Transhepatic intravascular ultrasound for evaluation of portal venous involvement in patients with cancer of the pancreatic head region.

The aim of this study was to evaluate the ability of intravascular ultrasound to diagnose tumor involvement of the portal and the superior mesenteric veins using the preoperative percutaneous, transhepatic approach, and to compare the findings with those made at concomitant direct portography, surgery, and histopathological examination. Ten patients with a preoperative diagnosis of a resectable tumor in the pancreatic head region were examined with percutaneous transhepatic portography (PTP) and intravascular ultrasound (IVUS). The surgeon's intraoperative evaluation and the histopathological examination in combination revealed tumor involvement of the portal or superior mesenteric veins in six of the ten patients. Percutaneous transhepatic portography suggested tumor involvement of the veins in six patients but two of the examinations were false positive and another two were false negative. Intravascular ultrasound showed signs of tumor involvement in eight patients. The examination was, however, false positive in two patients, but there were no false negatives. Complications of the percutaneous transhepatic procedure occurred in six patients including severe pain, bleeding, and related death. Percutaneous transhepatic IVUS of the portal vein may be a useful tool in the preoperative selection of the subgroup of patients with tumor of the pancreatic head region that could benefit from surgery. There is a need for technical improvement as well as studies with larger patient series to definitely decide the role of the technique.

Influence of bile duct stones on patient features and effect of endoscopic sphincterotomy on early outcome of edematous gallstone pancreatitis.

The objective of this study was to study the influence on patients' features and the effect on early outcome of the presence of bile duct stones and endoscopic sphincterotomy (EST), respectively, in the presence of edematous gallstone pancreatitis (GSP). It was a retrospective review of a patient series from 1981 through 1992 at a university teaching hospital. Altogether 96 patients aged 42 to 93 years (median 74 years) with edematous GSP were investigated by endoscopic retrograde cholangiography (ERC) in our department, 75 of whom underwent ERC at first admission. A total of 49 patients (group 1) had common bile duct (CBD) stones, and in 47 (group 2) no CBD stones were found. All patients in group 1 and 15 in group 2 underwent EST; 57 of the 75 patients had EST at first admission. The main outcome measures were pancreas-related complications and the length of the hospital stay. The early major complication and stone clearance rates of the EST procedure were 3.2% and 96%, respectively. Duration of symptoms prior to ERC was similar in groups 1 and 2 (median 5 and 8 days, respectively). Serum amylase activity was higher in group 2 patients than in group 1 patients [21-258 (median 75) microkat/L vs. 10-328 (median 48) microkat/L (p = 0.01)], but the length of hospital stay was similar: [4-39 (median 11) days vs. 4-19 (median 9) days (p = 0.05)]. Cholangitis at acute admission was more common in group 1 than in group 2 patients (31% vs. 7%; p = 0. 02), whereas a history of pancreatitis was noted more often in group 2 patients (49% vs. 8%; p< 0.001). ERC was done 1 to 18 days (median 2 days) and 1-16 days (median 5 days) (p = 0.02) after admission in groups 1 and 2 respectively, because of the more frequent cholangitis symptoms in group 1. It was concluded that the history and features at admission differed between patients with and without CBD stones at ERC done during an attack of GSP. Early EST had no influence on outcome or hospitalization. This study does not support routine EST in conjunction with mild GSP.

Effect of endoscopic sphincterotomy and interval cholecystectomy on late outcome after gallstone pancreatitis.

