MKCL: Medical Knowledge with Contrastive Learning model for radiology report generation.

Automatic radiology report generation has the potential to alert inexperienced radiologists to misdiagnoses or missed diagnoses and improve healthcare delivery efficiency by reducing the documentation workload of radiologists. Motivated by the continuous development of automatic image captioning, more and more deep learning methods have been proposed for automatic radiology report generation. However, the visual and textual data bias problem still face many challenges in the medical domain. Additionally, do not integrate medical knowledge, ignoring the mutual influences between medical findings, and abundant unlabeled medical images influence the accuracy of generating report. In this paper, we propose a Medical Knowledge with Contrastive Learning model (MKCL) to enhance radiology report generation. The proposed model MKCL uses IU Medical Knowledge Graph (IU-MKG) to mine the relationship among medical findings and improve the accuracy of identifying positive diseases findings from radiologic medical images. In particular, we design Knowledge Enhanced Attention (KEA), which integrates the IU-MKG and the extracted chest radiological visual features to alleviate textual data bias. Meanwhile, this paper leverages supervised contrastive learning to relieve radiographic medical images which have not been labeled, and identify abnormalities from images. Experimental results on the public dataset IU X-ray show that our proposed model MKCL outperforms other state-of-the-art report generation methods. Ablation studies also demonstrate that IU medical knowledge graph module and supervised contrastive learning module enhance the ability of the model to detect the abnormal parts and accurately describe the abnormal findings. The source code is available at: https://github.com/Eleanorhxd/MKCL.

Management of radiology services during the 2022 FIFA football (soccer) World Cup.

Diagnostic imaging is the predominant medical service sought for the assessment and staging of musculoskeletal injuries in professional sports events. During the 2022 FIFA Football (soccer) World Cup, a centralized medical care infrastructure was established. This article provides a comprehensive account of the radiological services implemented during this event, encompassing the deployment of equipment and human resources, the structuring of workflows to uphold athlete confidentiality, and initiatives aimed at enhancing communication. Communication channels were refined through radiology consultations held with national teams' health care providers and the adoption of audiovisual reports available in multiple languages, which could be accessed remotely by team physicians. Our established framework can be replicated in international professional football events for seamless integration and efficacy.

Structured reporting of computed tomography in the polytrauma patient assessment: a Delphi consensus proposal.

OBJECTIVES: To develop a structured reporting (SR) template for whole-body CT examinations of polytrauma patients, based on the consensus of a panel of emergency radiology experts from the Italian Society of Medical and Interventional Radiology. METHODS: A multi-round Delphi method was used to quantify inter-panelist agreement for all SR sections. Internal consistency for each section and quality analysis in terms of average inter-item correlation were evaluated by means of the Cronbach's alpha (Cα) correlation coefficient. RESULTS: The final SR form included 118 items (6 in the "Patient Clinical Data" section, 4 in the "Clinical Evaluation" section, 9 in the "Imaging Protocol" section, and 99 in the "Report" section). The experts' overall mean score and sum of scores were 4.77 (range 1-5) and 257.56 (range 206-270) in the first Delphi round, and 4.96 (range 4-5) and 208.44 (range 200-210) in the second round, respectively. In the second Delphi round, the experts' overall mean score was higher than in the first round, and standard deviation was lower (3.11 in the second round vs 19.71 in the first round), reflecting a higher expert agreement in the second round. Moreover, Cα was higher in the second round than in the first round (0.97 vs 0.87). CONCLUSIONS: Our SR template for whole-body CT examinations of polytrauma patients is based on a strong agreement among panel experts in emergency radiology and could improve communication between radiologists and the trauma team.

Artificial Intelligence to Improve Patient Understanding of Radiology Reports.

Diagnostic imaging reports are generally written with a target audience of other providers. As a result, the reports are written with medical jargon and technical detail to ensure accurate communication. With implementation of the 21st Century Cures Act, patients have greater and quicker access to their imaging reports, but these reports are still written above the comprehension level of the average patient. Consequently, many patients have requested reports to be conveyed in language accessible to them. Numerous studies have shown that improving patient understanding of their condition results in better outcomes, so driving comprehension of imaging reports is essential. Summary statements, second reports, and the inclusion of the radiologist's phone number have been proposed, but these solutions have implications for radiologist workflow. Artificial intelligence (AI) has the potential to simplify imaging reports without significant disruptions. Many AI technologies have been applied to radiology reports in the past for various clinical and research purposes, but patient focused solutions have largely been ignored. New natural language processing technologies and large language models (LLMs) have the potential to improve patient understanding of their imaging reports. However, LLMs are a nascent technology and significant research is required before LLM-driven report simplification is used in patient care.

