Cardiopulmonary Imaging Original Research

Transcription

1 Cardiopulmonary Imaging Original Research Sheh et al. CTA and V/Q Scintigraphy of Pulmonary Embolism Cardiopulmonary Imaging Original Research Steven H. Sheh 1,2 Eran Bellin 3,4 Katherine D. Freeman 5,6 Linda B. Haramati 1,3 Sheh SH, Bellin E, Freeman KD, Haramati LB Keywords: mortality, pulmonary CT angiography, pulmonary embolism, ventilation-perfusion scintigraphy DOI: /AJR Received January 4, 2011; accepted after revision May 24, Dr. Sheh was a student at Einstein s Clinical Research Training Program, which was funded by CTSA Grant UL1 RR025750, KL2 RR025749, and TL1 RR from the National Center for Research Resources (NCRR), a component of the National Institutes of Health (NIH), and NIH roadmap for Medical Research. The contents of this manuscript are solely the responsibility of the authors and do not necessary represent the official view of the NCRR or NIH. 1 Department of Radiology, Albert Einstein College of Medicine, Montefiore Medical Center, 111 E 210 St, Bronx, NY Address correspondence to L. B. Haramati (lharamati@gmail.com). 2 Present address: Department of Radiology, University of California, Davis, Health System, Sacramento, CA. 3 Department of Internal Medicine, Albert Einstein College of Medicine, Montefiore Medical Center, Bronx, NY. 4 Department of Outcomes Analysis and Decision Support, Albert Einstein College of Medicine, Montefiore Medical Center, Bronx, NY. 5 Department of Epidemiology and Population Health, Albert Einstein College of Medicine, Montefiore Medical Center, Bronx, NY. 6 Present address: Extrapolate, LLC, Delray Beach, CA. AJR 2012; 198: X/12/ American Roentgen Ray Society Pulmonary Embolism Diagnosis and Mortality With Pulmonary CT Angiography Versus Ventilation- Perfusion Scintigraphy: Evidence of Overdiagnosis With CT? OBJECTIVE. The purposes of this study were to determine whether pulmonary emboli diagnosed with pulmonary CT angiography (CTA) represent a milder disease spectrum than those diagnosed with ventilation-perfusion (V/Q) scintigraphy, to determine the trends in incidence and mortality among patients with the diagnosis of pulmonary embolism from 2000 to 2007, and to correlate incidence and mortality trends with imaging modality trends. MATERIALS AND METHODS. Diagnoses of pulmonary embolism from 2000 to 2007 at an urban academic medical center were retrospectively identified. Patient data were collected from the hospital database and the Social Security Death Index. Incident diagnoses, type of imaging used, and date of death were documented. Bivariate and multivariate analyses were used to explore the relations between imaging use and the incidence and mortality of pulmonary embolism. Logistic regression analysis was used to estimate the odds of death of pulmonary embolism diagnosed with pulmonary CTA versus V/Q scintigraphy. RESULTS. The cases of 2087 patients (1361 women, 726 men; mean age, 61.8 years) with pulmonary embolism were identified. From 2000 to 2007 the incidence of pulmonary embolism increased from 0.69 to 0.91 per 100 admissions in strong correlation with increased use of pulmonary CTA. There was no change in mortality, but the case-fatality rate decreased from 5.7% to 3.3%. On average, pulmonary emboli diagnosed with pulmonary CTA were one half as lethal as those diagnosed with V/Q scintigraphy (odds ratio, 0.538; 95% CI, ). CONCLUSION. The results of this study are evidence that the shift in imaging from V/Q scintigraphy to pulmonary CTA resulted in increased diagnosis of a less fatal spectrum of pulmonary embolic disease, raising the possibility of overdiagnosis. Outcome-based clinical trials with long-term follow-up would be helpful to further guide management. P ulmonary embolism (PE) is a potentially fatal form of venous thromboembolic disease and is responsible for approximately 250,000 hospitalizations annually in the United States [1]. The two main imaging modalities for suspected PE are pulmonary CT angiography (CTA) and ventilation-perfusion (V/Q) scintigraphy. Pulmonary CTA overtook V/Q scintigraphy as the most common imaging modality for suspected PE in 2001 [2]. Ongoing technologic advances have resulted in better delineation of the subsegmental pulmonary vasculature at pulmonary CTA [3, 4]. Anderson et al. [5] and the Prospective Investigation of Pulmonary Embolism Diagnosis II investigators [6] in prospective clinical trials found that pulmonary CTA showed more pulmonary emboli than did V/Q scintigraphy. They did not, however, find a clinical benefit of pulmonary CTA in terms of mortality or recurrent thromboembolic events. These results raise