Caroline Taylor

Semi-supervised clinical text classification with Laplacian SVMs

OBJECTIVE To compare linear and Laplacian SVMs on a clinical text classification task; to evaluate the effect of unlabeled training data on Laplacian SVM performance. BACKGROUND The development of machine-learning based clinical text classifiers requires the creation of labeled training data, obtained via manual review by clinicians. Due to the effort and expense involved in labeling data, training data sets in the clinical domain are of limited size. In contrast, electronic medical record (EMR) systems contain hundreds of thousands of unlabeled notes that are not used by supervised machine learning approaches. Semi-supervised learning algorithms use both labeled and unlabeled data to train classifiers, and can outperform their supervised counterparts. METHODS We trained support vector machines (SVMs) and Laplacian SVMs on a training reference standard of 820 abdominal CT, MRI, and ultrasound reports labeled for the presence of potentially malignant liver lesions that require follow up (positive class prevalence 77%). The Laplacian SVM used 19,845 randomly sampled unlabeled notes in addition to the training reference standard. We evaluated SVMs and Laplacian SVMs on a test set of 520 labeled reports. RESULTS The Laplacian SVM trained on labeled and unlabeled radiology reports significantly outperformed supervised SVMs (Macro-F1 0.773 vs. 0.741, Sensitivity 0.943 vs. 0.911, Positive Predictive value 0.877 vs. 0.883). Performance improved with the number of labeled and unlabeled notes used to train the Laplacian SVM (pearsons ρ=0.529 for correlation between number of unlabeled notes and macro-F1 score). These results suggest that practical semi-supervised methods such as the Laplacian SVM can leverage the large, unlabeled corpora that reside within EMRs to improve clinical text classification.

The Effect of a Lung Cancer Care Coordination Program on Timeliness of Care

BACKGROUND Timeliness of care improves patient satisfaction and might improve outcomes. The CCCP was established in November 2007 to improve timeliness of care of NSCLC at the Veterans Affairs Connecticut Healthcare System (VACHS). PATIENTS AND METHODS We performed a retrospective cohort analysis of patients diagnosed with NSCLC at VACHS between 2005 and 2010. We compared timeliness of care and stage at diagnosis before and after the implementation of the CCCP. RESULTS Data from 352 patients were analyzed: 163 with initial abnormal imaging between January 1, 2005 and October 31, 2007, and 189 with imaging conducted between November 1, 2007 and December 31, 2010. Variables associated with a longer interval between the initial abnormal image and the initiation of therapy were: (1) earlier stage (mean of 130 days for stages I/II vs. 87 days for stages III/IV; P < .0001); (2) lack of cancer-related symptoms (145 vs. 60 days; P < .0001); (3) presence of more than 1 medical comorbidity (123 vs. 82; P = .0002); and (4) depression (126 vs. 98 days; P = .029). The percent of patients diagnosed at stages I/II increased from 32% to 48% (P = .006) after establishment of the CCCP. In a multivariate model adjusting for stage, histology, reason for imaging, and presence of primary care provider, implementation of the CCCP resulted in a mean reduction of 25 days between first abnormal image and the initiation of treatment (126 to 101 days; P = .015). CONCLUSION A centralized, multidisciplinary, hospital-based CCCP can improve timeliness of NSCLC care, and help ensure that early stage lung cancers are diagnosed and treated.

Limitations of computed tomography in the recognition of delayed splenic rupture.

Delayed splenic rupture is a rare but life threatening complication of abdominal trauma. A case is reported in which initial evaluation by CT did not show hemoperitoneum or splenic laceration in a patient who ruptured his spleen 10 days following the initial injury. The limitations of current diagnostic methods are reviewed.

EMR-linked cancer tracker facilitates lung and liver cancer care.

