Interpretation Of Chest Radiographs Research Articles

Computer-aided diagnosis in chest radiography: results of large-scale observer tests at the 1996-2001 RSNA scientific assemblies.

Since 1996, computer-aided diagnosis (CAD) schemes have been presented as interactive demonstrations on computer workstations at each scientific assembly of the Radiological Society of North America. The schemes involved (a) detection of pulmonary nodules, (b) temporal subtraction, (c) detection of interstitial lung disease, (d) differential diagnosis of interstitial lung disease, and (e) distinction between benign and malignant pulmonary nodules on chest radiographs. Large-scale observer tests were carried out to examine how radiologists can benefit from CAD systems. Observer performance was evaluated by analysis of receiver operating characteristic (ROC) curves. The statistical significance of the difference between the areas under the ROC curves without and with CAD was analyzed with the Student t test. In all of the tests, the diagnostic accuracy of the radiologists in total improved significantly when CAD was used. This result provides additional evidence that CAD has the potential to improve the performance of radiologists in their decision-making process in interpreting chest radiographs.

A Latent Variable Model for the Analysis of Variability in the Classification of Radiographs of Pneumoconioses

The interpretation of chest radiographs for pneumoconioses is subjective and susceptible to inter- and intra-observer variability. The κ statistic has traditionally been used to assess interobserver agreement. In this paper a different paradigm for the analysis of radiographic readings for pneumoconiosis is introduced. This framework involves log-linear modelling and considers the radiograph and the reader together. The results of the analysis are estimates of the relative ‘abnormal appearance’ of each radiograph and the relative tendency of each reader to read ‘high’ or ‘low’. Log-linear analyses could be useful to ensure quality and to place a reader’s codings in the context of those of his peers.

Time course of perception and decision making during mammographic interpretation.

This article describes the time course of lesion detection on digital mammograms using data about both eye position and decision time to compare performance between experienced mammographers and trainees. Research indicates that a longer decision time works against performance in the interpretation of chest radiographs because the likelihood of error is increased, particularly for trainees. Is this relation between decision time and performance also true for interpreting mammograms? Is there an optimal decision time-performance trade-off for detecting breast lesions? Six radiology trainees (experience, 302-976 cases) and three mammographers (experience, 3000-5000 cases per year) reviewed 40 test cases. Each test case was represented by two mammograms that showed different views of the same breast. Twenty breasts contained suspicious lesions, and 20 were lesion-free. An interactive computer display system with an eye-head tracker measured the timing of decisions, where visual attention was directed, and how much time was spent fixating on a region of interest for each decision. Eye position was monitored during an initial-decision phase, and decision times were measured throughout a final-decision phase during which suspicious lesions recognized initially were interpreted and localized. Performance was analyzed using localization receiver operating characteristic curves. The time course of interpreting mammograms is similar to that for interpreting chest radiographs. Mammographers detected 71% of the true lesions within 25 sec, and trainees detected 46% within 40 sec. Both a fixation dwell time of 1000 msec and a high level of confidence in the decision were associated with the detection of true lesions for the mammographers but not for the trainees. Mammographers detected most breast lesions by global recognition within 25 sec, but trainees took more time. Prolonging one's search beyond the global recognition phase yielded few new lesions and increased the risk of error.

The use of pulse oximetry to exclude pneumonia in children

The objective of this study was to determine whether pulse oximetry alone or in conjunction with the clinical examination is predictive of pneumonia in children who present to the emergency department with respiratory complaints. A retrospective comparison of children with radiographic pneumonia with children with respiratory complaints and negative chest radiography was used. The study took place in an emergency department of a large academic, tertiary care hospital. All children less than 24 months of age who presented with a respiratory complaint and underwent chest radiography during a 1-year period were included. Charts of children with radiographic pneumonia were compared with charts of children without pneumonia, retrospectively. Data abstracted onto data collection forms included: pulse oximetry measurement, vital signs, general appearance, lung examination, and final radiology interpretation of chest radiographs. Pneumonia was defined as a chest radiograph showing any opacity consistent with pneumonia as read by a board-prepared or -certified radiologist. A total of 803 children qualified for the study. Radiograph interpretations were available for 762, and 10.5% were found to have radiographic pneumonia. The median pulse oximetry reading of children with radiographic pneumonia was 97% (interquartile range 95th-98th percentile) compared with 98% (interquartile range 96th-99th percentile) in the control group. Forty-five percent (35 of 78) of the children with radiographic pneumonia showed oxygen saturations of 98% or higher with greater than 10% (8 of 78) displaying oxygen saturations of 100%. By using logistic regression, pulse oximetry was not found to be a statistically significant predictive variable for radiographic pneumonia. Pulse oximetry could not be used to rule out the presence of radiographic pneumonia in children less than 2 years of age who presented with respiratory complaints. (Am J Emerg Med 2002;20:521-523. This is a US government work. There are no restrictions on its use.)