concern about overdiagnosis of PE, defined as the diagnosis of clinically unimportant disease [7, 8]. A disadvantage of pulmonary CTA compared with V/Q scintigraphy is its higher radiation exposure. Mettler et al. [9] estimated that performed with currently available, nonoptimized technology, pulmonary CTA is associated with a nearly sevenfold higher radiation burden than V/Q scintigraphy. Burge et al. [10], in an ecologic study, described an increase in the diagnosis of PE in the state of New York between 1994 and 2004 with the advent of pulmonary CTA with no change in PE mortality. In their age-adjusted multivariate analysis, those investigators did not find a significant association between the rates of PE diagnosis or mortality and rates of specific risk factors for PE, including malignancy, smoking, obesity, and surgery. Wiener et al. [11] 1340 AJR:198, June 2012

2 CTA and V/Q Scintigraphy of Pulmonary Embolism confirmed similar trends at a national level and found inferential evidence of an increase in complications of anticoagulation. Thus evidence in the literature raises the possibility that the widespread adoption of pulmonary CTA has led to diagnosis of a milder spectrum of PE disease with a natural history different from that historically associated with the diagnosis. We designed this outcome-based singlecenter retrospective cohort study to build on our previous work [10]. Using individual patient longitudinal data, we sought to verify the implications of our ecologic trends analysis by documenting an increase in the diagnosis of PE associated with increased use of pulmonary CTA without a change in PE mortality. To test the hypothesis that PE diagnosed with pulmonary CTA represents a milder disease spectrum, we compared the odds of death in our cohort of patients with PE diagnosed with pulmonary CTA and the odds among patients with PE diagnosed with V/Q scintigraphy, adjusting for risk. Materials and Methods Data Collection We identified the cases of all adults (18 years and older) admitted to our urban academic medical center over the 8-year period January 1, 2000, through December 31, 2007, with a primary or secondary diagnosis of PE based on the International Classification of Diseases, ninth revision, codes 415.1, , , and Only the first instance of PE was considered. For each year, the number of unique patients admitted to the hospital, the number with the diagnosis of PE, and the number with PE who died within 7 and 14 days of diagnosis was ascertained. We collected patient-level data from the hospital electronic medical record and the national Social Security Death Index using Clinical Looking Glass software, a user-friendly interactive application developed at our institution to evaluate health care quality, effectiveness, and efficiency. We noted the date of PE diagnosis, the type of imaging studies performed (pulmonary CTA, V/Q scintigraphy, both, or neither), and patient sex and age. To adjust mortality for baseline patient differences, we recorded serum albumin concentration and the component categories constituting the Charlson comorbidity score: sex, albumin concentration, and comorbid conditions (myocardial infarction, congestive heart failure, peripheral vascular disease, cerebrovascular disease, dementia, chronic pulmonary disease, rheumatologic disease, peptic ulcer disease, diabetes without complications, diabetes with chronic complications, hemiplegia or paraplegia, renal disease, mild liver disease, moderate or severe liver disease, malignancy, metastatic solid tumor, AIDS or HIV infection). We included the presence of any malignancy, which Cushman et al. [12] found to be a primary risk factor for PE. Outcome of Interest Our outcome of interest was death due to PE. Accurate attribution of cause of death is notoriously difficult and remains elusive because inpatient autopsy rates nationwide approach zero. Nijkeuter et al. [13] found that PE-related deaths occurred mostly within the first week after diagnosis. Carson et al. [14] found that most PE-related deaths occurred within the first 2 weeks after diagnosis. Their analyses excluded the subset of patients with PE who died suddenly at home. Because our analyses were focused on hospital patients, we were guided by the work of Nijkeuter et al. and Carson et al. and described PE mortality as death within 7 and 14 days after PE diagnosis. In truth, this rate represents allcause mortality enhanced by the presence of PE, a surrogate unbiased toward a specific imaging modality. At our institution, both pulmonary CTA and V/Q scintigraphy have consistently been readily available and routinely performed, even during