Dual-Energy Chest Radiography with a Flat-Panel Digital Detector: Revealing Calcified Chest Abnormalities

Transcription

1 Frank Fischbach 1 Torsten Freund Rainer Röttgen Ulrike Engert Roland Felix Jens Ricke Received March 31, 2003; accepted after revision June 17, All authors: Department of Radiology, Charité, Campus Virchow-Klinikum, Humboldt-University Medical School, Augustenburgerplatz 1, Berlin, Germany. Address correspondence to F. Fischbach (frank.fischbach@charite.de). AJR 2003;181: X/03/ American Roentgen Ray Society Dual-Energy Chest Radiography with a Flat-Panel Digital Detector: Revealing Calcified Chest Abnormalities OBJECTIVE. The aim of this study was to assess the value of dual-energy chest radiography obtained using a cesium iodide flat-panel detector in addition to standard posteroanterior chest radiography for the detection of calcified chest abnormalities. MATERIALS AND METHODS. The study included 20 patients with a total of 37 calcified chest lesions (16 pulmonary nodules, 17 mediastinal calcifications, and four pleural calcifications) as confirmed on CT. Twenty-eight locations in the chests of the same patients who were free of lesions were used as negative controls. Four radiologists reviewed posteroanterior chest radiographs in a blinded manner alone and in conjunction with dual-energy soft-tissue and bone images. We calculated sensitivity, specificity, the negative predictive value (NPV), and the positive predictive value (PPV) for lesion prediction. The Wilcoxon s and the Brunner and Langer s tests were performed for statistical analysis. RESULTS. For posteroanterior chest radiography, sensitivity was 36%, the PPV was 64%, and the NPV was 47%. When dual-energy images were added, sensitivity increased significantly to 66% (p < 0.05), the PPV to 76%, and the NPV to 62%. The specificity remained constant at 73%. Brunner and Langer s test revealed a highly significant difference between posteroanterior chest radiography and dual-energy imaging in the detection of calcified chest abnormalities (p < 0.01). CONCLUSION. Dual-energy images added to standard posteroanterior chest radiographs significantly improve the detection of calcified chest lesions. D etecting and categorizing solitary pulmonary nodules and pleural opacities are primary tasks for the chest radiologist. The presence of calcification in a nodule is the single most valuable sign that the nodule may be benign [1, 2]. Hence, reliable identification of calcifications in chest radiographs is essential. The Revolution XQ/i amorphous silicon based digital radiography system (General Electric Medical Systems, Milwaukee, WI) has recently been introduced. In observer preference studies, radiologists rated the quality of the amorphous silicon based digital radiography superior to that of conventional film and computed radiography [3 5]. The system can perform advanced applications such as dual energy chest radiography. In the system used for this study, a dual-exposure technique was used to generate a subtracted bone and a subtracted soft-tissue image from a high-energy (120 kv) and a low-energy (60 kv) exposure. In effect, the contrast of the bone (ribs and spine) was eliminated from the soft-tissue image, and the contrast of the soft tissue was greatly reduced in the bone image [6]. Use of such a technique permits searching for soft-tissue nodules on one image of the chest unobscured by ribs while determining calcification on another image. Dual energy as a single- or dual-shot technique has shown high sensitivity for the detection of lung nodules and the identification of calcifications in previous studies using computed radiography with phosphor plates or fan-beam digital detectors [7 9]. However, despite the proof of higher sensitivity for the detection of pulmonary nodules, those preliminary systems were hampered by poor subtraction (tissue contrast cancellation), workflow inconveniences, and detective quantum efficiency limitations (i.e., inefficient usage of the X-ray dose). Thus, dual-energy imaging based on computed radiography has never found its way into the clinical routine [8, 9]. AJR:181, December

