Objective　To summarize the methods of diagnosis and operation of carcinoma of the extrahepatic bile ducts. Methods　The clinical data of 100 cases of carcinoma of the extrahepatic bile ducts which were treated in our hospital from 1972 to 1999 were retrospectively analysed. In this series, there were 68 cases of the cancer located in the upper portion of exlrahepatic duct(proximal cancer), 12 in midder portion(midder cancer), 18 in lower portion(distal cancer), and 2 in whole bile duct. Results　The initial symptom was upper abdominal discomfort or vague pain, abdominal distension, weakness, weight loss and progressive jaundice. BUS, CT and MRI were scatheless. If the intrahepatic bile duct dilatation or extrahepatic cholestatic jaundice were revealed, PTC(13 cases in this series) or ERCP(42 cases in this series) were to further determine the location of tumor. According to the position and type of the tumor, the different operations were selected. Twenty-five cases(36.8%) of the proximal cancer were resected, including 15 cases of type Ⅰ treated with localresection or “skeletonization” resection, 9 cases of type Ⅱ treated with resection of the tumor and caudate lobe, 1 case of type Ⅲb treated with resection of the tumor, caudate lobe and left hepatic trisegmentectomy. Nine cases(75%) of midder cancer were resected. After resected the proximal and midder cancer, bile duct reconstruction by Roux-en-Y hepaticojejunostomy was performed on all the cases. Fourteen cases(77.8%) of distal cancer were treated by pancreatoduodenectomy. The total resection rate in this series was 48%. Of the cancer resected cases, 35 were followed up, the five-year survival rate was 58%. 32 of the 52 cases without cancer resection were followed up, and all of them died one to one and half year after operation. Conclusions　BUS, CT and MRI are the first selective methods for early diagnosis of the carcinomas of extrahepatic bile duct. If needed, PTC or ERCP should be done because of these methods have more accurate diagnostic value. Surgical resection of the tumor is the only likelihood for effective treatment.

Objective : To explore the predictive value of oxygenation index ( P/F) for clinical outcomes in adult community acquired pneumonia (CAP) patients . Methods : 493 patients with CAP eligible for inclusion were extracted from AHCCDM , and were divided into survival group ( n = 405 ) and death group ( n = 88) according to whether they survived at discharge ; multivariate Logistic regression analysis was used to identify independent risk factors associated with hospitalization death in CAP patients ; the trend of changes in oxygenation index (P/F) and hospital mortality risk was evaluated using a restricted cubic spline ( RCS) curve . Results : Multivariate Logistic regression analysis found that independent risk factors affecting hospitalization death in CAP patients included acute and chronic health scores (Apache II score)( OR = 1 . 11 , 95% CI : 1 . 07 - 1 . 15 , P = 0. 02) , P/F ( OR = 1 . 70 ,95% CI : 1 . 03 - 2. 82 , P = 0. 04) , and hemoglobin (Hb) and HCO3- concentration P > 0. 05 , which were not statistically significant; restrictive cubic spline regression analysis found that there was a non⁃linear response relation ship between P/F and hospital mortality (non⁃linear test P < 0. 001) , and a P/F of 250 mmHg was the cutoff point for the hospital mortality association analysis . P/F < 250 mmHg , for every 1 mmHg increased in P/F , the risk of in⁃hospital mortality was reduced by 1% , HR : 0. 99 (0. 99 - 1 . 00) , P < 0. 001 . P/F > 250 mmHg , there was no association between CAP patients and in⁃hospital mortality , HR : 1 . 00 ( 1 . 00 - 1 . 01) , P < 0. 001 , indicating that other factors played a dominant role . Conclusion A significantly reduced P/F has a high predictive value for mortality in patients with community⁃acquired pneumonia .