offhours and on weekends; thus in our analysis the potential for bias related to the availability of specific diagnostic imaging modality was reduced [15]. Analyses We documented annual trends in the incidence of PE, clinical imaging practices, 7- and 14-day mortality. We studied the annual case rate (number of admissions for PE per annual hospital admissions), mortality rate (number of patients who died with diagnosis of PE per annual admissions), and fatality rate (number of patients with PE who died per annual population of PE patients). Death was determined from electronic medical record data and matched to the Social Security Death Index. Patient-level data, including modality for diagnosis (pulmonary CTA or V/Q scintigraphy) and presence of comorbid conditions were used to build logistic models to explore the effect of diagnostic modality on 14-day death rate as a proxy for PE mortality rate. Poisson regression was used to calculate population trends in annual PE incidence rate, mortality rate, and rate of specific diagnostic tests adjusted by the total number of hospital admissions for each year. Nonparametric pairwise correlation was used to correlate the trends in PE incidence and pulmonary CTA use. Our second novel analysis focused on the differences in mortality rate between patients with PE diagnosed at our institution with pulmonary CTA and those with PE diagnosed with V/Q scintigraphy. To establish homogeneous groups suitable for this analysis, we included the subset of patients initially imaged with either pulmonary CTA or V/Q scintigraphy, but not both. Reports of all pulmonary CTA and V/Q scintigraphy scans were reviewed. To control for confounding, we adjusted for the nonrandom distribution of patients between pulmonary CTA and V/Q scintigraphy by calculating a propensity score for diagnostic modality that included patient age, sex, albumin concentration, and the all of the individual Charlson comorbid conditions [16]. The logistic regression model was used to estimate the odds ratio for death of PE. This study was approved by our institutional review board with waiver of consent. All statis- TABLE 1: Baseline Characteristics According to Imaging Modality Characteristic Pulmonary CTA Only (n = 832) V/Q Scintigraphy Only (n = 475) Both (n = 241) Neither (n = 539) Age (y) Mean SD Sex (%) Women Men Race (%) White Black Charlson index 4.0 ± ± ± ± 2.8 Age-related risk 2.5 ± ± ± ± 1.3 Combined score 5.5 ± ± ± ± 3.1 Note CTA = CT angiography, V/Q = ventilation-perfusion scintigraphy. AJR:198, June

3 Sheh et al. TABLE 2: Data by Year and Imaging Modality Year Total No. of Admissions No. of PEs tical analyses were performed with Stata/IC 10.1 software (StataCorp). A p value of less than 0.05 was considered significant. Results Among the 2087 adults with a primary or secondary diagnosis of PE during the study period were 1361 women and 726 men (mean age, 61.8 ± 17.8 [SD] years) (Table 1). Over the 8-year period, there was a significant increase in the rate of PE diagnosis from 0.69 to 0.91 per 100 admissions (incidence rate ratio, 1.056; 95% CI, ). The number of patients with an International Classification of Diseases, ninth revision, code diagnosis of PE increased from 219 in 2000 to 366 in 2007 (Table 2). Of the 2087 patients, 40% (n = 832) underwent imaging only with pulmonary CTA, 23% (n = 475) with only V/Q scintigraphy, No. of PEs Diagnosed With Modality 12% (n = 241) with both modalities, and 26% (n = 539) with neither modality at our institution. The number of cases of PE diagnosed with pulmonary CTA increased from 45 to 167, corresponding to a significant increase in incidence rate ratio from 0.14 to 0.41 (incidence rate ratio, 1.219; 95% CI, ) (Fig. 1). There was a trend toward a significant decrease in rate of V/Q scintigraphy diagnosed PE (incidence rate ratio, 0.921; 95% CI, ) over the study period. There was a significant and strong correlation between the trends in PE incidence and use of pulmonary CTA (r = 0.82, p < 0.01). There was no significant change in 7-day PE mortality rate over the study period (incidence rate ratio, 0.966; 95% CI, ) with nine deaths in 2000 and nine deaths in The 14-day mortality rate also did not change significantly over the study period (incidence rate No. of Deaths After PE Diagnosis Pulmonary CTA Only V/Q Scintigraphy Only Both Neither Within 7 d Within 14 d , , , , , , , , Total 282, Note PE = pulmonary embolism, CTA = CT angiography, V/Q = ventilation-perfusion scintigraphy. Admission Fraction Rate of pulmonary embolism diagnosis Rate of pulmonary CTA utilization Rate of ventilation-perfusion scintigraphy utilization Mortality