Advancements in Standardizing Radiological Reports: A Comprehensive Review.

Standardized radiological reports stimulate debate in the medical imaging field. This review paper explores the advantages and challenges of standardized reporting. Standardized reporting can offer improved clarity and efficiency of communication among radiologists and the multidisciplinary team. However, challenges include limited flexibility, initially increased time and effort, and potential user experience issues. The efforts toward standardization are examined, encompassing the establishment of reporting templates, use of common imaging lexicons, and integration of clinical decision support tools. Recent technological advancements, including multimedia-enhanced reporting and AI-driven solutions, are discussed for their potential to improve the standardization process. Organizations such as the ACR, ESUR, RSNA, and ESR have developed standardized reporting systems, templates, and platforms to promote uniformity and collaboration. However, challenges remain in terms of workflow adjustments, language and format variability, and the need for validation. The review concludes by presenting a set of ten essential rules for creating standardized radiology reports, emphasizing clarity, consistency, and adherence to structured formats.

Video Radiology Reports: A Valuable Tool to Improve Patient-Centered Radiology.

BACKGROUND. Improved communication between radiologists and patients is a key component of patient-centered radiology. OBJECTIVE. The purpose of this study was to create patient-centered video radiology reports using simple-to-understand language and annotated images and to assess the effect of these reports on patients' experience and understanding of their imaging results. METHODS. During a 4-month study period, faculty radiologists created video radiology reports using a tool integrated within the diagnostic viewer that allows both image and voice capture. To aid patients' understanding of cross-sectional images, cinematically rendered images were automatically created and made immediately available to radiologists at the workstation, allowing their incorporation into video radiology reports. Video radiology reports were made available to patients via the institutional health portal along with the written radiology report and the examination images. Patient views of the video report were recorded, and descriptive analyses were performed on radiologist and examination characteristics as well as patient demographics. A survey was sent to patients to obtain feedback on their experience. RESULTS. During the study period, 105 of 227 faculty radiologists created 3763 video radiology reports (mean number of reports per radiologist, 36 ± 27 [SD] reports). Mean time to create a video report was 238 ± 141 seconds. Patients viewed 864 unique video reports. The mean overall video radiology report experience rating based on 101 patient surveys was 4.7 of 5. The mean rating for how well the video report helped patients understand their findings was also 4.7 of 5. Of the patients who responded to the survey, 91% preferred having both written and video reports together over having written reports alone. CONCLUSION. Patient-centered video radiology reports are a useful tool to help improve patient understanding of imaging results. The mechanism of creating the video reports and delivering them to patients can be integrated into existing informatics infrastructure. CLINICAL IMPACT. Video radiology reports can play an important role in patient-centered radiology, increasing patient understanding of imaging results, and they may improve the visibility of radiologists to patients and highlight the radiologist's important role in patient care.

Development and application of an electronic synoptic report for reporting and management of low-dose computed tomography lung cancer screening examination.

BACKGROUND: Interpretation of Low Dose CT scans and protocol driven management of findings is a key aspect of lung cancer screening program performance. Reliable and reproducible methods are needed to communicate radiologists' interpretation to the screening program or clinicians driving management decision. METHODS: We performed an audit of a subset of dictated reports from the PANCAN study to assess for omissions. We developed an electronic synoptic reporting tool for radiologists embedded in a clinical documentation system software. The tool was then used for reporting as part of the Alberta Lung Cancer Screening Study and McGill University Health Centre Pilot Lung Cancer Screening Program. RESULTS: Fifty reports were audited for completeness. At least one omission was noted in 30 (70%) of reports, with a major omission (missing lobe, size, type of nodule in report or actionable incidental finding in recommendation section of report) in 24 (48%). Details of the reporting template and functionality such as automated nodule cancer risk assessment, Lung-RADS category assignment, auto-generated narrative type report as well as personalize participant results letter is provided. A description of the system's performance in its application in 2815 CT reports is then summarized. CONCLUSIONS: We found that narrative type radiologist reports for lung cancer screening CT examinations frequently lacked specific discrete data elements required for management. We demonstrate the successful implementation of a radiology synoptic reporting system for use in lung cancer screening, and the use of this information to drive program management and communications.