Chest radiograph interpretation skills of anesthesiologists

Objective: To assess the skills of anesthesiologists in the interpretation of chest radiographs. Design: Randomized evaluation conducted among anesthesiologists and radiologists. Setting: Postgraduate Assembly of the New York State Society of Anesthesiologists in 1999, and the Department of Radiology, New York University Medical Center. Participants: A total of 61 anesthesiologists (48 attending physicians; 13 residents); control group of 8 radiology residents (all participants volunteered). Interventions: After completing a demographic survey, participants were asked to review a series of 10 chest radiographs. A brief clinical scenario accompanied each radiograph. No time limit was set for these interpretations. Measurements and Main Results: The demographic characteristics of the anesthesiology participants included university faculty (46%), private group practitioners (41%), independent practitioners (11%), and 1 participant with an unspecified type of practice. Additional training among the participants included internal medicine (31%), surgery (19%), and pediatrics (3%); 34% did not specify any additional training. Of the participants, 92% were involved in cases requiring general anesthesia; 96% managed patients in the recovery room; and 34% managed patients in the intensive care unit. Of participants, 80% usually order chest radiographs, but only 42% interpret the films themselves. Misdiagnosed radiographs included pneumothorax by 11% of participants, free air under the diaphragm by 41%, bronchial perforation from a nasogastric tube by 28%, right mainstem intubation by 20%, superior vena cava perforation from a central venous catheter by 31%, normal film by 75%, negative pressure pulmonary edema by 16%, left lower lobe collapse by 80%, pulmonary infarction from a pulmonary artery catheter by 29%, and tension pneumothorax by 41%. Overall scores of the attending physicians were not significantly different from that of residents (p > 0.05). The control group of radiology residents scored significantly better (mean, 83.7; p = 0.009) than the anesthesia residents (mean, 62.8) and anesthesia attending physicians (mean, 62.5). Conclusion: Anesthesiologists are deficient in skills for the interpretation of chest radiographs. The skill level of university-based physicians is not greater than physicians in private practice, and skill level does not improve with level of training or experience. Most anesthesiologists rely on radiologists for interpretative results. Further training during the residency years may help improve diagnostic skills. Copyright © 2001 by W.B. Saunders Company

The development of ventilator-associated pneumonia does not change aspects of mechanical ventilation.

To evaluate whether the development of ventilator-associated pneumonia (VAP) is associated with changes in ventilation parameters. Matched case-control study. Mixed intensive care unit of a university hospital. From a large database we selected 33 patients with VAP, diagnosed with quantitative cultures of bronchoscopically obtained specimens. In addition, 33 other mechanically ventilated patients who did not develop VAP were selected (controls). Patients with VAP and controls were matched on seven variables representing severity of illness: duration of ventilation until matching, diagnosis on admission, renal function, liver function, preceding infection, preceding surgery and immunosuppressive therapy. Each patient with VAP was matched to a single control. Variables regarding type and mode of ventilation and interpretation of chest radiographs were not included in the matching procedure. Characteristics of mechanical ventilation (mode of ventilation, tidal volume, expired minute ventilation, peak airway pressures, mean airway pressures, level of positive end-expiratory pressure, arterial oxygen tension(PaO2)/fractional inspired oxygen (FIO2) ratio), were compared on the day of diagnosis of VAP (or matching for controls) and 2 and 4 days before. Although there was a significant difference in PaO2/FIO2 ratios between cases and controls on the day of diagnosis of VAP, the change in PaO2/FIO2 ratios during the days of study were not statistically different between patients developing VAP and controls. No significant differences were found for any of the other variables of ventilation at any of the three time points studied, nor were there significant differences in changes of these parameters within individual patients. Characteristics and parameters of mechanical ventilation are not influenced by the development of VAP. It is, therefore, unlikely that these variables are useful in the diagnostic work-up of VAP.

