Accuracy of CT Attenuation Measurement for Differentiating Treated Osteoblastic Metastases From Enostoses

Transcription

1 Musculoskeletal Imaging Original Research Elangovan and Sebro CT Differentiation of Treated Osteoblastic Metastases From Enostoses Musculoskeletal Imaging Original Research Stacey M. Elangovan 1 Ronnie Sebro Elangovan SM, Sebro R Keywords: bone island, enostosis, metastasis, osteoblastic, sclerotic doi.org/1.2214/ajr Received June 17, 217; accepted after revision August 16, 217. Both authors contributed equally to this work. 1 Both authors: Department of Radiology, University of Pennsylvania, 34 Spruce St, Philadelphia, PA Address correspondence to R. Sebro (ronnie.sebro@uphs.upenn.edu). AJR 218; 21: X/18/ American Roentgen Ray Society Accuracy of CT Attenuation Measurement for Differentiating Treated Osteoblastic Metastases From Enostoses OBJECTIVE. The objective of our study was to assess whether the maximum and mean CT attenuations are accurate for differentiating between enostoses and treated sclerotic metastases. MATERIALS AND METHODS. We retrospectively reviewed CT studies of 165 patients (167 lesions) that included 49 patients with 49 benign lesions, 69 patients with 71 sclerotic treated lesions, and 47 patients with 47 untreated lesions, and calculated the mean and maximum CT attenuations of each lesion. ROC curves were used to identify thresholds for differentiating enostoses from treated sclerotic metastases and from untreated sclerotic metastases. RESULTS. The maximum CT attenuation of enostoses (1212. HU) was higher from that of untreated (754.7 HU) (p = ) and that of treated (891.7 HU) (p = ) sclerotic metastases. The maximum CT attenuation of treated sclerotic metastases (891.7 HU) was higher than that of untreated sclerotic metastases (754.7 HU) (p =.3). Enostoses had higher mean CT attenuation (1123. HU) than untreated (62. HU) (p < ) and treated (731.7 HU) (p = ) sclerotic metastases. A threshold mean CT attenuation of 885 HU had an accuracy of 91.7% and 81.7% to differentiate enostoses from untreated and treated metastases, respectively, whereas a threshold maximum CT attenuation of 16. HU had an accuracy of 81.3% and 72.5% to differentiate enostoses from untreated and treated metastases. CONCLUSION. The mean and maximum CT attenuations can differentiate between enostoses and sclerotic metastases; however, the accuracy of both metrics decreases after treatment. C T is increasingly used for the evaluation and staging of patients with malignancy [1]. Over the past 15 years, the number of CT studies performed in the United States has increased severalfold [2 4]. As a result, the number of incidentally discovered osseous lesions has increased with the increase in CT usage [5]. Newly discovered sclerotic lesions may represent enostoses or osteoblastic bone metastases, among other causes, in patients with malignancies [6]. It is important to identify a bone metastasis because this finding potentially changes disease stage, patient prognosis, and management [7, 8]. The most common malignancies presenting with osteoblastic osseous metastases include prostate cancer, breast cancer, and lung cancer [9, 1]. Indeterminate sclerotic lesions identified on CT may further be evaluated by 18 F-FDG PET/CT to assess for increased glycolysis and metabolic activity [11]. However, 18 F-FDG PET/CT is less accurate for the evaluation of prostate cancer [12]. Bone scintigraphy performed using 99m Tc meth- ylene diphosphonate (MDP) or 18 F-NAF PET/CT may be more accurate than 18 F-FDG PET/CT for differentiating enostoses from osteoblastic metastases, although enostoses may show increased 99m Tc-MDP uptake on bone scintigraphy [13 15]. These additional tests often occur several days after the initial CT study and result in increased radiation to the patient [13, 16]. More recently, Ulano et al. [17] proposed using CT attenuation (maximum and mean values) to differentiate enostoses from untreated osteoblastic metastases and showed that both methods had high accuracy. Ulano et al. found a threshold mean CT attenuation of 885 HU had an AUC of.982, and a threshold maximum CT attenuation of 16 HU had an AUC of.976 [17]. However, radiologists are often asked to interpret imaging studies with limited clinical or treatment history and sometimes treatment history is unknown, even by the referring physician. It is well known that there is an osteoblastic reparative response after chemotherapy [18] and treatment with bisphosphonate therapy AJR:21, March