rate from pulmonary embolism Year Fig. 1 Graph shows calculated rates of pulmonary embolism (PE) diagnosis, clinical use of imaging, and PE mortality between 2000 and Rate of diagnosis of PE increased significantly from 0.69 to 0.91 (incidence rate ratio, 1.056; 95% CI, ), as did use of pulmonary CT angiography (CTA), from 0.14 to 0.41 (incidence rate ratio, 1.219; 95% CI, ). Use of ventilation-perfusion scintigraphy declined (incidence rate ratio, 0.921; 95% CI, ), and PE mortality did not change (0.028 in 200, in 2007; incidence rate ratio, 0.966; 95% CI, ). There was significant correlation between trends in pulmonary CTA utilization and PE incidence over study period (r = 0.82, p < 0.01). Case-Fatality Rate (%) ratio, 0.937; 95% CI, ). However, the case-fatality rate significantly decreased from 5.7% in 2000 to 3.3% in 2007 (incidence rate ratio, 0.90; 95% CI, ) (Fig. 2). The propensity score was used to adjust for the fact that patients evaluated with pulmonary CTA more likely had dementia, chronic pulmonary disease, liver disease, malignant disease, or HIV/AIDS. Those who underwent V/Q scintigraphy were more likely to be older patients with congestive heart failure, complicated diabetes, renal disease, or high albumin concentrations (Table 3). Logistic regression analysis with propensity adjustment showed that the relative odds of death of PE diagnosed with pulmonary CTA were approximately one half those of PE diagnosed with V/Q scintigraphy (odds ratio, 0.538; 95% CI, ) Year Fig. 2 Graph shows significant decrease in pulmonary embolism case-fatality rate over study period, from 5.7% in 2000 to 3.3% in 2007 (incidence rate ratio, 0.90; 95% CI, ) AJR:198, June 2012

4 CTA and V/Q Scintigraphy of Pulmonary Embolism Discussion In the current study, two analyses were used to test the hypothesis that the spectrum of PE disease diagnosed with pulmonary CTA is milder than that of disease diagnosed with V/Q scintigraphy. The results of both analyses support this hypothesis. First we used patientlevel data to confirm the results of an ecologic analysis by Burge et al. [10]. Like those investigators, we found a significant increase in the incidence rate of PE, mirroring and significantly correlating with an increase in use of pulmonary CTA. One might assume that this increase in the diagnosis of PE represents a clearcut positive development. However, it is increasingly emphasized that results of imaging should be evaluated with an eye toward improved patient outcome rather than simply detecting more disease. In evaluating the hard outcome of mortality, we did not find a decline in mortality. Our second analysis was focused specifically on the subgroup of patients with PE diagnosed solely with either pulmonary CTA or V/Q scintigraphy. Because the study had a retrospective cohort design and the patients were not randomly assigned to the respective imaging modality, we developed a propensity score to statistically adjust for the differences between the two populations. After this adjustment, logistic regression analysis showed a significant difference in odds of death between patients with PE diagnosed with V/Q scintigraphy and those with PE diagnosed with pulmonary CTA. The odds of death among patients with PE diagnosed with pulmonary CTA were approximately one half of those among patients with PE diagnosed with V/Q scintigraphy. In 2001, pulmonary CTA overtook V/Q scintigraphy as the most-used imaging modality to assess for suspected PE in the United States [2]. At our institution this shift occurred in June In general, introduction of a new diagnostic test results in detection of a different spectrum of disease than is detected with the traditional test. The clinical implications of a change in the sensitivity profile of a new diagnostic test require scrutiny. The results of the current study support the hypothesis that PE diagnosed with pulmonary CTA has a different, more benign natural history than PE diagnosed with V/Q scintigraphy and that overdiagnosis of clinically unimportant disease may be occurring. Our analysis proceeded with the premise that the prevalence of fatal PE, a hard outcome, is a marker for clinically important TABLE 3: Patient Characteristics by Imaging Modality as Calculated With Propensity Score Characteristic Pulmonary CT Angiography Ventilation-Perfusion Scintigraphy Age Younger Older Congestive heart failure Less More Dementia More Less Chronic pulmonary disease More Less Complicated diabetes mellitus Less More Renal disease Less More Liver disease More Less Malignancy More Less AIDS or HIV infection More Less Albumin concentration Lower Higher disease. Results