Advances in Pancreatic CT Imaging. AJR Am J Roentgenol. 2018;211:52-66. [PMID: 29629796 DOI: 10.2214/ajr.17.18665]

Pancreas segmentation using AI developed on the largest CT dataset with multi-institutional validation and implications for early cancer detection

Accurate and fully automated pancreas segmentation is critical for advancing imaging biomarkers in early pancreatic cancer detection and for biomarker discovery in endocrine and exocrine pancreatic diseases. We developed and evaluated a deep learning (DL)-based convolutional neural network (CNN) for automated pancreas segmentation using the largest single-institution dataset to date (n = 3031 CTs). Ground truth segmentations were performed by radiologists, which were used to train a 3D nnU-Net model through five-fold cross-validation, generating an ensemble of top-performing models. To assess generalizability, the model was externally validated on the multi-institutional AbdomenCT-1K dataset (n = 585), for which volumetric segmentations were newly generated by expert radiologists and will be made publicly available. In the test subset (n = 452), the CNN achieved a mean Dice Similarity Coefficient (DSC) of 0.94 (SD 0.05), demonstrating high spatial concordance with radiologist-annotated volumes (Concordance Correlation Coefficient [CCC]: 0.95). On the AbdomenCT-1K dataset, the model achieved a DSC of 0.96 (SD 0.04) and a CCC of 0.98, confirming its robustness across diverse imaging conditions. The proposed DL model establishes new performance benchmarks for fully automated pancreas segmentation, offering a scalable and generalizable solution for large-scale imaging biomarker research and clinical translation.

Insulinoma detection on low-dose pancreatic CT perfusion: comparing with conventional contrast-enhanced CT and MRI

OBJECTIVES

To evaluate the efficacy of low-dose pancreatic CT perfusion (pCTP) in detecting insulinomas in patients with recurrent hypoglycemia, and to compare its diagnostic performance with conventional contrast-enhanced CT (CECT) and MRI.

METHODS

This study retrospectively collected 53 patients with recurrent hypoglycemia (28 with insulinomas; 25 without insulinomas). PCTP image analysis was conducted by two radiologists. Quantitative perfusion parameters of insulinomas vs. tumor-free pancreatic parenchyma were analyzed. For cases where both pCTP and CECT/MRI were performed, six radiologists blinded to the patients' diagnosis independently evaluated the pCTP and CECT/MRI to determine the presence and location of insulinoma. The diagnostic performance of insulinoma detection between pCTP and CECT/MRI was compared.

RESULTS

For patients who underwent both CECT and pCTP, the sensitivity (CECT 0.167-0.333 vs. pCTP 0.667-1.000) of tumor detection was higher for five of six radiologists on pCTP than on CECT. For patients who underwent both MRI and pCTP, four radiologists showed higher sensitivity (MRI 0.400-600 vs. pCTP 0.700-0.800) of tumor detection on pCTP than on MRI, while two radiologists showed slightly lower sensitivity (MRI 0.800, 1.000 vs. pCTP 0.700, 0.900) on pCTP. Among perfusion parameters, peak enhancement, blood flow, and mean transit time exhibited higher AUC than blood volume and time to peak.

CONCLUSION

PCTP demonstrated superior diagnostic performance in insulinoma detection among less-experienced radiologists compared to CECT and MRI, while more-experienced radiologists achieved marginally better results with MRI. These findings suggest pCTP's potential as a complementary imaging modality, particularly beneficial for junior radiologists in insulinoma detection.

CRITICAL RELEVANCE STATEMENT

Pancreatic CT perfusion exhibited promising diagnostic performance in insulinoma detection, particularly among junior radiologists, demonstrating the potential to complement conventional imaging modalities and serve as a valuable clinical tool for the detection and localization of insulinoma.

KEY POINTS

Accurate preoperative identification and localization of insulinomas is important for appropriate treatment. Peak enhancement, blood flow, and mean transit time outperformed blood volume and time to peak in insulinoma detection. Pancreatic CT perfusion has the potential to complement conventional imaging modalities for insulinoma detection.

Use of Pancreatic Density on Computed Tomography to Predict Postendoscopic Retrograde Cholangiopancreatography Pancreatitis

INTRODUCTION

Postendoscopic retrograde cholangiopancreatography (ERCP) pancreatitis (PEP) is a frequent complication, and its pathogenesis remains unclear, with various patient and procedural factors proposed as potential contributors. This study aimed to assess the predictive value of pancreatic to splenic density ratio on computed tomography (CT) for PEP in patients with inadvertent pancreatic duct cannulation.

