Do radiologists still need to report chest x rays?

Use of artificial intelligence in triaging of chest radiographs to reduce radiologists' workload

OBJECTIVES

To evaluate whether deep learning-based detection algorithms (DLD)-based triaging can reduce outpatient chest radiograph interpretation workload while maintaining noninferior sensitivity.

METHODS

This retrospective study included patients who underwent initial chest radiography at the outpatient clinic between June 1 and June 30, 2017. Readers interpreted radiographs with/without a commercially available DLD that detects nine radiologic findings (atelectasis, calcification, cardiomegaly, consolidation, fibrosis, nodules, pneumothorax, pleural effusion, and pneumoperitoneum). The reading order was determined in a randomized, crossover manner. The radiographs were classified into negative and positive examinations. In a 50% worklist reduction scenario, radiographs were sorted in descending order of probability scores: the lower half was regarded as negative exams, while the remaining were read with DLD by radiologists. The primary analysis evaluated noninferiority in sensitivity between radiologists reading all radiographs and simulating a 50% worklist reduction, with the inferiority margin of 5%. The specificities were compared using McNemar's test.

RESULTS

The study included 1964 patients (median age [interquartile range], 55 years [40-67 years]). The sensitivity was 82.6% (195 of 236; 95% CI: 77.5%, 87.3%) when readers interpreted all chest radiographs without DLD and 83.5% (197 of 236; 95% CI: 78.8%, 88.1%) in the 50% worklist reduction scenario. The difference in sensitivity was 0.8% (95% CI: - 3.8%, 5.5%), establishing noninferiority of 50% worklist reduction (p = 0.01). The specificity increased from 86.7% (1498 of 1728) to 90.4% (1562 of 1728) (p < 0.001) with DLD-based triage.

CONCLUSION

Deep learning-based triaging may substantially reduce workload without lowering sensitivity while improving specificity.

CLINICAL RELEVANCE STATEMENT

Substantial workload reduction without lowering sensitivity was feasible using deep learning-based triaging of outpatient chest radiograph; however, the legal responsibility for incorrect diagnoses based on AI-standalone interpretation remains an issue that should be defined before clinical implementation.

KEY POINTS

• A 50% workload reduction simulation using deep learning-based detection algorithm maintained noninferior sensitivity while improving specificity. • The CT recommendation rate significantly decreased in the disease-negative patients, whereas it slightly increased in the disease-positive group without statistical significance. • In the exploratory analysis, the noninferiority of sensitivity was maintained until 70% of the workload was reduced; the difference in sensitivity was 0%.

Assessing and comparing chest radiograph interpretation in the Department of Internal Medicine at the University of the Witwatersrand medical school, according to seniority

Background

Chest radiographs are a common diagnostic tool in the internal medicine department, and correct interpretation is imperative for adequate patient management.

Objectives

To determine the diagnostic accuracy of common pathologies in South Africa that are evident on chest radiographs, and to determine whether there are discrepancies according to different levels of qualification of doctors rotating through the internal medicine department, and which factors contribute to an accurate diagnosis.

Methods

Fifteen chest radiographs with common pathologies were given to all doctors rotating through the Department of Internal Medicine at Chris Hani Baragwanath Academic Hospital, and they were asked to interpret them. Information pertaining to their experience, designation and confidence in chest radiograph interpretation was also obtained.

Results

Diagnostic accuracy according to years of experience was as follows: 0 - 5 years 27.0%, 6 - 10 years 43.0%, and >10 years 47.9%. For different designations, accuracy was as follows: consultants 50.5%, registrars 40.9%, medical officers 36.4%, and interns 19.5%. Participants who were confident obtained a mean score of 39.4% and those who were not, a mean score of 31.6%.

Conclusion

Chest radiographs are readily accessible and used daily in clinical practice in numerous facilities. An accurate diagnosis is important to provide quality healthcare. Improved training in interpretation for all, but especially for junior doctors, should be a priority in our training facilities.

Study synopsis

What the study adds. This study tested the diagnostic accuracy with regard to common pathologies present on chest X ray by doctors rotating through, or stationed at the internal medicine department at an academic hospital. Implications of the findings. Interpretation of chest X-rays was generally poor but the study did find that this improves with experience and confidence in diagnostic ability. These findings are significant in that they indicate a need to implement improved teaching programs in radiological interpretation, especially at an undergraduate level.

