Radiology residents' on-call interpretation of chest radiographs for pneumonia

RATIONALE AND OBJECTIVES: This study is designed to assess the performance of radiology residents in interpreting emergency department chest radiographs for pneumonia and to characterize chest radiographic findings in patients for which interpretation was amended by an attending radiologist. MATERIALS AND METHODS: We retrospectively reviewed all amended reports for chest radiographs performed on emergency department patients July 2002-June 2003. Reports preliminarily interpreted by residents and amended by a board-certified staff radiologist for the presence or absence of pneumonia were identified. A panel of three experienced radiologists, blinded to reports, jointly reviewed each chest radiograph. If the panel diagnosed pneumonia, the chest radiograph was evaluated for the projection that best showed the pneumonia, its size and location, and the presence or absence of the following features: increased opacity, air bronchograms, loss of vascular markings, silhouette sign, and linear opacities. The resident's post-graduate year (PGY) training level was noted. RESULTS: One percent (134/12,600 reports) of chest radiographic reports were amended for the presence or absence of pneumonia. One hundred chest radiographs were available and comprised the series. There were 56 females and 44 males with a mean age of 45 years (range, 1-99 years). The staff radiologist diagnosed pneumonia in 79% (79/100 radiographs). The panel agreed with the staff in 77% (kappa = 0.76) and the resident in 23% (kappa = 0.43). The panel diagnosed pneumonia in 60% (60/100 radiographs) with the following chest radiographic findings: 100% (60/60), increased opacity; 37% (22/60), air bronchograms; 72% (43/60), loss of vascular markings; 40% (24/60), silhouette sign; and 20% (12/60), linear opacities. The pneumonia was right sided in 52% (31/60), left sided in 37% (22/60), and bilateral in 11% (7/60). Right-sided pneumonias were equally distributed among the three lobes, and left-sided pneumonias had a lower-lobe predominance of 77% (17/22). Seventy-five percent (45/60) of pneumonias were segmental or smaller, and 82% (49/60) of chest radiographs showing pneumonia had both posteroanterior and lateral projections. The pneumonia was conspicuous on only one projection in 43% (21/49); the posteroanterior view in 22% (11/49), and the lateral view in 20% (10/49). Eighty-one percent (81/100) of interpreting residents were PGY-3. CONCLUSION: Interpretation of chest radiographs for pneumonia by PGY-3 residents has a low error rate. Missed pneumonias often were segmental or smaller and conspicuous on only one projection.     

Inter-observer variation in the interpretation of chest radiographs for pneumonia in community-acquired lower respiratory tract infections

AIM: To assess inter-observer variation in the interpretation of chest radiographs of individuals with pneumonia versus those without pneumonia. MATERIALS AND METHODS: Chest radiographs of out-patients with a lower respiratory tract infection (LRTI) were assessed for the presence of infiltrates by radiologists from three local hospitals and were reassessed by one university hospital radiologist. Various measures of inter-observer agreement were calculated. RESULTS: The observed proportional agreement was 218 in 243 patients (89.7%). Kappa was 0.53 (moderate agreement) with a 95% confidence interval of 0.37 to 0.69. The observed positive agreement (59%) was much lower than for negative agreement (94%). Kappa was considerably lower, if chronic obstructive pulmonary disease was present (kappa = 0.20) or Streptococcus pneumoniae (kappa = -0.29) was the infective agent. CONCLUSION: The overall inter-observer agreement adjusted for chance was moderate. Inter-observer agreement in cases with pneumonia was much worse than the agreement in negative (i.e. non-pneumonia) cases. A general practitioner's selection of patients with a higher chance of having pneumonia for chest radiography would thus not improve the observer agreement.     

High-resolution CT diagnosis of emphysema in symptomatic patients with normal chest radiographs and isolated low diffusing capacity.