METHODS

This retrospective study involved 2556 patients undergoing ERCP from January 2014 to December 2020. Inclusion criteria comprised patients with choledocholithiasis, preprocedural CT imaging, and inadvertent pancreatic duct cannulation during ERCP. Demographics, preprocedural laboratory values, pancreatic to splenic density ratios from CT scans, and pancreatic stent usage were analyzed in relation to the development of PEP.

RESULTS

A total of 90 patients were included in the study. Of all patients, 51.1% were female (n = 46), and 48.9% were male (n = 44). The mean (±standard deviation) age was 58.93 (±17.01). Significant differences in sodium levels and the pancreatic to splenic density ratio were noted between the PEP and non-PEP groups. Pancreatic to splenic density ratio <0.74 (odds ratio: 8.253; P = 0.020) was identified as an independent risk factor for PEP.

CONCLUSIONS

Pancreas to spleen density ratio on CT imaging serves as a potential predictive marker for PEP, offering insights into risk stratification and guiding prophylactic measures in high-risk patients.

Attention-enhanced multiscale feature fusion network for pancreas and tumor segmentation

BACKGROUND

Accurate pancreas and pancreatic tumor segmentation from abdominal scans is crucial for diagnosing and treating pancreatic diseases. Automated and reliable segmentation algorithms are highly desirable in both clinical practice and research.

PURPOSE

Segmenting the pancreas and tumors is challenging due to their low contrast, irregular morphologies, and variable anatomical locations. Additionally, the substantial difference in size between the pancreas and small tumors makes this task difficult. This paper proposes an attention-enhanced multiscale feature fusion network (AMFF-Net) to address these issues via 3D attention and multiscale context fusion methods.

METHODS

First, to prevent missed segmentation of tumors, we design the residual depthwise attention modules (RDAMs) to extract global features by expanding receptive fields of shallow layers in the encoder. Second, hybrid transformer modules (HTMs) are proposed to model deep semantic features and suppress irrelevant regions while highlighting critical anatomical characteristics. Additionally, the multiscale feature fusion module (MFFM) fuses adjacent top and bottom scale semantic features to address the size imbalance issue.

RESULTS

The proposed AMFF-Net was evaluated on the public MSD dataset, achieving 82.12% DSC for pancreas and 57.00% for tumors. It also demonstrated effective segmentation performance on the NIH and private datasets, outperforming previous State-Of-The-Art (SOTA) methods. Ablation studies verify the effectiveness of RDAMs, HTMs, and MFFM.

CONCLUSIONS

We propose an effective deep learning network for pancreas and tumor segmentation from abdominal CT scans. The proposed modules can better leverage global dependencies and semantic information and achieve significantly higher accuracy than the previous SOTA methods.

Imaging the pancreas with photon-counting CT - A review of normal pancreatic anatomy

PURPOSE

Compared to conventional energy integrating detector CT, Photon-Counting CT (PCCT) has the advantage of increased spatial resolution. The pancreas is a highly complex organ anatomically. The increased spatial resolution of PCCT challenges radiologists' knowledge of pancreatic anatomy. The purpose of this review was to review detailed macroscopic and microscopic anatomy of the pancreas in the context of current and future PCCT.

METHOD

This review is based on a literature review of all parts of pancreatic anatomy and a retrospective imaging review of PCCT scans from 20 consecutively included patients without pancreatic pathology (mean age 61.8 years, 11 female), scanned in the workup of pancreatic cancer with a contrast enhanced multiphase protocol. Two radiologists assessed the visibility of the main and accessory pancreatic ducts, side ducts, ampulla, major papilla, minor papilla, pancreatic arteries and veins, regional lymph nodes, coeliac ganglia, and coeliac plexus.

RESULTS

The macroscopic anatomy of the pancreas was consistently visualized with PCCT. Visualization of detailed anatomy of the ductal system (including side ducts), papillae, arteries, vein, lymph nodes, and innervation was possible in 90% or more of patients with moderate to good interreader agreement.

CONCLUSION

PCCT scans of the pancreas visualizes previously unseen or inconsistently seen small anatomical structures consistently. Increased knowledge of pancreatic anatomy could have importance in imaging of pancreatic cancer and other pancreatic diseases.

