Browsing by Author "Schmidt, C. Max"
Now showing 1 - 10 of 43
Results Per Page
Sort Options
Item Advanced natural language processing and temporal mining for clinical discovery(2015-08-17) Mehrabi, Saeed; Jones, Josette F.; Palakal, Mathew J.; Chien, Stanley Yung-Ping; Liu, Xiaowen; Schmidt, C. MaxThere has been vast and growing amount of healthcare data especially with the rapid adoption of electronic health records (EHRs) as a result of the HITECH act of 2009. It is estimated that around 80% of the clinical information resides in the unstructured narrative of an EHR. Recently, natural language processing (NLP) techniques have offered opportunities to extract information from unstructured clinical texts needed for various clinical applications. A popular method for enabling secondary uses of EHRs is information or concept extraction, a subtask of NLP that seeks to locate and classify elements within text based on the context. Extraction of clinical concepts without considering the context has many complications, including inaccurate diagnosis of patients and contamination of study cohorts. Identifying the negation status and whether a clinical concept belongs to patients or his family members are two of the challenges faced in context detection. A negation algorithm called Dependency Parser Negation (DEEPEN) has been developed in this research study by taking into account the dependency relationship between negation words and concepts within a sentence using the Stanford Dependency Parser. The study results demonstrate that DEEPEN, can reduce the number of incorrect negation assignment for patients with positive findings, and therefore improve the identification of patients with the target clinical findings in EHRs. Additionally, an NLP system consisting of section segmentation and relation discovery was developed to identify patients' family history. To assess the generalizability of the negation and family history algorithm, data from a different clinical institution was used in both algorithm evaluations.Item Assessment of a Universal Preprocedural Screening Program for COVID-19(Cambridge, 2021) Dbeibo, Lana; Kuebler, Kari; Keen, Alyson; George, Annie; Kelley, Kristen; Sadowski, Josh; Basham, Laura; Beeson, Terrie; Schmidt, C. Max; Beeler, Cole; Webb, Douglas; Medicine, School of MedicineObjectives Study objectives were to: (1) Determine the value of a COVID-19 universal preprocedural screening program; and (2) Using the results of asymptomatic positive screens, determine the safety of resuming elective procedures. Design This was a descriptive study detailing the process and findings from implementation of a COVID-19 universal preprocedural screening program. Setting An adult academic tertiary center in Indiana. Patients Patients were included in the analysis if they were screened 96 hours prior to or within 24 hours after undergoing a procedure in the operating room, cardiac catheterization lab, or endoscopy. Methods A report was generated from the electronic health record of patients undergoing procedures from a six week period of time (May 4th-June 14th, 2020). Health records for positive screens were reviewed and classified as symptomatic if they met either criteria: (1) screen performed due to presence of COVID-19 symptoms; (2) documentation of symptoms at the time of the screen. Patients with a positive screen that did not meet symptomatic criteria were classified as asymptomatic. Descriptive statistics were used to calculate frequencies and percentages for the included sample. Results The initial sample included 2,194 patients, comprised of 46 positive and 2,148 negative screens. Out of the 46 patients who had a positive test, 17 were asymptomatic, resulting in an asymptomatic rate of 0.79% (17/2165). Conclusion Findings validated the value of the program through identification of a low rate of asymptomatic positive screens and procedural team adoption and sustainment. Findings may help inform decision making of like organizations attempting to enhance safety while resuming elective procedures.Item Automated pancreatic cyst screening using natural language processing: a new tool in the early detection of pancreatic cancer(Elsevier, 2015-05) Roch, Alexandra M.; Mehrabi, Saeed; Krishnan, Anand; Schmidt, Heidi E.; Kesterson, Joseph; Beesley, Chris; Dexter, Paul R.; Palakal, Matthew; Schmidt, C. Max; Department of Surgery, IU School of MedicineINTRODUCTION: As many as 3% of computed tomography (CT) scans detect pancreatic cysts. Because pancreatic cysts are incidental, ubiquitous and poorly understood, follow-up is often not performed. Pancreatic cysts may have a significant malignant