ASO Author Reflections: Reshaping Tumor Size through Insights from Pancreatic Cancer Imaging and Pathology.

Pancreatic cancer pathology image segmentation with channel and spatial long-range dependencies

A suspected pancreatic lesion can be a difficult challenge for the clinician. In the last years we have witnessed tumultuous technological improvements of the radiological and nuclear medicine imaging. Taking this into account, we will try to delineate the new role of endoscopic ultrasound (EUS) in pancreatic imaging and to place it in a shareable diagnostic and staging algorithm of pancreatic cancer (PC). To date the most accurate imaging techniques for the PC remain contrast-enhanced computed tomography (CT) and EUS. The latter has the highest accuracy in detecting small lesions, in assessing tumor size and lymph nodes involvement, but helical CT or an up-to-date magnetic resonance imaging (MRI) must be the first choice in patients with a suspected pancreatic lesion. After this first step there is place for EUS as a second diagnostic level in several cases: negative results on CT/MRI scans and persistent strong clinical suspicion of PC, doubtful results on CT/MRI scans or need for cyto-histological confirmation. In the near future there will be great opportunities for the development of diagnostic and therapeutic EUS and pancreatic pathology could be the best testing bench.

Screening for Early Pancreatic Neoplasia in High-Risk Individuals: A Prospective Controlled Study

Caldwell and colleagues (1) highlight several challenges associated with employing a pancreatic cancer screening test. Several questionable assumptions in their study, especially regarding individuals at high risk of pancreatic cancer, limit their conclusions.First, patients with a positive screening test would not proceed immediately to endoscopic ultrasound/fine-needle aspiration (EUS/FNA). Before clinical implementation, the main causes of false-positive screening tests would need to have been determined. Before broad-scale implementation of a screening test, studies would be undertaken to evaluate strategies to limit complications that arise from aggressively pursuing positive screening tests, such as identifying comorbidities that cause false positives and/or repeating screening tests. Furthermore, a test with 90% specificity would generate too many false positives and would be unlikely implemented for pancreatic cancer screening. If a 99%-specificity test was used, patients testing positive could proceed to EUS, where an FNA, which is the main cause of EUS-related complications, would only be performed if a suspicious lesion was detected. Even with a high specificity test, many positive screening tests would be false positives; therefore suspicious-looking, false-positive lesions would rarely be seen on EUS. Second, the authors oversimplify by imposing a scenario whereby all patients with a positive EUS, including false positives, would proceed directly to surgical resection. Pancreatic surgeons do not resect the pancreas indiscriminately. In some cases, pancreatic surveillance imaging identifies worrisome lesions, but clinicians are well aware of the risk benefits of proceeding to surgery. In this setting, the decision to undertake surgery is best decided in a multidisciplinary discussion. As a result, the authors overestimate the morbidity of pancreatic cancer screening. Third, the authors assume all high-risk individuals with germline mutations would undergo pancreatic cancer screening. Pancreatic cancer surveillance is not recommended until middle age (age 50+ for most mutation carriers) when the incidence of pancreatic cancer is greater than the 0.2% the authors estimate.Currently, high-risk individuals typically undergo annual surveillance with pancreatic imaging, with acceptable morbidity, as outlined in the most recent CAPS consensus conference (2). This surveillance strategy is associated with downstaging to resectable disease of diagnosed pancreatic cancers (3, 4), including the detection of stage I cancer, with preliminary evidence of improved survival (3). The pursuit of early detection of pancreatic cancer is still a work in progress, but patients diagnosed with stage I pancreatic cancer who undergo surgical resection have excellent 5- and 10-year survival (5). Therefore, early detection research remains an important strategy to improve the prognosis of pancreatic cancer, especially amongst high-risk individuals.See the Response, p. 975B.W. Katona reports grants from NIH/NIDDK during the conduct of the study; other support from Janssen and Exact Sciences outside the submitted work. R.E. Brand reports grants from Immunovia and Freenome during the conduct of the study. E.M. Stoffel reports the CAPS5 pancreatic cancer screening consortium is funded by NIH (principal investigator: M. Goggins). Many of the authors of this letter are part of this consortium. S. Syngal reports personal fees from Myriad Genetics outside the submitted work. No disclosures were reported by the other authors.This study is supported by Smith Family Research Fund (B.W. Katona), Bowen-Chapman Family Fund (S. Syngal), NIH/NCI grant U01CA210170 (M. Goggins).

