| BackgroundPancreatic cancer is globally thought to be the fourth killer among tumors due to rare identification at early stage, deadly malignancy, lower resection rate, high rate of relapse and strong invasiveness. Because of its biological characteristic of early metastasis, the cancer is affirmatively diagnosed at advanced stage among the patients by80%so that the surgical excision rate is only10%-20%and the median survival time only six months. But, the5-year rate among the patients undergoing surgical resection grows up to20%. At present, radical surgical resection is commonly believed to be the only effective way to prolong their survival for those affirmatively diagnosed pancreatic cancer patients. To achieve this goal, however, we have to solve the problems in preoperative location and credible assessment on the resectablility.To make an affirmative diagnosis, clinically, noninvasive means like ultrasounds, CT, MRI and serologic tumor marker are preferable for the population at high risk. Among them, imageology has its critical role in the diagnosis, staging and prognosis of pancreatic cancer. But the current imageological limitations like defects in imaging reconstruction and incapability of presenting the imaging features among some patients and even the indefinite diagnosis from some imageologists lead to abandonment from surgical therapy. The resectability assessment to the pancreatic cancer, therefore, is critical in that the accurate judgment and location of the cancer has multiple benefits in minimal wounds, rational utility of medical resources, and proper and economical treatment.Funded by the national863project, this study aimed to reconstruct the3D model of pancreatic and periampullary neoplasms by way of threshold value method and injection test method on the Medical Image Three-Dimensional Visualization System (MI-3DVS) and then assess the respectability of the neoplasms.Objective1. To acquire the high quality submillimeter CT data of pancreatic and peripancreatic blood vessels;2. To establish a new set of criteria for assessing the resectability of pancreatic and periampullary neoplasms using the Medical Image Three-Dimensional Visualization System (MI-3DVS);3. To investigate the clinical value of the criteria for assessing the resectablity of pancreatic and periampullary neoplasms.Methods1. Materials:1.1Instruments:(1) Philips Brilliance64-slice helical CT and images processing workstation;(2) Binocular tube high pressure injector;(3) Computer;(4) DICOM Viewer;(5) ACDSee;(6) Medical Image3D Visualization System (MI-3DVS);(7) FreeForm Modeling System and PHANTOM.1.2Agent:Ultravist (300mg I/ml)(Schering AG, Germany).2. Subjects:80patients (male49, female31; aged from15-91y, averaged57.9±1.70y) with confirmed pancreatic and periampullary neoplasms from the Southwest Hospital of the Third Military Medical University involved in the study between November2009and August2011. The imaging data were acquired from their64-MSCT-enhanced epigastric scans. Among them,45contracted pancreatic head carcinomas,14periampullary carcinomas and21tumors in the body and tail of the pancreas. Preoperative resectability assessment for the patients was performed using the results of CTA imaging (referenced by radiologists’reports) and MI-3DVS3D reconstruction. All80cases were examined surgically. Those cases with lesions in distant organs like liver and bones as well as those receiving no surgical examination were all excluded from the study group.3. The acquisition of high quality upper abdomen CT image data:Abdominal CTA, tube voltage of120KV,300mAs tube current,0.5s per lap, pitch (pitch)0.984,5mm slice thickness. scan is divided into four phases:precontrast, arterial phase, portal phase, delayed phase, image data which will be cut a thin slice thickness from5mm to1mm after scaning. Enhanced scanning of the contrast agent injection rate of5ml/s, every scan time is5s, according to measuring the peak of aortic agent contrast when automatically triggers the arterial phase scan, first scan time is usually20-25s after the start of injection of agent contrast, the door pulse of the scan time is generally at the50～55s, the portal venous phase scan immediately start after the delayed phase scan. The image data was transported to the Mxview image post-processing workstation.CT image data collection:The image data from epigastric CT plain scanning and the enhanced scanning for arterial phase, pancreatic phase and venous phase were transmitted from Mxview images processing station to the HP blade servers of our Clinical Medicine Department of the Graduate Center and then exported to store them as documents.4. Image Reconstruction:The original data was inputted into the personal computer and translated into JPG format through DICOM viewer; then the JPG format was translated into JPG format and the size of the images were