Differentiation Of Osteolytic Bone Tumors: Quantitative Assessment With CT Perfusion

Objective: On the rat osteosarcoma model: To observe the relationship between the CT perfusion parameters, the maximal enhancement rate (MER) and the microvessel (MVD) of the tumors; To compare the difference of the CT perfusion parameters, MER and MVD between the rim and the center of the rat osteosarcoma. For the clinical CT examination of the patients with osteolytic bone tumors: To analysis the relationship between the CT perfusion parameters, MER and MVD of the tumors; To investigate the diagnostic value of MER in osteolytic bone tumors; To assess diagnostic value of rim-to-center differential blood flow ratio[R_rim-cenBF] and rim-to-center differential blood volume ratio[R_rim-cenBV] in differentiating benign from malignant osteolytic bone tumors.Part One: The CT perfusion study of the rat osteosarcomaMaterials and methods : Fifteen male and 15 female Wistar rats weighted 56.1 g⁶4.8g were used in this experiment. Rat osteosarcoma cell line named UMR106 was injected into the distal femoral or proximal tibia bone marrow cavity of the rats. One month later, except 4 accidentally died implanted rats and another 4 implanted rats without tumor growth, osteosarcoma was present in remained 22 implanted rats. Subsequently , The CT scans were performed on the grafted osteosarcoma model.CT examination: A SIEMENS multi-slice spiral CT scanner (SOMATOM Sensation10slice) was used in the present study. The scan series including: scout view, plain scan, enhanced dynamic scan and enhanced scan. The scan parameters of the dynamic perfusion scan are as follows: cine acquisition, 80kV, 100mA, 512X 512matrix, 9.6 X 9.6cm FOV, 2mm slice thickness(4slices per rotation), lsec per 360 revolution, total scan time 50sec. 200 images were acquired. The nonionic contrast agent (Ultravist 300) were injected through the tail vein by an injection pump with 1.5-2.0 ml dosage at rate of 0.5ml/sec. Data were transferred to a workstation by means of an optical disk or Digital Imaging and Communication on Medicine network transport. The data were evaluated at the workstation with PERFUSION and DYNEVA CT software (Siemens). DYNEVA software was used to generate time-density curves (TDC) in the regions of interest (ROI) which covered the whole lesion. The steepest slope of TDC was calculated to obtain MER; PERFUSION CT software allowed the calculation of a variety of perfusion maps and perfusion parameters such as BF and BV. Small ROIs were placed on the periphery and the center part of the lesion respectively to acquire corresponding BF, BV values.The rats were sacrificed after the CT examination, and the tumors of the rats were taken out for normal HE and immunohistochemistry staining, The SP method was used for the immunohistochemistry staining. The anti-CD34 monoclonal antibodies were used to mark microvessels. The MVD of the tumors were calculated using Weidner technique.Results: The CT scans demonstrated expansive or osteolytic bony destruction with periosteal reaction; The parenchyma of the tumors enhanced obviously on the enhanced images. Pathological changes included the erosion of bone tissue and epiphyseal plate, and destruction of cartilage by osteosarcoma cell. The karyokinesis was common and there were plenty of blood vessels, osteonecrosis, calcified granules and neogenetic osteoid tissue in the tumor.The contour of the tumors were well describes on the BF and BV mapping imaging. Quantitive study shows that the mean values of BF and BV were 100.09ml/min/100g and 10.81ml/100g which were much higher than those of normal muscle tissues which were 60.45 ml/min/lOOg, 1.89 ml/lOOg respectfully.There were positive linear correlation between the tumor MVD and tumor BF, BV and MER. The results demonstrated that the bigger the MVD of the tumor, the higher the BF, BV and MER. The difference between BF,BV and MVD in rim and center part of the tumors had statistically significance (P<0.05) .Part two: The application of CT perfusion in differentiating benign from malignant osteolytic bone tumorsMaterials and methods : During a period from August 2003 to December 2004,32 patients with a osteolytic bone tumor or tumorlike mass were examined with CT perfusion imaging at Tianjin hospital. Of the 32 patients, 18 were malignant (7 metastasis, 4 osteolytic osteosarcoma, 2 chondrosarcoma, 2 recrudescent grade-Ill giant-cell tumor, 1 fibrosarcoma, 1 myeloma and 1 fibrous histiocytoma) and 14 were benign (including 4 grade-I giant-cell tumor, 3 aneurysmal bone cyst, 2 nonossifying fibroma, fibrous dysplasia , chondromyxoid fibroma, desmoplastic fibroma , osteochondroma and eosinophilic granuloma in one patient respectively).The malignant masses had an average transection diameter of 4.52±1.75cm (range 2.53⁹.22cm) and that of the benign ones was 3.24±1.26 cm (range2.05 ⁴.98cm).CT examination: The scan series including: scout view, plain scan, enhanced dynamic scan and enhanced scan. A SIEMENS multi-slice spiral CT scanner (SOMATOM Sensation) was used in the present study. The scan parameters of the dynamic perfusion scan were as follows: cine acquisition, 80kV, 190mA, 512 matrix, 30³5cm FOV, 5mm slice thickness(4slices per rotation), lsec per 360 revolution, lsec interval, total scan time lOOsec. 