BACKGROUND: Endoscopic sphincterotomy alone, or followed by cholecystectomy, are options in patients with gallstone pancreatitis. METHODS: Ninety-six patients of median age 74 (range 30-93) years with gallstone pancreatitis had endoscopic retrograde cholangiography and were followed for a median of 84 (range 33-168) months. Forty-eight of 49 patients with, and nine of 47 without, common bile duct (CBD) stones had urgent endoscopic sphincterotomy. One patient with, and six without, CBD stones had delayed endoscopic sphincterotomy a median of 35 (range 12-111) days after acute pancreatitis. Thus, 64 patients had endoscopic sphincterotomy (group 1) and 32 did not (group 2). Fifteen and 16 patients in each group respectively had interval cholecystectomy after a median of 3 months and 1 month. RESULTS: Patients in groups 1 and 2 had similar rates of interval cholecystectomy (15 of 64 versus 16 of 32 patients respectively) or required cholecystectomy (15 of 49 versus five of 16 patients), recurrent CBD calculi (three of 64 versus three of 32 patients) or total length of hospitalization after interval cholecystectomy (median 15.5 and 15 days) or required (median 22 and 24 days) cholecystectomy. The overall incidence of recurrent pancreatitis was one of 64 patients in group 1 and five of 32 in group 2 (P = 0.02), but after interval cholecystectomy the recurrence rate of biliopancreatic symptoms was similar (one of 15 patients versus three of 16 patients respectively). CONCLUSION: Endoscopic sphincterotomy, but not interval cholecystectomy, reduced the overall incidence of recurrent pancreatitis, but not of late biliary complications. Some 31 per cent of the patients required cholecystectomy, suggesting that routine cholecystectomy should be considered in fit patients following acute pancreatitis.

Factors predictive of early complications of endoscopic treatment of bile duct calculi.

BACKGROUND/AIMS: Factors associated with an increased early complication rate of the endoscopic sphincterotomy procedure have been identified. Precut or needle knife papillotomy has been shown to improve the success rate of endoscopic retrograde cholangiography and endoscopic sphincterotomy, but has often been reported to be hazardous. In order to identify patients with bile duct stones at risk for a complicated course in connection with endoscopic clearance of the calculi, factors predictive of early complications were sought. METHODOLOGY: 417 consecutive patients with bile duct calculi at endoscopic retrograde cholangiography were considered for endoscopic treatment in our department from 1981 to 1992. Endoscopic sphincterotomy was performed in 246 patients with intact gallbladders and in 147 with prior cholecystectomy, 55 of whom had retained calculi. RESULTS: There was a 9.4% overall and 7.1% major early complication rate of the EST procedure and a 30-day mortality of 0.5% (2 patients, non-procedure related). In 22% (6/27) of the patients with major complications, surgery was required or preferred to additional endoscopic measures. Complete stone removal failed in 35/393 patients (8.9%). The immediate and early complication rate of standard sphincterotomy was not found to be increased in patients with prior or present biliopancreatic complications, failed bile duct clearance at first attempt, or juxtapapillary diverticula. It was the same after standard sphincterotomy as after precut papillotomy followed by immediate or delayed sphincterotomy. No increased morbidity was found after failed therapy as compared to failed diagnostic precut papillotomy. There was neither a greater need for, nor an increased complication rate following, precut papillotomy in patients with, as compared to those without, juxtapapillary diverticula. Endoscopic experience did not influence the complication rate. There were no significant differences regarding outcome or risk factors associated morbidity between patients with and without intact gallbladder. CONCLUSIONS: These findings confirm that endoscopic treatment is safe and that precut papillotomy can be performed without increased morbidity. Furthermore, none of the commonly identified factors associated with increased morbidity were found to be risk factors in this study.

Endoscopic drainage in benign pancreatic disease: immediate and medium term outcome.

OBJECTIVE: To elucidate further the role of endoscopy in the treatment of benign pancreatic disease. DESIGN: Retrospective study. SETTING: University hospital, Sweden. SUBJECTS: 136 of 319 patients who underwent endoscopic retrograde cholangiopancreatography (ERCP) for known or suspected pancreatic disease had abnormal findings at duodenoscopy or ductography, or both. In 28 patients endoscopic treatment was considered because of recurrent acute pancreatitis (n = 6), chronic pancreatitis (n = 5), pancreatic pain syndrome (n = 5), pancreatic fistula with ascites (n = 1), and pseudocyst (n = 1), or appreciable biliary obstruction from chronic pancreatitis (n = 10). INTERVENTIONS: Pancreatic duct drainage was attempted in 18 patients and successful in 13 (72%). Endoscopic sphincterotomy (EST) alone was done in 7/13 patients and an endoprosthesis (stent) was placed in 6/13. Bile duct drainage was attempted and successful in 10 patients by EST together with dilatation with (n = 6) or without (n = 4) simultaneous placement of a stent. MAIN OUTCOME MEASURES: Relief of pain and cholestasis. RESULTS: Immediate (1-30 days after initial treatment) and medium term (median 48 months after initial treatment) outcome after pancreatic duct drainage was excellent (no pain) or good (occasional mild pain) in 62% (8/13) and 67% (8/12) of the patients, respectively. Complications of the endoscopic procedure were encountered in four patients (31%) and comprised infection with abscess formation (n = 1), repeated stent clogging (n = 1) or stent migration (n = 2). Surgery was subsequently required in three patients (23%) because of intraabdominal abscess (n = 1), recurrent pain (n = 1), or no pain relief (n = 1). At medium term follow-up (median 68 months) after biliary drainage 7/10 patients had liver function tests within the reference ranges. Only two patients required subsequent biliodigestive shunts 7 and 13 months after EST, respectively. CONCLUSION: Our findings favour endoscopic drainage as a safe and effective method for temporary and medium term relief of pain and biliary obstruction in selected patients with benign pancreatic disease.