Radiographic and CT Appearances of the Major Fissures

The major fissure is an important anatomic landmark in the interpretation of chest radiographs and computed tomographic (CT) scans. At radiography, the major fissures normally appear as hairlines of soft-tissue density; at conventional CT, they typically appear as lucent, hypovascular bands; and at high-resolution CT, they most often appear as sharp lines. The superolateral major fissure usually manifests as a curving edge at the upper lateral lung field with lateral opacity and medial lucency. The vertical fissure line appears as a fine, linear shadow, commencing in or near the costophrenic angle and coursing upward. The superomedial major fissure manifests as a short, obliquely oriented straight line. Progressive widening of the major fissure inferiorly manifests as a triangular area of increased opacity and represents intrafissural fat. Various inflammatory, granulomatous, neoplastic, and abnormal hemodynamic conditions involving the major fissure can affect its imaging appearance. Oblique orientation of the major fissure may complicate radiographic interpretation. The fissure may be incomplete or absent, complicating identification of various diseases. An incomplete major fissure may lead to disease spread, collateral air drift, or the "incomplete fissure sign," a sign that may, however, also be present in cases of complete fissure. Knowledge of the anatomy and normal variants of the major fissures is essential for recognizing their variable imaging appearances as well as related abnormalities.

Mediastinal interfaces and lines in children: radiographic-CT correlation.

Various mediastinal interfaces and lines are well known in adults, but not fully understood in children. To review the mediastinal interfaces and lines on plain radiographs in the paediatric age group with CT correlation. Soft copies of 195 sets of CT examinations and concurrent chest radiographs in 180 paediatric patients (age 1 month to 15 years) were retrospectively reviewed. The frequency of visualisation and the anatomical basis of the mediastinal interfaces and lines were assessed. The evaluated mediastinal interfaces and lines were superior vena cava interface, descending aorta interface, left paraspinal interface, azygoesophageal recess interface, right paratracheal stripe, anterior junction line and posterior junction line. Chest radiographs showed the SVC interface in 161, the descending aorta interface in 155, the left paraspinal interface in 98, the azygoesophageal recess in 94, the right paratracheal stripe in 53, the anterior junction line in 15 and the posterior junction line in 10. Non-visualisation of mediastinal interfaces and lines on plain radiographs was explained by normal anatomical difference compared with the adult and underlying intrathoracic abnormalities when they were correlated with CT. The frequency of visualisation of the SVC interface, descending aortic interface, left paraspinal interface and azygoesophageal recess fluctuated with age, while the frequency of the right paratracheal stripe, anterior junction line and posterior junction line increased with age. Awareness of the frequency of visualisation and the anatomical basis of the mediastinal interfaces and lines in paediatric patients may be helpful for interpretation of chest radiographs.

Black, white, and shades of gray: common abnormalities in chest radiographs.

Chest radiographs remain one of the oldest, noninvasive methods for identifying chest diseases. This tool is relatively inexpensive, low risk, and important for the advanced practice nurse to use in making an initial differential diagnosis. Timely intervention depends on the advanced practice nurse's ability to recognize the radiographic abnormalities commonly associated with acute pulmonary disease. When linked with the history, presenting signs and symptoms, and a thorough physical assessment, the chest radiograph is key to appropriate diagnosis and management. This article reviews the basic radiographic lung densities, describes a systematic approach to ensure thorough chest radiographic interpretation, and presents characteristic chest radiographs found in the acute care setting. Selected examples of diffuse lung diseases, localized alveolar opacities, pleural effusions, and hyperlucency of the lungs will also be presented.

Evaluation of Competence in the Interpretation of Chest Radiographs

Rationale and Objectives The purpose of this study was to determine relative rates of missed diagnoses for radiologists as a measure of competence in interpreting chest radiographs. Materials and Methods Cases involving differing interpretations of chest radiographs were collected from January 1994 through December 1999 by faculty (chest and nonchest radiology specialists) in an academic radiology department. A quarterly peer-review process designated cases months after the fact, and anonymously, as no miss or as class I (nondiagnosable), class II (very difficult diagnosis), class III (should be diagnosed most of time), or class IV (should almost always be diagnosed) missed diagnoses. The rates and classes of missed diagnoses were compared among chest faculty and for the nonchest radiology specialists as a group. Results Chest radiologists read 184,977 studies, and nonchest radiologists read 300,684 studies. Of these, 243 missed diagnoses were classified (classes I and II, 184 cases; class III, 50; and class IV, nine). No difference was detected in the rate of class III and IV misses among chest faculty, but nonchest faculty had significantly more class III ( P = .022) and class IV misses ( P = .016). Conclusion Random sampling of differing interpretations can yield a relative rate of missed diagnoses for radiologists. No difference was detected in clinically important misses (ie, classes III and IV) among chest radiologists, but a statistically significantly higher rate of seemingly obvious misdiagnoses was found for nonchest specialty radiologists. Potential biases may have influenced this analysis, including disease prevalence, sampling, clinical factors, observer variability, and truth-in-diagnosis.