in the literature support this premise [6, 17, 18]. Mayo et al. [19] found that 34% more cases of PE were diagnosed with pulmonary CTA than with V/Q scintigraphy. Anderson et al. [5] performed a landmark randomized prospective clinical trial comparing pulmonary CTA and V/Q scintigraphy in a highrisk population and found that 52% more cases of PE were diagnosed with pulmonary CTA than with V/Q scintigraphy without a significant difference in the major clinical outcomes of subsequent thromboembolic disease and death. However, because even a small pulmonary embolus can present considerable risk to a patient with poor cardiopulmonary reserve, in clinical practice these results must be tailored to each patient s care. Even so, Carrier et al. [20], in a meta-analysis of 22 clinical trials, documented an increase in the proportion of patients with subsegmental PE at pulmonary CTA without an effect on the rate of subsequent thromboembolic disease. They flatly concluded that the results of their analysis suggest that subsegmental PE may not be clinically relevant. If there is an increasing trend toward overdiagnosis, then we would expect the PE casefatality rate to decline over time while the mortality rate remains stable. Results of the current study met these expectations and showed that the case-fatality rate significantly decreased approximately 10 11% per year. This finding mirrors the national trends from 1998 to 2005 reported by Park et al. [21], who did not consider the possibility of overdiagnosis in their analysis, and confirms the findings of Wiener et al. [11]. In older studies [22 25] investigators attributed decreasing trends in the casefatality of PE to the advent of interventions that changed the natural history of the disease. However, since the late 1990s, the two most important changes in the diagnosis and treatment of PE have been increased use of lowmolecular-weight heparin and increased use of pulmonary CTA [26, 27]. Numerous studies have shown that low-molecular-weight heparin does not reduce mortality compared with standard heparin therapy [26, 28, 29]. The more likely reason for the decrease in case-fatality is related to the increased use of pulmonary CTA. Wiener et al. documented a national increase in hemorrhagic events, which they attributed to anticoagulation for PE. Goodman [17] noted that incidental PE are found in a large minority of autopsies and that a normal function of the pulmonary capillary bed is to filter small clots to protect the systemic circulation. Engleke et al. [30] found no difference between the 1-year survival rate of patients with small PE on CT images that were missed, and therefore not treated, and the survival rate of patients with PE that was diagnosed and treated. Suh et al. [31] described a series of patients with PE who underwent pulmonary CTA and lower extremity ultrasound. They found that 58% of the patients with central PE had lower extremity deep venous thrombosis, in contrast to none of the patients with small peripheral emboli. These data in aggregate suggest that small PE may be physiologic and that the phenomenon of overdiagnosis, defined as the diagnosis of clinically unimportant disease, may be relevant to PE. As CT technology advances, a larger number of pathologic changes are becoming visible. The conundrum of overordering for patients with mild symptoms may accentuate this phenomenon. Although intuitively it seems that seeing and diagnosing more disease should lead to improvement in patient AJR:198, June

5 Sheh et al. outcome, in many cases this perception has not been substantiated by data. Little supportive outcome research has been performed, and the body of evidence suggesting that overdiagnosis may be a real problem for imaging is small [10, 11, 20, 32, 33]. Anticoagulation was the mainstay of treatment of PE before the advent of pulmonary CTA as the primary imaging modality. The risk-to-benefit ratio for the treatment of disease in the pre-ct era has been toward benefit in clinical trials, despite the considerable risk of bleeding related to anticoagulation [29, 34 36]. The risk-to-benefit ratio for treating patients with a milder spectrum of PE disease may favor no treatment or shorter treatment of patients in stable condition with small emboli [11, 20]. Further study with prospective clinical trials is needed. Patients with PE diagnosed with pulmonary CTA are exposed to higher radiation doses than those with PE diagnosed with V/Q scintigraphy, although dose-reduction techniques for CT are developing rapidly. We [37] documented an increased rate of imaging in our population of patients with PE, and Mettler et