adjusted from512px×512px into304px×304px by ACDSee.5. Image Segmentation:The adjusted images were imported into Medical Image3D Visualization System (MI-3DVS) and underwent procedure segmentation and three-dimensional reconstruction. The reconstructed models were outputted to STL format; and then STL models were imported into FreeForm Modeling System to be smoothed, denoised and digital painting, etc. The result was not only able to shown as three-dimensional combined model but also independent vessel model.6. Diagnosis and resectability evaluation of pancreatic neoplasms:Based on the3D reconstruction results, the cases were classified as follows (the large vessels described below include the portal vein, superior mesenteric artery, inferior vena cava, superior mesenteric vein, left renal vein, right renal vein, hepatic artery, celiac trunk, and abdominal aorta):Type Ⅰ:Apparent space was viewed between the tumor and the large vessels, but no lumen was deformed from extrusion. Type Ⅱ:The primary tumor or lymph nodes were attached to the large vessels. Morphologically the lumen remained round or roundish (with or without extrusion), and the cross-sectional area of the blood vessels adjacent to the tumors was not decreased. Type Ⅲ:The primary tumor or lymph nodes were attached to and compressed the large vessels, the lumen deformed and the cross-sectional area of the vessels adjacent to the tumors decreased and the surface of the vessel walls attached to the tumor was smooth. Type Ⅳ:The primary tumor or lymph nodes were attached to and compressed the large vessels. The vessels were stiff, the lumen narrowed or the surface unsmooth. Type Ⅴ:The primary tumor enveloped the large blood vessels, with or without significant dilation of the small peripancreatic veins. Or extensive metastases into celiac lymph nodes were observed.The tumors of Type Ⅰ-Ⅱ were assessed with definite recectability, those of Type Ⅲ with possible resectability, or under the conditions of combined resection or reconstruction of blood vessels and those of types IV and V with unresectability.7. Preoperative assessment via CTA imaging All cases in this study were subject to MSCT examinations, in which pancreatic resectability was assessed based on CTA imaging (referenced by reports issued by radiologists). The CTA criteria for surgical unresectability included:(1) The tumor was intricately associated with the celiac trunk and its main branches, the abdominal aorta, the inferior vena cava, the portal vein, the superior mesenteric artery, the inferior mesenteric vein, no visible space between them.(2) The low-density tumor completely enveloped its neighboring blood vessels without morphological changes in the lumen;(3)The low density tumor enveloped the blood vessels, fat free between them or luminal narrowing from their invasion; and (4) The low-density tumors produced vascular occlusion or stenosis.8. The preoperative simulation surgery The reconstructed3D models were imported to the FreeForm Modeling System to perform a variety of simulation surgeries for the determination of optimal program for the actual surgery. Then the final scheme was determined via comparisons of different simulation surgeries so as to provide preoperative surgical planning for the actual operations as well as the real-time guidance.9. Actual surgeries for removing the tumors Experienced surgeons performed the surgical explorations and resections of the tumors assessed to be resectable or potentially resectable, the procedures videotaped all through the course. The surgical specimens were then submitted for pathological examinations.10. Postoperative follow-up The patients were followed up3-24months after operation for CT examinations and MI-3DVS reconstructions.11. Data processing The aforementioned evaluation results were compared with the actual vascular invasions during laparotomy and the actual tumor removal to calculate the positive predictive value, false-positive rate, and false-negative rate of surgical resectability. SPSS13.0was used to perform the statistical analysis, P<0.05considered statistically significant.Results1. Acquisition of high quality upper abdomen CT image data collection The enhanced CT scanning was completed on the80patients, the images indicated pancreatic head carcinoma in47patients, periampullary carcinoma in12ones and pancreatic body and tail carcinoma in21ones. The image data had clear displays of the pancreas, the tumor and peripancreatic blood vessels. Significantly, the small vascular branches from the enhanced scanning made it feasible to reconstruct the pancreatic as well as the peripancreatic blood vessels.2. Individual3D reconstruction of pancreatic tumor The3D models for the80patients were successfully reconstructed, including the models of47pancreatic head cancers,12periampullary carcinomas and21pancreatic body and tail carcinomas. All