Performance and Agreement When Annotating Chest X-ray Text Reports—A Preliminary Step in the Development of a Deep Learning-Based Prioritization and Detection System

A chest X-ray report is a communicative tool and can be used as data for developing artificial intelligence-based decision support systems. For both, consistent understanding and labeling is important. Our aim was to investigate how readers would comprehend and annotate 200 chest X-ray reports. Reports written between 1 January 2015 and 11 March 2022 were selected based on search words. Annotators included three board-certified radiologists, two trained radiologists (physicians), two radiographers (radiological technicians), a non-radiological physician, and a medical student. Consensus labels by two or more of the experienced radiologists were considered “gold standard”. Matthew’s correlation coefficient (MCC) was calculated to assess annotation performance, and descriptive statistics were used to assess agreement between individual annotators and labels. The intermediate radiologist had the best correlation to “gold standard” (MCC 0.77). This was followed by the novice radiologist and medical student (MCC 0.71 for both), the novice radiographer (MCC 0.65), non-radiological physician (MCC 0.64), and experienced radiographer (MCC 0.57). Our findings showed that for developing an artificial intelligence-based support system, if trained radiologists are not available, annotations from non-radiological annotators with basic and general knowledge may be more aligned with radiologists compared to annotations from sub-specialized medical staff, if their sub-specialization is outside of diagnostic radiology.

Added Value of Bone Suppression Image in the Detection of Subtle Lung Lesions on Chest Radiographs with Regard to Reader's Expertise

BACKGROUND

Chest radiographs (CXR) are the most commonly used imaging techniques by various clinicians and radiologists. However, detecting lung lesions on CXR depends largely on the reader's experience level, so there have been several trials to overcome this problem using post-processing of CXR. We investigated the added value of bone suppression image (BSI) in detecting various subtle lung lesions on CXR with regard to reader's expertise.

METHODS

We applied a software program to generate BSI in 1,600 patients in the emergency department. Of them, 80 patients with subtle lung lesions and 80 patients with negative finding on CXR were retrospectively selected based on the subtlety scores on CXR and CT findings. Ten readers independently rated their confidence in deciding the presence or absence of a lung lesion at each of 960 lung regions on the two separated imaging sessions: CXR alone vs. CXR with BSI.

RESULTS

The additional use of BSI for all readers significantly increased the mean area under the curve (AUC) in detecting subtle lung lesions (0.663 vs. 0.706; P < 0.001). The less experienced readers were, the more AUC differences increased: 0.067 (P < 0.001) for junior radiology residents; 0.064 (P < 0.001) for non-radiology clinicians; 0.044 (P < 0.001) for senior radiology residents; and 0.019 (P = 0.041) for chest radiologists. The additional use of BSI significantly increased the mean confidence regarding the presence or absence of lung lesions for 213 positive lung regions (2.083 vs. 2.357; P < 0.001) and for 747 negative regions (1.217 vs. 1.195; P = 0.008).

CONCLUSION

The use of BSI increases diagnostic performance and confidence, regardless of reader's expertise, reduces the impact of reader's expertise and can be helpful for less experienced clinicians and residents in the detection of subtle lung lesions.

Does the Use of a Checklist Help Medical Students in the Detection of Abnormalities on a Chest Radiograph?

The interpretation of chest radiographs is a complex task that is prone to diagnostic error, especially for medical students. The aim of this study is to investigate the extent to which medical students benefit from the use of a checklist regarding the detection of abnormalities on a chest radiograph. We developed a checklist based on literature and interviews with experienced thorax radiologists. Forty medical students in the clinical phase assessed 18 chest radiographs during a computer test, either with (n = 20) or without (n = 20) the checklist. We measured performance and asked participants for feedback using a survey. Participants that used a checklist detected more abnormalities on images with multiple abnormalities (M = 50.1%) than participants that could not use a checklist (M = 41.9%), p = 0.04. The post-experimental survey shows that on average, participants considered the checklist helpful (M = 3.25 on a five-point scale), but also time consuming (M = 3.30 on a five-point scale). In conclusion, a checklist can help medical students to detect abnormalities in chest radiographs. Moreover, students tend to appreciate the use of a checklist as a helpful tool during the interpretation of a chest radiograph. Therefore, a checklist is a potentially important tool to improve radiology education in the medical curriculum.

Impact of donor chest radiography on clinical outcome after lung transplantation

Background

Organ donation guidelines recommend a “clear” conventional bedside chest radiograph before lung transplantation despite only moderate accuracy for cardiopulmonary abnormalities.