Intravascular ultrasound for evaluation of portal venous involvement in pancreatic cancer.

The aim of this study was to evaluate the accuracy of intravascular ultrasound (IVUS) in diagnosing tumour involvement of the portal vein in patients with exocrine cancer of the head of the pancreas. Seven consecutive patients with a preoperative diagnosis of carcinoma, preoperatively deemed to be resectable, were examined with IVUS of the portal vein during surgery. The IVUS catheters were 6.2 F (2.0 mm) in diameter with a 20-MHz transducer and were introduced into the portal vasculature through the mesenteric superior vein. All patients had tumour extending to the portal vein as demonstrated at histopathological examinations in six cases and at surgical dissection in one case. The IVUS technique correctly identified all these patients, whereas five patients were incorrectly deemed at surgery not to have tumour involvement of the portal vein. These results indicate that IVUS is a very sensitive method for the evaluation of tumour involvement of the portal vein.

Isolated liver perfusion using percutaneous methods: an experimental study in the pig.

PURPOSE: To develop a method for isolated perfusion of the liver using radiological methods. METHODS: Twenty-one pigs, weighing about 20 kg, were divided into three groups. By transjugular and transfemoral approaches two occlusion balloons were placed in the inferior vena cava cranial and caudal, respectively, to the origin of the hepatic veins. One occlusion balloon was placed transfemorally in the common hepatic artery. Another occlusion balloon was inserted in the main branch of the portal vein via the transjugular-transhepatic approach in 11 pigs (groups 1 and 2), and in 10 pigs (group 3) by a percutaneous transhepatic route. After inflation of the balloons, patency of the isolated liver circulation was evaluated by recirculation of 99Tcm-labelled human albumin during 30 min. Blood tests were obtained after 1, 3, 5, 10, 15, and 30 min to evaluate leakage from the liver to the systemic circulation. RESULTS: Increasing leakage to the systemic circulation from the isolated liver circulation was observed in groups 1 and 2. In the third group the leakage was less than 10%. CONCLUSION: In an experimental animal model, isolated perfusion of the liver with minor leakage to the systemic circulation may be achieved using radiological methods.

Routine preoperative infusion cholangiography at elective cholecystectomy: a prospective study in 694 patients.

The role of preoperative infusion cholangiography (PIC) before elective cholecystectomy has yet to be defined. Between 1985 and 1991 PIC was performed routinely in 694 patients with median (range) age of 52 (17-85) years with biliary calculus disease 1 or 4 days before elective cholecystectomy. Satisfactory opacification of the biliary system was achieved in 90.1 per cent and mild to moderate adverse reactions to the infusion of contrast medium occurred in 0.9 per cent of patients. Peroperative cholangiography (POC) in 43 of 499 patients with a normal PIC was normal. The presence of common bile duct stones was confirmed in 34 of 36 patients with prominent stones at bile duct exploration and in 8 of 26 with suspected stones at PIC. Bile duct dilatation (median diameter 10 mm) without concomitant stones was seen at PIC and confirmed at POC in 20 patients. Bile duct anomalies were found or suspected in 35 patients at PIC and were confirmed in six by POC (n = 5), or during the cholecystectomy making the anomaly incidence 0.9 per cent. Information provided by the PIC was not of crucial importance for safe execution of cholecystectomy in any patient, while a single (minor) bile duct lesion (1 of 694; 0.1 per cent) could be ascribed to severe pericholecystitis. The results of this study suggest that routine PIC is not warranted. However, PIC is an alternative to selective preoperative endoscopic retrograde cholangiography or selective POC. PIC needs to be supported by POC in approximately 20 per cent of patients (in case of absent or poor opacification of the bile ducts or suspicion of stones or anomalies at PIC).