the reconstructed3D models, clear, vivid and solid, displayed the shape and scope of the tumors, their legible adjacency to the great blood vessels, the clear margins of lesions with a diameter of2.0-10.5cm. The models were rotatable and convenient for zooming and combining. The target organs could be displayed transparent or concealing on the models in order to protrude the tumors in view of tumor size and shape, vascular size and trend, anatomical relations with the organ and the vessels, and invasive scope and position.3. Tumor staging based on MI-3DVS Based on MI-3DVS, the tumors in60patients were assessed to be resectable or potentially resectable, including21cases of type Ⅰ,22cases of type Ⅱ, and17cases of type Ⅲ and the rest20cases (5cases of pancreatic head tumors and5of pancreatic body and tail tumors) unresectable including14cases of type IV and6cases of type V. 4. Resectability by CTA assessment The preoperative CTA assessment determined50cases to be resectable, including25pancreatic head carcinomas,11periampullary carcinomas and14tumors of the pancreatic body and tail. CTA assessment over30unresectable cases including7tumors enveloping celiac axis,6invading into main portal vein and superior mesenteric vein leading to long-distance interruption and17invading into portal veins and causing stenosis in main portal vein.5. Simulation surgery:Visual simulation operations like virtual pancreaticoduodenectomy, virtual medial pancrectomy and virtual pancrecaudectomy were performed on the FreeForm Modeling System using Phantom with virtual surgical instruments. Through visual simulation operations, the relations between the tumors and main portal vein, splenic vein and superior mesenteric veins were displayed anatomically, avoiding the possible injuries at the actual operations. The trends of gastroduodenal artery were displayed on the models to avoid the possible injuries of hepatic artery. Moreover, the simulation surgeries based on the3D models directed the standardized clearance of lymph nodes by displaying the turgescent lymph nodes. Eventually, an optimal scheme was determined by screening the virtual operation schemes.6. Operation results:All of the80cases were subject to surgical explorations, only60of them undergoing excision of pancreatic tumor, with a rate of75%. For the60receiving operations,44underwent pancreaticoduodenectomies,14distal pancreatectomies and2medial pancreatectomies. The20cases, all unresectable by surgical explorations, included4tumors enveloping celiac trunk,9ones enveloping portal veins and superior mesenteric vein leading to long distance linear stenosis of veins and6ones invading in main portal veins and superior mesenteric veins leading to long distance interruption, and1invading in superior mesenteric vein, distal end of splenic vein, proximal end of portal vein and arteries of celiac trunk. They all were subject to palliative surgery for improving their quality of life.7. Comparisons between MI-3DVS and CTA assessments and intraoperative findings:The MI-3DVS assessment indicated respectability in60cases (21cases of type I,22ones of type II and17ones of type III) and unresectability in20cases (14cases of type IV and6cases of type V). The intraoperative findings conformed to the results from the MI-3DVS assessments. In comparison, CTA identified the resectability in50cases, but intraoperative findings proved the unresectability in2cases, with a false-positive rate of10%(2/20). Among the60cases identified to be resectable by MI-3DVS assessments,12were assessed to be unresectable by CTA, while the20cases, unresectable by MI-3DVS, were assessed to be resectable by CTA. The intraoperative findings indicated that the accurate rate of MI-3DVS assessment was significantly higher than that of CTA.8. Postoperative follow-up The postoperative follow-ups lasted from3months to2years. The image data from the follow-ups were rendered for MI-3DVS reconstructions, which showed the preoperative portal vein or inferior vena cava concave from pressure restored to their normal forms and trends.Conclusions1. We collected the high quality submillimeter images of pancreatic and peripancreatic vascular successfully.2. MI-3DVS three-dimensional reconstructed models is accurate in displaying the relationship between the pancreas and periampullary cancer, mesenteric vein, inferior vena cava, celiac trunk, abdominal aorta, hepatic artery, splenic vein, renal vein and other large vessels, and significantly presenting tumor’s shape and scope. They are of great value for the assessment of resectability of tumors.3. The resectability criteria based on MI-3DVS reconstructed models is of great value for assessing the resectability of the pancreas and periampullary cancers clinically.4. MI-3DVS is valuable for clinical diagnosis and treatment of pancreatic and periampullary neoplasms... |
PDF Full Text Request