Computer-Aided Detection of Acute Pulmonary Embolism With 64-Slice MultiYDetector Row Computed Tomography: Impact of the Scanning Conditions and Overall Image Quality in the Detection of Peripheral Clots Marion Dewailly, MD,* Martine Re´my-Jardin, MD, PhD,* Alain Duhamel, PhD,Þ Jean-Baptiste Faivre, MD,* Fran0ois Pontana, MD,* Vale´rie Deken, MD,Þ Anne-Marie Bakai, PhD,þ and Jacques Remy, MD* Purpose: To evaluate the performance of a computer-aided detection (CAD) system for diagnosing peripheral acute pulmonary embolism (PE) with a 64-slice multiYdetector row computed tomography (CT). Materials and Methods: Two radiologists investigated the accuracy of a software aimed at detecting peripheral clots (PECAD prototype, version 7; Siemens Medical Systems, Forchheim, Germany) by applying this tool for the analysis of the pulmonary arterial bed of 74 CT an- giograms obtained with 64-slice dual-source CT (Definition; Siemens Medical Systems). These cases were retrospectively selected from a database of CT studies performed on the same CT unit, with a similar collimation (64 0.6 mm) and similar injection protocols. Patient selection was based on a variety of (1) scanning conditions, namely, nongated (n = 30), electrocardiography-gated (n = 30), and dual-energy CT angiograms (n = 14), and (2) image quality (IQ), namely, scans of excellent IQ (n = 53) and lower IQ due to lower levels of arterial en- hancement and/or presence of noise (n = 21). The standard of truth was based on the 2 radiologists’ consensus reading and the results of CAD. Results: The software detected 80 of 93 peripheral clots present in the 21 patients (42 segmental and 38 subsegmental clots). The over- all sensitivity (95% confidence interval) of the CAD tool was 86% (77%Y92%) for detecting peripheral clots, 78% (64.5%Y88%) at the segmental level and 97% (85.5%Y99.9%) at the subsegmental level. Assuming normal vascular anatomy with 20 segmental and 40 sub- segmental arteries, overall specificity and positive and negative predic- tive values (95% confidence interval) of the software were 91.8% (91%Y92.6%), 18.4% (15%Y22.4%), and 99.7% (99.5%Y99.8%), res- pectively. A mean of 5.4 false positives was found per patient (total, 354 false positives), mainly linked to the presence of perivascular con- nective tissue (n = 119; 34%) and perivascular airspace consolidation (n = 97; 27%). The sensitivities (95% confidence interval) for the CAD tool were 91% (69.8%Y99.3%) for dual-energy, 87% (59.3%Y93.2%) for electrocardiography-gated, and 87% (73.5%Y95.3%) for nongated scans (P 9 0.05). No significant difference was found in the sensitivity of the CAD software when comparing the scans according to the scan- ning conditions and image quality. Conclusions: The evaluated CAD software has a good sensitivity in detecting peripheral PE, which is not influenced by the scanning con- ditions or the overall image quality. Key Words: pulmonary arteries, pulmonary embolism, CT angiography, CAD, dual-energy CT (J Comput Assist Tomogr 2010;34: 23Y30) S ince the introduction of multiYdetector row computed to- mography (MDCT) with high spatial and temporal resolu- tions, CT angiography (CTA) has become the method of choice for imaging the pulmonary vasculature for suspected pulmo- nary embolism (PE) in routine clinical practice. 1 Compared with single-slice CT, this technology has been shown to improve the evaluation of peripheral pulmonary arteries and thus the accu- racy of CT in the diagnostic workup of acute PE. 2Y6 Despite this improvement in the overall image quality, assessment or exclu- sion of peripheral clots on MDCTAs is not always an easy task in daily clinical routine owing to the huge number of thin- collimated CT scans generated from each MDCT data set. In the absence of central PE, missing peripheral clots may lead to a false-negative (FN) diagnosis of acute PE with therapeutic and prognostic consequences. This situation is particularly criti- cal when considering the specific category of isolated subseg- mental PE whose frequency has been reported to vary between 1% and 5.4% in study populations with suspected PE. 6Y9 The clinical relevance of small peripheral pulmonary emboli and the need to anticoagulate these patients remain a subject of debate. As recently emphasized, 10 there are 3 clinical scenarios in which most would agree that even small PE require treatment: (a) in patients with small PE and inadequate cardiopulmonary reserve, (b) in patients who have a small embolus and coexisting acute deep venous thrombosis, and (c) in patients who have recurrent small PE possibly due to thrombophilia, to prevent chronic PE and pulmonary artery hypertension. Lastly, one should keep in mind the dramatic decrease in the prevalence of acute PE that has dropped from 33% in angiographic studies 11 to 6% in recent studies reporting MDCT use in emergency departments. 12,13 In this context, the introduction of computer- aided diagnostic systems for detection of acute PE could help improve the radiologist’s level of confidence in excluding PE, especially in the peripheral pulmonary arterial bed. All the pre- viously mentioned reasons have led to the investigation of several CAD systems over the last few years that showed that they are able to detect filling defects with high sensitivity. 14Y21 Because the performance of a CAD algorithm depends strongly on the degree of difficulty of the lesions to be detected in the data set, we undertook the present study to evaluate the impact of the scanning conditions and overall image quality of 64-slice MDCT scans in the detection of peripheral clots by a prototype system. ORIGINAL ARTICLE J Comput Assist Tomogr & Volume 34, Number 1, January/February 2010 www.jcat.org 23 From the Departments of *Thoracic Imaging, Hospital Calmette, and †Medi- cal Statistics, University Center of Lille, Lille, France; and ‡Department of Research and Development, Siemens Medical Systems, Forchheim, Germany. Received for publication February 9, 2009; accepted June 11, 2009. Reprints: Martine Re´my-Jardin, MD, PhD, Department of Thoracic Imaging, Hospital Calmette, University Center of Lille, Blvd Jules Leclercq, 59037, Lille cedex, France (e-mail: Martine.REMY@CHRU-LILLE.FR). The institutional review board and ethics committee of the University Center of Lille approved this retrospective study with waiver of patient’s informed consent, in agreement with French regulations. Copyright * 2010 by Lippincott Williams & Wilkins Copyright @ 20 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. 10