Routine preoperative infusion cholangiography versus intraoperative cholangiography at elective cholecystectomy: a prospective study in 995 patients.

BACKGROUND: There has been a resurgence of interest in recent years in preoperative infusion cholangiography (PIC). The role of routine PIC compared to routine intraoperative cholangiography (IOC) has not been clearly defined. STUDY DESIGN: In our department between 1985 and 1991, 1,042 of 1,576 consecutive patients with biliary calculous disease had elective cholecystectomy: 694 patients were prospectively scheduled for PIC, and 348 patients were randomly allocated to IOC. The patients in the PIC and IOC groups were similar with regard to age, history of biliopancreatic complications, and laboratory findings. The cost of PIC in Sweden is nearly five times greater than the cost of IOC. RESULTS: Satisfactory opacification of the biliary system was obtained in 90.1 and 96.8 percent of patients who underwent PIC and IOC, respectively. Preoperative infusion cholangiography required support by IOC in 19.5 percent of patients. There were no statistically significant differences between the PIC and IOC groups with regard to the incidence (7 percent in both groups) of or positive predictive value (68 and 80 percent, respectively) for bile duct stones, rate of retained stones (6 and 20 percent, respectively), intraoperative (5.6 and 6.3 percent, respectively) or postoperative (13.3 and 15.9 percent, respectively) morbidity, or incidence of bile duct anomalies (0.9 and 0.3 percent, respectively). Median operative time was longer in patients with (95 minutes) compared to those without (75 minutes) IOC (p < 0.001). More postoperative complications occurred after bile duct exploration (26 of 75 patients) compared to cholecystectomy alone (114 of 917 patients, p < 0.001). The 30-day mortality was zero. Minor bile duct injuries occurred in two patients (0.2 percent) at cholecystectomy, (one with and one without bile duct exploration). In no patient was the cholangiographic finding of a biliary anomaly crucial for the safe execution of cholecystectomy. CONCLUSIONS: In our study, PIC and IOC were comparable, but routine use of either method did not promote the safety of cholecystectomy and thus their routine use is not warranted. The shorter operative time and preoperative identification of common bile duct (CBD) stones provided by PIC might favor this examination when applied selectively in patients with increased risk of having CBD stones. However, this potential advantage is offset by the need for PIC to be supported by IOC in approximately 20 percent of patients. Also, the cost of PIC is greater than the cost of IOC.

Endoscopic treatment of bile duct calculi in patients with gallbladder in situ: long-term outcome and factors.

BACKGROUND: Whether endoscopic sphincterotomy (EST) in elderly and/or high-risk patients with common bile duct calculi (CBD) and the gallbladder in situ should be followed by routine cholecystectomy is still a subject of controversy. METHODS: To identify factors predictive of subsequent biliary tract symptoms after EST and bile duct clearance, we reviewed 265 patients with intact gallbladder and CBD calculi who were considered for EST in our department from 1981 to 1992. In 15 of 265 patients endoscopic treatment was not carried out, and the records of 4 patients were missing. RESULTS: Complete removal of all bile duct calculi failed in 27 patients (11%). Cholecystectomy was performed in 35 patients (16%) with cleared bile ducts 1-765 days (median, 60 days) after EST, in spite of absence of recurrent symptoms from the biliary tract. The remaining 184 patients have been retrospectively followed up for 14-150 months (median, 69 months). Cholecystectomy was required in 35 because of acute cholecystitis (n = 23) or biliary colic (n = 12). Of the cholecystectomies 86% were performed within 24 months after EST and only one after 4 years of follow-up. Increased frequency of cholecystectomy was found in patients with complete opacification of the gallbladder at endoscopic cholangiography (p = 0.005). This was especially evident in patients younger than 80 years (p = 0.002). Cholecystectomy was also required more often in patients with gallbladder calculi (p = 0.02). The risk of cholangitis in patients without recurrent stones was higher in those with juxtapapillary diverticula (p = 0.02). Fifty-nine patients without and 17 with mild to moderate symptoms from the biliary tract died after a median time of 39 and 46 months, respectively. Seventy-three patients are alive, and 59 are symptom-free. Ten patients have had and four still have complaints of mild to moderate biliary tract symptoms. They have been followed for up to 16-146 months (median, 40 months). CONCLUSIONS: These findings confirm that endoscopic treatment alone in this group of patients is a feasible treatment principle. Recognition of the registered risk factors might be helpful when selecting patients for subsequent cholecystectomy.