To determine the prevalence of "nonobstructive" (impairment of gas transfer) emphysema in a select population of smokers with dyspnea, a retrospective study of patients with emphysema evident at high-resolution computed tomography (HRCT) was undertaken. Four hundred seventy HRCT studies were reviewed. In 47 cases, centrilobular emphysema was the dominant or sole parenchymal abnormality. Concomitant chest radiographs were available in 41 of these cases; 16 of the 41 lacked radiographic findings of emphysema. Among these 16 patients, pulmonary function testing revealed 10 to have normal flow rates (ratio of forced expiratory volume in 1 second to forced vital capacity and forced expiratory volume in 1 second greater than 80% predicted) and impaired gas transfer (single-breath carbon monoxide diffusing capacity [DLCOSB] less than 80% predicted). With the exclusion of one patient with congestive heart failure from the group of 10, the severity of emphysema at HRCT correlated inversely with DLCOSB (r = -.643). These results indicate that HRCT allows detection of emphysema in symptomatic patients when chest radiographs and pulmonary function tests are nondiagnostic.     

Chest radiographs in congestive heart failure: response to therapy in acute and chronic heart disease

Thirty-four men with left ventricular mechanical dysfunction were admitted to an intensive care unit with either an acute myocardial infarction (Group 1, n = 18) or worsening of clinical respiratory signs and symptoms in the setting of a chronic congestive cardiomyopathy (Group 2, n = 16). On admission, all individuals had pulmonary venous hypertension classified as at least Grade 3 by standard radiographic criteria. In each subject, mean pulmonary capillary wedge pressure (mm Hg), extravascular lung water (EVLW) (ml/kg), and chest radiographs were serially evaluated. In the patients in whom pharmacologic therapy successfully returned left ventricular filling pressures to near normal levels (less than or equal to 15 mm Hg), the chest radiograph returned to its baseline level (defined by the discharge radiograph) later in the patients with chronic heart failure (5.1 +/- 1.0 days) than in the patients with acute myocardial infarctions (2.1 +/- 1.2 days, p less than 0.01). Radiographic changes in extravascular water (interstitial and alveolar edema) mirrored changes in EVLW, although EVLW was initially greater in Group 2 (16.3 +/- 1.8 ml/kg) than in Group 1 (10.7 +/- 1.3 ml/kg, p less than 0.01). In the patients in whom filling pressures either worsened or changed less than 3 mm Hg, EVLW and chest radiographs did not markedly change. It is concluded that changes in radiographic pulmonary edema mirror changes in indicator-dilution measurements of EVLW. Radiographic phase lag represents a slow decline in EVLW after therapy for heart failure, which is prolonged in patients with chronic failure and greater EVLW.     

Underreporting of vertebral fractures on routine chest radiography

OBJECTIVE: Osteoporosis is underdiagnosed and therefore undertreated. We determined the potential usefulness of chest radiography for detecting clinically important vertebral fractures by performing semiquantitative reviews and quantitative digital morphometry on 100 routine chest radiographs taken in the emergency department and comparing the yield of these independent reviews with official radiology reports. MATERIALS AND METHODS: One hundred randomly selected chest radiographs of patients 60 years or older who presented to the emergency department of a tertiary care hospital were evaluated. Radiographs were selected without knowledge of the presenting complaint and were independently reviewed by two board-certified radiologists and a radiology resident. A validated semiquantitative method was used to assess lateral chest radiographs for vertebral fracture. In addition, quantitative digital morphometry was undertaken. A clinically important vertebral fracture was defined as one that was at least moderate to severe (loss of height >or=> 25%). RESULTS: Mean age of the population was 75 years, 47% were women, and 46% were admitted to the hospital. According to the reference radiologist, prevalence of moderate to severe vertebral fractures was 22%. Simple agreement was 87-88% among reviewers; kappa values were moderate (0.56-0.58). The greatest agreement was between the reference standard radiologist and quantitative digital morphometry (89% agreement; kappa = 0.67). Only 55% (12/22) of vertebral fractures we identified were mentioned in the official radiology reports. CONCLUSION: Chest radiography has potential as a screening tool for revealing previously undiagnosed vertebral fractures, although in this study only half of moderate to severe fractures that we identified were mentioned in official reports.     

The chest radiograph in legionnaires'' disease

The chest radiographs of 24 patients with documented Legionnaires' disease were evaluated. Twenty-two of the 24 patients had positive findings for the disease initially. There was unilateral involvement in 68% and the most common lung shadows were poorly marginated round opacities (46%), diffuse patchy (25%) and peripheral opacities (21%). At peak, 70% of patients had a lobar shadow. Pleural effusions were present in 39% of cases but could be explained by underlying congestive heart failure or renal failure in 7 of the 9. Although the findings are not specific, the radiologist should consider this diagnosis in a patient with compatible clinical history, a pneumonia of obscure etiology, and these radiographic manifestations.     