2 Elangovan and Sebro [19] that results in increased sclerosis of a lesion and a higher CT attenuation [19]. With this knowledge, we hypothesized that the accuracy of CT attenuation measurements would decrease after a lesion had been treated. The aim of this study was to determine whether CT attenuation thresholds could be used to reliably differentiate enostoses from treated osteoblastic metastases and to assess how treatment changes affect the accuracy of the techniques proposed by Ulano et al. [17]. Materials and Methods This retrospective HIPAA-compliant study was approved by the local institutional review board, and the need for signed informed consent was waived. This study evaluated patients evaluated and treated at our institution between January 1, 26, and May 31, 217. Enostoses We retrospectively identified patients between 18 and 3 years old with enostoses by searching radiology text reports using Montage software (Montage Healthcare Solutions) to find CT studies of cases with bone islands or enostoses reported. Because most lesions were not biopsied, we considered a sclerotic lesion to be a bone island or enostosis if it remained unchanged in size or appearance over at least 2 years in a patient with no known history of malignancy. Patients were excluded if they had any history of malignancy. Untreated Osteoblastic Metastases Sclerotic lesions were identified by retrospective review of radiology text reports using Montage software to identify potential CT cases by searching for the following terms: sclerotic metastasis or sclerotic metastases ; sclerotic lesion or sclerotic lesions ; or osseous metastasis or osseous metastases ; or osteoblastic metastasis or osteoblastic metastases. The index CT study of a patient with no history of chemotherapy, antiandrogen therapy, or bisphosphonate therapy who had sclerotic lesions that subsequently grew in size or changed in appearance was used. These sclerotic lesions were considered untreated metastases in concordance with the clinical history from the medical record. We also used lesions that were biopsied and histologically confirmed as metastases. Treated Osteoblastic Metastases We identified patients with treated osteoblastic metastases using the same search terms as those used for untreated osteoblastic metastases, but we chose patients who had a history of systemic chemotherapy or antiandrogen therapy of at least 3 months. We also recorded whether patients used bisphosphonate therapy or not. All clinical histories were verified by a musculoskeletal radiologist and a radiology resident for confirmation. CT Examination and Acquisition Parameters Lesions were evaluated using one of the following MDCT units: a 192-MDCT unit (Somatom Force, Siemens Healthcare), a 64-MDCT unit with 128 channels (Somatom Definition Edge, Siemens Healthcare), a 128-MDCT unit (Somatom Definition Flash, Siemens Healthcare), a 16-MDCT unit (Sensation, Siemens Healthcare), a 64-MDCT unit (Discovery 75 HD, GE Healthcare), or a 256- MDCT unit (Revolution HD, GE Healthcare). Lesions were identified on CT urography (n = 1,.6%) or CT of the head (n = 3, 1.8%), neck and cervical spine (n = 7, 4.2%), face (n = 2, 1.2%), chest (n = 55, 32.9%), thoracic spine (n = 1,.6%), abdomen and pelvis (n = 97; 58.1%), or lumbar spine (n = 1,.6%). CT urography was performed with a slice thickness of 2 mm and table feed of 15 mm/s at a tube voltage