potential and their identification represents a 'window of opportunity' for the early detection of pancreatic cancer. The purpose of this study was to implement an automated Natural Language Processing (NLP)-based pancreatic cyst identification system. METHOD: A multidisciplinary team was assembled. NLP-based identification algorithms were developed based on key words commonly used by physicians to describe pancreatic cysts and programmed for automated search of electronic medical records. A pilot study was conducted prospectively in a single institution. RESULTS: From March to September 2013, 566,233 reports belonging to 50,669 patients were analysed. The mean number of patients reported with a pancreatic cyst was 88/month (range 78-98). The mean sensitivity and specificity were 99.9% and 98.8%, respectively. CONCLUSION: NLP is an effective tool to automatically identify patients with pancreatic cysts based on electronic medical records (EMR). This highly accurate system can help capture patients 'at-risk' of pancreatic cancer in a registry.Item Biliary Stricture after Necrotizing Pancreatitis: An Underappreciated Challenge(Wolters Kluwer, 2020-10) Maatman, Thomas K.; Ceppa, Eugene P.; Fogel, Evan L.; Easler, Jeffrey J.; Gromski, Mark A.; House, Michael G.; Nakeeb, Attila; Schmidt, C. Max; Sherman, Stuart; Zyromski, Nicholas J.; Surgery, School of MedicineObjective: Biliary stricture in necrotizing pancreatitis (NP) has not been systematically categorized; therefore, we sought to define the incidence and natural history of biliary stricture caused by NP. Summary/Background Data: Benign biliary stricture occurs secondary to bile duct injury, anastomotic narrowing, or chronic inflammation and fibrosis. The profound loco-regional inflammatory response of NP creates challenging biliary strictures. Methods: NP patients treated between 2005–2019 were reviewed. Biliary stricture was identified on cholangiography as narrowing of the extrahepatic biliary tree to < 75% of the diameter of the unaffected duct. Biliary stricture risk factors and outcomes were evaluated. Results: Among 743 NP patients, 64 died, 13 were lost to follow up; therefore, a total of 666 patients were included in the final cohort. Biliary stricture developed in 108 (16%) patients. Mean follow up was 3.5 ± 3.3 years. Median time from NP onset to biliary stricture diagnosis was 4.2 months (IQR, 1.8–10.9). Presentation was commonly clinical or biochemical jaundice, n = 30 (28%) each. Risk factors for stricture development were splanchnic vein thrombosis and pancreatic head parenchymal necrosis. Median time to stricture resolution was 6.0 months after onset (2.8–9.8). A mean of 3.3 ± 2.3 procedures were performed. Surgical intervention was required in 22 (20%) patients. Endoscopic treatment failed in 17% (17/99) of patients and was not associated with stricture length. Operative treatment of biliary stricture was more likely in patients with infected necrosis or NP disease duration ≥6 months. Conclusion: Biliary stricture occurs frequently after necrotizing pancreatitis and is associated with splanchnic vein thrombosis and pancreatic head necrosis. Surgical correction was performed in 20%.Item Cancer History: A Predictor of IPMN Subtype and Dysplastic Status?(Elsevier, 2017) Carr, Rosalie A.; Kiel, Brandon A.; Roch, Alexandra M.; Ceppa, Eugene P.; House, Michael G.; Zyromski, Nicholas J.; Nakeeb, Attila; Schmidt, C. Max; Surgery, School of MedicineIntroduction The aim of this study was to determine the association of PMH and FH of pancreatic (PDAC) and non-pancreatic cancers with IPMN malignant risk. Methods A retrospective review of a prospective database of IPMN patients undergoing resection was performed to assess FH and PMH. Results FH of PDAC was present in 13% of 362 included patients. Of these, 8% had at least one first degree relative (FDR) with PDAC. The rate of PDAC positive FH in non-invasive versus invasive IPMN patients was 14% and 8%, respectively (p = 0.3). In main duct IPMN patients, FH (44%) and PMH of non-pancreatic cancer (16%) was higher than that seen in branch duct IPMN (FH 29%; PMH 6%; p = 0.004 and 0.008). Conclusions FH of PDAC is not associated with IPMN malignant progression. FH and PMH of non-pancreatic cancer is associated with main duct IPMN, the subtype with the highest rate of invasive transformation.Item Circulating Leptin and Branched Chain Amino Acids – Correlation with Intraductal Papillary Mucinous Neoplasm Dysplastic Grade(Springer Verlag, 2019-05) Yip-Schneider, Michele T.; Simpson, Rachel; Carr, Rosalie A.; Wu, Huangbing; Fan, Hao; Liu, Ziyue; Korc, Murray; Zhang, Jianjun; Schmidt, C. Max; Surgery, School