Browsing by Author "Rubin, Geoffrey D"

Now showing 1 - 3 of 3
  • Thumbnail Image
    ItemOpen Access
    Computed tomography angiography in microsurgery: indications, clinical utility, and pitfalls.
    (Eplasty, 2013) Lee, Gordon K; Fox, Paige M; Riboh, Jonathan; Hsu, Charles; Saber, Sepideh; Rubin, Geoffrey D; Chang, James
    OBJECTIVE: Computed tomographic angiography (CTA) can be used to obtain 3-dimensional vascular images and soft-tissue definition. The goal of this study was to evaluate the reliability, usefulness, and pitfalls of CTA in preoperative planning of microvascular reconstructive surgery. METHODS: A retrospective review of patients who obtained preoperative CTA in preparation for planned microvascular reconstruction was performed over a 5-year period (2001-2005). The influence of CTA on the original operative plan was assessed for each patient, and CTA results were correlated to the operative findings. RESULTS: Computed tomographic angiography was performed on 94 patients in preparation for microvascular reconstruction. In 48 patients (51%), vascular abnormalities were noted on CTA. Intraoperative findings correlated with CTA results in 97% of cases. In 42 patients (45%), abnormal CTA findings influenced the original operative plan, such as the choice of vessels, side of harvest, or nature of the reconstruction (local flap instead of free tissue transfer). Technical difficulties in performing CTA were encountered in 5 patients (5%) in whom interference from external fixation devices was the main cause. CONCLUSIONS: This large study of CTA obtained for preoperative planning of reconstructive microsurgery at both donor and recipient sites study demonstrates that CTA is safe and highly accurate. Computed tomographic angiography can alter the surgeon's reconstructive plan when abnormalities are noted preoperatively and consequently improve results by decreasing vascular complication rates. The use of CTA should be considered for cases of microsurgical reconstruction where the vascular anatomy may be questionable.
  • Thumbnail Image
    ItemOpen Access
    Multi-label annotation of text reports from computed tomography of the chest, abdomen, and pelvis using deep learning.
    (BMC medical informatics and decision making, 2022-04) D'Anniballe, Vincent M; Tushar, Fakrul Islam; Faryna, Khrystyna; Han, Songyue; Mazurowski, Maciej A; Rubin, Geoffrey D; Lo, Joseph Y

    Background

    There is progress to be made in building artificially intelligent systems to detect abnormalities that are not only accurate but can handle the true breadth of findings that radiologists encounter in body (chest, abdomen, and pelvis) computed tomography (CT). Currently, the major bottleneck for developing multi-disease classifiers is a lack of manually annotated data. The purpose of this work was to develop high throughput multi-label annotators for body CT reports that can be applied across a variety of abnormalities, organs, and disease states thereby mitigating the need for human annotation.

    Methods

    We used a dictionary approach to develop rule-based algorithms (RBA) for extraction of disease labels from radiology text reports. We targeted three organ systems (lungs/pleura, liver/gallbladder, kidneys/ureters) with four diseases per system based on their prevalence in our dataset. To expand the algorithms beyond pre-defined keywords, attention-guided recurrent neural networks (RNN) were trained using the RBA-extracted labels to classify reports as being positive for one or more diseases or normal for each organ system. Alternative effects on disease classification performance were evaluated using random initialization or pre-trained embedding as well as different sizes of training datasets. The RBA was tested on a subset of 2158 manually labeled reports and performance was reported as accuracy and F-score. The RNN was tested against a test set of 48,758 reports labeled by RBA and performance was reported as area under the receiver operating characteristic curve (AUC), with 95% CIs calculated using the DeLong method.

    Results

    Manual validation of the RBA confirmed 91-99% accuracy across the 15 different labels. Our models extracted disease labels from 261,229 radiology reports of 112,501 unique subjects. Pre-trained models outperformed random initialization across all diseases. As the training dataset size was reduced, performance was robust except for a few diseases with a relatively small number of cases. Pre-trained classification AUCs reached > 0.95 for all four disease outcomes and normality across all three organ systems.

    Conclusions

    Our label-extracting pipeline was able to encompass a variety of cases and diseases in body CT reports by generalizing beyond strict rules with exceptional accuracy. The method described can be easily adapted to enable automated labeling of hospital-scale medical data sets for training image-based disease classifiers.
  • Thumbnail Image
    ItemOpen Access
    Right coronary wall CMR in the older asymptomatic advance cohort: positive remodeling and associations with type 2 diabetes and coronary calcium.
    (J Cardiovasc Magn Reson, 2010-12-30) Terashima, Masahiro; Nguyen, Patricia K; Rubin, Geoffrey D; Meyer, Craig H; Shimakawa, Ann; Nishimura, Dwight G; Ehara, Shoichi; Iribarren, Carlos; Courtney, Brian K; Go, Alan S; Hlatky, Mark A; Fortmann, Stephen P; McConnell, Michael V
    BACKGROUND: Coronary wall cardiovascular magnetic resonance (CMR) is a promising noninvasive approach to assess subclinical atherosclerosis, but data are limited in subjects over 60 years old, who are at increased risk. The purpose of the study was to evaluate coronary wall CMR in an asymptomatic older cohort. RESULTS: Cross-sectional images of the proximal right coronary artery (RCA) were acquired using spiral black-blood coronary CMR (0.7 mm resolution) in 223 older, community-based patients without a history of cardiovascular disease (age 60-72 years old, 38% female). Coronary measurements (total vessel area, lumen area, wall area, and wall thickness) had small intra- and inter-observer variabilities (r = 0.93~0.99, all p < 0.0001), though one-third of these older subjects had suboptimal image quality. Increased coronary wall thickness correlated with increased coronary vessel area (p < 0.0001), consistent with positive remodeling. On multivariate analysis, type 2 diabetes was the only risk factor associated with increased coronary wall area and thickness (p = 0.03 and p = 0.007, respectively). Coronary wall CMR measures were also associated with coronary calcification (p = 0.01-0.03). CONCLUSIONS: Right coronary wall CMR in asymptomatic older subjects showed increased coronary atherosclerosis in subjects with type 2 diabetes as well as coronary calcification. Coronary wall CMR may contribute to the noninvasive assessment of subclinical coronary atherosclerosis in older, at-risk patient groups.