Angiomyolipoma with Minimal Fat: Differentiation from Renal Cell Carcinoma at Helical CT
Zhao Xiao-jun,Pu Jin-xian,Ping Ji-gen,Zang Jin,Lu Yong,Xi Qi-lin,Hou Wen-jie
DOI: https://doi.org/10.3760/cma.j.issn.0366-6999.20120672
IF: 6.133
2013-01-01
Chinese Medical Journal
Abstract:Renal cell carcinoma (RCC) accounts for 85%&90% of all kidney tumors. Although the majority of renal tumors in surgical series are malignant, a number of patients do undergo nephrectomy for benign lesions.1 Such minimal fat angiomyolipomas (AML) are visually indistinguishable from RCC and present particular diagnostic difficulties.2,3 In this study, CT number in region of whole mass (ROW) is advocated. METHODS Patients A retrospective study was conducted; we classified 117 patients into two groups, 96 were RCC and 21 were AML by pathologic diagnosis through post-operative examination of specimen. All patients who had undergone a dedicated renal CT examination in our hospital over the study period (from March 2006 to March 2010) were selected. CT examination and image analysis A standard protocol was used in all cases. All patients received 500A1000 ml oral contrast material 30 minutes before CT. All patients underwent biphasic CT, including unenhanced, corticomedullary phase (CMP) and excretory phase (EP) scanning. CT of the entire kidney was performed in every phase during approximately 20P30 seconds of the patient holding breath. Scanning for the CMP and EP was started 30 and 300 seconds after contrast injection, respectively. An experienced radiologist reviewed CT images at a special workstation. The CT number of the lesion in a particular ROI was recorded in unenhanced CMP and EP. The ROI and ROW were placed on the same selected slice. CT number in ROW includes the mean, maximum, minimum and the range between the maximum and minimum. Normal renal parenchyma and perinephric/sinus fat were scrupulously avoided (Figure 1).Figure 1.: The CT number of ROI is -1 (A), minimum CT number in ROW is -50 (B). Pathologic diagnosis is AML.Statistical analysis Results were tested for normality of distribution before statistical comparison. A commercially available statistics program (SPSS 13.0, SPSS Inc., USA) was used and P <0.05 was taken to indicate a significant difference greater than expected by chance. Area under the the receiver operating characteristics (ROC) curves, the sensitivity, and specificity of these indexes were compared for statistical significance using the same statistics program. RESULTS The age range was 19186 years in 96 RCC, mean age was 58.4 year, 43% were female. The age range was 24-69 years in 15 AML, mean age was 51.8 year, and 53% were female. CT number in ROI had difference between groups RCC and AML only in unenhanced (P<0.05), and was 10.80±12.10 in AML and 31.42±11.58 in RCC (Table 1). Enhancement of CT number in ROI had no difference between two groups (Table 2). CT number of six indexes in ROW had difference between two groups (P <0.05) (Table 3).Table 1: CT number in ROITable 2: Enhancement of CT numberTable 3: CT number in ROWArea under the curve of CT number in ROI in unenhanced was 0.633. Every area under ROC curve of 6 indexes in ROW was beyond 0.633. Top 3 of these indexes in descending order was minimum in unenhanced (0.849), minimum in CMP (0.793), and minimum in EP (0.712). When the most optimal diagnostic threshold for AML using CT number in ROI in unenhanced was ≤19.5 Hu, sensitivity was 50% and specificity was 71.4%; the minimum in ROW in unenhanced scans was ≤-32 Hu, its sensitivity was 76.7% and specificity was 98.7%; minimum in ROW in CMP was ≤-22.5 Hu, its sensitivity was 54.5%, specificity was 98%; minimum in ROW in EP was ≤-7.5 Hu, its sensitivity was 73.3%, specificity was 83.8% (Table 4).Table 4: Sensitivity and specificity for the diagnosis of AML using diagnostic thresholdsDISCUSSION Imaging plays a central role in renal mass categorization and suspicious renal masses are best evaluated using CT. The degree of enhancement is a consistent parameter for differentiating RCC and AML; RCC are distinguished by a positive pre-contrast density (usually 30 Hu) with a post-contrast incremental rise 20 Hu.4 In this study, enhancement of CT number had no difference between the two groups, on histopathological examination these angiomyolipomas were composed mostly of smooth muscle and blood vessels. Management of the incidentally found renal tumors, especially when their size is small, has been problematic. The management of renal tumor has been different from that of tumors in most other organs. Instead of attempting to make a pre-operative diagnosis by imaging or biopsy, the standard of care in patients who are found to have a solid non fat-containing renal mass has been straight surgical resection. The main reason for this management approach is that a vast majority of solid renal tumors are malignant. A number of patients, however, do have benign renal masses, and surgical resection of an entire or part of a kidney in these patients implies unnecessary surgery and potential complications, including long-term renal failure. In addition, because earlier detection and modern surgical techniques have been associated with longer patient survival, the long-term renal health of patients treated for renal masses has become a greater concern. Using CT, fat is identified with a ROI and in routine practice its presence is sufficient to confirm an AML, but some angiomyolipomas contain only tiny amounts of fat, and could, thus, be diagnosed as RCC, because fat can not be found in a small ROI circle. The presence of fat can only be found in whole region of mass. When imaging results do not show typical patterns of RCC or angiomyolipoma, careful sampling must be performed with unenhanced thin-section CT to identify the minimal fat components within the mass. Our study evaluated the diagnostic performance of CT number in ROW for differentiating AML without visible fat from RCC. Analysis of data showed that AML group counts a significantly greater number than RCC group in unenhanced with CT number in ROI less than 19.5 Hu, the area under ROC curve was 0.633, this corresponded with a specificity of 89.7% but sensitivity of only 50%. In comparison of the chosen threshold of every index in ROW, AML group counts a significantly greater number than RCC group with six indexes. Every area under ROC curve beyond 0.633. Top 3 of these indexes in descending order was minimum in unenhanced (0.849), minimum in CMP (0.793) and minimum in EP (0.712). Their sensitivity were 76.7%, 54.5%, and 73.3% and specificity were 98.7%, 98%, and 83.8%, respectively. Thus effectively improving the sensitivity and specificity in diagnosing minimal fat AML and RCC. CT number recorded in ROW is better able to resolve these small fat clusters and has a higher sensitivity and specificity. It may be a new way in the differential diagnosis between minimal fat AML and RCC. Minimum in ROW on unenhanced scans uu32 Hu may be accurate diagnostic threshold for AML.