Determining the applicability of the RSNA radiology lexicon (RadLex) in high-grade glioma MRI reporting-a preliminary study on 20 consecutive cases with newly diagnosed glioblastoma.

BACKGROUND: The implementation of a collective terminology in radiological reporting such as the RSNA radiological lexicon (RadLex) yields many benefits including unambiguous communication of findings, improved education, and fostering data mining for research purposes. While some fields in general radiology have already been evaluated so far, this is the first exploratory approach to assess the applicability of the RadLex terminology to glioblastoma (GBM) MRI reporting. METHODS: Preoperative brain MRI reports of 20 consecutive patients with newly diagnosed GBM (mean age 68.4 ± 10.8 years; 12 males) between January and October 2010 were retrospectively identified. All terms related to the tumor as well as their frequencies of mention were extracted from the MRI reports by two independent neuroradiologists. Every item was subsequently analyzed with respect to an equivalent RadLex representation and classified into one of four groups as follows: 1. verbatim RadLex entity, 2. synonymous/multiple equivalent(s), 3. combination of RadLex concepts, or 4. no RadLex equivalent. Additionally, verbatim entities were categorized using the hierarchical RadLex Tree Browser. RESULTS: A total of 160 radiological terms were gathered. 123/160 (76.9%) items showed literal RadLex equivalents, 9/160 (5.6%) items had synonymous (non-verbatim) or multiple counterparts, 21/160 (13.1%) items were represented by means of a combination of concepts, and 7/160 (4.4%) entities could not eventually be transferred adequately into the RadLex ontology. CONCLUSIONS: Our results suggest a sufficient term coverage of the RadLex terminology for GBM MRI reporting. If applied extensively, it may improve communication of radiological findings and facilitate data mining for large-scale research purposes.

Automatic text classification of actionable radiology reports of tinnitus patients using bidirectional encoder representations from transformer (BERT) and in-domain pre-training (IDPT).

BACKGROUND: Given the increasing number of people suffering from tinnitus, the accurate categorization of patients with actionable reports is attractive in assisting clinical decision making. However, this process requires experienced physicians and significant human labor. Natural language processing (NLP) has shown great potential in big data analytics of medical texts; yet, its application to domain-specific analysis of radiology reports is limited. OBJECTIVE: The aim of this study is to propose a novel approach in classifying actionable radiology reports of tinnitus patients using bidirectional encoder representations from transformer BERT-based models and evaluate the benefits of in domain pre-training (IDPT) along with a sequence adaptation strategy. METHODS: A total of 5864 temporal bone computed tomography(CT) reports are labeled by two experienced radiologists as follows: (1) normal findings without notable lesions; (2) notable lesions but uncorrelated to tinnitus; and (3) at least one lesion considered as potential cause of tinnitus. We then constructed a framework consisting of deep learning (DL) neural networks and self-supervised BERT models. A tinnitus domain-specific corpus is used to pre-train the BERT model to further improve its embedding weights. In addition, we conducted an experiment to evaluate multiple groups of max sequence length settings in BERT to reduce the excessive quantity of calculations. After a comprehensive comparison of all metrics, we determined the most promising approach through the performance comparison of F1-scores and AUC values. RESULTS: In the first experiment, the BERT finetune model achieved a more promising result (AUC-0.868, F1-0.760) compared with that of the Word2Vec-based models(AUC-0.767, F1-0.733) on validation data. In the second experiment, the BERT in-domain pre-training model (AUC-0.948, F1-0.841) performed significantly better than the BERT based model(AUC-0.868, F1-0.760). Additionally, in the variants of BERT fine-tuning models, Mengzi achieved the highest AUC of 0.878 (F1-0.764). Finally, we found that the BERT max-sequence-length of 128 tokens achieved an AUC of 0.866 (F1-0.736), which is almost equal to the BERT max-sequence-length of 512 tokens (AUC-0.868,F1-0.760). CONCLUSION: In conclusion, we developed a reliable BERT-based framework for tinnitus diagnosis from Chinese radiology reports, along with a sequence adaptation strategy to reduce computational resources while maintaining accuracy. The findings could provide a reference for NLP development in Chinese radiology reports.

Structured reporting of computed tomography in the staging of colon cancer: a Delphi consensus proposal.

