Neumonía por COVID19: Valoración por imagen, lo básico / COVID­19 pneumonia: Imaging evaluation, the basics

La pandemia de COVID­19 ha resultado en una emergencia de salud global. Los estu­dios de imagen utilizados en esta enfermedad son la radiografía de tórax (RX) y la to­mografía computarizada (TC). Ambas modalidades tienen sus hallazgos descritos, pero no son específicos dado que muchas enfermedades pueden producir patrones simila­res, particularmente las neumonías virales. Los RX de tórax muestra hallazgos consis­tentes en opacidades alveolares las cuales son múltiples, periféricas, bilaterales y basales, mientras que la tomografía de tórax sus hallazgos más frecuentes son presen­cia de patrón en vidrio deslustrado, consolidaciones, engrosamiento septal, patrón en empedrado, dilatación bronquial y engrosamiento peri bronquial, broncograma, patrón de halo invertido y patrón de neumonía organizada. Los hallazgos por imagen depen­den del tiempo de evolución de la enfermedad ya que en etapas tempranas puede ser normal tanto en la RX como la TC. El riesgo de trombo embolismo pulmonar es alto y más frecuente que en pacientes con COVID­19 negativo

The COVID­19 pandemic has resulted in a global health emergency. The imaging stu­dies used in this disease are chest radiography (CXR) and computed tomography (CT). Both imaging modalities findings have had their findings. These findings described are not specific since many diseases can produce similar patterns. CXR shows somewhat consistent findings consisting of alveolar opacities which are multiple, peripheral, bilate­ral and basal, while CT the most frequent findings are the presence of grounded glass pattern, consolidations, septal thickening, crazy paving pattern, bronchial dilation and peribronchial thickening, air bronchograms, inverted halo sign and organized pneumo­nia. Imaging findings depends on the evolution time of the disease since in the early sta­ges both chest radiography and tomography may be normal. The risk for pulmonary embolism is high and more frequent than in patients with negative COVID­19

A dataset of clinically generated visual questions and answers about radiology images.

Radiology images are an essential part of clinical decision making and population screening, e.g., for cancer. Automated systems could help clinicians cope with large amounts of images by answering questions about the image contents. An emerging area of artificial intelligence, Visual Question Answering (VQA) in the medical domain explores approaches to this form of clinical decision support. Success of such machine learning tools hinges on availability and design of collections composed of medical images augmented with question-answer pairs directed at the content of the image. We introduce VQA-RAD, the first manually constructed dataset where clinicians asked naturally occurring questions about radiology images and provided reference answers. Manual categorization of images and questions provides insight into clinically relevant tasks and the natural language to phrase them. Evaluating with well-known algorithms, we demonstrate the rich quality of this dataset over other automatically constructed ones. We propose VQA-RAD to encourage the community to design VQA tools with the goals of improving patient care.

Proposing New RadLex Terms by Analyzing Free-Text Mammography Reports.

After years of development, the RadLex terminology contains a large set of controlled terms for the radiology domain, but gaps still exist. We developed a data-driven approach to discover new terms for RadLex by mining a large corpus of radiology reports using natural language processing (NLP) methods. Our system, developed for mammography, discovers new candidate terms by analyzing noun phrases in free-text reports to extend the mammography part of RadLex. Our NLP system extracts noun phrases from free-text mammography reports and classifies these noun phrases as "Has Candidate RadLex Term" or "Does Not Have Candidate RadLex Term." We tested the performance of our algorithm using 100 free-text mammography reports. An expert radiologist determined the true positive and true negative RadLex candidate terms. We calculated precision/positive predictive value and recall/sensitivity metrics to judge the system's performance. Finally, to identify new candidate terms for enhancing RadLex, we applied our NLP method to 270,540 free-text mammography reports obtained from three academic institutions. Our method demonstrated precision/positive predictive value of 0.77 (159/206 terms) and a recall/sensitivity of 0.94 (159/170 terms). The overall accuracy of the system is 0.80 (235/293 terms). When we ran our system on the set of 270,540 reports, it found 31,800 unique noun phrases that are potential candidates for RadLex. Our data-driven approach to mining radiology reports can identify new candidate terms for expanding the breast imaging lexicon portion of RadLex and may be a useful approach for discovering new candidate terms from other radiology domains.

Diagnostic performance of the Gynecology Imaging Reporting and Data System for malignant adnexal masses.