of MedicineBackground: The most common type of mucinous pancreatic cyst that may progress to pancreatic cancer is intraductal papillary mucinous neoplasm (IPMN). Low-risk IPMN with low-/moderate-grade dysplasia may be safely watched, whereas high-risk IPMN with high-grade dysplasia or invasive components should undergo resection. However, there is currently no reliable means of making this distinction. We hypothesize that blood concentrations of insulin resistance biomarkers may aid in the differentiation of low- and high-risk IPMN. Methods: Plasma/serum was collected from consented patients undergoing pancreatic resection. IPMN diagnosis and dysplastic grade were confirmed by surgical pathology. The study included 235 IPMN (166 low/moderate grade, 39 high grade, 30 invasive). Circulating levels of leptin, branched chain amino acids (BCAA), and retinol-binding protein-4 (RBP-4) were measured by enzyme-linked immunoassay and correlated with surgical pathology. Results: Circulating leptin levels (mean ± SE) were significantly higher in patients with low/moderate IPMN than in high-grade/invasive IPMN (15,803 ± 1686 vs. 10,275 ± 1228 pg/ml; p = 0.0086). Leptin levels were positively correlated with BMI (r = 0.65, p < 0.0001) and were higher in females (p < 0.0001). Stratified analysis showed that mean leptin levels were significantly different between low/moderate and high/invasive IPMNs only in females (24,383 ± 2748 vs. 16,295 ± 2040 pg/ml; p = 0.020). Conversely, circulating BCAA levels were lower in low/moderate IPMN than in high-grade/invasive IPMN (0.38 ± 0.007 vs. 0.42 ± 0.01 mM; p = 0.011). No significant differences in RBP-4 levels were observed. Conclusions: Circulating leptin in females and BCAA correlates with IPMN dysplastic grade and, if combined with clinical characteristics, have the potential to improve clinical decision-making.Item Circulating Thrombospondin-2 enhances prediction of malignant intraductal papillary mucinous neoplasm(Elsevier, 2018) Simpson, Rachel E.; Yip-Schneider, Michele T.; Wu, Huangbing; Fan, Hao; Liu, Ziyue; Korc, Murray; Zhang, Jianjun; Schmidt, C. Max; Surgery, School of MedicineBackground IPMNs are cystic pancreatic lesions with variable malignant potential. Thrombospondin-2 (THBS2)—an endogenous, anti-angiogenic matrix glycoprotein—may modulate tumor progression. We hypothesized that circulating levels of THBS2 could aid in preoperative prediction of malignant IPMN. Methods Preoperative serum/plasma samples were procured from patients undergoing surgery. Circulating levels of THBS2 were measured (enzyme-linked immunosorbent assay) and compared to surgical pathology IPMN dysplastic grade. Results 164 patients underwent THBS2 testing (100 Low/Moderate-IPMN; 64 High-Grade/Invasive-IPMN). Circulating THBS2 (mean ± SD) was greater in High-Grade/Invasive-IPMN than Low/Moderate-grade IPMN (26.6 ± 12.7 ng/mL vs. 20.4 ± 8.2 ng/mL; P < 0.001). THBS2 (AUC = 0.65) out-performed CA19-9 (n = 144; AUC = 0.59) in predicting IPMN grade. The combination of THBS2, CA19-9, radiographic main-duct involvement, main-duct diameter, age, sex, and BMI (AUC 0.82; n = 137) provided a good prediction model for IPMN grade. Conclusion Circulating THBS2 is correlated with IPMN dysplasia grade. THBS2 alone did not strongly predict IPMN grade but rather strengthened prediction models for High-Grade/Invasive IPMN when combined with other clinical/biomarker data.Item The continuum of complications in survivors of necrotizing pancreatitis(Elsevier, 2020-12) Maatman, Thomas K.; Roch, Alexandra M.; Ceppa, Eugene P.; Easler, Jeffrey J.; Gromski, Mark A.; House, Michael G.; Nakeeb, Attila; Schmidt, C. Max; Sherman, Stuart; Zyromski, Nicholas J.; Surgery, School of MedicineBackground: Necrotizing pancreatitis survivors develop complications beyond infected necrosis that often require invasive intervention. Remarkably few data have cataloged these late complications after acute necrotizing pancreatitis resolution. We sought to identify the types and incidence of complications after necrotizing pancreatitis. Design: An observational study was performed evaluating 647 patients with necrotizing pancreatitis captured in a single-institution database between 2005 and 2017 at a tertiary care hospital. Retrospective review and analysis of newly diagnosed conditions attributable to necrotizing pancreatitis was performed. Exclusion criteria included the following: death before disease resolution (n = 57, 9%) and patients lost to follow-up (n = 12, 2%). Results: A total of 578 patients were followed for a median of 46 months (range, 8 months to 15 y) after necrotizing pancreatitis. In 489 (85%) patients 