Xin Bai1, Gumuyang Zhang1, Lili Xu1, Xiaoxiao Zhang1, Xuebin Zhang2, Zhengyu Jin3, Hao Sun4. 1. Department of Radiology, Peking Union Medical College Hospital, Peking Union Medical College, Chinese Academy of Medical Sciences, Shuaifuyuan No.1, Wangfujing Street, Dongcheng District, Beijing, 100730, China. 2. Department of Urology Surgery, Peking Union Medical College Hospital, Peking Union Medical College, Chinese Academy of Medical Sciences, Shuaifuyuan No.1, Wangfujing Street, Dongcheng District, Beijing, 100730, China. 3. Department of Radiology, Peking Union Medical College Hospital, Peking Union Medical College, Chinese Academy of Medical Sciences, Shuaifuyuan No.1, Wangfujing Street, Dongcheng District, Beijing, 100730, China. jinzy@pumch.cn. 4. Department of Radiology, Peking Union Medical College Hospital, Peking Union Medical College, Chinese Academy of Medical Sciences, Shuaifuyuan No.1, Wangfujing Street, Dongcheng District, Beijing, 100730, China. sunhao_robert@126.com.
Abstract
PURPOSE: To predict the diagnosis of pelvic lipomatosis (PL) by assessing CT imaging parameters. METHODS: Fifty clinically proven PL patients and 50 controls were included. Two radiologists recorded the types of bladder shapes and measured the following CT imaging parameters: (1) pelvic fat volume (PFV); (2) a set of morphological indices of bladder and rectosigmoid, including the ratio of craniocaudal to anterior-posterior length of bladder (CC/AP), angle between anterior and posterior wall (AAP), relative length of posterior urethra (rLPU), right and left angles between bladder and seminal vesicle (RABS and LABS), distance from verumontanum to anterior wall of rectum (DVR) and rectosigmoid morphological index (RMI). Secondary complications were also noted. RESULTS: The results were evaluated by an unpaired t test, ROC analysis and logistic regression. All CT imaging parameters except RMI (p = 1.000) indicated a statistical difference between PL and controls (p < 0.01). The AUCs of these parameters were between 0.724 (for rLPU) and 0.993 (for AAP). The sensitivity and specificity were 94% and 98% for AAP and 92% and 94% for the combined model, respectively. 23 patients were pathologically diagnosed with cystitis glandular. The incidence of hydronephrosis in the PL group was higher than that in the control group (p < 0.01). CONCLUSION: CT imaging parameters can predict the diagnosis of PL. The combined model of objective parameters including PFV, rLPU and DVR showed the best diagnostic performance for PL.
PURPOSE: To predict the diagnosis of pelvic lipomatosis (PL) by assessing CT imaging parameters. METHODS: Fifty clinically proven PL patients and 50 controls were included. Two radiologists recorded the types of bladder shapes and measured the following CT imaging parameters: (1) pelvic fat volume (PFV); (2) a set of morphological indices of bladder and rectosigmoid, including the ratio of craniocaudal to anterior-posterior length of bladder (CC/AP), angle between anterior and posterior wall (AAP), relative length of posterior urethra (rLPU), right and left angles between bladder and seminal vesicle (RABS and LABS), distance from verumontanum to anterior wall of rectum (DVR) and rectosigmoid morphological index (RMI). Secondary complications were also noted. RESULTS: The results were evaluated by an unpaired t test, ROC analysis and logistic regression. All CT imaging parameters except RMI (p = 1.000) indicated a statistical difference between PL and controls (p < 0.01). The AUCs of these parameters were between 0.724 (for rLPU) and 0.993 (for AAP). The sensitivity and specificity were 94% and 98% for AAP and 92% and 94% for the combined model, respectively. 23 patients were pathologically diagnosed with cystitis glandular. The incidence of hydronephrosis in the PL group was higher than that in the control group (p < 0.01). CONCLUSION: CT imaging parameters can predict the diagnosis of PL. The combined model of objective parameters including PFV, rLPU and DVR showed the best diagnostic performance for PL.