The role of preliminary interpretation of chest radiographs by radiographers in the management of acute medical problems within a cardiothoracic centre

Radiographic staff in a regional cardiothoracic centre were asked to assess all pre- and post-operative chest radiographs over a 6-month period. Radiographs showing new, acute changes were noted and a red dot placed on the film. Medical staff were notified of radiographs with red dots, and these were subsequently reported by radiologists. Using reports by radiologists as the gold standard, an audit was performed of the radiographers' accuracy in identifying new abnormalities. The absence of a necessary red dot as well as inappropriate use were noted. To enhance the accuracy of radiograph interpretation, a series of lectures on the chest radiograph and a protocol for red dot use were developed by senior radiologists. During this 6-month period 8614 chest radiographs were taken; red dots were applied to 464 (5%). These red dots were considered incorrect in 100 radiographs. Radiographers misinterpreted or missed potentially important changes in 38 of the remaining 8150 radiographs without red dots (sensitivity and specificity of 90% and 99%, respectively). Radiographers appeared to err on the side of caution when confronted with an abnormal chest radiograph, especially when previous radiographs and reports were unavailable. This resulted is a relatively high false positive rate. Future audits will show whether this rate can be reduced by continued training. Subtle interpretation is crucial to distinguish between an abnormal chest radiograph needing urgent medical attention and an abnormal chest radiograph with normal post-operative changes. The opinion of the experienced and trained radiographer is immediate and may be invaluable to the diagnostic care of the patient.

Image compression and chest radiograph interpretation: image perception comparison between uncompressed chest radiographs and chest radiographs stored using 10:1 JPEG compression.

We have assessed the effect of 10:1 lossy (JPEG) compression on six board-certified radiologists' ability to detect three commonly seen abnormalities on chest radiographs. The study radiographs included 150 chest radiographs with one of four diagnoses: normal (n = 101), pulmonary nodule (n = 19), interstitial lung disease (n = 19), and pneumothorax (n = 11). Before compression, these images were printed on laser film and interpreted in a blinded fashion by six radiologists. Following an 8-week interval, the images were reinterpreted on an image display workstation after undergoing 10:1 lossy compression. The results for the compressed images were compared with those of the uncompressed images using receiver operating characteristic (ROC) analyses. For five of six readers, the diagnostic accuracy was higher for the uncompressed images than for the compressed images, but the difference was not significant (P > .1111). Combined readings for the uncompressed images were also more accurate when compared with the compressed images, but this difference was also not significant (P = .1430). The sensitivity, specificity, and accuracy values were 81.5%, 89.2%, and 86.7% for the compressed images, respectively, as compared with 78.9%, 94.5%, and 89.3% for the uncompressed images. There was no correlation between the readers' accuracy and their experience with soft-copy interpretation; the extent of radiographic interpretation experience had no correlation with overall interpretation accuracy. In conclusion, five of six radiologists had a higher diagnostic accuracy when interpreting uncompressed chest radiographs versus the same images modified by 10:1 lossy compression, but this difference was not statistically significant.

Interobserver Variation in Interpreting Chest Radiographs for the Diagnosis of Acute Respiratory Distress Syndrome

To measure the reliability of chest radiographic diagnosis of acute respiratory distress syndrome (ARDS) we conducted an observer agreement study in which two of eight intensivists and a radiologist, blinded to one another's interpretation, reviewed 778 radiographs from 99 critically ill patients. One intensivist and a radiologist participated in pilot training. Raters made a global rating of the presence of ARDS on the basis of diffuse bilateral infiltrates. We assessed interobserver agreement in a pairwise fashion. For rater pairings in which one rater had not participated in the consensus process we found moderate levels of raw (0.68 to 0.80), chance-corrected (kappa 0.38 to 0.55), and chance-independent (Phi 0. 53 to 0.75) agreement. The pair of raters who participated in consensus training achieved excellent to almost perfect raw (0.88 to 0.94), chance-corrected (kappa 0.72 to 0.88), and chance-independent (Phi 0.74 to 0.89) agreement. We conclude that intensivists without formal consensus training can achieve moderate levels of agreement. Consensus training is necessary to achieve the substantial or almost perfect levels of agreement optimal for the conduct of clinical trials.