Nonenhanced limited CT in children suspected of having appendicitis: prospective comparison of attending and resident interpretations.

PURPOSE: To prospectively compare resident and attending radiologic interpretations of nonenhanced limited computed tomographic (CT) scans obtained in children suspected of having appendicitis. MATERIALS AND METHODS: Seventy-five consecutive children underwent nonenhanced limited CT for suspected appendicitis. The scans were prospectively interpreted by a resident and an attending radiologist, each unaware of the other's interpretation. The probability that the findings indicated a diagnosis of appendicitis, level of certainty in the interpretation, and presence of an alternate diagnosis were statistically analyzed. RESULTS: Nineteen children (25%) had appendicitis. The area under the receiver operating characteristic curve was not significantly different between residents (0.97 +/- 0.02) and attendings (0.95 +/- 0.04). The percentage agreement between residents and attendings was 91% (kappa = 0.73 +/- 0.095). The average level of certainty tended to be higher for attendings (93% +/- 15) than residents (89% +/- 12). The sensitivity, specificity, and accuracy of resident interpretations were 63%, 96%, and 88%, respectively, compared with those of attending interpretations--95%, 98%, and 97%, respectively. Residents and attendings noted alternate diagnoses in 30% of children without appendicitis. CONCLUSION: A high level of agreement exists between resident and attending radiologists in the interpretation of nonenhanced limited CT scans in children suspected of having appendicitis. Residents, however, tend to be less confident in their interpretations.     

Interpretation of chest radiographs in the intensive care unit: a hard- versus soft-copy comparative study

PURPOSE: To compare the reliability of hard- versus soft-copy interpretation of intensive care unit chest radiographs, using a non-specialized standard resolution computer screen. MATERIALS AND METHODS: 104 chest radiographs were included in this study. Three physicians (one radiology resident, one intensive care unit resident and one experienced intensive care unit physician) gave their interpretations on computerized grids. Results were analyzed statistically using ROC curves and Kappa (kappa) index of concordance with experts. RESULTS: Results for reanimation equipment detection are almost independent from the modality (kappa(soft-copy)=0.891+/-0.037, kappa(hard-copy)=0.899+/-0.037). Regarding pathology detection, a global analysis only shows a difference at the limits of significance to the advantage of hard-copy films (kappa(soft)=0.514 +/-0.028, kappa(hard)=0.572+/-0.028). Overall results were significantly better for the radiologist compared to the intensive care unit physicians (kappa(radiologist)=0.751+/-0.048, kappa(intensive-care)=0.405+/-0.048). CONCLUSION: Concerning the task that is studied here, which requires only routine computer equipment, our results suggest that human factors can be more important than material factors.     

From Being a Radiologist to Watching a Radiologist: Impact of Filmless Operation on the Training of Radiology Residents

RATIONALE AND OBJECTIVES: The authors performed this study to investigate the impact of changing from a film-based image interpretation system to one using digital image workstations on the training of radiology residents in the interpretation of radiographs. MATERIALS AND METHODS: Data were collected during a period when a conventional system of image interpretation with hard-copy images and multiviewers was used and during a period when digital image workstations were used. During each period, it was noted whether the first interpretation of the radiographs was performed by a radiology resident, by an attending radiologist, or as a group effort including both an attending radiologist and a radiology resident(s). In addition, it was noted whether a radiology resident or an attending radiologist dictated the report. RESULTS: The proportion of images first interpreted by the radiology resident alone decreased from 38% (53 of 139) when using the conventional system to 17% (34 of 199) after the switch to interpreting images on the workstations (P = .001). During the film-based period, radiology residents dictated 45% of reports (141 of 312), but during the workstation period, radiology residents dictated only 4% of reports (24 of 667; P = .001). CONCLUSION: The authors observed a decrease in autonomous participation by radiology residents in image interpretation and dictation of reports and an increase in "group reading" after the switch from a film-based system to a workstation system.     