BACKGROUND: Structured reporting (SR) in radiology is becoming increasingly necessary and has been recognized recently by major scientific societies. This study aims to build structured CT-based reports in colon cancer during the staging phase in order to improve communication between the radiologist, members of multidisciplinary teams and patients. MATERIALS AND METHODS: A panel of expert radiologists, members of the Italian Society of Medical and Interventional Radiology, was established. A modified Delphi process was used to develop the SR and to assess a level of agreement for all report sections. Cronbach's alpha (Cα) correlation coefficient was used to assess internal consistency for each section and to measure quality analysis according to the average inter-item correlation. RESULTS: The final SR version was built by including n = 18 items in the "Patient Clinical Data" section, n = 7 items in the "Clinical Evaluation" section, n = 9 items in the "Imaging Protocol" section and n = 29 items in the "Report" section. Overall, 63 items were included in the final version of the SR. Both in the first and second round, all sections received a higher than good rating: a mean value of 4.6 and range 3.6-4.9 in the first round; a mean value of 5.0 and range 4.9-5 in the second round. In the first round, Cronbach's alpha (Cα) correlation coefficient was a questionable 0.61. In the first round, the overall mean score of the experts and the sum of scores for the structured report were 4.6 (range 1-5) and 1111 (mean value 74.07, STD 4.85), respectively. In the second round, Cronbach's alpha (Cα) correlation coefficient was an acceptable 0.70. In the second round, the overall mean score of the experts and the sum of score for structured report were 4.9 (range 4-5) and 1108 (mean value 79.14, STD 1.83), respectively. The overall mean score obtained by the experts in the second round was higher than the overall mean score of the first round, with a lower standard deviation value to underline greater agreement among the experts for the structured report reached in this round. CONCLUSIONS: A wide implementation of SR is of critical importance in order to offer referring physicians and patients optimum quality of service and to provide researchers with the best quality data in the context of big data exploitation of available clinical data. Implementation is a complex procedure, requiring mature technology to successfully address the multiple challenges of user-friendliness, organization and interoperability.

Extracting clinical terms from radiology reports with deep learning.

Extracting clinical terms from free-text format radiology reports is a first important step toward their secondary use. However, there is no general consensus on the kind of terms to be extracted. In this paper, we propose an information model comprising three types of clinical entities: observations, clinical findings, and modifiers. Furthermore, to determine its applicability for in-house radiology reports, we extracted clinical terms with state-of-the-art deep learning models and compared the results. We trained and evaluated models using 540 in-house chest computed tomography (CT) reports annotated by multiple medical experts. Two deep learning models were compared, and the effect of pre-training was explored. To investigate the generalizability of the model, we evaluated the use of other institutional chest CT reports. The micro F1-score of our best performance model using in-house and external datasets were 95.36% and 94.62%, respectively. Our results indicated that entities defined in our information model were suitable for extracting clinical terms from radiology reports, and the model was sufficiently generalizable to be used with dataset from other institutions.

Effect of AI Explanations on Human Perceptions of Patient-Facing AI-Powered Healthcare Systems.

Ongoing research efforts have been examining how to utilize artificial intelligence technology to help healthcare consumers make sense of their clinical data, such as diagnostic radiology reports. How to promote the acceptance of such novel technology is a heated research topic. Recent studies highlight the importance of providing local explanations about AI prediction and model performance to help users determine whether to trust AI's predictions. Despite some efforts, limited empirical research has been conducted to quantitatively measure how AI explanations impact healthcare consumers' perceptions of using patient-facing, AI-powered healthcare systems. The aim of this study is to evaluate the effects of different AI explanations on people's perceptions of AI-powered healthcare system. In this work, we designed and deployed a large-scale experiment (N = 3,423) on Amazon Mechanical Turk (MTurk) to evaluate the effects of AI explanations on people's perceptions in the context of comprehending radiology reports. We created four groups based on two factors-the extent of explanations for the prediction (High vs. Low Transparency) and the model performance (Good vs. Weak AI Model)-and randomly assigned participants to one of the four conditions. Participants were instructed to classify a radiology report as describing a normal or abnormal finding, followed by completing a post-study survey to indicate their perceptions of the AI tool. We found that revealing model performance information can promote people's trust and perceived usefulness of system outputs, while providing local explanations for the rationale of a prediction can promote understandability but not necessarily trust. We also found that when model performance is low, the more information the AI system discloses, the less people would trust the system. Lastly, whether human agrees with AI predictions or not and whether the AI prediction is correct or not could also influence the effect of AI explanations. We conclude this paper by discussing implications for designing AI systems for healthcare consumers to interpret diagnostic report.