Purpose

To evaluate the influence of donor image interpretation on lung transplantation outcome in recipients by following early and late complications, one-year survival, and to correlate imaging findings and blood gas analysis with lung transplantation outcome in recipients.

Material and Methods

In 35 lung donors from a single institution clinical reports and study reviews of imaging findings of the mandatory bedside chest radiographs and blood gas analyses were compared with clinical outcome in 38 recipients. Hospitalization time, peri- and postoperative complications, early complications (primary graft dysfunction, infection), 30-day and one-year survival, and forced expiratory volume in 1 s percentage of predicted normal value (FEV1%) at one-year follow-up were analyzed.

Results

Findings in clinical reports and study reviews differed substantially, e.g. regarding reported decompensation, edema, infection, and atelectasis. No correlation was shown between imaging findings in clinical report or study review and blood gas analyses in the lung donors compared to postoperative outcome in recipients.

Conclusion

The interpretation of the mandatory chest radiograph in its present form does not influence one-year outcome in lung transplantation. Larger imaging studies or a change in clinical routine including computed tomography may provide evidence for future guidelines.

Individualised perioperative open-lung approach versus standard protective ventilation in abdominal surgery (iPROVE): a randomised controlled trial

BACKGROUND

The effects of individualised perioperative lung-protective ventilation (based on the open-lung approach [OLA]) on postoperative complications is unknown. We aimed to investigate the effects of intraoperative and postoperative ventilatory management in patients scheduled for abdominal surgery, compared with standard protective ventilation.

METHODS

We did this prospective, multicentre, randomised controlled trial in 21 teaching hospitals in Spain. We enrolled patients who were aged 18 years or older, were scheduled to have abdominal surgery with an expected time of longer than 2 h, had intermediate-to-high-risk of developing postoperative pulmonary complications, and who had a body-mass index less than 35 kg/m². Patients were randomly assigned (1:1:1:1) online to receive one of four lung-protective ventilation strategies using low tidal volume plus positive end-expiratory pressure (PEEP): open-lung approach (OLA)-iCPAP (individualised intraoperative ventilation [individualised PEEP after a lung recruitment manoeuvre] plus individualised postoperative continuous positive airway pressure [CPAP]), OLA-CPAP (intraoperative individualised ventilation plus postoperative CPAP), STD-CPAP (standard intraoperative ventilation plus postoperative CPAP), or STD-O₂ (standard intraoperative ventilation plus standard postoperative oxygen therapy). Patients were masked to treatment allocation. Investigators were not masked in the operating and postoperative rooms; after 24 h, data were given to a second investigator who was masked to allocations. The primary outcome was a composite of pulmonary and systemic complications during the first 7 postoperative days. We did the primary analysis using the modified intention-to-treat population. This trial is registered with ClinicalTrials.gov, number NCT02158923.

FINDINGS

Between Jan 2, 2015, and May 18, 2016, we enrolled 1012 eligible patients. Data were available for 967 patients, whom we included in the final analysis. Risk of pulmonary and systemic complications did not differ for patients in OLA-iCPAP (110 [46%] of 241, relative risk 0·89 [95% CI 0·74-1·07; p=0·25]), OLA-CPAP (111 [47%] of 238, 0·91 [0·76-1·09; p=0·35]), or STD-CPAP groups (118 [48%] of 244, 0·95 [0·80-1·14; p=0·65]) when compared with patients in the STD-O₂ group (125 [51%] of 244). Intraoperatively, PEEP was increased in 69 (14%) of patients in the standard perioperative ventilation groups because of hypoxaemia, and no patients from either of the OLA groups required rescue manoeuvres.

INTERPRETATION

In patients who have major abdominal surgery, the different perioperative open lung approaches tested in this study did not reduce the risk of postoperative complications when compared with standard lung-protective mechanical ventilation.

FUNDING

Instituto de Salud Carlos III of the Spanish Ministry of Economy and Competitiveness, and Grants Programme of the European Society of Anaesthesiology.

Added Value of Bone Subtraction in Dual-energy Digital Radiography in the Detection of Pneumothorax: Impact of Reader Expertise and Medical Specialty

RATIONALE AND OBJECTIVES

This study aimed to determine the value of dual-energy thoracic radiography in the diagnosis of pneumothorax considering the reader's experience.