1 or more complications developed and included symptomatic disconnected pancreatic duct syndrome (285 of 578, 49%), splanchnic vein thrombosis (257 of 572, 45%), new endocrine insufficiency (195 of 549, 35%), new exocrine insufficiency (108 of 571, 19%), symptomatic chronic pancreatitis (93 of 571, 16%), incisional hernia (89 of 420, 21%), biliary stricture (90 of 576, 16%), chronic pain (44 of 575, 8%), gastrointestinal fistula (44 of 578, 8%), pancreatic duct stricture (30 of 578, 5%), and duodenal stricture (28 of 578, 5%). During the follow-up period, a total of 340 (59%) patients required an invasive intervention after necrotizing pancreatitis resolution. Invasive pancreatobiliary intervention was required in 230 (40%) patients. Conclusion: Late complications are common in necrotizing pancreatitis survivors. A broad variety of problems manifest themselves after resolution of the acute disease process and often require invasive intervention. Necrotizing pancreatitis patients should be followed lifelong by experienced clinicians.Item Cyclooxygenase-2 Expression in Hamster and Human Pancreatic Neoplasia(Elsevier, 2006-06) Crowell, Pamela L.; Schmidt, C. Max; Yip-Schneider, Michele T.; Savage, Jesse J.; Hertzler II, Dean A.; Cummings, William O.; Biochemistry and Molecular Biology, School of MedicineCyclooxygenase-2 (COX-2) has been implicated in the development of gastrointestinal malignancies. The aim of the present study was to determine COX-2 expression/activity throughout stages of experimental and human pancreatic neoplasia. COX-2 immunohistochemistry was performed in pancreata of hamsters subjected to the carcinogen N-nitrosobis-(2-oxopropyl)amine (BOP) and in human pancreatic tumors. COX-2 activity was determined by prostaglandin E2 assay in tumor versus matched normal pancreatic tissues. The activity of the COX inhibitor sulindac was tested in the PC-1 hamster pancreatic cancer model. COX-2 expression was elevated in all pancreatic intraepithelial neoplasias (PanINs) and adenocarcinomas. In BOP-treated hamsters, there were significant progressive elevations in COX-2 expression throughout pancreatic tumorigenesis. In human samples, peak COX-2 expression occurred in PanIN2 lesions and remained moderately elevated in PanIN3 and adenocarcinoma tissues. COX-2 activity was significantly elevated in hamster and human pancreatic cancers compared to pair-matched normal pancreas. Furthermore, hamster pancreatic tumor engraftment/formation in the PC-1 hamster pancreatic cancer model was reduced 4.9-fold by oral administration of sulindac. Increased COX-2 expression is an early event in pancreatic carcinogeneses. The BOP-induced hamster carcinogenesis model is a representative model used to study the role of COX-2 in well-differentiated pancreatic tumorigenesis. COX inhibitors may have a role in preventing tumor engraftment/formation.Item DEEPEN: A negation detection system for clinical text incorporating dependency relation into NegEx(Elsevier, 2015-04) Mehrabi, Saeed; Krishnan, Krishnan; Sohn, Sunghwan; Roch, Alexandra M; Schmidt, Heidi; Kesterson, Joe; Beesley, Chris; Dexter, Paul; Schmidt, C. Max; Liu, Hongfang; Palakal, Mathew; Surgery, School of MedicineIn Electronic Health Records (EHRs), much of valuable information regarding patients’ conditions is embedded in free text format. Natural language processing (NLP) techniques have been developed to extract clinical information from free text. One challenge faced in clinical NLP is that the meaning of clinical entities is heavily affected by modifiers such as negation. A negation detection algorithm, NegEx, applies a simplistic approach that has been shown to be powerful in clinical NLP. However, due to the failure to consider the contextual relationship between words within a sentence, NegEx fails to correctly capture the negation status of concepts in complex sentences. Incorrect negation assignment could cause inaccurate diagnosis of patients’ condition or contaminated study cohorts. We developed a negation algorithm called DEEPEN to decrease NegEx’s false positives by taking into account the dependency relationship between negation words and concepts within a sentence using Stanford dependency parser. The system was developed and tested using EHR data from Indiana University (IU) and it was further evaluated on Mayo Clinic dataset to assess its generalizability. The evaluation results demonstrate DEEPEN, which incorporates dependency parsing into NegEx, can reduce the number of incorrect negation assignment for patients with positive findings, and therefore improve the identification of patients with the target clinical findings in EHRs.