Use of an Online Crowdsourcing Platform to Assess Patient Comprehension of Radiology Reports and Colloquialisms.

OBJECTIVE. The purpose of this study was to use an online crowdsourcing platform to assess patient comprehension of five radiology reporting templates and radiology colloquialisms. MATERIALS AND METHODS. In this cross-sectional study, participants were surveyed as patient surrogates using a crowdsourcing platform. Two tasks were completed within two 48-hour time periods. For the first crowdsourcing task, each participant was randomly assigned a set of radiology reports in a constructed reporting template and subsequently tested for comprehension. For the second crowdsourcing task, each participant was randomly assigned a radiology colloquialism and asked to indicate whether the phrase indicated a normal, abnormal, or ambivalent finding. RESULTS. A total of 203 participants enrolled for the first task and 1166 for the second within 48 hours of task publication. The payment totaled $31.96. Of 812 radiology reports read, 384 (47%) were correctly interpreted by the patient surrogates. Patient surrogates had higher rates of comprehension of reports written in the patient summary (57%, p < 0.001) and traditional unstructured in combination with patient summary (51%, p = 0.004) formats than in the traditional unstructured format (40%). Most of the patient surrogates (114/203 [56%]) expressed a preference for receiving a full radiology report via an electronic patient portal. Several radiology colloquialisms with modifiers such as "low," "underdistended," and "decompressed" had low rates of comprehension. CONCLUSION. Use of the crowdsourcing platform is an expeditious, cost-effective, and customizable tool for surveying laypeople in sentiment- or task-based research. Patient summaries can help increase patient comprehension of radiology reports. Radiology colloquialisms are likely to be misunderstood by patients.

Understanding spatial language in radiology: Representation framework, annotation, and spatial relation extraction from chest X-ray reports using deep learning.

Radiology reports contain a radiologist's interpretations of images, and these images frequently describe spatial relations. Important radiographic findings are mostly described in reference to an anatomical location through spatial prepositions. Such spatial relationships are also linked to various differential diagnoses and often described through uncertainty phrases. Structured representation of this clinically significant spatial information has the potential to be used in a variety of downstream clinical informatics applications. Our focus is to extract these spatial representations from the reports. For this, we first define a representation framework based on the Spatial Role Labeling (SpRL) scheme, which we refer to as Rad-SpRL. In Rad-SpRL, common radiological entities tied to spatial relations are encoded through four spatial roles: Trajector, Landmark, Diagnosis, and Hedge, all identified in relation to a spatial preposition (or Spatial Indicator). We annotated a total of 2,000 chest X-ray reports following Rad-SpRL. We then propose a deep learning-based natural language processing (NLP) method involving word and character-level encodings to first extract the Spatial Indicators followed by identifying the corresponding spatial roles. Specifically, we use a bidirectional long short-term memory (Bi-LSTM) conditional random field (CRF) neural network as the baseline model. Additionally, we incorporate contextualized word representations from pre-trained language models (BERT and XLNet) for extracting the spatial information. We evaluate both gold and predicted Spatial Indicators to extract the four types of spatial roles. The results are promising, with the highest average F1 measure for Spatial Indicator extraction being 91.29 (XLNet); the highest average overall F1 measure considering all the four spatial roles being 92.9 using gold Indicators (XLNet); and 85.6 using predicted Indicators (BERT pre-trained on MIMIC notes). The corpus is available in Mendeley at http://dx.doi.org/10.17632/yhb26hfz8n.1 and https://github.com/krobertslab/datasets/blob/master/Rad-SpRL.xml.

Assisting radiologists with reporting urgent findings to referring physicians: A machine learning approach to identify cases for prompt communication.

Radiologists are expected to expediently communicate critical and unexpected findings to referring clinicians to prevent delayed diagnosis and treatment of patients. However, competing demands such as heavy workload along with lack of administrative support resulted in communication failures that accounted for 7% of the malpractice payments made from 2004 to 2008 in the United States. To address this problem, we have developed a novel machine learning method that can automatically and accurately identify cases that require prompt communication to referring physicians based on analyzing the associated radiology reports. This semi-supervised learning approach requires a minimal amount of manual annotations and was trained on a large multi-institutional radiology report repository from three major external healthcare organizations. To test our approach, we created a corpus of 480 radiology reports from our own institution and double-annotated cases that required prompt communication by two radiologists. Our evaluation on the test corpus achieved an F-score of 74.5% and recall of 90.0% in identifying cases for prompt communication. The implementation of the proposed approach as part of an online decision support system can assist radiologists in identifying radiological cases for prompt communication to referring physicians to avoid or minimize potential harm to patients.