2 Fischbach et al. Dual-energy imaging with flat-panel digital detectors has the potential to be superior to previous methods. The flat-panel detectors have shown reduced noise and the potential for improved image quality [3, 10, 11]. Low noise may be a significant advantage because noise is increased in the dual-energy subtraction process. The fast image readout of this detector enables the use of the dual-exposure technique with individually optimized kilovoltages for the two exposures. Thus, dual-energy imaging can be performed with a wide energy separation between the two images as the critical point for effective tissue cancellation [12, 13]. The aim of this study was to assess the value of dual-energy subtracted bone images generated by the new flat-panel detector system for the detection of calcifications in chest radiographs. Materials and Methods Of 20 CT examinations randomly selected from our clinical routine program, 10 depicted various calcifications of the chest. All patients were older than 50 years. Patients with other confounding diseases and opacities (such as pulmonary fibrosis and consolidation) were excluded. The patients were asked to voluntarily undergo posteroanterior digital chest radiography including a dual-exposure dualenergy examination. Written informed consent was obtained from all patients. The study was approved by the local ethics committee. Chest radiographs were acquired using a flatpanel digital chest system (XQ/i Revolution, General Electric Medical Systems) within 2 weeks after the CT images were obtained. The chest system includes a cesium iodide (CsI) scintillator and an amorphous silicon photodiode transistor array. The detector has an image size of cm and a pixel dimension of mm [14]. The standard digital radiography examination consisted of the posteroanterior exposure. The dual energy examination consisted of the standard digital posteroanterior radiograph as well as the soft-tissue image and the bone image. Dual-energy images were acquired in a dual-exposure technique with 200 msec between the high- and low-energy exposures. The imaging parameters included a 120 kv image at a speed equivalent of approximately 400, and a 60 kv image at a speed equivalent of approximately 1,000. In addition to subtraction techniques generating the isolated soft-tissue image and the isolated bone image, postprocessing algorithms with pixel shifting and noise reduction were used [12]. CT examinations were performed using a multidetector CT (Somatom Plus 4 Volume Zoom, Siemens, Erlangen, Germany) at 140 kv, 120 mas; collimation, mm; table feed, 15 mm/sec; increment, 5 mm; window/center, 1,600/ 600. The entire lung was scanned in single breathhold mode in a caudocranial direction using 80 ml of IV iodine contrast media (iopromide, Ultravist 370, Schering, Berlin, Germany), flow rate of 2 ml/sec, and scanning delay of 60 sec for a better differentiation of lymph nodes in the mediastinum. The CT examination served as the gold standard to determine the exact sizes and locations of the calcified lesions. Two experienced radiologists reviewed the CT images using consensus to determine truth. Calcifications were determined by comparing the lesion with nearby ribs on the tissue window images. A total of 37 chest locations of pathologic calcifications were visible in CT images. Specifically, 16 calcified pulmonary nodules measured 0.6 cm on average (range, cm; median, 0.4 cm). In addition, 17 mediastinal calcified lymph nodes were found with an average diameter of 0.9 cm (range, cm; median, 1 cm), and four pleural calcifications (0.5, 1, 1, and 4 cm) were identified. Chest radiographs were reviewed by four radiologists who had not been members of the group for CT review but were similarly experienced in chest radiography. CT findings were not known to the radiography reviewers. The chest radiographs were presented as standard posteroanterior images alone or as standard posteroanterior radiographs in conjunction with dual energy images. All images were printed to film (Drystar 3000, Agfa, Köln, Germany) and reviewed on a viewbox under ambient room light conditions. Review was undertaken in a blinded manner by each radiologist independently. Standard and dual-energy sets were viewed in two sessions separated by 1 week. Images were presented in random order. Two of the radiologists had gained extensive experience with the dual-energy technique during a clinical development period; the other two gained basic experience with the technique by reviewing 20 dual-energy examinations before the study. Reviewers were given a schematic diagram of the lungs. They were asked to document the exact location and size of each calcified lesion they had detected on the chest radiographs and the dual-energy images, and to identify the corresponding abnormalities on CT scan. Other noncalcified lesions were ignored. For data analysis, each location of a calcification indicated on the schematic diagram was correlated individually with the corresponding CT images by the two independent graders of the CT review in consensus. These two radiologists also determined 28 locations on the CT scans to serve as the negative control group. Twelve of the locations were defined prospectively by determining chest regions in the medial and lateral upper, middle, and lower lung fields, and 16 locations were defined retrospectively by taking individual chest regions that had been identified mistakenly as false-positive