Intraobserver and interobserver agreement of the interpretation of pediatric chest radiographs

The objective of this study is to quantify the magnitude of intraobserver and interobserver agreement among physicians for the interpretation of pneumonia on pediatric chest radiographs. Chest radiographs that produced discordant interpretations between the emergency physician and the radiologist's final interpretation were identified for patients aged 1–4 years. From 24 radiographs, eight were randomly selected as study radiographs, and 16 were diversion films. Study participants included two pediatric radiologists, two senior emergency medicine physicians, and two junior fellowship-trained pediatric emergency medicine physicians. Each test included 12 radiographs: the eight study radiographs and four randomly interspersed diversion radiographs, and each radiograph was paired with a written clinical vignette. Testing was repeated on four occasions, separated by ≥2 weeks. The dependent variable was the interpretation of presence or absence of pneumonia; primary analysis done with Cohen's kappa (95% confidence intervals). Intraobserver agreement was good for pediatric radiologists (kappa = 0.87; 95% CI 0.60–0.99) for both but was lower for senior emergency physicians (mean kappa = 0.68; 95% CI 0.40–0.95) and junior pediatric emergency physicians (mean kappa = 0.62; 95% CI 0.35–0.98). Interobserver agreement was fair to moderate overall; between pediatric radiologists, kappa = 0.51 (0.39–0.64); between senior emergency physicians, kappa = 0.55 (0.41–69), and between junior pediatric emergency medicine physicians, kappa = 0.37 (0.25–0.51). Practicing emergency clinicians demonstrate considerable intraobserver and interobserver variability in the interpretation of pneumonia on pediatric chest radiographs.     

MRI of fistula in ano: inter- and intraobserver agreement and effects of directed education

OBJECTIVE: Preoperative MRI of fistula in ano is becoming more common. This prospective study aimed to determine if a significant difference occurred in interpretation between one expert and one novice observer and to assess inter- and intraobserver agreement after both observers underwent a period of directed education. SUBJECTS AND METHODS. An outcome-derived reference standard was defined in 100 patients with suspected fistula in ano via a combination of preoperative MRI, surgical findings, and clinical outcome. The performances of a single expert and a single novice interpreter were compared with this reference standard both before and after a period of directed education, and inter- and intraobserver agreement was determined. RESULTS: Initially the expert correctly classified significantly more fistulas than the novice (85% vs 63%, p = 0.024), but after directed education there was no significant difference, with good agreement for both the classification of the primary track (kappa = 0.71) and identification of extensions (k = 0.61). Intraobserver agreement was very good for the expert (kappa = 0.92) and novice (kappa = 0.88) for classification of the primary track and good (kappa = 0.64 and 0.74, respectively) for identification of extensions. CONCLUSION: The diagnostic accuracy for fistula in ano classification using MRI was significantly higher for one expert than for one novice, though this was rectified by a short period of directed education.     

The utility of routine emergency department chest radiography in a population at high risk for tuberculosis

The purpose of this study was to determine the utility of performing routine screening chest radiography on all asymptomatic admissions from the emergency department in an urban population at high risk for contracting tuberculosis (TB). Chest radiographs were obtained on all asymptomatic patients admitted through the emergency department for drug detoxification or psychiatric illness at two urban hospitals in an area endemic for multidrug-resistant TB. The chest radiographs were interpreted prospectively by one of two radiologists, and the results were correlated with age, sex, and clinical outcome. A total of 481 chest radiographs were evaluated (407 men, 74 women; average age, 38 years). Of these, 436 (91%) were negative. Of the 45 with abnormalities, 35 (7%) were chronic, requiring no further work-up, whereas 10 (2%) had changes considered of immediate consequence. Of those with acute abnormalities, five patients presented with nodular densities, ranging from 5 mm to 3 cm in diameter; two patients had lobar infiltrates; and subsegmental atelectasis, congestive changes and an abnormal aortic contour were noted on one examination each. Of the six patients not lost to follow-up, five improved with medical therapy, and one was scheduled for surgical excision. The individuals with radiographic findings were significantly older than those with negative chest radiographs (47 years vs. 37 years). Only one patient had active TB; three others had chronic calcified granulomas. We conclude that routine chest radiography in young, asymptomatic individuals considered at high risk for contracting TB rarely detects significant pulmonary abnormalities or evidence of active TB.     