of 12 kvp. CT scans of the head were obtained with a slice thickness of 5 mm, table feed of 15 mm/s, and tube voltage of 12 kvp. CT scans of the neck were obtained with a slice thickness of 3 mm and table feed of 15 mm/s at 1 kvp, and scanning of the cervical spine was performed with a slice thickness of 2.5 mm and table feed of 15 mm/s at 12 kvp. CT angiography of the neck was performed with a slice thickness of.625 mm, table feed of 15 mm/s, and 12 kvp. CT scans of the face were obtained with a slice thickness of 1 mm, table feed of 15 mm/s, and tube voltage of 12 kvp. CT scans of the chest were obtained with a slice thickness of 3 mm, table feed of 15 mm/s, and tube voltage of 12 kvp. CT of the thoracic spine was performed with a slice thickness of 2.5 mm and table feed of 15 mm/s at 12 kvp. CT scans of the abdomen and pelvis were obtained with a slice thickness of 3 mm, table feed of 15 mm/s, and tube voltage of 12 kvp. CT of the lumbar spine was performed with a slice thickness of 2 mm, table feed of 15 mm/s, and tube voltage of 12 kvp. One hundred thirty (77.8%) lesions were identified on contrast-enhanced CT studies and 37 (22.2%) lesions were identified on unenhanced CT studies. Patients who had contrast-enhanced examinations were injected IV with iopamidol (3 mg/ml [Isovue 37, Bracco Diagnostics]) approximately 45 9 seconds before acquisition of CT images. Patient age, sex, malignancy type (if there was a history of malignancy), lesion location, history of chemotherapy, and history of bisphosphonate use and receptor activator of nuclear factor kappa-b ligand (RANKL) inhibitor use were recorded at the time of the CT study. CT Attenuation Measurements Attenuation measurements were performed using a PACS workstation (GE Centricity, GE Healthcare). An ROI was placed around the lesion in question using the circle ROI tool, which generated the minimum, maximum, mean, and SD of the lesion s CT attenuation in Hounsfield units (Fig. 1). All lesions had to be at least 5 mm when measured along the maximum dimension. Measurements were maximized to fit as much of the Fig. 1 7-year-old woman with no history of cancer who presented with abdominal pain. Axial contrast-enhanced CT scan of abdomen and pelvis shows 1.7-cm indeterminate sclerotic lesion in left sacral ala. Circle shows placement of ROI over sclerotic lesion. Area of ROI measures 69.2 mm 2. Attenuation values of ROI are as follows: mean, HU; SD, HU; maximum, 1438 HU; and minimum, 929 HU. Both maximum and mean CT attenuation values suggest this lesion is enostosis. 616 AJR:21, March 218

3 CT Differentiation of Treated Osteoblastic Metastases From Enostoses lesion as possible. Measurements were repeated in a random sample of 25 subjects to calculate intrarater reliability and interrater reliability. Statistics A priori we hypothesized that a sample size of 135 subjects (46 in each group) would have 8% power to detect a 2-HU CT attenuation difference between treated or untreated osteoblastic lesions and enostoses, assuming a SD of 34 HU and a type I error rate, alpha value of.5. Summary statistics for demographic and clinical data were TABLE 1: Clinical and Demographic Characteristics of Three Groups Variable Enostoses (n = 49) Untreated Sclerotic Metastases (n = 47) calculated. The mean and SD CT attenuations for each patient s largest lesion were calculated. Intrarater and interrater reliabilities were assessed using the intraclass correlation coefficient (ICC). The mean lesion size and maximum and mean CT attenuations were compared between groups using twosample t tests with unequal variances. ROC curve analysis was performed to identify the best threshold to differentiate untreated metastases from treated metastases and enostoses. The DeLong test was used to compare ROC curves. We compared the sensitivity, specificity, positive predictive value, and negative predictive value of the thresholds provided by Ulano et al. [17] to differentiate enostoses