findings by at least one of the four observers during the review of chest radiographs and dual energy images. For statistical analysis, we compared the results of the review of chest radiographs alone versus chest radiographs in conjunction with dual energy images, using the Wilcoxon s test and a nonparametric measure described by Brunner and Langer [14]. For Brunner and Langer s test, we introduced a quality factor, which was the ratio of the true-positive plus true-negative results divided by the maximum of all true results. This procedure basically counts all correct decisions in identifying calcified lesions and rejecting controls. The quality factor for standard posteroanterior chest radiography was compared with that of the standard posteroanterior radiograph with the addition of dual-energy images for each observer, applying Brunner and Langer s test of significance [14]. The confidence interval was 95%, thus the critical p value was 5% (< 0.05). To determine the relative dose applied by the dual-energy system in comparison with standard posteroanterior radiography, a phantom entrance dose was measured using an anthropomorphic Rando phantom (Alderson Research Laboratories, Stamford, CT) and a commercial 2026c radiation meter (Radcal, Monrovia, CA). Results Image Review For review of standard posteroanterior chest radiographs by all reviewers, the average sensitivity was 36%, the average positive predictive value (PPV) was 64%, the average specificity was 73%, and the average negative predictive value (NPV) was 47%. Further details are shown in Table 1. For review of standard posteroanterior chest radiographs in conjunction with dualenergy images, the sensitivity was 66%, the PPV was 76%, the specificity was 73%, and the NPV was 62% (Table 1). The Wilcoxon s test showed a significant improvement in the sensitivity of the detection of calcified lesions when dual-energy images were added for review (p < 0.05). The average value for the quality factor was 0.52 for the chest radiography review. Adding dual-energy images increased the value to Applying the methodology of Brunner and Langer, a p value of less than 0.01 was calculated for the difference between the two methods, showing significant improvement for the dual-energy system. No significant differences were determined among the four observers. Dose Measurements The phantom entrance dose for standard chest radiographs was 110 µgy for the posteroanterior shot and 680 µgy for the lateral shot using an automatic exposure control at a speed equivalent of 400. Applying dual energy, the entrance dose for the high-kilovolt image at a speed equivalent of 400 was 110 µgy. The low-kilovoltage shot at a speed equivalent of 1,000 used 110 µgy also. For the dual-energy examination the high-kilo AJR:181, December 2003

3 Dual-Energy Radiography of Chest Abnormalities TABLE 1 Detection of Calcified Lesions with Standard Digital Chest Radiography (CXR) Alone and with Dual-Energy Imaging (DE) for All Observers (A D) Compared with CT, the Gold Standard Observer True-Positive False-Positive False-Negative True-Negative Sensitivity Specificity PPV NPV Quality Factor a CXR A B C D Mean CXR + DE A B C D Mean Note. PPV = positive predictive value, NPV = negative predictive value. a Sum of the true-positives and true-negatives divided by the maximum of all true results. voltage image was used as the standard chest image so that the only additional image was the low-kilovoltage image. Hence, the total dose for the chest examination including posteroanterior and lateral shots increased 14% by applying the dual-energy mode versus the standard chest radiography examination. Discussion Discovery of a pulmonary nodule on conventional radiography of the chest is a common finding, and the presence of calcification within a nodule is the single most reliable evidence that the nodule is benign [1, 2]. Also, rib abnormalities (such as bone islands) can mimic lung nodules, and calcified pleural plaques can be mistaken for pulmonary consolidation. Radiologists are often required to make an initial assessment of the presence or absence of calcium on the basis solely of the radiographic findings. If one could reliably and accurately define the presence of calcification, additional workup might be avoided. In this study, we sought to determine the benefit of dual-exposure dual-energy images generated by a CsI detector to determine the presence or absence of calcifications in chest radiographs. CT findings served as the gold standard. The results (Fig. 1) indicate that when dualenergy images are added to standard chest radiographs, sensitivity, PPN, NPV, and level of confidence in a correct diagnosis improved significantly although it has to be considered that because the four radiologists could not be blinded from knowing which