Long-term follow-up after endoscopic treatment of bile duct calculi in cholecystectomized patients.

Endoscopic sphincterotomy (EST) is an established method for treatment of retained or recurrent common bile duct (CBD) calculi after cholecystectomy. Present experience shows that few patients have recurrent biliary tract complications, but follow-up periods are most often short. EST was performed in 147 patients with bile duct calculi and remote cholecystectomy in our department from 1981 to 1992. In 8 of 147 patients (5.4%) complete removal of calculi failed. A total of 135 patients with a median age of 71 years (range 24-96 years) were eligible for a follow-up of 23 to 153 months (median 86 months). Thirty-seven patients have died without recurrent symptoms (a recurrent stone was revealed at postmortem examination in one patient), and four patients (two with calculi and two with cholangiocarcinoma) died with recurrent symptoms from the biliary tract. Ninety-four patients are alive; and with the exception of two who have had cholangitis without or with post-EST stenosis, respectively, they are all symptom-free. Jaundice, cholangitis, and biliary pancreatitis prior to EST were the only factors that significantly (p = 0.006, Fisher's exact test) predicted late biliary complications after EST in patients with recurrent calculi. These findings confirm that endoscopic treatment of CBD calculi in cholecystectomized patients has a low long-term rate (5 of 135; 3.7%) of recurrent nonmalignant bile duct disease (three patients with CBD calculi and two with cholangitis).

Endoscopic sphincterotomy for bile duct calculi-factors influencing the success rate.

BACKGROUND/AIMS: Endoscopic sphincterotomy is an established treatment of retained or recurrent common bile duct calculi after cholecystectomy and in the majority of patients with an intact gallbladder. In order to identify patients ultimately requiring additional endoscopic procedures or surgery, factors predictive of decreased endoscopic bile duct clearance were sought. MATERIAL AND METHODS: Between 1981 and 1992 endoscopic sphincterotomy was performed in 393 consecutive patients with either intact gallbladders (n = 246), recurrent (n = 92) or retained (n = 55) calculi. RESULTS: There was a 9.4% overall early complication rate for the procedure and a 30-day mortality rate of 0.5% (two patients, non-procedure related). Complete removal of all bile duct calculi failed in 35/393 patients (8.9%). In patients with intact gallbladder the clearance rate at first attempt was higher (p = 0.002; Chi-square test) in the presence of solitary as compared to multiple stones. The same was found in patients with recurrent stones as well (p < 0.001). In non-cholecystectomized patients with single or multiple stones smaller than 10 mm, the clearance rate was higher at first (p = 0.02) as well as at final (p < 0.002) attempt as compared to patients with larger stones. Single small ( < 10 mm) compared to single large stones in patients with intact gallbladder had a higher clearance rate at final (p = 0.002) but not at first (p = 0.18) attempt. Patients with intact gallbladder and pancreatitis had higher clearance rate at first attempt compared to patients with jaundice (p = 0.001) or patients without concomitant pancreatobiliary disease (p = 0.002). Clearance rate was the same in patients with and without endoscopic sphincterotomy complications, except for patients with basket impaction, in whom the clearance rate was decreased (p = 0.02). CONCLUSIONS: Thus, in patients with intact gallbladder surgery should be considered after failed complete bile duct clearance at first attempt.

Long-term follow-up of a prospective randomized study of endoscopic versus surgical treatment of bile duct calculi in patients with gallbladder in situ.