Pediatric diffuse lung disease: diagnosis and classification using high-resolution CT.

Our purpose was to categorize high-resolution CT findings in children with diffuse lung disease and to evaluate the accuracy of diagnoses made using CT. The chest radiographs and high-resolution CT scans of 20 children (1-16 years old; median, 9 years old) with biopsy-proven chronic diffuse lung diseases were reviewed separately by two independent chest radiologists. Thirteen types of diffuse lung disease were included in the study. Radiographic and CT features were noted, and three choices of diagnosis were recorded, with the confidence level. Diagnoses were made with a high degree of confidence (definite or probable) in 25 of 40 interpretations of CT scans, compared with only five of 40 interpretations of chest radiographs (p < .001). Fourteen (56%) of the 25 confident first-choice diagnoses on CT scans were correct, compared with two (40%) of the five interpretations on chest radiographs. Diseases were classified as belonging to one of five distinct groups on the basis of dominant CT features. Airway disease (n = 5) (bronchiolitis obliterans or bronchocentric granulomatosis) showed geographic hyperlucency on CT. Septal disease (n = 4) (lymphangiomatosis, hemangiomatosis, or microlithiasis) showed septal thickening. Infiltrative lung disease (n = 7) (desquamative interstitial pneumonitis, hypersensitivity pneumonitis, or lymphoid interstitial pneumonitis) showed ground-glass opacity. Air-space disease (n = 3) (aspiration, vasculitis, or bronchiolitis obliterans organizing pneumonia) showed lung consolidation. Langerhans' histiocytosis (n = 1) showed cysts and nodules. Surprisingly little overlap was seen among these groups. CT increases the level of diagnostic confidence for pediatric infiltrative lung disease, improves diagnostic accuracy, and provides a useful classification system.

Computer-aided diagnosis for detection of interstitial opacities on chest radiographs.

Our objective was to evaluate the impact of a computer-aided diagnostic scheme on radiologists' interpretations of chest radiographs with interstitial opacities by performing an observer test using receiver operating characteristic (ROC) analysis. Twenty chest radiographs with normal findings and 20 chest radiographs with abnormal findings were used. Each radiograph was divided into four quadrants. One hundred twenty-nine quadrants (80 normal and 49 abnormal quadrants) were used for testing because we excluded 31 equivocal quadrants. Sixteen independent observers (10 residents and six attending radiologists) participated in this study. The radiologists' performance without and with computer assistance, which indicated cases with normal and abnormal findings by various markers, was evaluated by ROC analysis. The diagnostic accuracy of the observers improved by a statistically significant magnitude when computer-aided diagnosis was used. Thus, the values for the area under the ROC curve obtained with and without the computer-aided diagnostic output were .970 and .948 (p = .0002), respectively, for all observers; .969 and .943 (p = .0006), respectively, for the residents' subgroup; and .972 and .960 (p = .162), respectively, for the attending radiologists' subgroup. The value for the area under the ROC curve for the computerized scheme by itself was .943. Our computer-aided diagnostic scheme can assist radiologists in the diagnosis or exclusion of interstitial disease on chest radiographs.

Distinguishing hantavirus pulmonary syndrome from acute respiratory distress syndrome by chest radiography: are there different radiographic manifestations of increased alveolar permeability?

Hantavirus infection may cause diffuse air space disease, termed hantavirus pulmonary syndrome (HPS). The authors sought to determine if chest radiographs could differentiate HPS from typical acute respiratory distress syndrome (ARDS). The authors identified patients with either HPS (n = 11) or acute ARDS (n = 32) and selected the earliest chest radiograph showing diffuse airspace disease, and a chest radiograph taken 24 to 48 hours previously. Thoracic and general radiologists first viewed the chest radiograph showing diffuse air space disease, and ranked the likelihood that each case represented HPS versus ARDS. Afterward, readers viewed earlier chest radiographs and rescored each case. Receiver operating characteristic (ROC) curves from both scoring sessions were generated. The mean areas under the ROC curves for the entire group was 0.83 +/- 0.12 initially, and improved to 0.87 +/- 0.09 (p < 0.05) after viewing prior chest radiographs. Receiver operating characteristic curves of thoracic radiologists described greater areas than those of general radiologists both before and after viewing prior chest radiographs; 0.95 +/- 0.01 versus 0.78 +/- 0.08 (p < 0.05) and 96 +/- 0.02 versus 0.80 +/- 0.05 (p < 0.05). The mean sensitivity and specificity of chest radiograph interpretation for HPS was 86 +/- 13% and 74 +/- 11%, respectively. Chest radiographs can differentiate HPS from ARDS. Accuracy is improved by the use of serial radiographs and more highly trained readers. The chest radiograph findings may represent differences in the extent of alveolar epithelial damage seen in HPS and ARDS.