from untreated and treated osteoblastic metastases for comparison. Statistics were calculated using R software (version 3.2, The R Foundation). All tests were two-sided, and p values <.5 were considered statistically significant. Results The study group was composed of 167 lesions (patient age range, years; 92 men [55.1%] and 75 women [44.9%]). There were Treated Sclerotic Metastases (n = 71) p a (95% CI) p b (95% CI) p c (95% CI) < Age (y), mean (SD) 27.6 (3.3) 66.7 (13.3) 62.7 (11.2) < ( 45.1 to 35.1) d ( 37.8 to 32.2) d (.6 to 8.8) d Male, no. of patients 23 (46.9) 3 (63.8) 39 (54.9).16 (.8 4.9) e.459 (.6 3.1) e.349 (.3 1.6) e Maximum size (mm), mean (SD) 7.4 (3.1) 18.4 (1.3) 17.1 (8.) ( 14.2 to 7.9) d ( 11.8 to 7.7) d.467 ( 2.2 to 4.8) d Maximum CT attenuation (HU), mean (SD) (247.6) (215.3) (272.8) ( ) d ( ) d.3 ( to 47.7) d Mean CT attenuation (HU), mean (SD) (225.9) 62. (172.2) (248.2) < ( ) e ( ) e.1 ( 26.3 to 53.) e Malignancy < < None 49 (1.) (.) (.) Prostate (.) 16 (34.) 24 (33.8) Breast (.) 9 (19.1) 18 (25.4) Lung (.) 8 (17.) 12 (16.9) Other (.) 14 (29.8) 17 (23.9) Treatment f 1. < < None 49 (1.) 47 (1.) (.) Chemotherapy (.) (.) 48 (67.6) Antiandrogen (.) (.) 18 (25.4) Antiestrogen (.) (.) 8 (11.3) Radioactive iodine (.) (.) 2 (2.8) Bisphosphonate or RANKL inhibitor therapy (.) (.) 22 (31.) Indications < Cancer screening (.) 29 (61.7) 63 (88.7) Pain 17 (34.7) 9 (19.1) 2 (2.8) Dyspnea 1 (2.) 1 (2.1) 2 (2.8) Renal calculus 6 (12.2) (.) (.) Crohn disease 14 (28.6) (.) (.) Trauma 5 (1.2) (.) (.) Other 6 (12.2) 8 (17.) 4 (5.6) Note RANKL = receptor activator of nuclear factor kappa-b ligand. a Comparison between enostoses and untreated osteoblastic metastases. b Comparison between enostoses and treated osteoblastic metastases. c Comparison between untreated osteoblastic metastases and treated osteoblastic metastases. d The 95% CI of the mean difference. e The 95% CI of the odds ratio. f Numbers will not add up to the total due to combination therapy for some patients. AJR:21, March

4 Elangovan and Sebro 49 (29.3%) lesions with no malignancy and enostosis and 118 (7.7%) lesions with a history of malignancy and osteoblastic osseous metastases. Of the 118 metastatic lesions, 71 (6.2%) were treated osteoblastic metastases. The most common malignancies were prostate (n = 4, 33.9%), breast (n = 27, 22.9%), and lung (n = 2, 16.9%). Clinical and demographic summary statistics are shown in Table 1. The intrarater reliability for the maximum CT attenuation was an ICC of 1. (p < ), and the interrater reliability for the maximum CT attenuation was an ICC of.95 (p = ) (Table 2). The mean CT attenuation had an ICC for intrarater reliability of 1. (p < ) and an ICC for interrater reliability of.95 (p = ). The maximum CT attenuation (p =.148) and the mean CT attenuation (p =.265) were not significantly different between lesions that underwent unenhanced CT. We found that the maximum CT attenuation was significantly different between enostoses (1212. HU) and untreated osteoblastic metastases (754.7 HU) (p = ; 95% CI, HU). Similarly, the maximum CT attenuation was significantly different between enostoses (1212. HU) and treated osteoblastic metastases (891.7 HU) (p = ; 95% CI, HU). Untreated osteoblastic metastases (754.7 HU) had a lower maximum CT attenuation Sensitivity HU TABLE 2: Interrater and Intrarater Reliabilities of CT Attenuation T hresholds CT Attenuation Threshold ICC p 95% CI Intrarater reliability Maximum CT attenuation 1. < Mean CT attenuation 1. < Interrater reliability Specificity Maximum CT attenuation Mean CT attenuation Note ICC = intraclass correlation coefficient. than treated osteoblastic metastases (891.7 HU) (p =.3; 95% CI, to 47.7 HU). We noted similar findings when evaluating the mean CT attenuation (Table 1). Lesions in patients treated with bisphosphonate therapy had, on average, higher maximum CT attenuation (p =.32; 95% CI, HU) and, on average, higher mean CT attenuation (p =.8; 