200 images were acquired. Nonionic contrast agent (Ultravist 300) 1.5ml/kg was injected through the forearm vein by an injection pump at the rate of 6ml/sec and the injection rate was 1 ml/sec²ml/sec in children. The delay time was selected according to the lesion position, from 5sec¹5sec.Data were transferred to a workstation by means of an optical disk orDigital Imaging and Communication on Medicine network transport. The data were evaluated at the workstation with PERFUSION and DYNEVA CT software (Siemens). A large ROI covered the whole lesion was used with DYNEVA software to generate TDC. The steepest slope of TDC was calculated to obtain MER; PERFUSION CT software allowed calculation of a variety of perfusion maps and perfusion parameters such as BF and BV. Small ROIs were placed on the periphery and the center of the lesion to obtain the Rrjm_cenBF and Rrim-cenBV .Fourteen malignant andlO benign bone tumors selected from 32 histologically confirmed osteolytic bone tumors were underwent SP immunohistochemistry staining. The anti-CD34 monoclonal antibodies were used to mark the microvessels. The MVD of the tumors were calculated using Weidner technique.The differences between BF and BV in the rim and center of benign and malignant osteolytic bone tumors were analysed; The correlations between MER, BF, BV and MVD of the tumors were analysed; The MER, Rnm.cenBF and Rrim-cenBV of benign and malignant lesions were compared. If significant difference was existent between them, the receiver operating characteristic (ROC) was used to evaluate their values in distinguishing the benign from malignant masses and the cutoff values for differentiation were obtained accordingly.Results: Among the 18 malignant bone tumors, 4 osteolytic osteosarcoma, 7 metastasis, 2 recrudescent grade-Ill giant-cell tumors, one fibrosarcoma, one myeloma and one fibrous histiocytoma showed strong enhancement; one chondrosarcoma showed slight enhancement and one chondrosarcoma showed no enhancement. While among the 14 benign bone tumors, moderate or strong enhancement was seen in 4 grade-I giant-cell tumor, 3 aneurysmal bone cyst and one eosinophilic granuloma; slight enhancement in one nonossifying fibroma, one chondromyxoid fibroma and one osteochondroma; no enhancement in one nonossifying fibroma, one desmoplastic fibroma and one fibrous dysplasia. Ofthe 18 malignant bone tumors, 16 tumors demonstrate high perfusion with well-defined the contour on the BF and BV mapping imaging except 2 chondrosarcomas which showed low perfusion. Among 14 benign bone tumors, those that high perfusion was present in 8 cases which showed moderate or strong enhancement, whereas the rest with mild or non-enhancement manifested low perfusion.There was positive linear correlation between the tumor MVD and tumor BF, BV and MER. The results demonstrated that the bigger the MVD of the tumor, the higher the BF, BV and MER. MER can reflect the vascularity of tissue accurately. Because of extensive overlap of MER between benign and malignant bone tumors, it can't be used as a parameter to distinguish benign from malignant. The Rrim.CenBF and Rrim-cenBV of benign and malignant bone tumors differed significantlyC P < 0.01 )and the Rrim.cenBF and Rrim-cenBV of malignant bone tumors tend to be bigger than that of benign ones. The areas under ROC curve were 0.966 and 0.895(There would be a better diagnosis accuracy when the area was closer to 1 ), when a cutoff value was 0.11 for Rhm.cenBF (The sensitivity was 94.4and the specificity was 85.7%)and when a cutoff value was 0.08 for Rrjm.cenBV(The sensitivity was 88.9%and The specificity was 78.5%). CT perfusion not only demonstrates the blood supply of osteolytic bone tumors, but also correctly predicts its benignity and malignancy. Conclusions:1. This study has demonstrated that there was positive correlation between BF, BV and the MVD of the rat osteosarcoma and osteolytic bone tumors. CT perfusion imaging was feasible for evaluation of microvessel angiogenesis of bone tumors in vivo.2. There was positive correlation between MER and MVD of the osteolytic bone tumors, so the MER reflected the amount of vascularity of bone tumors accurately and estimated the degree of blood supply of them impersonally.