MATERIALS AND METHODS

Forty patients with a suspected pneumothorax, imaged with dual-energy chest radiographs, were divided into two groups: those with pneumothorax as the final diagnosis (n = 19) and those without (n = 21). The images were analyzed by 36 readers (5 interns, 16 residents, 15 senior physicians) for the presence or absence of pneumothorax during three readout sessions at 2-week intervals: standard images alone (session 1), dual-energy images with bone subtraction alone (session 2), and a combination of the two (session 3).

RESULTS

The number of correct responses increased 13.3% between sessions 1 and 2 (P < .001) and 9.4% between sessions 1 and 3 (P < .001). The mean sensitivity for pneumothorax detection was higher in sessions 2 (82%) and 3 (79%) compared to session 1 (70%). There was no statistically significant difference in specificity between the sessions. The number of correct responses for small volume pneumothoraces was higher in sessions 2 (10.6 ± 1.8) and 3 (10.1 ± 2.0) than in session 1 (8.9 ± 2.3), with a statistically significant difference between sessions 1 and 2 (P = .002) and between sessions 1 and 3 (P = .048).

CONCLUSION

Bone subtracted dual-energy thoracic radiographs improve the detection sensitivity of pneumothorax, including in cases of small pneumothoraces, regardless of the reader's level or expertise.

Final year physiotherapy student's reliability in chest X-ray interpretation

BACKGROUND

Newly qualified physiotherapists may be required to interpret an unreported chest X-ray (CXR) as part of their physiotherapy assessment in "on call" situations. Their interpretation may influence the patient management strategies they adopt. There is no research published which have tested the reliability of final year physiotherapy students in CXR interpretation.

METHODS

Twenty-five final year physiotherapy students were asked to view and interpret the findings of six CXRs, together with a brief vignette, typical of a single commonly encountered diagnosis. Students were also asked if they had received additional CXR training on placement or had a desire to specialize in respiratory care.

RESULTS

The CXR interpretations were scored as incorrect 0, partially correct 1 (abnormality detected but not able to diagnose or missed some detail) and 2 correct. Scores for each of the six CXRs were added to give a total score (out of 12). The median score was 3 out of 12, (range 0-9). Median scores were slightly higher at 4 out of 12 in those students with additional training or a desire to specialize (range 1-7), but this was not statistically significant (p = 0.43).

CONCLUSIONS

Final year physiotherapy students were not able to reliably interpret CXRs. These findings were consistent with previous published research involving medical students. Therefore on graduation before starting "on call" duties it is recommended newly qualified physiotherapists receive additional training in CXR interpretation.

Imaging of the Lungs in Organ Donors and its Clinical Relevance: A Retrospective Analysis

PURPOSE

The aim of the study was to retrospectively evaluate the diagnostic imaging that potential lung donors undergo, the reader variability of image interpretation and its relevance for donation, and the potential information gained from imaging studies not primarily intended for lung evaluation but partially including them.

MATERIALS AND METHODS

Bedside chest radiography and computed tomography (CT), completely or incompletely including the lungs, of 110 brain-dead potential organ donors in a single institution during 2007 to 2014 were reviewed from a donation perspective. Two chest radiologists in consensus analyzed catheters and cardiovascular, parenchymal, and pleural findings. Clinical reports and study review were compared for substantial differences in findings that could have led to a treatment change, triggered additional examinations such as bronchoscopy, or were considered important for donation.

RESULTS

Among 136 bedside chest radiographs, no differences between clinical reports and study reviews were found in 37 (27%), minor differences were found in 28 (21%), and substantial differences were found in 71 (52%) examinations (P<0.0001). In 31 of 42 (74%) complete or incomplete CT examinations, 50 of 74 findings with relevance for lung donation were not primarily reported (P<0.0001).

CONCLUSIONS

The majority of donor patients undergo only chest radiography. A targeted imaging review of abnormalities affecting the decision to use donor lungs may be useful in the preoperative stage. With a targeted list, substantial changes were made from initial clinical interpretations. CT can provide valuable information on donor lung pathology, even if the lungs are only partially imaged.

Comparing posteroanterior with lateral and anteroposterior chest radiography in the initial detection of parapneumonic effusions

BACKGROUND

It is unclear whether anteroposterior (AP) or posteroanterior with lateral (PA/Lat) chest radiographs are superior in the early detection of clinically relevant parapneumonic effusions (CR-PPEs). The objective of this study was to identify which technique is preferred for detection of PPEs using chest computed tomography (CCT) as a reference standard.