OBJECTIVE: To evaluate the Gynecology Imaging Reporting and Data System (GI-RADS) for diagnosis of malignant adnexal masses in a Chinese population. METHODS: A retrospective study was conducted of patients who underwent evaluation of suspected adnexal masses at a hospital in Tianjin, China, between January 1, 2015, and January 31, 2016. Ultrasonographic diagnosis was based on the GI-RADS classification-a standardized summary of imaging data that estimates the risk of malignancy-and compared with the final pathological diagnosis. RESULTS: Among 242 patients, thick wall, solid papillary projection, solid area, central blood flow, ascites, and GI-RADS classification were associated with malignancy (P<0.05 for all variables). The 263 masses evaluated were classified as GI-RADS 2 (functional cyst; n=65), GI-RADS 3 (benign neoplasm; n=68), GI-RADS 4 (one or two morphological findings suggestive of malignancy; n=101), and GI-RADS 5 (≥3 morphological findings suggestive of malignancy; n=28). Four malignant cases with false-negative findings were misclassified as GI-RADS 3, whereas 24 benign cases with false-positive findings were misclassified as GI-RADS 4. The sensitivity, specificity, false-positive rate, false-negative rate, accuracy, and Youden index of the GI-RADS classification were 96.4%, 84.3%, 18.5%, 3.0%, 89.3%, and 80.7%, respectively. CONCLUSION: The GI-RADS classification performed well in the diagnosis of malignant adnexal masses.

The Role of Report Comparison, Analysis, and Discrepancy Categorization in Resident Education.

OBJECTIVE: The purpose of this study was to show the value of automated radiology report comparison and analysis in resident education by providing qualitative and quantitative feedback on the discrepancies between preliminary and finalized reports. MATERIALS AND METHODS: Anonymous surveys on dictation practices and the process of reviewing reports were completed by consenting radiology residents and faculty. All 277 reports obtained across all modalities during the 4-week study were retrieved from the dictation server in both their preliminary and finalized states, for a total of 544 reports. Disparities between these reports were automatically compared side by side and were categorized according to clinical relevance, report quality, or report structure. The frequency of report corrections was compared between junior (postgraduate years [PGYs] 2 and 3) and senior (PGYs 4 and 5) residents. Residents were surveyed regarding the usefulness of the feedback. RESULTS: Eighty-six reports (31%) were verified as unchanged, with no statistically significant difference noted between junior and senior residents (33.2% and 25.9%, respectively; p = 0.03). Of the 370 discrepancies noted in the 191 edited reports, 81 (21.9%) were discrepancies in clinically relevant findings; 106 (28.6%) were discrepancies in report quality; and 183 (49.5%) were discrepancies in report structure, syntax, or both. Although senior residents had a lower rate of discrepancies in the clinical relevance category than did junior residents (12.8% and 26.5%; p = 0.004), they had a higher rate of discrepancies in the report quality category (58.4% and 44.9%; p = 0.02). Surveys of both residents and faculty showed strong support for the project. CONCLUSION: Categorization of corrections was deemed useful by residents and can be helpful in assessing elements of reporting accuracy for individual feedback. Quantitative report comparison and analysis show promise in tailoring resident education at the programmatic level as cumulative data are gathered and trends are analyzed.

JOURNAL CLUB: Assessment of Interobserver Variability in the Peer Review Process: Should We Agree to Disagree?

OBJECTIVE: Peer review is an important and necessary part of radiology. There are several options to perform the peer review process. This study examines the reproducibility of peer review by comparing two scoring systems. MATERIALS AND METHODS: American Board of Radiology-certified radiologists from various practice environments and subspecialties were recruited to score deidentified examinations on a web-based PACS with two scoring systems, RADPEER and Cleareview. Quantitative analysis of the scores was performed for interrater agreement. RESULTS: Interobserver variability was high for both the RADPEER and Cleareview scoring systems. The interobserver correlations (kappa values) were 0.17-0.23 for RADPEER and 0.10-0.16 for Cleareview. Interrater correlation was not statistically significantly different when comparing the RADPEER and Cleareview systems (p = 0.07-0.27). The kappa values were low for the Cleareview subscores when we evaluated for missed findings (0.26), satisfaction of search (0.17), and inadequate interpretation of findings (0.12). CONCLUSION: Our study confirms the previous report of low interobserver correlation when using the peer review process. There was low interobserver agreement seen when using both the RADPEER and the Cleareview scoring systems.

