165 Background: Academic institutions are adopting automated tools to help develop and maintain large databases. Leveraging advances in clinical data warehouses and artificial intelligence can potentially decrease resource utilization and the time required to conduct research by reducing intensive manual human curation of data. Methods: This research was performed as part of the institutional Data-Driven Determinants for COVID-19 Oncology Discovery Effort (D3CODE), IRB-approved protocol 2020-0348. Data were obtained from both structured and unstructured data sources within the MD Anderson electronic health record to validate a retrospective analysis for COVID-19 positive patients undergoing radiation treatments. Multiple data sources were identified, integrated and analyzed using the Syntropy Foundry platform, part of the Context Engine Data Management System at M.D. Anderson Cancer Center (MDACC) to include patient demographics, clinical notes, laboratory values, radiology reports, and oncology treatments. The platform allows real-time updating of these data-sources with integrated statistical packages to streamline rapid statistical analyses. We conducted a retrospective review of patients treated within 3 weeks of COVID infection between March 2020 and November 2022. Patient data was obtained using Foundry. Patients who tested positive for COVID-19 and had their treatment postponed were categorized as having a treatment delay. Patients who started radiation therapy and subsequently had an interruption in treatment due to contracting COVID-19 were categorized as having a treatment break. The Kaplan-Meier method was used to compare survival outcomes and was conducted within Foundry. Results: Syntropy Foundry helped develop a database comprising of 380 COVID-19 positive patients treated with radiation therapy between March 2020 through November 2022. It validated patient demographic information and research variables including patient age, histology, cancer stage, treatment plans, radiographic scans, recurrences, and follow up time that were separately performed manually. It reduced the time and human resources required to retrospectively collect and review these patient charts. Conclusions: Syntropy Foundry and ongoing efforts to leverage machine learning models to facilitate the interpretation of the large amount of accessible clinical data can potentially improve quality and reduce resources needed to generate large research databases. Future applications using artificial intelligence with Foundry to identify clinical recurrence and report patient outcome are under development.
Infection caused regarding vascular grafts in vascular surgery still remains a major problem. To reduce this problem and the complications which follow, the surgeon must be able to apply the best surgical management and also be confident with the vascular substitute used. There are two important factors to be considered: the biomaterial must have low propensity to infection and good stability if and when infected. In an attempt to verify this problem, 93 vascular grafts surgically excised for overt infection were examined. Techniques used for examinations were gross morphology, histopathology and scanning electron microscopy (SEM) evaluation. There were 23 human umbilical vein (HUV) grafts, 51 Dacron grafts and 19 expanded polytetrafluoroethylene (ePTFE) grafts. Histopathological signs of infection were absent in 57% of the ePTFE and Dacron grafts and in 17.4% of the HUV grafts. The latter were more heavily histologically infected and in some cases the walls were destructed. Histopathological signs of infection were seen on all the prosthetic walls in 36% of all the specimens and were mainly on the external portion of the grafts for the remaining prostheses. Bacteria were seen in respectively 21.7, 15.7 and 20% of the HUV, Dacron and ePTFE grafts with the Gram stain and in 86.9, 84.3 and 94.7% with SEM. The implantation period was shorter for the bioprostheses compared to that of the synthetic grafts because of the site and the indication of implantation. The stability of the bioprostheses was lower compared to that of the synthetic grafts when infected, leading to a breakdown of the wall along the length of the graft. The infection was found on the external capsule of the grafts rather than on the luminal surface.
816 Background: More than 44,000 patients are seen at MD Anderson Cancer Center annually. However, using the diverse, mostly unstructured, data from these patient encounters has been challenging and required manual chart reviews. In 2018, MD Anderson and Palantir Technologies (Denver, CO) began developing a unified, cloud-based, graphical user interface clinical informatics platform to extract, structure, and integrate data from the different data sources that comprise the electronic health record (EHR). Here, we describe our experience using this novel platform to apply a real-world evidence (RWE) approach to study patients with gastrointestinal (GI) malignancies. Methods: Institutional Review Board approval for retrospective chart review of patients with GI malignancies was previously obtained. The Foundry platform was used to incorporate more than 150 datasets, including structured data elements like lab values, unstructured data as the full text of clinical notes, and natural language processing (NLP) derived datasets. The datasets include unique patient identifiers to allow the merging of demographic, clinical, molecular, and outcomes information. The platform allows processing of the note text through NLP to extract non-discrete data elements into a discrete form. In addition, it continuously updates new data on daily bases, allowing the inclusion of new patients' information in an automated fashion. Results: From 2,013,048 patients with date of diagnosis ranging from 1944 to 2024, we have created datasets for colorectal adenocarcinoma (CRC, >50,000 patients, >8,000 with molecular data), pancreatic adenocarcinoma (PDAC, >13,000 patients), and appendiceal adenocarcinoma (AA, >3,000 patients). More than 50 variables have been integrated, including demographic information, stage, grade, overall survival, and molecular information. Focused manual validation of the automated extraction across the cohorts consistently demonstrated an accuracy of over 94%. Work is underway to extract additional features including DFS and PFS, sites of metastasis, and to build out additional cohorts for biliary tract, upper GI, and neuroendocrine tumors. Initial discovery efforts have already led to multiple publications including discovery of molecular causes of racial and ethnic disparities in CRC, survival impact of KRAS and co-mutations in PDAC, and prognostic utility of serum tumor markers in AA. Conclusions: Utilizing an automated, highly dynamic platform allowed integration of comprehensive datasets for multiparameter oncology data in patients with GI malignancies. This resulted in dramatic acceleration of cohort identification, outcomes analysis, and enabled utilizing a data-driven approach to guide decision making in an effort to enhance and optimize outcomes.