95% CI, HU) than lesions in patients who did not receive bisphosphonate therapy. ROC curve analysis showed that both the maximum CT attenuation (AUC = 92.2%; 95% CI, %) and the mean CT attenuation (AUC = 96.4%; 95% CI, %) can be reliably used to differentiate enostoses from untreated osteoblastic metastases (Figs. 2 and 3). We also found that the maximum CT attenuation (AUC = 81.7%; 95% CI, %) and the mean CT attenuation Untreated osteoblastic metastases versus enostoses: maximum CT attenuation, AUC = 92.2% (95% CI, %) Treated osteoblastic metastases versus enostoses: maximum CT attenuation, AUC = 81.7% (95% CI, %) Sensitivity HU (AUC = 88.4%; 95% CI, %) can be reliably used to differentiate enostoses from treated osteoblastic metastases (Figs. 2 and 3). The DeLong test comparing the ROC curves differentiating enostoses from untreated osteoblastic metastases to ROC curves differentiating enostoses from treated osteoblastic metastases using the maximum CT attenuation gave a p value of.31 and using the mean CT attenuation gave a p value of.34. We found that a threshold maximum CT attenuation of 16 HU had a sensitivity of 73.5%, specificity of 89.4%, and accuracy of 81.3% for differentiating enostoses from untreated osteoblastic metastases; however, this threshold had a sensitivity of 73.5%, specificity of 71.8%, and accuracy of 72.5% for differentiating enostoses from treated osteoblastic metastases (Table 3). We noted similar findings when using the threshold mean Untreated osteoblastic metastases versus enostoses: mean CT attenuation, AUC = 96.4% (95% CI, %) Treated osteoblastic metastases versus enostoses: mean CT attenuation, AUC = 88.4% (95% CI, %) Specificity Fig. 2 ROC curve shows results when maximum CT attenuation of lesion is used to differentiate enostoses from treated and untreated sclerotic metastases. Threshold maximum CT attenuation of 91 HU had sensitivity of 93.9% and specificity of 78.7% for differentiating enostoses from untreated osteoblastic metastases. Diagonal line is where sensitivity = 1 - specificity. Fig. 3 ROC curve shows results when mean CT attenuation of lesion is used to differentiate enostoses from treated and untreated sclerotic metastases. Threshold mean CT attenuation of 864 HU had a sensitivity of 93.9% and specificity of 93.6% for differentiating enostoses from untreated osteoblastic metastases. Diagonal line is where sensitivity = 1 - specificity. 618 AJR:21, March 218

5 CT Differentiation of Treated Osteoblastic Metastases From Enostoses TABLE 3: Accuracy of Maximum CT Attenuation Threshold of 16 HU to Differentiate Enostoses From Treated and Untreated Osteoblastic Metastases Diagnosis Using Maximum CT Attenuation Threshold of 16 HU CT attenuation of 885 HU. A threshold mean CT attenuation of 885 HU had a sensitivity of 89.8%, specificity of 93.6%, and accuracy of 91.7% for differentiating enostoses from untreated osteoblastic metastases. However, this same threshold had a sensitivity of 89.8%, specificity of 76.1%, and accuracy of 81.7% for differentiating enostoses from treated osteoblastic metastases (Table 4). Sensitivity Specificity Positive Predictive Value Negative Predictive Value Accuracy Enostoses vs untreated osteoblastic metastases 73.5 (36/49) 89.4 (42/47) 87.8 (36/41) 76.4 (42/55) 81.3 (78/96) Enostoses vs treated osteoblastic metastases 73.5 (36/49) 71.8 (51/71) 64.3 (36/56) 79.7 (51/64) 72.5 (87/12) Note Values in parentheses are number of lesions / total number of lesions. Discussion The results of our data show that both the maximum and mean CT attenuations can be reliably used to differentiate enostoses from untreated and treated osteoblastic metastases; however, the accuracy of both methods decreases for differentiating enostoses from treated osteoblastic metastases. A threshold maximum CT attenuation of 91 HU had a sensitivity of 93.9% and specificity of 78.7% for differentiating enostoses from untreated osteoblastic metastases, and a threshold maximum