Radiology resident interpretation of on-call CT pulmonary angiograms

RATIONALE AND OBJECTIVES: To evaluate the interpretation of computed tomographic pulmonary angiograms performed outside of regular reporting hours, comparing the initial interpretation by the radiology resident to the attending radiologist. MATERIALS AND METHODS: Records for 840 consecutive computed tomographic pulmonary angiograms (CTPA) performed outside of regular reporting hours at two tertiary referral centers from January 1, 2004-December 31, 2005 were reviewed. The preliminary interpretation by the on-call radiology resident was compared to the subsequent final report issued by a subspecialty trained chest radiologist. Studies were stratified as positive, negative, or equivocal for pulmonary embolus. Cases with discordant interpretations or negative CTPA were reviewed to determine impact on clinical outcome. Patients were followed up to 12 months after CTPA to document any subsequent thromboembolic event. RESULTS: Sixteen percent (131/840) of CTPAs were reported positive by the staff radiologist. There was agreement in 90% (752/840) of studies (P = .76, 95% confidence interval, 0.71-0.81) with 86% (114/133) agreement for studies interpreted as positive by residents, 95% (582/612) for studies interpreted as negative by residents, and 63% (60/95) for studies interpreted as equivocal by residents. Studies of optimal quality had higher interobserver agreement than studies of suboptimal quality (P < .0001). In-patient studies were more likely to be positive than emergency room patients (20% vs. 13%) (P = .004). No adverse clinical outcomes were attributed to discordant interpretations. CONCLUSIONS: Radiology residents provide a high level interpretation of on-call CTPA studies, achieving good concordance with the attending radiologists' assessment.     

Aortic dissection: a statistical analysis of the usefulness of plain chest radiographic findings

Findings on plain chest radiographs of patients with aortic dissection are variable and often overlap those of patients without dissection. To determine which findings were most useful in predicting aortic dissection, plain chest radiographs from 36 patients with aortographically proven aortic dissection and 36 patients from a control population were randomized and analyzed independently by five radiologists for the presence of various radiographic features associated with this condition. A widened aortic knob, widened descending aorta, and widened mediastinum showed the greatest interobserver agreement (p less than .001) although the overall interobserver agreement was poor. The final conclusion of the radiologists was a better predictor of dissection than any of the individual radiographic features alone. Widening of the mediastinum (p less than .001) and widening of the aortic knob (p less than .012) were the only two radiographic features of significance in predicting dissection. In a stepwise multiple logistic regression model, the radiologists achieved an overall accuracy of 85%, a sensitivity of 81%, and a specificity of 89%. Although this illustrates the usefulness of plain chest radiographs in diagnosing aortic dissection, poor interobserver agreement dictates that further definitive investigation be undertaken.     

Extracorporeal membrane oxygenation: radiographic appearance of the neonatal chest

The chest radiographs of 26 newborns treated with extracorporeal membrane oxygenation (ECMO) for intractable respiratory failure were reviewed. The typical radiographic appearance of the lungs in these patients is that of diffuse pulmonary opacification with variable volume loss. Air bronchograms and patchy basilar atelectasis are also common findings. Generally, decreasing ECMO requirements were reflected in improving chest radiographs with radiographic improvement lagging behind clinical improvement. Of 167 chest radiographs available for evaluation, 105 (62.8%) reflected changes in ECMO flow rates. Radiographs in patients with individual diagnoses of hyaline membrane disease, meconium aspiration syndrome and sepsis showed the best correlation with clinical improvement (95 [69%] of 137 radiographs). Those obtained in patients with congenital diaphragmatic hernia and persistent pulmonary hypertension of the newborn alone showed the poorest correlation (10 [30%] of 30 of radiographs). Neither the absolute degree of radiographic abnormality nor degree of radiographic improvement correlated well with ECMO requirements. Initial radiographs were useful in confirming the position of bypass cannulae and respiratory tubes. Routine daily examinations did not reveal unexpected abnormalities. However, radiographs taken during periods of increased ECMO requirements due to patent ductus arteriosus or volume overload showed worsening lung opacification.     