Classification of radiology reports for falls in an HIV study cohort.

OBJECTIVE: To identify patients in a human immunodeficiency virus (HIV) study cohort who have fallen by applying supervised machine learning methods to radiology reports of the cohort. METHODS: We used the Veterans Aging Cohort Study Virtual Cohort (VACS-VC), an electronic health record-based cohort of 146 530 veterans for whom radiology reports were available (N=2 977 739). We created a reference standard of radiology reports, represented each report by a feature set of words and Unified Medical Language System concepts, and then developed several support vector machine (SVM) classifiers for falls. We compared mutual information (MI) ranking and embedded feature selection approaches. The SVM classifier with MI feature selection was chosen to classify all radiology reports in VACS-VC. RESULTS: Our SVM classifier with MI feature selection achieved an area under the curve score of 97.04 on the test set. When applied to all the radiology reports in VACS-VC, 80 416 of these reports were classified as positive for a fall. Of these, 11 484 were associated with a fall-related external cause of injury code (E-code) and 68 932 were not, corresponding to 29 280 patients with potential fall-related injuries who could not have been found using E-codes. DISCUSSION: Feature selection was crucial to improving the classifier's performance. Feature selection with MI allowed us to select the number of discriminative features to use for classification, in contrast to the embedded feature selection method, in which the number of features is chosen automatically. CONCLUSION: Machine learning is an effective method of identifying patients who have suffered a fall. The development of this classifier supplements the clinical researcher's toolkit and reduces dependence on under-coded structured electronic health record data.

Identification of Incidental Pulmonary Nodules in Free-text Radiology Reports: An Initial Investigation.

Advances in image quality produced by computed tomography (CT) and the growth in the number of image studies currently performed has made the management of incidental pulmonary nodules (IPNs) a challenging task. This research aims to identify IPNs in radiology reports of chest and abdominal CT by Natural Language Processing techiniques to recognize IPN in sentences of radiology reports. Our preliminary analysis indicates vastly different pulmonary incidental findings rates for two different patient groups.

Follow-up Recommendation Detection on Radiology Reports with Incidental Pulmonary Nodules.

The management of follow-up recommendations is fundamental for the appropriate care of patients with incidental pulmonary findings. The lack of communication of these important findings can result in important actionable information being lost in healthcare provider electronic documents. This study aims to analyze follow-up recommendations in radiology reports containing pulmonary incidental findings by using Natural Language Processing and Regular Expressions. Our evaluation highlights the different follow-up recommendation rates for oncology and non-oncology patient cohorts. The results reveal the need for a context-sensitive approach to tracking different patient cohorts in an enterprise-wide assessment.

Automated Image Retrieval of Chest CT Images Based on Local Grey Scale Invariant Features.

Textual-based tools are regularly employed to retrieve medical images for reading and interpretation using current retrieval Picture Archiving and Communication Systems (PACS) but pose some drawbacks. All-purpose content-based image retrieval (CBIR) systems are limited when dealing with medical images and do not fit well into PACS workflow and clinical practice. This paper presents an automated image retrieval approach for chest CT images based local grey scale invariant features from a local database. Performance was measured in terms of precision and recall, average retrieval precision (ARP), and average retrieval rate (ARR). Preliminary results have shown the effectiveness of the proposed approach. The prototype is also a useful tool for radiology research and education, providing valuable information to the medical and broader healthcare community.

A Statistical Analysis of Term Occurrences in Radiology Reporting.

To compare term occurrences in free-text radiology reports and RSNA reporting templates, we selected five templates from an RSNA reporting template library and their corresponding free-text reports as a test set, and employed the Wilcoxon signed-rank test to find out whether the terms in RSNA reporting templates match those terms appearing in corresponding free-text radiology reports. The results show that most terms in free-text radiology reports are covered by RSNA reporting templates. By assessing the terminology coverage of existing templates, this study may benefit the growth of the RSNA reporting template library.

An Approach for Automatic Classification of Radiology Reports in Spanish.