Automated Detection of Measurements and Their Descriptors in Radiology Reports Using a Hybrid Natural Language Processing Algorithm.

Radiological measurements are reported in free text reports, and it is challenging to extract such measures for treatment planning such as lesion summarization and cancer response assessment. The purpose of this work is to develop and evaluate a natural language processing (NLP) pipeline that can extract measurements and their core descriptors, such as temporality, anatomical entity, imaging observation, RadLex descriptors, series number, image number, and segment from a wide variety of radiology reports (MR, CT, and mammogram). We created a hybrid NLP pipeline that integrates rule-based feature extraction modules and conditional random field (CRF) model for extraction of the measurements from the radiology reports and links them with clinically relevant features such as anatomical entities or imaging observations. The pipeline was trained on 1117 CT/MR reports, and performance of the system was evaluated on an independent set of 100 expert-annotated CT/MR reports and also tested on 25 mammography reports. The system detected 813 out of 806 measurements in the CT/MR reports; 784 were true positives, 29 were false positives, and 0 were false negatives. Similarly, from the mammography reports, 96% of the measurements with their modifiers were extracted correctly. Our approach could enable the development of computerized applications that can utilize summarized lesion measurements from radiology report of varying modalities and improve practice by tracking the same lesions along multiple radiologic encounters.

Evaluation of Audiovisual Reports to Enhance Traditional Emergency Musculoskeletal Radiology Reports.

Traditional radiology reports are narrative texts that include a description of imaging findings. Recent implementation of advanced reporting software allows for incorporation of annotated key images and hyperlinks directly into text reports, but these tools usually do not substitute in-person consultations with radiologists, especially in challenging cases. Use of on-demand audio/visual reports with screen capture software is an emerging technology, providing a more engaged imaging service. Our study evaluates a video reporting tool that utilizes PACS integrated screen capture software for musculoskeletal imaging studies in the emergency department. Our hypothesis is that referring orthopedic surgeons would find that recorded audio/video reports add value to conventional reports, may increase engagement with radiology staff, and also facilitate understanding of imaging findings from urgent musculoskeletal cases. Seven radiologists prepared a total of 47 audiovisual reports for 9 attending orthopedic surgeons from the emergency department. We applied two surveys to evaluate the experience of the referring physicians using audio/visual reports as a complementary material from the conventional text report. Positive responses were statistically significant in most questions including: if the clinical suspicion was answered in the video; willingness to use such technology in other cases; if the audiovisual report made the imaging findings more understandable than the traditional report; and if the audiovisual report is faster to understand than the traditional text report. Use of audiovisual reports in emergency musculoskeletal cases is a new approach to evaluate potentially challenging cases. These results support the potential of this technology to re-establish the radiologist's role as an essential member of patient care and also provide more engaging, precise, and personalized reports. Further studies could streamline these methods in order to minimize work redundancy with traditional text reporting or even evaluate acceptance of using only audiovisual radiology reports. Additionally, widespread adoption would require integration with the entire radiology workflow including non-urgent cases and other medical specialties.

Incomplete Thyroid Ultrasound Reports for Patients With Thyroid Nodules: Implications Regarding Risk Assessment and Management.

OBJECTIVE: The purpose of this study was to determine the completeness of thyroid ultrasound (US) reports, assess for differences in report interpretation by clinicians, and evaluate for implications in patient care. MATERIALS AND METHODS: We retrospectively reviewed thyroid US examinations performed between January and June 2013 in Nova Scotia, Canada. Baseline examinations that identified a nodule were evaluated for 10 reporting elements. Reports that lacked a comment regarding malignancy risk or a recommendation for biopsy were considered unclassified and were graded by three clinical specialists in accordance with the 2015 American Thyroid Association management guidelines. Interrater agreement was assessed using the Cohen kappa statistic. A radiologist reviewed the images of unclassified nodules, and on the basis of radiologic grading, biopsy rates and pathologic findings were compared between nodules that did and did not warrant biopsy. RESULTS: Of 971 first-time thyroid US studies, 478 detected a nodule. The number of reports lacking a comment on the 10 elements ranged from 154 to 433 (32-91%). A total of 222 nodules (46%) were unclassified, and agreement in assigned grading by the clinical specialists was very poor (κ = 0.07; p < 0.05). According to radiologist grading, only 57 of 127 biopsies were performed on nodules that warranted biopsy, and 16 of 95 biopsies were performed unnecessarily. On the basis of the three clinical specialists' interpretation, 10, 31, and 33 reports were considered too incomplete to assign a grade; 40, 10, and four biopsies would have been unnecessarily ordered; and zero, three, and four cancers would have been missed. CONCLUSION: There is widespread underreporting of established elements in thyroid US reports, and this causes confusion and discrepancy among clinical specialists regarding the risk of malignancy and the need for biopsy.