METHODS

A secondary analysis of a pneumonia database was conducted to identify patients who received a CCT within 24 hours of presentation and also received AP or PA/Lat chest radiographs within 24 hours of CCT. Sensitivity and specificity were then calculated by comparing the radiographic diagnosis of PPEs of both types of radiographs compared with CCT by using the existing attending radiologist interpretation. Clinical relevance of effusions was determined by CCT effusion measurement of >2.5 cm or presence of loculation.

RESULTS

There was a statistically significant difference between the sensitivity of AP (67.3%) and PA/Lat (83.9%) chest radiography for the initial detection of CR-PPE. Of 16 CR-PPEs initially missed by AP radiography, 7 either required drainage initially or developed empyema within 30 days, whereas no complicated PPE or empyema was found in those missed by PA/Lat radiography.

CONCLUSIONS

PA/Lat chest radiography should be the initial imaging of choice in pneumonia patients for detection of PPEs because it appears to be statistically superior to AP chest radiography.

Clinical communication in diagnostic imaging studies: mixed-method study of pre- and post-implementation of a hospital information system

OBJECTIVE

To examine how and why the quality of clinical communication between radiologists and referring physicians was changed in the inpatient imaging process after implementation of a hospital information system (HIS).

METHODS

A mixed-method study of the chest X-ray (CXR) requests and reports, and their involved processes within a pre- and post-HIS implementation setting.

RESULTS

Documentation of patient age, patient ward, and name and signature of requesting physician decreased significantly in post-HIS CXR requests (P<0.05). However, documentation of requested position and technique increased significantly (P<0.05). In post-HIS CXR reports, documentation of patient age, patient chart number, urgent/normal status of requisition, position and technique of CXR, name of referring physician, and date of request were increased significantly (P<0.05). However, documentation of discussion for important findings was decreased significantly (P<0.05). The mean number of words in the body text of post-HIS reports was increased significantly (18.65 vs. 16.3, P = 0.00).Our qualitative findings highlighted that involving nursing and radiology staff in the communication loop between physicians and radiologists after the implementation resulted in extra steps in the workflow and more workload for them. To cope with the new workload, they adopted different workarounds that could explain the results seen in the quantitative study.

CONCLUSION

The HIS improved communication of administrative and identification information but did not improve communication of clinically relevant information. The reason was traced to the complications that the inappropriate implementation of the system brought to clinical workflow and communication loop.

Assessing the accuracy and certainty in interpreting chest X-rays in the medical division

The chest X-ray (CXR) is an important diagnostic tool in diagnosing and monitoring a spectrum of diseases. Despite our universal reliance on the CXR, our ability to confidently diagnose and accurately document our findings can be unreliable. We sought to assess the diagnostic accuracy and certainty of making a diagnosis based on 10 short clinical histories with one CXR each. We conclude from our study that specialist registrars (StRs) and consultants scored the highest marks with the highest average certainty levels. Junior trainees felt least certain about making their diagnosis and were less likely to be correct. We recommend that StRs and consultants review all the CXRs requested to ensure accuracy of diagnosis. There also needs to be discussion with the Joint Royal Colleges of Physicians Training Board (JRCPTB) about the need of including a separate CXR competency as part of a trainee's generic curriculum on the e-portfolio, something which is currently lacking.

The impact of an early-morning radiologist work shift on the timeliness of communicating urgent imaging findings on portable chest radiography

PURPOSE

The aim of this study was to assess the potential impact of staggered radiologist work shifts on the timeliness of communicating urgent imaging findings that are detected on portable overnight chest radiography of hospitalized patients.

METHODS

The authors conducted a retrospective study that compared the interval between the acquisition and communication of urgent findings on portable overnight critical care chest radiography detected by an early-morning shift for radiologists (3 am to 11 am) with historical experience with a standard daytime shift (8 am to 5 pm) in the detection and communication of urgent findings in a similar patient population a year earlier.

RESULTS

During a 4-month period, 6,448 portable chest radiographic studies were interpreted on the early-morning radiologist shift. Urgent findings requiring immediate communication were detected in 308 (4.8%) studies. The early-morning shift of radiologists, on average, communicated these findings 2 hours earlier compared with the historical control group (P < .001).

CONCLUSION

Staggered radiologist work shifts that include an early-morning shift can improve the timeliness of reporting urgent findings on overnight portable chest radiography of hospitalized patients.