Automatic detection of relevant terms in medical reports is useful for educational purposes and for clinical research. Natural language processing (NLP) techniques can be applied in order to identify them. In this work we present an approach to classify radiology reports written in Spanish into two sets: the ones that indicate pathological findings and the ones that do not. In addition, the entities corresponding to pathological findings are identified in the reports. We use RadLex, a lexicon of English radiology terms, and NLP techniques to identify the occurrence of pathological findings. Reports are classified using a simple algorithm based on the presence of pathological findings, negation and hedge terms. The implemented algorithms were tested with a test set of 248 reports annotated by an expert, obtaining a best result of 0.72 F1 measure. The output of the classification task can be used to look for specific occurrences of pathological findings.

Automated classification of radiology reports to facilitate retrospective study in radiology.

Retrospective research is an import tool in radiology. Identifying imaging examinations appropriate for a given research question from the unstructured radiology reports is extremely useful, but labor-intensive. Using the machine learning text-mining methods implemented in LingPipe [1], we evaluated the performance of the dynamic language model (DLM) and the Naïve Bayesian (NB) classifiers in classifying radiology reports to facilitate identification of radiological examinations for research projects. The training dataset consisted of 14,325 sentences from 11,432 radiology reports randomly selected from a database of 5,104,594 reports in all disciplines of radiology. The training sentences were categorized manually into six categories (Positive, Differential, Post Treatment, Negative, Normal, and History). A 10-fold cross-validation [2] was used to evaluate the performance of the models, which were tested in classification of radiology reports for cases of sellar or suprasellar masses and colloid cysts. The average accuracies for the DLM and NB classifiers were 88.5% with 95% confidence interval (CI) of 1.9% and 85.9% with 95% CI of 2.0%, respectively. The DLM performed slightly better and was used to classify 1,397 radiology reports containing the keywords "sellar or suprasellar mass", or "colloid cyst". The DLM model produced an accuracy of 88.2% with 95% CI of 2.1% for 959 reports that contain "sellar or suprasellar mass" and an accuracy of 86.3% with 95% CI of 2.5% for 437 reports of "colloid cyst". We conclude that automated classification of radiology reports using machine learning techniques can effectively facilitate the identification of cases suitable for retrospective research.

Automatically pairing measured findings across narrative abdomen CT reports.

Radiological measurements are one of the key variables in widely adopted guidelines (WHO, RECIST) that standardize and objectivize response assessment in oncology care. Measurements are typically described in free-text, narrative radiology reports. We present a natural language processing pipeline that extracts measurements from radiology reports and pairs them with extracted measurements from prior reports of the same clinical finding, e.g., lymph node or mass. A ground truth was created by manually pairing measurements in the abdomen CT reports of 50 patients. A Random Forest classifier trained on 15 features achieved superior results in an end-to-end evaluation of the pipeline on the extraction and pairing task: precision 0.910, recall 0.878, F-measure 0.894, AUC 0.988. Representing the narrative content in terms of UMLS concepts did not improve results. Applications of the proposed technology include data mining, advanced search and workflow support for healthcare professionals managing radiological measurements.

Automatic classification of mammography reports by BI-RADS breast tissue composition class.

Because breast tissue composition partially predicts breast cancer risk, classification of mammography reports by breast tissue composition is important from both a scientific and clinical perspective. A method is presented for using the unstructured text of mammography reports to classify them into BI-RADS breast tissue composition categories. An algorithm that uses regular expressions to automatically determine BI-RADS breast tissue composition classes for unstructured mammography reports was developed. The algorithm assigns each report to a single BI-RADS composition class: 'fatty', 'fibroglandular', 'heterogeneously dense', 'dense', or 'unspecified'. We evaluated its performance on mammography reports from two different institutions. The method achieves >99% classification accuracy on a test set of reports from the Marshfield Clinic (Wisconsin) and Stanford University. Since large-scale studies of breast cancer rely heavily on breast tissue composition information, this method could facilitate this research by helping mine large datasets to correlate breast composition with other covariates.

Automatic identification of critical follow-up recommendation sentences in radiology reports.

Communication of follow-up recommendations when abnormalities are identified on imaging studies is prone to error. When recommendations are not systematically identified and promptly communicated to referrers, poor patient outcomes can result. Using information technology can improve communication and improve patient safety. In this paper, we describe a text processing approach that uses natural language processing (NLP) and supervised text classification methods to automatically identify critical recommendation sentences in radiology reports. To increase the classification performance we enhanced the simple unigram token representation approach with lexical, semantic, knowledge-base, and structural features. We tested different combinations of those features with the Maximum Entropy (MaxEnt) classification algorithm. Classifiers were trained and tested with a gold standard corpus annotated by a domain expert. We applied 5-fold cross validation and our best performing classifier achieved 95.60% precision, 79.82% recall, 87.0% F-score, and 99.59% classification accuracy in identifying the critical recommendation sentences in radiology reports.