SEY‐Net: Semantic edge Y‐shaped network for pancreas segmentation

Pancreas segmentation has great significance in computer‐aided diagnosis of pancreatic diseases. The small size of the pancreas, high variability in shape, and blurred edges make the task of pancreas segmentation challenging. A new model called SEY‐Net is proposed to solve the above problems, which is a one‐stage model with multi‐inputs. SEY‐Net is composed of three main components. Firstly, the edge information extraction (EIE) module is designed to improve the segmentation accuracy of the pancreas boundary. Then, the SE_ResNet50 is selected as the encoder's backbone to fit the size of the pancreas. Finally, the dual cross‐attention is integrated into the skip connection to better focus on the variable shape of the pancreas. The experimental results shows that the proposed method has better performance and outperforms the other existing state‐of‐the‐art pancreas segmentation methods.

Hidden in plain sight: commonly missed early signs of pancreatic cancer on CT

Pancreatic ductal adenocarcinoma (PDAC) has poor prognosis mostly due to the advanced stage at which disease is diagnosed. Early detection of disease at a resectable stage is, therefore, critical for improving outcomes of patients. Prior studies have demonstrated that pancreatic abnormalities may be detected on CT in up to 38% of CT studies 5 years before clinical diagnosis of PDAC. In this review, we highlight commonly missed signs of early PDAC on CT. Broadly, these commonly missed signs consist of small isoattenuating PDAC without contour deformity, isolated pancreatic duct dilatation and cutoff, focal pancreatic enhancement and focal parenchymal atrophy, pancreatitis with underlying PDAC, and vascular encasement. Through providing commentary on demonstrative examples of these signs, we demonstrate how to reduce the risk of missing or misinterpreting radiological features of early PDAC.

Customization of Computed Tomography Radio-Opacity in 3D-Printed Contrast-Injectable Tumor Phantoms

Medical Imaging Phantoms (MIPs) calibrate imaging devices, train medical professionals, and can help procedural planning. Traditional MIPs are costly and limited in customization. Additive manufacturing allows for customizable, patient-specific phantoms. This study examines the CT attenuation characteristics of contrast-injectable, chambered 3D-printed phantoms to optimize tissue-mimicking capabilities. A MIP was constructed from a CT of a complex pelvic tumor near the iliac bifurcation. A 3D reconstruction of these structures composed of three chambers (aorta, inferior vena cava, tumor) with ports for contrast injection was 3D printed. Desired attenuations were 200 HU (arterial I), 150 HU (venous I), 40 HU (tumor I), 150 HU (arterial II), 90 HU (venous II), and 400 HU (tumor II). Solutions of Optiray 350 and water were injected, and the phantom was scanned on CT. Attenuations were measured using ROIs. Mean attenuation for the six phases was as follows: 37.49 HU for tumor I, 200.50 HU for venous I, 227.92 HU for arterial I, 326.20 HU for tumor II, 91.32 HU for venous II, and 132.08 HU for arterial II. Although the percent differences between observed and goal attenuation were high, the observed relative HU differences between phases were similar to goal HU differences. The observed attenuations reflected the relative concentrations of contrast solutions used, exhibiting a strong positive correlation with contrast concentration. The contrast-injectable tumor phantom exhibited a useful physiologic range of attenuation values, enabling the modification of tissue-mimicking 3D-printed phantoms even after the manufacturing process.

Cross-sectional imaging after pancreatic surgery: The dialogue between the radiologist and the surgeon

Pancreatic surgery is nowadays considered one of the most complex surgical approaches and not unscathed from complications. After the surgical procedure, cross-sectional imaging is considered the non-invasive reference standard to detect early and late compilations, and consequently to address patients to the best management possible. Contras-enhanced computed tomography (CECT) should be considered the most important and useful imaging technique to evaluate the surgical site. Thanks to its speed, contrast, and spatial resolution, it can help reach the final diagnosis with high accuracy. On the other hand, magnetic resonance imaging (MRI) should be considered as a second-line imaging approach, especially for the evaluation of biliary findings and late complications. In both cases, the radiologist should be aware of protocols and what to look at, to create a robust dialogue with the surgeon and outline a fitted treatment for each patient.