The unfolded protein response (UPR) is activated in pancreatic pathologies and suggested as a target for therapeutic intervention. In this study, we examined activating transcription factor 3 (ATF3), a mediator of the UPR that promotes acinar-to-ductal metaplasia (ADM) in response to pancreatic injury. Since ADM is an initial step in the progression to pancreatic ductal adenocarcinoma (PDAC), we hypothesized that ATF3 is required for initiation and progression of PDAC. We generated mice carrying a germline mutation of Atf3 (Atf3-/-) combined with acinar-specific induction of oncogenic KRAS (Ptf1acreERT/+KrasG12D/+). Atf3-/- mice with (termed APK) and without KRASG12D were exposed to cerulein-induced pancreatitis. In response to recurrent pancreatitis, Atf3-/- mice showed decreased ADM and enhanced regeneration based on morphological and biochemical analysis. Similarly, an absence of ATF3 reduced spontaneous pancreatic intraepithelial neoplasia (PanIN) formation and PDAC in Ptf1acreERT/+KrasG12D/+ mice. In response to injury, KRASG12D bypassed the requirement for ATF3 with a dramatic loss in acinar tissue and PanIN formation observed regardless of ATF3 status. Compared to Ptf1acreERT/+KrasG12D/+ mice, APK mice exhibited a significant decrease in pancreatic and total body weight, did not progress through to PDAC, and showed altered pancreatic fibrosis and immune cell infiltration. These findings suggest a complex, multifaceted role for ATF3 in pancreatic cancer pathology.

Recent studies have shown evaluation of the small peripancreatic veins to have potential in improving pancreatic cancer staging. This study was performed to determine the effectiveness of thin-section pancreatic phase helical CT images in visualizing these veins. Seventy-two patients (30 with pancreatic adenocarcinoma and 42 with no pancreatic disease) underwent dual-phase helical CT with thin-section pancreatic phase acquisition (40-70 sec after i.v. contrast initiation at 3 ml/sec) and hepatic phase acquisition (70-100 sec). Visualization (with diameter measurement) or nonvisualization of the posterior superior pancreaticoduodenal vein (PSPDV), anterior superior pancreaticoduodenal vein (ASPDV), and gastrocolic trunk was recorded for both acquisitions. We also correlated surgical tumor resectability with the status of the small peripancreatic veins. Visualization of peripancreatic veins was significantly better on pancreatic phase images than on hepatic phase images for both healthy individuals (PSPDV, 88% of the veins visualized on the pancreatic phase images versus 50% on the hepatic phase images; ASPDV, 93% on the pancreatic phase images versus 48% on the hepatic phase images; gastrocolic trunk, 98% on the pancreatic phase images versus 76% on the hepatic phase images) and for pancreatic cancer patients (PSPDV, 97% on the pancreatic phase images versus 57% on the hepatic phase images; ASPDV, 77% on the pancreatic phase images versus 43% on the hepatic phase images) (p < .05). The exception was the gastrocolic trunk in cancer patients (83% on the pancreatic phase images versus 77% on the hepatic phase images) (p > .05). In pancreatic cancer patients, 11 dilated peripancreatic veins were identified on the pancreatic phase images compared with six on the hepatic phase images. However, only one of the 11 dilated peripancreatic veins was in a patient with surgically resectable disease. In a dual-phase helical CT protocol, thin-section pancreatic phase images provided visualization of the small peripancreatic veins that was superior to hepatic phase images, providing further support for the use of this protocol in pancreatic cancer evaluation.

Chapter 11 - Hepatobiliary and pancreatic imaging

Synchronous primary epithelial tumors of the pancreas

Ultrasonography is probably the most widely used medical imaging technique. In fact, its non-invasivity, portability, widespread distribution and good clinical results have made it the preferred method for a first imaging approach to patients in a large variety of clinical settings. Ultrasound technology is advancing rapidly, aimed to increase image quality. In particular, increased color-Doppler sensitivity is contributing to the diagnosis and staging of pancreatic disease. Doppler studies are in fact integral part of ultrasound examination of the pancreas evaluating the peripancreatic vascular structures enclose portal vein, splenic artery and vein, superior mesenteric artery and vein, aorta, and inferior vena cava. While flows in all peri-pancreatic vessels are well recognized, only few parenchymal vessels are usually appreciable in normal conditions. Anyway the increased Doppler sensitivity allows the visualization of smaller peri-pancreatic and intra-pancreatic vessels. This article will review the technical background underlying the clinical application of Doppler ultrasonography in pancreatic imaging, with particular emphasis to normal vascular anatomy of the pancreas, and to the main applications of Doppler techniques in the different pancreatic pathologies.