radiography method was used, they may have had a bias to try harder with the dual energy images that resulted in improved accuracy for this method. We did not perform a subclassification by lesion size, but large calcifications can be overlooked with the standard technique. The bone selective image reveals those lesions with higher confidence, as illustrated in Figures 2 and 3. However, small lesions detected easily in the dual energy bone image must be interpreted with caution because of possible misinterpretation of artifacts or cross-sections of nonsubtracted pulmonary vessels. Quality Factor CXR The significant improvement of nodule detection by the addition of dual-energy images is supported by the results of previous studies that compared dual-energy methods with computed radiography or film-screen systems. Kelcz et al. [7] evaluated single-exposure phosphor plate dual-energy radiography and found a significant improvement in the characterization of pulmonary nodules for example, whether the nodules carried calcifications. Niklason et al. [8] showed that dualenergy digital chest radiographs obtained from a fan-beam system visualized calcified and noncalcified simulated pulmonary nod- CXR + DE Fig. 1. Graph shows results of Brunner and Langer s test [14] comparing accuracy of standard digital chest radiography (CXR) alone and in combination with dual-energy (CXR + DE) review. Values between 0 and 1 are calculated for each individual reviewer (A D). Improvement contributed by adding DE images was significant (p < 0.01). Slope of line connecting reviewers scores inclines upwards, reflecting increase in accuracy of detection. AJR:181, December

4 Fischbach et al. A B Fig year-old woman with medical history of small cell bronchial carcinoma. A and B, Standard chest radiograph (A) is less effective than dual-energy bone image (B) in clearly showing calcifications in pulmonary nodule of upper left field and multiple mediastinal calcified lymph nodes (arrows, B). Note motion artifacts visible as dense line along left heart border and outline of pacemaker cables. C, Dual-energy standard posteroanterior soft-tissue image has potential to show lesions obscured by ribs. (Fig. 2 continues on next page) C ules significantly better than conventional film-screen radiographs did. To our knowledge, no clinical results of a CsI flat-panel detector based dual-energy system have been reported in the literature. The intrinsic superiority of this detector over film-screen or computed radiography benefits the image quality of dual-energy techniques for several reasons. First, the image quality for flatpanel digital radiography is better than that for standard chest imaging [4, 5, 10, 11], in spite of 1522 the reduced radiation dose. Next, the wider energy separation possible through the dual-exposure technique is advantageous [12, 13, 15 17]. Finally, the CsI detector based application used in this study comprises a dual-exposure technique enhanced by postprocessing algorithms such as pixel shifting and noise reduction [17]. Reliable assessment of chest calcifications in pulmonary nodules, pleura, and mediastinum is not a trivial task. Previous authors have shown that sensitivity using a standard chest radiograph for the detection of calcium may be as low as 50%, with a specificity of 78% [17]. The increase in confidence levels for the detection of calcifications, mediastinal calcified lymph nodes, and pleural calcified plaques shows that they are significantly better visualized on dual-energy images. A positive clinical impact may also be possible in the diagnosis of granulomatous infection or exposure to environmental toxins such as asbestos. AJR:181, December 2003

5 Dual-Energy Radiography of Chest Abnormalities D E Fig. 2. (continued) 75-year-old woman with medical history of small cell bronchial carcinoma. D and E, CT scans confirm diagnosis of calcifications (arrows) in mediastinum and upper left lung field shown in B. A B Fig year-old woman with pulmonary calcifications. A and B, Standard chest radiograph (A) is less effective than dual-energy bone image (B) in clearly showing pleural calcification and calcified nodule (arrows, B) in upper right lung field. Note motion artifacts along aortic arc. C, Dual-energy standard posteroanterior soft-tissue image has potential to show lesions obscured by ribs. (Fig. 3 continues on next page) C AJR:181, December