Pancreatic Iodine Density and Fat Fraction on Dual-Energy Computed Tomography in Acute Pancreatitis

The aim of our study was to investigate iodine density (ID) and fat fraction (FF) on dual-energy computed tomography (DECT) in patients with acute pancreatitis (AP). This retrospective study included 72 patients with clinically confirmed AP and 62 control subjects with DECT of the abdomen. Two radiologists assessed necrosis and measured attenuation values, ID, and FF in three pancreatic segments. We used receiver operating characteristic (ROC) analysis to determine the optimal threshold for ID for the differentiation between AP groups. The ID was significantly higher in interstitial edematous AP compared to necrotizing AP and the control group (both p < 0.05). The ROC curve analysis revealed the thresholds of ID for detecting pancreatic necrosis ≤ 2.2, ≤2.3, and ≤2.4 mg/mL (AUC between 0.880 and 0.893, p > 0.05) for the head, body, and tail, respectively. The FF was significantly higher for pancreatitis groups when compared with the control group in the head and body segments (both p < 0.001). In the tail, the difference was significant in necrotizing AP (p = 0.028). The ID values were independent of attenuation values correlated with the FF values in pancreatic tissue. Iodine density values allow for differentiation between morphologic types of AP.

Pancreatic congenital anomalies and their features on CT and MR imaging: a pictorial review

OBJECTIVE

To review the congenital anomalies of the pancreas with their main clinical manifestations and key imaging findings on CT and MRI.

BACKGROUND AND CLINICAL SIGNIFICANCE

Anomalies of pancreatic development are frequent and generally asymptomatic, but can mimic and predispose individuals to pancreatic or peripancreatic pathologies, such as pancreatitis or malignancy. Their correct diagnosis may help avoid unnecessary further investigations and procedures, or establish adequate treatment when they manifest clinically. Differentiating pancreatic congenital anomalies from their main radiological mimics constitutes a challenge for the radiologist and requires familiarity with key imaging findings.

CONCLUSION

The imaging findings of CT and MRI are essential for the correct diagnosis of congenital pancreatic anomalies.

Pancreatic Adenocarcinoma: Imaging Modalities and the Role of Artificial Intelligence in Analyzing CT and MRI Images

Pancreatic ductal adenocarcinoma (PDAC) stands out as the predominant malignant neoplasm affecting the pancreas, characterized by a poor prognosis, in most cases patients being diagnosed in a nonresectable stage. Image-based artificial intelligence (AI) models implemented in tumor detection, segmentation, and classification could improve diagnosis with better treatment options and increased survival. This review included papers published in the last five years and describes the current trends in AI algorithms used in PDAC. We analyzed the applications of AI in the detection of PDAC, segmentation of the lesion, and classification algorithms used in differential diagnosis, prognosis, and histopathological and genomic prediction. The results show a lack of multi-institutional collaboration and stresses the need for bigger datasets in order for AI models to be implemented in a clinically relevant manner.

Predicting resection margin status of pancreatic neuroendocrine tumors on CT: performance of NCCN resectability criteria

OBJECTIVE

To test the performance of the National Comprehensive Cancer Network (NCCN) CT resectability criteria for predicting the surgical margin status of pancreatic neuroendocrine tumor (PNET) and to identify factors associated with margin-positive resection.

METHODS

Eighty patients with pre-operative CT and upfront surgery were retrospectively enrolled. Two radiologists assessed the CT resectability (resectable [R], borderline resectable [BR], unresectable [UR]) of the PNET according to NCCN criteria. Logistic regression was used to identify factors associated with resection margin status. κ statistics were used to evaluate interreader agreements. Kaplan-Meier method with log-rank test was used to estimate and compare recurrence-free survival (RFS).

RESULTS

Forty-five patients (56.2%) received R0 resection and 35 (43.8%) received R1 or R2 resection. R0 resection rates were 63.6-64.2%, 20.0-33.3%, and 0% for R, BR, and UR diseases, respectively (all p ≤ 0.002), with a good interreader agreement (κ, 0.74). Tumor size (<2 cm, 2-4 cm, and >4 cm; odds ratio (OR), 9.042-18.110; all p ≤ 0.007) and NCCN BR/UR diseases (OR, 5.918; p = 0.032) were predictors for R1 or R2 resection. The R0 resection rate was 91.7% for R disease <2 cm and decreased for larger R disease. R0 resection and smaller tumor size in R disease improved RFS.

CONCLUSION

NCCN resectability criteria can stratify patients with PNET into distinct groups of R0 resectability. Adding tumor size to R disease substantially improves the prediction of R0 resection, especially for PNETs <2 cm.