Eighty-three patients with bile duct calculi were entered in a prospective randomized study of endoscopic sphincterotomy (ES) and stone removal (group 1) versus surgery alone (group 2), and were followed for more than 5 years. In group 1 endoscopic stone clearance was successful in 35 of 39 patients. Thirteen patients subsequently had cholecystectomy with (n = 7) or without (n = 6) biliary symptoms and one had a cholecystostomy for acute cholecystitis. Two patients have had mild biliary colic or pancreatitis. Two patients died from gallbladder carcinoma after 9 days and 18 months. In group 2 bile duct stones were cleared surgically in 37 of 41 patients. Late complications occurred in two patients (incisional hernia and recurrent stone). One patient with gallbladder carcinoma was cured and another died after 16 months. Early major and minor complications occurred in three and four respectively of 39 patients in group 1, and in three and six respectively of 41 patients in group 2. There were no deaths. During follow-up the total morbidity rate reached 28 percent (11 of 39) and 5 percent (two of 41) (P = 0.005) and the non-biliary related mortality rate was 31 percent (12 of 39) and 10 percent (four of 41) (P = 0.02) in groups 1 and 2 respectively. Nine patients in group 1 and two in group 2 died from heart disease (P = 0.02). Total hospital stay was 2-42 (median 13) days and 6-36 (median 16) days in groups 1 and 2 respectively (P not significant). Endoscopic and surgical treatment of bile duct calculi in middle-aged and elderly patients with gallbladder in situ are equally effective in the long term. However, the significantly increased mortality rate from heart disease in patients treated endoscopically compared with those treated surgically might speak in favour of operation.

Intravascular ultrasound of the portal vein--normal anatomy.

The purpose of this study was to evaluate the appearance of the normal portal vein with intravascular ultrasound. The portal vein was studied in 10 patients with colorectal carcinoma without known liver or pancreatic disease. For the intravascular portovenous examination, a 2.0-mm, 20-MHz ultrasound catheter was used. The field of view was 30 mm. The wall of the portal vein appeared as a single hyperechoic layer 0.5 to 0.8 mm in thickness. Normal structures adjacent to the portal vein, such as the common bile duct, the hepatic artery or small lymph nodes, were clearly separated from the lumen of the portal vein by the wall of the vein together with some periportal fat. In most cases the parenchyma of the pancreas could be separated from the wall of the vein.

Self-expanding stent in the treatment of benign esophageal strictures: experimental study in pigs and presentation of clinical cases.

The reaction of the normal esophageal wall to inserted self-expanding nitinol stents was studied in pigs. An inflammatory reaction with increasing fibrotic activity and degeneration of the muscular layers in the esophageal wall was demonstrated. Five patients with severe dysphagia secondary to benign esophageal strictures also underwent insertion of self-expanding nitinol stents. All of the stents expanded completely, with subsequent regression of dysphagia. One treated esophagus was resected and showed deep implantation of the stent meshwork in the esophageal wall. Significant stenoses secondary to tissue hyperplasia, located at the edges of the stent, occurred in two patients. These results show that self-expanding nitinol stents may be used for palliation of dysphagia in patients with benign esophageal strictures. Because of the observed reactions in the esophageal wall, such treatment should be restricted to selected patients until more experience has been gained.

Malignant esophageal strictures: treatment with a self-expanding nitinol stent.

A self-expanding esophageal nitinol stent was implanted under fluoroscopic guidance in 40 patients with malignant esophageal strictures and clinically significant dysphagia. The strictures were caused by squamous cell carcinoma (n = 14), adenocarcinoma (n = 12), recurrent anastomotic carcinoma (n = 8), and mediastinal tumors (n = 6). Eight stents were balloon dilated to maximum diameter immediately after insertion. Sixteen stents self-expanded to maximum diameter within 24 hours, and the other stents expanded to maximum diameter during further observation. There were no serious stent-related complications, and the dysphagia was reduced considerably in all patients immediately after stent insertion. Persistent tumor bleeding occurred in two patients, and ingrowth of tumor into the stent was seen in eight patients. Two stents occluded due to tumor ingrowth but were successfully recanalized with endoscopic laser coagulation. At the end of the study, 28 patients were dead with a mean survival of 2.9 months (range, 0.1-7.0 months), and 12 patients were alive with a mean follow-up of 8.8 months (range, 4.0-15.0 months).

Electrolytic stents to inhibit tumor growth. An experimental study in vitro and in rats.

Obstruction of metallic stents used in the management of malignant biliary strictures is common and in several cases due to tumor growth through the stent wall. In an experimental animal model in rats, a tumor was implanted subcutaneously adjacent to different metallic stents. Ingrowth of tumor through the wall of these stents was frequent. A stent generating direct electric current on the basis of electrolysis was developed and tested in vitro and in the experimental model. The study confirms that an "electrolytic" stent can inhibit ingrowth of tumor.