al. [9] reported that 30% of patients who underwent CT underwent at least three examinations, 7% at least five, and 4% at least nine. Kline et al. [38] similarly reported repeat pulmonary CTA examinations of one third of patients evaluated for suspected PE in the emergency department. These data portend a public health risk if imaging continues unchecked. The current study differs from other PE trends research in that we used patient-level data and repeat admissions were excluded. We therefore were able to perform propensity and clinical adjustments and to track deaths outside the hospital using the Social Security Death Index. Limitations of this study include its retrospective nature and the lack of randomization of the comparison groups, pulmonary CTA versus V/Q scintigraphy. Although we adjusted statistically for potential bias using a propensity score of known baseline characteristics, statistical adjustment always represents an imperfect model. We described all deaths within 7 and 14 days after diagnosis of PE as attributable mortality. Although this attributable mortality is surely an overestimate, it does not bias our analysis in favor of a specific imaging modality and represents a surrogate in the timeframe during which PE is expected to have the greatest mortality influence [13, 14]. We also did not perform a more detailed review of the medical records of those who died, precluding a detailed individual analysis. Another limitation was that approximately one fourth of the patients with newly diagnosed PE underwent neither pulmonary CTA nor V/Q scintigraphy at our institution, and we were unable to identify the method of diagnosis. This group was included in our analysis of trends in PE imaging, mortality, and case-fatality rate, which confirmed previously observed trends [11, 20, 36]. However, only patients who underwent initial imaging with either pulmonary CTA or V/Q scintigraphy were included in our novel analysis of adjusted odds of death among patients with PE. Other clinically important outcomes that were not assessed and merit future investigation are the effect of diagnostic modality on treatment and its complications, the development of chronic thromboembolic disease during long-term follow-up, and deaths of PE more than 14 days after diagnosis. Conclusion The results of the current study are evidence that the shift in imaging from V/Q scintigraphy to pulmonary CTA has resulted in diagnosis of a different, less fatal spectrum of PE disease. This finding raises the possibility of overdiagnosis. Because current treatment recommendations were designed before the widespread adoption of pulmonary CTA, outcome-based clinical trials with long-term follow-up would be helpful to further guide management. References 1. Anderson FA Jr, Wheeler HB, Goldberg RJ, et al. A population-based perspective of the hospital incidence and case-fatality rates of deep vein thrombosis and pulmonary embolism: the Worcester DVT study. Arch Intern Med 1991; 151: Stein PD, Kayali F, Olson RE. Trends in the use of diagnostic imaging in patients hospitalized with acute pulmonary embolism. Am J Cardiol 2004; 93: Schoepf UJ, Holzknecht N, Helmberger TK, et al. Subsegmental pulmonary emboli: improved detection with thin-collimation multi-detector row spiral CT. Radiology 2002; 222: Patel S, Kazerooni EA, Cascade PN. Pulmonary embolism: optimization of small pulmonary artery visualization at multi-detector row CT. Radiology 2003; 227: Anderson DR, Kahn SR, Rodger MA, et al. Computed tomographic pulmonary angiography vs ventilation-perfusion lung scanning in patients with suspected pulmonary embolism: a randomized controlled trial. JAMA 2007; 298: Stein PD, Fowler SE, Goodman LR, et al.; PIOPED II Investigators. Multidetector computed tomography for acute pulmonary embolism. N Engl J Med 2006; 354: Perrier A, Bounameaux H. Accuracy or outcome in suspected pulmonary embolism. N Engl J Med 2006; 354: Glassroth J. Imaging of pulmonary embolism too much of a good thing? JAMA 2007; 298: Mettler FA Jr, Huda W, Yoshizumi TT, Mahesh M. Effective doses in radiology and diagnostic nuclear medicine: a catalog. Radiology 2008; 248: Burge AJ, Freeman KD, Klapper PJ, Haramati LB. Increased diagnosis of pulmonary embolism without a corresponding decline in mortality during the CT era. Clin Radiol 2008; 63: Wiener RS, Schwartz LM, Woloshin S. Time trends in pulmonary embolism in the United States: evidence of overdiagnosis. Arch Intern Med 2011; 171: Cushman M, Tsai AW, White RH, et al. Deep vein thrombosis and pulmonary embolism in two cohorts: the longitudinal investigation of thromboembolism etiology. Am J Med 2004; 117: Nijkeuter M, Söhne M, Tick LW, et al; Christopher Study Investigators. The natural course of hemodynamically stable pulmonary embolism: clinical outcome and risk factors in a large prospective cohort study. Chest 2007; 131: Carson JL, Kelley MA, Duff A, et al. The clinical course of pulmonary embolism. N Engl J Med 1992; 326: Stein EG, Haramati LB, Chamarthy M, Sprayregen S, Davitt MM, Freeman LM. Success of a safe and simple algorithm to reduce use of CT pulmonary angiography in the emergency department. AJR 2010; 194: D Agostino RB Jr. Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group. Stat Med 1998; 17: Goodman LR. Small pulmonary emboli: what do we know? Radiology 2005; 234: Eyer BA, Goodman LR, Washington L. Clinicians response to radiologists reports of isolated subsegmental pulmonary embolism or inconclusive interpretation of pulmonary embolism using MDCT. AJR 2005; 184: Mayo JR, Remy-Jardin M, Müller NL, et al. Pulmonary embolism: prospective comparison of spiral CT with ventilation-perfusion scintigraphy. Radiology 1997; 205: Carrier M, Righini M, Wells PS, et al. Subsegmental pulmonary embolism diagnosed by computed tomography: incidence and clinical implications a systematic review and meta-analysis 1344 AJR:198, June 2012

6 CTA and V/Q Scintigraphy of Pulmonary Embolism of the management outcome. J Thromb Haemost 2010; 8: Park B, Messina L, Dargon P, Huang W, Ciocca R, Anderson FA. Recent trends in clinical outcomes and resource utilization for pulmonary embolism in the United States: findings from the nationwide inpatient sample. Chest 2009; 136: Horlander KT, Mannino DM, Leeper KV. Pulmonary embolism mortality in the United States, : an analysis using multiple-cause mortality data. Arch Intern Med 2003; 163: Lilienfeld DE, Chan E, Ehland J, Godbold JH, Landrigan PJ, Marsh G. Mortality from pulmonary embolism in the United States: 1962 to Chest 1990; 98: Lilienfeld DE, Godbold JH, Burke GL, Sprafka JM, Pham DL, Baxter J. Hospitalization and case fatality for pulmonary embolism in the twin cities: Am Heart J 1990; 120: Stein PD, Kayali F, Olson RE. Estimated case fatality rate of pulmonary embolism, 1979 to Am J Cardiol 2004; 93: [No authors listed]. Low-molecular-weight heparin in the treatment of patients with venous thromboembolism. The Columbus Investigators. N Engl J Med 1997; 337: Prologo JD, Gilkeson RC, Diaz M, Asaad J. CT pulmonary angiography: a comparative analysis of the utilization patterns in emergency department and hospitalized patients between 1998 and AJR 2004; 183: Hull RD, Raskob GE, Brant RF, et al. Low-molecular-weight heparin vs heparin in the treatment of patients with pulmonary embolism. American- Canadian Thrombosis Study Group. Arch Intern Med 2000; 160: Quinlan DJ, McQuillan A, Eikelboom JW. Lowmolecular-weight heparin compared with intravenous unfractionated heparin for treatment of pulmonary embolism: a meta- analysis of randomized, controlled trials. Ann Intern Med 2004; 140: Engelke C, Rummeny EJ, Marten K. Pulmonary embolism at multi detector row CT of chest: one year survival of treated and untreated patients. Radiology 2006; 239: Suh JM, Cronan JJ, Healey TT. Dots are not clots: the over-diagnosis and over-treatment of pulmonary embolism. Emerg Radiol 2010; 17: Warren R, Eleti A. Overdiagnosis and overtreatment of breast cancer: is overdiagnosis an issue for radiologists? Breast Cancer Res 2006; 8: Yankelevitz DF. Quantifying overdiagnosis in lung cancer screening. (comment) Radiology 2008; 246: ; author reply Donato AA, Khoche S, Santora J, et al. Clinical outcomes in patients with isolated subsegmental pulmonary emboli diagnosed by multidetector CT pulmonary angiography. Thromb Res 2010; 126:e266 e Barritt DW, Jordan SC. Anticoagulant drugs in the treatment of pulmonary embolism: a controlled trial. Lancet 1960; 1: Haspel J, Bauer K, Goehler A, et al. Long-term anticoagulant therapy for idiopathic pulmonary embolism in the elderly: a decision analysis. Chest 2009; 135: Stein EG, Haramati LB, Bellin E, et al. Radiation exposure from medical imaging in patients with chronic and recurrent conditions. J Am Coll Radiol 2010; 7: Kline JA, Courtney DM, Beam DM, King MC, Steuerwald M. Incidence and predictors of repeated computed tomographic pulmonary angiography in emergency department patients. Ann Emerg Med 2009; 54:41 48 FOR YOUR INFORMATION The AJR has made getting the articles you really want really easy with an online tool, Really Simple Syndication, available at It s simple. Click the RSS button located in the menu on the right side of the page. You ll be on your way to syndicating your AJR content in no time. AJR:198, June