC: 0.80, sensitivity 88.30%, specificity 61.45%, Youden index 49.74). The stable BIS ranged from 80 to 88, while RSS varied from 1 to 3. The BIS and RSS correlated significantly (r: 0.83). The stable BIS was independent of any baseline characteristics. Six patients had transient untoward events, none requiring any intervention. ConclusionBIS serves as an effective objective tool for titrating the depth of conscious sedation. We advocate a BIS range of 80–86 for feasibly performing the AFNI procedure.

Progressive rehabilitation of the sprained ankle: A novel treatment method

ObjectivesThis randomised, single blinded cohort study was designed to assess the immediate effect of manual fascial manipulation on walking pain and the range of ankle dorsiflexion within the first 4 days after ankle trauma. MethodsMeasurements were taken from 19 subjects, 5 female and 14 male, who presented with grade I–III ankle sprains. Ankle dorsiflexion was photographed in a standardised position and calculated by means of the Dartfish® Advanced Video Analysis Software and SPSS® (version 17) was used to compare the pre- and post-treatment data. ResultsAfter one treatment session 13 of the 19 subjects were walking pain free and 3 of the 19 where walking with only little pain. The highly significant (p<0.001) mean improvement of ankle dorsiflexion was 7.9° (±5.8°). All, apart from one subject, whom were walking pain free after treatment showed a minimum of 4° increased dorsiflexion. ConclusionEarly fascia work around the injured ankle improves ankle dorsiflexion and reduces walking pain. It may reduce the delay of tissue healing and, thus, optimise further rehabilitation of the sprained ankle which may also reduce socio-economic costs.

Developing and validating an approach for diagnosing and prognosticating cancer from biochemical motifs in T-cell receptors.

e15260 Background: Immune repertoire deep sequencing allows profiling T-cell populations and enables novel approaches to diagnose and prognosticate cancer by identifying T-cell receptor sequence patterns associated with clinical phenotypes and outcomes. Methods: Our goal is to develop a method to diagnose and prognosticate cancer using sequenced T-cell receptors. To determine how to profile the specificity of a T-cell receptor, we analyze 3D X-ray crystallographic structures of T-cell receptors bound to antigen. We observe a contiguous strip typically 4 amino acid residues in length from the complimentary determining region 3 (CDR3) lying in direct contact with the antigen. Based on this observation, we extract 4 residue long snippets from every receptor’s CDR3 and represent each snippet using biochemical features encoded by its amino acid sequence. The biochemical features are combined with information about the abundance of the snippet or the receptor and scored using a machine learning based approach. Each predictive model is fitted and validated under the requirement that at least one positively labelled snippet appears per tumor and no positively labelled snippets appear in healthy tissue. Results: Using a patient-holdout cross-validation, we fit predictive models to distinguish: 1. colorectal tumors from healthy tissue matched controls with 93% accuracy, 2. breast tumors from healthy tissue matched controls with 94% accuracy, 3. ovarian tumors from non-cancer patient ovarian tissue with 95% accuracy (80% accuracy on a blinded follow-up cohort) 4. and regression of preneoplastic cervical lesions over 1 year in advance with 96% accuracy. Conclusions: Immune repertoires can be used to diagnose and prognosticate cancer.

Developing &amp; Validating an Approach for Diagnosing and Prognosticating Cancer from Biochemical Motifs in T-cell Receptors

Abstract Immune repertoire deep sequencing allows profiling T-cell populations and enables novel approaches to diagnose and prognosticate diseases by identifying T-cell receptor sequence patterns associated with clinical phenotypes and outcomes. Our study objective is to develop a method to diagnose and prognosticate cancer using T-cell receptors sequenced from tissue biopsies. To determine how to profile the specificity of a T-cell receptor, we analyze 3D X-ray crystallographic structures of T-cell receptors bound to antigen. We observe a contiguous strip typically 4 amino acid residues in length from the complimentary determining region 3 (CDR3) lying in direct contact with the antigen. Based on this observation, we extract 4 residue long snippets from every receptor’s CDR3 and represent each snippet using biochemical features encoded by its amino acid sequence. The biochemical features are combined with information about the abundance of the snippet or the receptor and scored using a logistic regression model. Each logistic regression model is fitted and validated under the requirement that at least one positively labelled snippet appears per tumor and no positively labelled snippets appear in healthy tissue. Using a patient-holdout cross-validation, we fit logistic regression models to distinguish colorectal tumors from healthy tissue matched controls with 93% accuracy, breast tumors from healthy tissue matched controls with 94% accuracy, ovarian tumors from non-cancer patient ovarian tissue with 95% accuracy (80% accuracy on a blinded follow-up cohort), and regression of preneoplastic cervical lesions over 1 year in advance with 96% accuracy. In conclusion, immune repertoires can be used to diagnose and prognosticate disease.

SMN1 gene copy number analysis for spinal muscular atrophy (SMA) in a Turkish cohort by CODE-SEQ technology, an integrated solution for detection of SMN1 and SMN2 copy numbers and the "2+0" genotype.

Spinal muscular atrophy (SMA) is one of the common autosomal recessive disorders with global heterozygous carrier frequency of 1:50. Due to high carrier frequency, significant morbidity associated with the infantile onset disease and prohibitive cost of recently approved therapy, American College of Medical Genetics and Genomics (ACMG) recommends population based screening for SMA carrier status in eligible individuals in the reproductive age group. CODE-SEQ is a novel proprietary next generation sequencing (NGS) based assay, which is capable of detecting homozygous as well as heterozygous SMN1 exon 7 deletions. Along with the copy number estimation, this assay is capable of detecting single nucleotide polymorphisms (SNPs) associated with silent SMA carrier status or "2+0" genotype. We have validated a proprietary CODE-SEQ technology in a blinded cohort of 80 clinically well characterized samples from Turkish population for the detection of SMA carriers as well as affected cases. The results were correlated with gold standard MLPA assay. The copy numbers in exon 7 of SMN1 gene matched with MLPA results in all 80 samples giving 100% correlation. The assay accurately detected the presence/ absence of SNPs associated with "2+0" genotype in the reference samples. None of the tested clinical samples had these SNPs. The results of this study support the notion that CODE-SEQ will be extremely useful in detecting SMA genotypes in large-scale population-based screening studies.