The Yale cTAKES extensions for document classification: architecture and application.

BACKGROUND: Open-source clinical natural-language-processing (NLP) systems have lowered the barrier to the development of effective clinical document classification systems. Clinical natural-language-processing systems annotate the syntax and semantics of clinical text; however, feature extraction and representation for document classification pose technical challenges. METHODS: The authors developed extensions to the clinical Text Analysis and Knowledge Extraction System (cTAKES) that simplify feature extraction, experimentation with various feature representations, and the development of both rule and machine-learning based document classifiers. The authors describe and evaluate their system, the Yale cTAKES Extensions (YTEX), on the classification of radiology reports that contain findings suggestive of hepatic decompensation. RESULTS AND DISCUSSION: The F(1)-Score of the system for the retrieval of abdominal radiology reports was 96%, and was 79%, 91%, and 95% for the presence of liver masses, ascites, and varices, respectively. The authors released YTEX as open source, available at http://code.google.com/p/ytex.

Radiological reports: a comparison between the transmission efficiency of information in free text and in structured reports.

INTRODUCTION: This work proposes to improve the transmission of information between requiring physicians and radiologists. OBJECTIVES: Evaluate the implementation of a structured report (SR) in a university hospital. METHODS: A model of a structured report for thyroid sonography was developed according to information gathered from radiologists and endocrinologists working in this field. The report was based on a web platform and installed as a part of a Radiological Information System (RIS) and a Hospital Information System (HIS). The time for the report generation under the two forms was evaluated over a four-month period, two months for each method. After this period, radiologists and requiring physicians were questioned about the two methods of reporting. RESULTS: For free text, 98 sonograms were reported to have thyroids with nodules in an average time of 8.71 (+/-4.11) minutes, and 59 sonograms of thyroids without nodules were reported in an average time of 4.54 (+/- 3.97) minutes. For SR, 73 sonograms in an average time of 6.08 (+/-3.8) minutes for thyroids with nodules and 3.67 (+/-2.51) minutes for thyroids without nodules. Most of the radiologists (76.2%) preferred the SR, as originally created or with suggested changes. Among endocrinologists, 80% preferred the SR. DISCUSSION: From the requiring physicians' perspective, the SR enabled standardization and improved information transmission. This information is valuable because physicians need reports prepared by radiologists. CONCLUSIONS: The implementation of a SR in a university hospital, under an RIS/HIS system, was viable. Radiologists and endocrinologists preferred the SR when compared to free text, and both agreed that the former improved the transmission of information.

Radiological reports: a comparison between the transmission efficiency of information in free text and in structured reports

Introduction: This work proposes to improve the transmission of information between requiring physicians and radiologists. Objectives: Evaluate the implementation of a structured report (SR) in a university hospital. Methods: A model of a structured report for thyroid sonography was developed according to information gathered from radiologists and endocrinologists working in this field. The report was based on a web platform and installed as a part of a Radiological Information System (RIS) and a Hospital Information System (HIS). The time for the report generation under the two forms was evaluated over a four-month period, two months for each method. After this period, radiologists and requiring physicians were questioned about the two methods of reporting. Results: For free text, 98 sonograms were reported to have thyroids with nodules in an average time of 8.71 (+/-4.11) minutes, and 59 sonograms of thyroids without nodules were reported in an average time of 4.54 (+/- 3.97) minutes. For SR, 73 sonograms in an average time of 6.08 (+/-3.8) minutes for thyroids with nodules and 3.67 (+/-2.51) minutes for thyroids without nodules. Most of the radiologists (76.2 percent) preferred the SR, as originally created or with suggested changes. Among endocrinologists, 80 percent preferred the SR. Discussion: From the requiring physicians' perspective, the SR enabled standardization and improved information transmission. This information is valuable because physicians need reports prepared by radiologists. Conclusions: The implementation of a SR in a university hospital, under an RIS/HIS system, was viable. Radiologists and endocrinologists preferred the SR when compared to free text, and both agreed that the former improved the transmission of information.