ADVANCES IN KNOWLEDGE

Tumor size and radiologic resectability independently predicted margin status of PNETs.

Pancreatic Cancer and Its Mimics

Pancreatic ductal adenocarcinoma (PDAC) is the most common primary pancreatic malignancy, ranking fourth in cancer-related mortality in the United States. Typically, PDAC appears on images as a hypovascular mass with upstream pancreatic duct dilatation and abrupt duct cutoff, distal pancreatic atrophy, and vascular encasement, with metastatic involvement including lymphadenopathy. However, atypical manifestations that may limit detection of the underlying PDAC may also occur. Atypical PDAC features include findings related to associated conditions such as acute or chronic pancreatitis, a mass that is isointense to the parenchyma, multiplicity, diffuse tumor infiltration, associated calcifications, and cystic components. Several neoplastic and inflammatory conditions can mimic PDAC, such as paraduodenal "groove" pancreatitis, autoimmune pancreatitis, focal acute and chronic pancreatitis, neuroendocrine tumors, solid pseudopapillary neoplasms, metastases, and lymphoma. Differentiation of these conditions from PDAC can be challenging due to overlapping CT and MRI features; however, certain findings can help in differentiation. Diffusion-weighted MRI can be helpful but also can be nonspecific. Accurate diagnosis is pivotal for guiding therapeutic planning and potential outcomes in PDAC and avoiding biopsy or surgical treatment of some of these mimics. Biopsy may still be required for diagnosis in some cases. The authors describe the typical and atypical imaging findings of PDAC and features that may help to differentiate PDAC from its mimics. ©RSNA, 2023 Online supplemental material is available for this article. Quiz questions for this article are available through the Online Learning Center. See the invited commentary by Zins in this issue.

Current state of radiomic research in pancreatic cancer: focusing on study design and reproducibility of findings

OBJECTIVES

To critically appraise methodology and reproducibility of published studies on CT radiomics of pancreatic ductal adenocarcinoma (PDAC).

METHODS

PRISMA literature search of MEDLINE, PubMed, and Scopus databases was conducted from June to August 2022 relating to CT radiomics human research articles pertaining to PDAC diagnosis, treatment, and/ or prognosis, utilising Image Biomarker Standardisation Initiative-compliant (IBSI) radiomic software. Keyword search included [pancreas OR pancreatic] AND [radiomic OR [quantitative AND imaging] OR [texture AND analysis]]. Analysis included cohort size, CT protocol used, radiomic feature (RF) extraction, segmentation, and selection, software used, outcome correlation, and statistical methodology, with focus on reproducibility.

RESULTS

Initial search yielded 1112 articles; however, only 12 articles met all inclusion/exclusion criteria. Cohort sizes ranged from 37 to 352 (median = 106, mean = 155.8). CT slice thickness varied among studies (4 using ≤ 1 mm, 5 using > 1 to 3 mm, 2 using > 3 to 5 mm, 1 not specifying). CT protocol varied (5 using a single portal-venous (pv)-phase, 5 using a pancreas protocol, 1 study using a non-contrast protocol). RF extraction and segmentation were heterogeneous (RF extraction: 5 using pv-phase, 2 using late arterial, 4 using multi-phase, 1 using non-contrast phase; RF selection: 3 pre-selected, 9 software-selected). 2D/3D RF segmentation was diverse (2D in 6, 3D in 4, 2D and 3D in 2 studies). Six different radiomics software were used. Research questions and cohort characteristics varied, ultimately leading to non-comparable outcome results.

CONCLUSION

The current twelve published IBSI-compliant PDAC radiomic studies show high variability and often incomplete methodology resulting in low robustness and reproducibility.

CLINICAL RELEVANCE STATEMENT

Radiomics research requires IBSI compliance, data harmonisation, and reproducible feature extraction methods for non-invasive imaging biomarker discoveries to be valid. This will ensure a successful clinical implementation and ultimately an improvement of patient outcomes as part of precision and personalised medicine.

KEY POINTS

• Current state of radiomics research in pancreatic cancer shows low software compliance to the Image Biomarker Standardisation Initiative (IBSI). • IBSI-compliant radiomics studies in pancreatic cancer are heterogeneous and not comparable, and the majority of study designs showed low reproducibility. • Improved methodology and standardisation of practice in the emerging field of radiomics has the potential of this non-invasive imaging biomarker in the management of pancreatic cancer.