Pitfalls in the interpretation of pediatric chest and airways radiographs

Pitfalls in the interpretation of pediatric chest and airways radiographs

Can Quantitative Capnometry Differentiate Between Cardiac and Obstructive Causes of Respiratory Distress?

To determine whether quantitative measurement of end-tidal carbon dioxide (ETCO2) can differentiate between cardiac and obstructive causes of respiratory distress. Prospective observational study. Emergency department (ED) of a tertiary care hospital. Adult patients who presented to the ED with moderate-to-severe dyspnea. Patients were excluded if they were unable to cooperate with the performance of peak expiratory flow rate (PEFR) or ETCO2 tests, were younger than 18 years of age, or had received prehospital intervention for their respiratory distress. Physicians obtained an ETCO2 level and PEFR prior to ED pharmacologic intervention. A hand-held capnometer with digital read-out was used to obtain the ETCO2 level. The patient's age, sex, initial vital signs, breath sounds and medication history, the presence or absence of diaphoresis and/or orthopnea, the duration of symptoms, the chest radiograph interpretation, and final diagnosis were also recorded. Forty-two patients were eligible for inclusion in the analysis. The mean ETCO2 level was 31.1+/-9.4 mm Hg; the mean PEFR was 161.3+/-53.1 L/min. The ETCO2 levels for pulmonary edema/congestive heart failure (CHF) patients differed significantly from those of asthma/COPD patients (27.1+/-7.8 mm Hg vs 33.4+/-9.6 mm Hg; p=0.0375). However, no single ETCO2 level was found to be a reliable predictor of diagnosis. ETCO2 levels for pulmonary edema/CHF patients differ significantly from those of asthma/COPD patients. However, no single ETCO2 level reliably differentiates between the two disease processes.

Influence of image processing on chest radiograph interpretation and decision changes.

The authors investigate the effect of image processing on diagnostic performance in the reading of computed chest radiographs. Six radiologists read 168 chest images with and without the use of image processing. Diagnostic performance was measured by means of receiver operating characteristic analysis, and changes made in diagnostic decisions with the use of image processing were evaluated. There were no statistically significant differences between reader performance with image processing and reader performance without image processing (P < .05). Readers' decisions were just as likely to change from false-negative to true-positive as from true-positive to false-negative with the use of image processing. More decisions changed from true-negative to false-positive than from false-positive to true-negative with processing, The effect of image processing does not greatly influence diagnostic performance in chest radiography.

Do pediatric intensivists and radiologists concur on the interpretation of chest radiographs?

BACKGROUND: Therapeutic decisions in the pediatric intensive care unit are made by pediatric intensivists (PI) based on their interpretation of chest radiographs before the formal interpretation by a pediatric radiologist (PR). This study was designed to determine the adequacy of chest radiograph interpretations by pediatric intensivists and the effects on patient care. The PI recorded their chest radiograph interpretations, documenting support devices and thoracic abnormalities. Concordance and discordance were determined by the pediatric pulmonologist who was not involved in the care of the patient by comparing the interpretations of the PI and PR. Clinically significant discordance was defined as interpretations by the radiologist that differed to those from the PI that may have required therapeutic intervention. RESULTS: The evaluation of 291 chest radiographs demonstrated an overall concordance rate of 82.5% (240 out of 291; P < 0.05). There was no significant difference in the ability of critical care medicine physicians to identify atelectasis, infiltrates, pleural effusions, or airleaks (P > 0.05). Support devices were correctly identified in 100% of the cases. Discordant interpretations included 20 that were clinically significant, 17 insignificant findings and 14 films over-interpreted by the PI. A chart review of the patients with discordant findings revealed only one finding that required an alteration in therapy. CONCLUSIONS: These findings demonstrate significant agreement between the interpretation of chest radiographs by PI and PR in selected clinical situations. These data support the current practice of the PI making therapeutic decisions based on their interpretations of chest radiographs.