6 Fischbach et al. In addition to the evaluation of the imaging capabilities of the dual-energy system, a skin entrance dose was estimated using a chest phantom. The total dose of the chest examination, including posteroanterior and lateral images, was increased by 14% when the dual-energy mode was used. Previous studies have indicated that CsI flat-panel detectors provide a dose reduction of up to 50% compared with computed radiography [9 11]; thus, a dual-energy examination using a flat-panel detector may require significantly less radiation dose than a standard chest examination using a computed radiography detector. Overall, dual-energy chest radiography is inferior to CT in its detection rate of calcified chest abnormalities, so it is not a substitute for CT [18 21]. However, dual-energy improves the capabilities of chest radiography when it is applied as an inexpensive detection method because it facilitates the characterization of pulmonary, mediastinal, and pleural pathology. This study presents evidence that detector-based dual-energy imaging enhances the diagnostic accuracy of chest radiography. References Fig. 3. (continued) 73-year-old woman with pulmonary calcifications. D and E, CT scans confirm diagnosis of pleural and pulmonary calcifications (arrows) shown in B. 1. Fraser RG, Pare VP, Pare PD, Fraser RS, Genereux GP, eds. Diagnosis of diseases of the chest, 3rd ed. Philadelphia: Saunders, 1988: Webb WR. The solitary pulmonary nodule. In: Freundlich IM, Bragg DG, eds. A radiological approach to diseases of the chest, 2nd ed. Baltimore: Williams & Wilkins, 1997: D 3. Aufrichtig R. Comparison of low contrast detectability between a digital amorphous silicon and a screenfilm based imaging system for thoracic radiography. Med Phys 1999;26: Fischbach F, Ricke J, Freund T, et al. Flat panel digital radiography compared to storage phosphor computed radiography: assessment of dose versus image quality in phantom studies. Invest Radiol 2002;37: Fink C, Hallscheidt PJ, Noeldge G, et al. Clinical comparative study with a large-area amorphous silicon flat-panel detector: image quality and visibility of anatomic structures on chest radiography. AJR 2002;178: Brody WR, Butt G, Hall A, Macovski A. A method for selective tissue and bone visualization using dualenergy scanned projection radiography. Med Phys 1981;8: Kelcz F, Zink FE, Peppler WW, Kruger DG, Ergun DL, Mistretta CA. Conventional chest radiography vs dual-energy computed radiography in the detection and characterization of pulmonary nodules. AJR 1994;162: Niklason LT, Hickey NM, Chakraborty DP, et al. Simulated pulmonary nodules: detection with dualenergy digital versus conventional radiography. Radiology 1986;160: Oestmann JW, Greene R, Rhea JT, et al. Single exposure dual-energy digital radiography in the detection of pulmonary nodules and calcifications. Invest Radiol 1989;24: Strotzer M, Gmeinwieser JK, Volk M, Frund R, Seitz J, Feuerbach S. Detection of simulated chest lesions with normal and reduced radiation dose: comparison of conventional screen-film radiography and a flatpanel X-ray detector based on amorphous silicon. Invest Radiol 1998;33: Spahn M, Strotzer M, Volk M, et al. Digital radiography with a large-area, amorphous-silicon, flatpanel X-ray detector system. Invest Radiol 2000; 35: Sabol JM, Avinash GB, Nicolas F, Claus B, Zhao J, Dobbins JT III. Development and characterization of a dual-energy subtraction imaging system for chest radiography based on CsI:Tl amorphous silicon flat-panel technology. In: Antonuk LE, Yaffe MJ, eds. Medical imaging 2002: physics of medical imaging proceedings, vol Bellingham, WA: Society of Photo- Optical Instrumentation Engineers, 2001: Shaw CG, Gur D. Comparison of three different schemes for dual-energy subtraction imaging in digital radiography: a signal-to-noise analysis. In: Shaw R, ed. Medical imaging VI: instrumentation proceedings, vol Bellingham, WA: Society of Photo- Optical Instrumentation Engineers, 1992: Brunner E, Langer F. Nonparametrical analysis of longitudinal data. Munich: Oldenbourg, 1999: Granfors PR, Aufrichtig R. Performance of a cm 2 amorphous silicon flat panel X-ray detector for radiographic imaging applications. Med Phys 2000;27: Ho JT, Kruger RA, Sorenson JA. Comparison of dual and single exposure techniques in dual-energy chest radiography. Med Phys 1989;16: Avinash GB, Jabri KN, Uppaluri R, et al. Effective dose reduction in dual-energy flat panel X-ray imaging: technique and clinical evaluation. In: Sonka M, Fitzpatrick JM, eds. Medical imaging 2002 proceedings, vol Bellingham, WA: Society of Photo- Optical Instrumentation Engineers, 2002: Berger WG, Erly WK, Krupinski EA, Standen JR, Stern RG. The solitary pulmonary nodule on chest radiography: can we really tell if the nodule is calcified? AJR 2001;176: Kundel HL. Predictive value and threshold detectability of lung tumors. Radiology 1981;139: Diederich S, Lenzen H, Windmann R, et al. Pulmonary nodules: experimental and clinical studies at low-dose CT. Radiology 1999;213: Henschke CI, McCauley DI, Yankelevitz DF, et al. Early Lung Cancer Action Project: overall design and findings from baseline screening. Lancet 1999;354: E 1524 AJR:181, December 2003