Nonstandardized Terminology to Describe Focal Liver Lesions in Patients at Risk for Hepatocellular Carcinoma: Implications Regarding Clinical Communication.

OBJECTIVE: The purpose of this study is to determine the correlation between malignancy risk of focal liver observations in patients at risk for hepatocellular carcinoma (HCC) implied by phrases used in nonstructured radiology reports with the risk inferred by hepatologists. MATERIALS AND METHODS: We performed a retrospective review of nonstructured radiology reports issued before Liver Imaging and Reporting Data System (LI-RADS) adoption from four-phase liver CT examinations of patients at risk for HCC. The phrase used by the radiologist in the report impression to describe each focal liver observation was recorded. Five hepatologists independently inferred the LI-RADS category from each phrase. Two abdominal radiologists independently reviewed the images and, blinded to all other information, assigned a LI-RADS category to each observation. Discrepancies were resolved by consensus. RESULTS: One hundred five observations in 77 patients were reported by 23 radiologists using 29 phrases. The most common phrase, "consistent with HCC" (n = 20), was applied to radiologist-assigned LR-3 (n = 1), LR-4 (n = 5), LR-5 (n = 11), and LR-5V (n = 3) observations. Eleven phrases were used more than once. Sixteen phrases were associated with LR-4 or higher observations; among these, hepatologists misinterpreted 37% of LR-4 or lower observations as definitely HCC and 46% of LR-5 and LR-5V observations as not definitely HCC. Overall, there was modest correlation (r = 0.69) between radiologist-assigned and hepatologist-inferred categories. CONCLUSION: Nonstandardized terminology results in inaccurate communication of HCC risk. Structured reporting systems such as LI-RADS may improve communication by conveying unambiguous estimates of malignancy risk.

Transferability of artificial neural networks for clinical document classification across hospitals: A case study on abnormality detection from radiology reports.

OBJECTIVE: Application of machine learning techniques for automatic and reliable classification of clinical documents have shown promising results. However, machine learning models require abundant training data specific to each target hospital and may not be able to benefit from available labeled data from each of the hospitals due to data variations. Such training data limitations have presented one of the major obstacles for maximising potential application of machine learning approaches in the healthcare domain. We investigated transferability of artificial neural network models across hospitals from different domains representing various age demographic groups (i.e., children, adults, and mixed) in order to cope with such limitations. MATERIALS AND METHODS: We explored the transferability of artificial neural networks for clinical document classification. Our case study was to detect abnormalities from limb X-ray reports obtained from the emergency department (ED) of three hospitals within different domains. Different transfer learning scenarios were investigated in order to employ a source hospital's trained model for addressing a target hospital's abnormality detection problem. RESULTS: A Convolutional Neural Network (CNN) model exhibited the best effectiveness compared to other networks when employing an embedding model trained on a large corpus of clinical documents. Furthermore, CNN models derived from a source hospital outperformed a conventional machine learning approach based on Support Vector Machines (SVM) when applied to a different (target) hospital. These models were further improved by leveraging available training data in target hospitals and outperformed the models that used only the target hospital data with F1-Score of 0.92-0.96 across three hospitals. DISCUSSION: Our transfer learning model used only simple vector representations of documents without any task-specific feature engineering. Transferring the CNN model significantly improved (approx.10% in F1-Score) the state-of-the-art approach for clinical document classification based on a trivial transferred model. In addition, the results showed that transfer learning techniques can further improve a CNN model that is trained only on either a source or target hospital's data. CONCLUSION: Transferring a pre-trained CNN model generated in one hospital to another facilitates application of machine learning approaches that alleviate both hospital-specific feature engineering and training data.