The Interobserver Agreement between Residents and Experienced Radiologists for Detecting Pulmonary Embolism and DVT with Using CT Pulmonary Angiography and Indirect CT Venography

Objective

We wanted to prospectively evaluate the interobserver agreement between radiology residents and expert radiologists for interpreting CT images for making the diagnosis of pulmonary embolism (PE).

Materials and Methods

We assessed 112 consecutive patients, from April 2007 to August 2007, who were referred for combined CT pulmonary angiography and indirect CT venography for clinically suspected acute PE. CT scanning was performed with a 64×0.5 collimation multi-detector CT scanner. The CT studies were initially interpreted by the radiology residents alone and then the CT images were subsequently interpreted by a consensus of the resident plus an experienced general radiologist and an experienced chest radiologist.

Results

Two of the 112 CTs were unable to be interpreted (1.7%). Pulmonary artery clots were seen on 36 of the thoracic CT angiographies (32%). The interobserver agreement between the radiology residents and the consensus interpretation was good (a kappa index of 0.73). All of the disagreements (15 cases) were instances of overcall by the resident on the initial interpretation. Deep venous thrombosis was detected in 72% (26 of 36) of the patients who had PE seen on thoracic CT. The initial and consensus interpretations of the CT venography images disagreed for two cases (kappa statistic: 0.96).

Conclusion

It does not seem adequate to base the final long-term treatment of PE on only the resident''s reading, as false positives occurred in 13% of such cases. Timely interpretation of the CT pulmonary angiography and CT venography images should be performed by experienced radiologists for the patients with suspected PE.     

Digital analysis of chest radiographs in pulmonary vascular congestion

Twenty-five erect postero-anterior chest radiographs of 7 patients in and out of congestive heart failure, as well as 6 healthy controls, were graded as to severity of pulmonary vascular congestion. These visual gradings were compared with a vascular complexity index (VCI) computed after digitization of each radiograph. Good correlation was found between visual grading and the VCI (R = 0.77) and for assessing changes in a given patient (R = 0.95). In parallel animal experiments, correlation between the VCI and left-ventricular end-diastolic pressure (LVEDP) was also evaluated. Hypervolemia was induced in 5 dogs by infusion of dextran. LVEDP was measured in each catheterized erect animal and radiographs taken before and after infusion. Correlation between the VCI and LVEDP during infusion was good (R = 0.99).     

Clinical trial of digital teleradiology in the practice of emergency room radiology

The application of digital teleradiology (DTR) to radiologic examinations performed in the emergency room was evaluated. A total of 919 examinations (ten computed tomographic; the rest, radiographic) were transmitted to a radiology resident at another hospital emergency room. The 512 X 512 images were reviewed by an attending radiologist and compared with another attending radiologist's interpretation of the original films. Cases with discrepant interpretations were analyzed. Inadequate DTR image quality was responsible for clinically significant discrepancies in 14 of 897 cases (1.6%) available for follow up. Problem areas such as the detection of pneumothorax and abdominal calcifications were identified. Retransmission of optically zoomed images of areas of concern and repeat radiographs of overpenetrated films are suggested to improve DTR performance.     

Accuracy of chest radiographs in the emergency diagnosis of heart failure

The purpose of this study was to determine the diagnostic accuracy of chest radiographic findings of heart failure (HF) in current patients presenting with dyspnea in the emergency department. In a secondary analysis of the BASEL study, initial chest radiographs of 277 patients with acute dyspnea were evaluated by two radiologists blinded to the adjudicated diagnosis (56% had the final diagnosis of HF). Predefined radiographic criteria of HF were used. Statistical analysis included receiver-operating characteristic (ROC) analysis and calculation of a logistic regression model including B-type natriuretic peptide (BNP) levels. The reader's overall impression showed the highest area under the ROC curve for the diagnosis of HF in both supine and erect patient positions (0.855 and 0.857). Among individual radiographic findings, peribronchial cuffing in the supine position (0.829) showed the highest accuracies. The lowest accuracy was found for the vascular pedicle width in the supine position (0.461). Logistic regression analysis showed no significant differences between the reader's overall impression, the radiographic model, and BNP testing. In our study, the combination of radiographic features provided valuable information and was of comparable accuracy as BNP-testing for the diagnosis of HF.