CT attenuation of HU had a sensitivity of 71.4% and specificity of 83.1% for differentiating enostoses from treated osteoblastic metastases. A threshold mean CT attenuation of 864 HU had a sensitivity of 93.9% and specificity of 93.6% for differentiating enostoses from untreated osteoblastic metastases, and a threshold mean CT attenuation of 871 HU had a sensitivity of 93.9% and specificity of 76.1% for differentiating enostoses from treated osteoblastic metastases. Our study results closely mirror those of Ulano et al. [17] and show that both the maximum and mean CT attenuations can be used to differentiate between enostoses and untreated osteoblastic metastases. We found that the maximum and mean CT attenuations can also be used to differentiate between enostoses and untreated osteoblastic metastases. A threshold maximum CT attenuation of 16 HU had a sensitivity of 73.5% and specificity of 89.4% and a threshold mean CT attenuation of 885 HU had a sensitivity of 89.8% and specificity of 93.6% for differentiating between enostoses and untreated osteoblastic metastases, which are similar to the findings reported by Ulano and colleagues. We found that treated osteoblastic metastases had increased maximum and mean CT attenuations compared with untreated osteoblastic metastases, which agrees with reports by Stattaus et al. [18] and Quattrocchi et al. [19]. We also noted that bisphosphonate therapy also resulted in a significant increase in the maximum and mean CT attenuations compared with untreated osteoblastic metastases. Enostoses are benign lesions that comprise compact bone within the cancellous bone and are thought to be a skeletal dysplasia [2]. Bone islands and enostoses therefore mimic cortical bone signal intensity on both radiographs and MRI sequences. The osseous skeleton is a relatively frequent site of cancer metastasis, and prostate, breast, and lung cancers are the most common osteoblastic osseous metastases. These osteoblastic metastases may have an appearance similar to that of bone islands and enostoses on CT studies. There is a need to reliably differentiate bone islands and enostoses from osteoblastic metastases because the presence of osseous metastases increases cancer stage and often affects prognosis and treatment. Our results show that the previously described maximum and mean CT attenuation thresholds for differentiating enostoses from untreated osteoblastic metastases are less accurate if lesions have been treated either with systemic chemotherapy or with bisphosphonate therapy or RANKL inhibitor therapy. If treatment history is unknown, we suggest using our more conservative thresholds to result in a smaller chance of misclassifying a treated metastasis as a bone island or enostosis. This study has a few limitations. It was a retrospective study and therefore subject to ascertainment bias. The enostoses were not biopsied for histologic confirmation; however, all patients with enostoses had no history of malignancy, were between 18 and 3 years of age, and the lesions were unchanged in size or appearance for more than 2 years. On the basis of these criteria, we think it is reasonable to conclude that these lesions were enostoses. Untreated osteoblastic metastases had no history of treatment; however, these data were obtained from the electronic medical record, which is sometimes incomplete or is not entirely accurate. The duration of treatment was not constant for all treated individuals, and the chemotherapy used varied depending on the primary malignancy, which potentially may add to the variation of CT attenuations noted in treated lesions. In summary, we find that the maximum and mean CT attenuations are metrics that can be used to differentiate between enostoses and osteoblastic osseous metastases; however, the accuracy of these metrics decreases for treated metastases. Radiologists should use caution when using these metrics to evaluate sclerotic lesions in patients when limited