Pancreatic and Extrapancreatic Features in Autoimmune Pancreatitis

Pancreatic Ductal Adenocarcinoma (PDAC) is one of the most common types of pancreatic cancer and one of the malignant cancers, with an overall five-year survival rate of 5%. CT is the most important imaging examination method for pancreatic diseases with high resolutions. Due to the subtle texture changes of PDAC, single-phase pancreatic imaging is not sufficient to assist doctors in diagnosis. Therefore, dual- phase pancreatPancreatic Ductal Adenocarcinoma (PDAC) is one of the most common types of pancreatic cancer and one of the malignant cancers, with an overall five-year survival rate of 5%. CT is the most important imaging examination method for pancreatic diseases with high resolutions. Due to the subtle texture changes of PDAC, single-phase pancreatic imaging is not sufficient to assist doctors in diagnosis. Therefore, dual- phase pancreatic imaging is recommended for better diagnosis of pancreatic disease. However, since manual labeling requires a lot of time and efforts for experienced physicians, and dual-phase images are often not aligned and largely different in texture, it is difficult to combine cross-phase images. Therefore, in this study, we aim to enhance PDAC automatic segmentation by integrating multi-phase images (i.e. arterial and venous phase) through transfer learning. Therefore, we first transform the image in source domain into the image in target domain through CycleGAN. Secondly, we propose an uncertainty loss to auxiliary training of pseudo target domain images by using pseudo images of different qualities generated during CycleGAN training. Finally, a feature fusion block is designed to compensate for the loss of details caused by downsampling. Experimental results show that the proposed method can obtain more accurate segmentation results than the existing methods.c imaging is recommended for better diagnosis of pancreatic disease. However, since manual labeling requires a lot of time and efforts for experienced physicians, and dual-phase images are often not aligned and largely different in texture, it is difficult to combine cross-phase images. Therefore, in this study, we aim to enhance PDAC automatic segmentation by integrating multi-phase images (i.e. arterial and venous phase) through transfer learning.

Diversity and Inclusion in Pancreatic Cancer Clinical Trials

Therapeutic Advances in Pancreatic Cancer

613 Background: Pancreatic adenocarcinoma (PDAC) is commonly diagnosed at advanced stages, leading to high mortality rates. Earlier detection of PDAC is associated with improved long-term survival, but effective population-level screening is not available. There is therefore an urgent need to improve early detection of PDAC. Wholomics has developed a PDAC-specific proprietary early detection test, using cancer-specific molecular signatures in peripheral blood. Methods: We conducted a retrospective analysis of serum samples from the Amsterdam UMC Liquid Biopsy Center. Three cohorts were included in this study: 1) A cohort of healthy control subjects, 2) patients with a histopathologically confirmed diagnosis of PDAC spanning stages I through IV, and 3) patients with a confirmed diagnosis of chronic pancreatitis. Peripheral blood samples of cancer patients were taken before treatment or surgery. Wholomics used its proprietary technology to quantitatively analyse all serum specimens. Disease-specific molecular signatures were identified and used to develop computational biomarkers specific to both pancreatic pathologies (chronic pancreatitis and pancreatic cancer) and to pancreatic cancer individually. A comparative analysis was conducted employing multiple machine learning algorithms, which were subjected to a 10-fold cross-validation procedure. The accuracy, specificity and sensitivity of the best-performing algorithm is reported. Results: This retrospective study included 50 healthy controls, 12 patients with chronic pancreatitis, and 42 patients with pancreatic cancer (12 stage I, 15 stage II, 9 stage III, 6 stage IV). Applying Wholomics´ technology, an accuracy of 99.1% in differentiating the chronic pancreatitis and pancreatic cancer patients from the healthy controls was achieved. A specificity of 98.0% and a sensitivity of 100.0% were reached. Furthermore, pancreatic cancer patients and chronic pancreatitis patients were differentiated with 95.2% accuracy. Wholomics was also capable of differentiating between early stage (stages I and II) and late-stage (stages III and IV) pancreatic cancer with an accuracy of 100%. Conclusions: Taken together, this study demonstrates that disease-specific molecular signatures in peripheral blood can distinguish patients with PDAC from healthy individuals and patients with pancreatitis with very high accuracy. If validated in additional cohorts, this method could be evaluated as a screening test for PDAC to allow for early detection and improve patient outcomes.

Objective To develop and evaluate the digital discrimination system for pancreatic ultrasound endoscopy images. Methods EUS images of 153 pancreatic cancer and 63 non-cancer cases were selected. According to the multi-fractal feature vectors based on the M-band wavelet transform, we acquired the fractal features with lower dimension with the feature screening algorithm. With the optimal feature com- bination, cases were classified into pancreatic cancer group and non-pancreatic cancer group automatically. Then the sensitivity, specificity and accuracy of this method were calculated, and compared with those of tra- ditional 9 dimension fractal feature vectors. Results Three kinds of muhi-fractal dimensions were intro- duced to the framework of M-band wavelet transform according to the EUS images to form fractal vectors of 18 dimension. With the selection by sequence forward search （SFS） algorithm, 7 dimension of feature vectors were chosen and were combined with bi-order muhi-fractal dimension to a better feature combination. The Bayes, support vector machine （SVM） and ModestAdaBoost classifiers were introduced to evaluate the clas- sification efficiency, resulting in a classification accuracy of 97.98% and short running time of 0. 49 s with lower feature dimension. Conclusion These data suggest the feasibility, accuracy, noninvasiveness and efficacy of classification of EUS images to differentiate pancreatic cancer from normal tissue based on the Mband wavelet transform algorithm. It is a new and valuable research area in diagnosis of pancreatic cancer. Key words: Pancreatic cancer; Endoscopic ultrasonography; M-band wavelet transform; Multifractal