77 Background: VA Connecticut Healthcare System has developed a web-based, EMR-linked Cancer Care Tracking System (CCTS) to facilitate tracking and follow-up of patients with imaging abnormalities concerning for lung or liver cancer. The tracker was developed to facilitate the efforts of a multidisciplinary team at the center of which is a cancer navigator. METHODS CCTS was first envisioned in 2007 when VACT hired a care navigator and implemented a radiology coding system to identify potential cancers. This created the need for a tool to process abnormal images and track the clinical steps required to reach a definitive diagnosis and treatment plan. CCTS was initially used for lung cancers and was expanded to track hepatocellular carcinoma (HCC) in 2009 with additional funding. In addition to case discovery, it offers easy access to patient information with live links to the VA EMR, a surveillance feature, and scheduling, alerting, and reporting functions. In 2011, the system was enhanced with a natural language processing (NLP) program that automatically identifies radiology reports describing potentially malignant lung or liver lesions. RESULTS CCTS has been in daily operation since February 2010, with 1,778 patients and 2,503 patients tracked in 2010 and 2011, respectively. Addition of NLP technology significantly increases the accuracy of identification of patients with lung or liver nodules. The NLP system identified 21% of all new cases with potential malignancies whose management could have been delayed through coding omissions or errors. Benefits of CCTS and our cancer care coordination program have included a decrease of 25 days in the time from abnormal image to treatment of lung cancer, a significant increase in the diagnosis of stage I/II lung cancers from 32% to 48%, and an increase in the incidence of liver cancer from 1% to 5% of all cancers at VACT. CONCLUSIONS A web-based, EMR-linked cancer care tracking system (CCTS) improves cancer detection, prevents loss to follow-up, provides a safety net for radiology coding omissions or errors, and improves provider efficiency. CCTS is an innovative tool to support multidisciplinary cancer care and has broad applicability to any electronic medical record.

Periostitis and osteomyelitis in chronic drug addicts

Periostitis and osteomyelitis can occur in drug addicts not only by hematogenous dissemination of the infecting organisms, but as a result of introduction of bacteria by direct injection into periosteum or injection through infected skin and subcutaneous tissues. A spectrum of examples of osteomyelitis of the bones of the forearm in drug addicts is presented to illustrate this phenomenon. Neglect of these infections and the trauma of continued injections can lead to extensive tissue and bone loss.

Comparison of Natural Language Processing and Manual Coding for the Identification of Cross-Sectional Imaging Reports Suspicious for Lung Cancer

Purpose To compare the accuracy and reliability of a natural language processing (NLP) algorithm with manual coding by radiologists, and the combination of the two methods, for the identification of patients whose computed tomography (CT) reports raised the concern for lung cancer. Methods An NLP algorithm was developed using Clinical Text Analysis and Knowledge Extraction System (cTAKES) with the Yale cTAKES Extensions and trained to differentiate between language indicating benign lesions and lesions concerning for lung cancer. A random sample of 450 chest CT reports performed at Veterans Affairs Connecticut Healthcare System between January 2014 and July 2015 was selected. A reference standard was created by the manual review of reports to determine if the text stated that follow-up was needed for concern for cancer. The NLP algorithm was applied to all reports and compared with case identification using the manual coding by the radiologists. Results A total of 450 reports representing 428 patients were analyzed. NLP had higher sensitivity and lower specificity than manual coding (77.3% v 51.5% and 72.5% v 82.5%, respectively). NLP and manual coding had similar positive predictive values (88.4% v 88.9%), and NLP had a higher negative predictive value than manual coding (54% v 38.5%). When NLP and manual coding were combined, sensitivity increased to 92.3%, with a decrease in specificity to 62.85%. Combined NLP and manual coding had a positive predictive value of 87.0% and a negative predictive value of 75.2%. Conclusion Our NLP algorithm was more sensitive than manual coding of CT chest reports for the identification of patients who required follow-up for suspicion of lung cancer. The combination of NLP and manual coding is a sensitive way to identify patients who need further workup for lung cancer.

Giant spinal cord angioma diagnosed by digital subtraction intravenous angiography and computerized tomography

Abstract The case of a giant spinal cord angioma initially diagnosed by digital subtraction intravenous angiography is reported. A brief discussion of this lesion is presented.