clinical history is provided and should perhaps use the more conservative measures (i.e., the measures to differentiate treated osteoblastic metastases from enostoses) rather than those used to distinguish between untreated osteoblastic lesions and enostoses. TABLE 4: Accuracy of Mean CT Attenuation Threshold of 885 HU to Differentiate Enostoses From Treated and Untreated Osteoblastic Metastases Diagnosis Using Mean CT Attenuation Threshold of 885 HU Sensitivity Specificity Positive Predictive Value Negative Predictive Value Accuracy Enostoses vs untreated osteoblastic metastases 89.8 (44/49) 93.6 (44/47) 93.6 (44/47) 89.8 (44/49) 91.7 (88/96) Enostoses vs treated osteoblastic metastases 89.8 (44/49) 76.1 (54/71) 72.1 (44/61) 91.5 (54/59) 81.7 (98/12) Note Values in parentheses are number of lesions / total number of lesions. AJR:21, March

6 Elangovan and Sebro References 1. Litière S, Collette S, de Vries EG, Seymour L, Bogaerts J. RECIST: learning from the past to build the future. Nat Rev Clin Oncol 217; 14: Salmon JM, Kilpatrick SE. Pathology of skeletal metastases. Orthop Clin North Am 2; 31: , vii viii 8. Sobin LH, Hermanek P, Hutter RV. TNM classifi- 14. Bastawrous S, Bhargava P, Behnia F, Djang DS, Haseley DR. Newer PET application with an old tracer: role of 18 F-NaF skeletal PET/CT in oncologic practice. RadioGraphics 214; 34: cation of malignant tumors: a comparison be- 15. Greenspan A, Stadalnik RC. Bone island: scinti- 2. Smith-Bindman R, Lipson J, Marcus R, et al. Radiation dose associated with common computed tomography examinations and the associated lifetime attributable risk of cancer. Arch Intern Med 29; 169: Amis ES Jr, Butler PF, Applegate KE, et al. American College of Radiology white paper on radiation dose in medicine. J Am Coll Radiol 27; 4: Berrington de González A, Mahesh M, Kim KP, et al. Projected cancer risks from computed tomographic scans performed in the United States in 27. Arch Intern Med 29; 169: Berland LL, Silverman SG, Gore RM, et al. Managing incidental findings on abdominal CT: white paper of the ACR incidental findings committee. J Am Coll Radiol 21; 7: Greenspan A, Jundt G, Remagen W. Differential diagnosis in orthopaedic oncology. New York, NY: Lippincott Williams & Wilkins, 27:529 tween the new (1987) and the old editions. Cancer 1988; 61: Suva LJ, Washam C, Nicholas RW, Griffin RJ. Bone metastasis: mechanisms and therapeutic opportunities. Nat Rev Endocrinol 211; 7: Coleman RE. Metastatic bone disease: clinical features, pathophysiology and treatment strategies. Cancer Treat Rev 21; 27: Peterson JJ. F-18 FDG-PET for detection of osseous metastatic disease and staging, restaging, and monitoring response to therapy of musculoskeletal tumors. Semin Musculoskelet Radiol 27; 11: Bouchelouche K, Tagawa ST, Goldsmith SJ, Turkbey B, Capala J, Choyke P. PET/CT imaging and radioimmunotherapy of prostate cancer. Semin Nucl Med 211; 41: Langsteger W, Rezaee A, Pirich C, Beheshti M. 18 F-NaF-PET/CT and 99m Tc-MDP bone scintigraphy in the detection of bone metastases in prostate cancer. Semin Nucl Med 216; 46: graphic findings and their clinical application. Can Assoc Radiol J 1995; 46: Quinn B, Dauer Z, Pandit-Taskar N, Schoder H, Dauer LT. Radiation dosimetry of 18 F-FDG PET/CT: incorporating exam-specific parameters in dose estimates. BMC Med Imaging 216; 16: Ulano A, Bredella MA, Burke P, et al. Distinguishing untreated osteoblastic metastases from enostoses using CT attenuation measurements. AJR 216; 27: Stattaus J, Hahn S, Gauler T, et al. Osteoblastic response as a healing reaction to chemotherapy mimicking progressive disease in patients with small cell lung cancer. Eur Radiol 29; 19: Quattrocchi CC, Santini D, Dell aia P, et al. A prospective analysis of CT density measurements of bone metastases after treatment with zoledronic acid. Skeletal Radiol 27; 36: Greenspan A. Bone island (enostosis): current concept a review. Skeletal Radiol 1995; 24: AJR:21, March 218