Confirmed and presumed canine insulinomas and their presumed metastases are most conspicuous in the late arterial phase in a triple arterial phase CT protocol

Arterial enhancement is the commonly described characteristic of canine insulinomas in contrast-enhanced computed tomography (CECT). However, this finding is also reported as inconsistent. The main aim of this single-center retrospective observational study was to describe the contrast enhancement (CE) pattern of canine presumed and confirmed insulinomas and presumed metastases in three consecutive (early, mid, and late) arterial phases. Included dogs had a medical-record-based clinical or cytological/histopathological diagnosis of insulinoma and quadruple-phase CECT. The arterial phases were identified according to published literature. The arterial enhancement of confirmed and presumed lesions was assessed using a visual grading score. Twelve dogs with a total of 17 pancreatic nodules were analyzed. Three dogs had multiple pancreatic nodules and nine had solitary findings. Four insulinomas were histopathologically confirmed. Late arterial phase (LAP) images demonstrated the largest number of pancreatic nodules reaching the highest enhancement scores (n = 13, 76%). All analyzed dogs had CT evidence of arterially enhancing nodules in the liver (n = 12), seven in the hepatic, splenic, or colic lymph nodes, and three in the spleen. Three out of five sampled livers and three lymph nodes were metastatic. All sampled spleens were benign. Avid arterial enhancement was the most dominant feature of canine presumed and confirmed insulinomas and presumed metastases in quadruple-phase CECT. The highest enhancement scores were observed primarily in LAP, followed by MAP. Authors, therefore, recommend including LAP in the standard CT protocol for dogs with suspected pancreatic insulinomas.

Bounding box-based 3D AI model for user-guided volumetric segmentation of pancreatic ductal adenocarcinoma on standard-of-care CTs

OBJECTIVES

To develop a bounding-box-based 3D convolutional neural network (CNN) for user-guided volumetric pancreas ductal adenocarcinoma (PDA) segmentation.

METHODS

Reference segmentations were obtained on CTs (2006-2020) of treatment-naïve PDA. Images were algorithmically cropped using a tumor-centered bounding box for training a 3D nnUNet-based-CNN. Three radiologists independently segmented tumors on test subset, which were combined with reference segmentations using STAPLE to derive composite segmentations. Generalizability was evaluated on Cancer Imaging Archive (TCIA) (n = 41) and Medical Segmentation Decathlon (MSD) (n = 152) datasets.

RESULTS

Total 1151 patients [667 males; age:65.3 ± 10.2 years; T1:34, T2:477, T3:237, T4:403; mean (range) tumor diameter:4.34 (1.1-12.6)-cm] were randomly divided between training/validation (n = 921) and test subsets (n = 230; 75% from other institutions). Model had a high DSC (mean ± SD) against reference segmentations (0.84 ± 0.06), which was comparable to its DSC against composite segmentations (0.84 ± 0.11, p = 0.52). Model-predicted versus reference tumor volumes were comparable (mean ± SD) (29.1 ± 42.2-cc versus 27.1 ± 32.9-cc, p = 0.69, CCC = 0.93). Inter-reader variability was high (mean DSC 0.69 ± 0.16), especially for smaller and isodense tumors. Conversely, model's high performance was comparable between tumor stages, volumes and densities (p > 0.05). Model was resilient to different tumor locations, status of pancreatic/biliary ducts, pancreatic atrophy, CT vendors and slice thicknesses, as well as to the epicenter and dimensions of the bounding-box (p > 0.05). Performance was generalizable on MSD (DSC:0.82 ± 0.06) and TCIA datasets (DSC:0.84 ± 0.08).

CONCLUSION

A computationally efficient bounding box-based AI model developed on a large and diverse dataset shows high accuracy, generalizability, and robustness to clinically encountered variations for user-guided volumetric PDA segmentation including for small and isodense tumors.

CLINICAL RELEVANCE

AI-driven bounding box-based user-guided PDA segmentation offers a discovery tool for image-based multi-omics models for applications such as risk-stratification, treatment response assessment, and prognostication, which are urgently needed to customize treatment strategies to the unique biological profile of each patient's tumor.

Prognostic value of tumor-to-parenchymal contrast enhancement ratio on portal venous-phase CT in pancreatic neuroendocrine neoplasms

OBJECTIVES

We aimed to evaluate the prognostic value of tumor-to-parenchymal contrast enhancement ratio on portal venous-phase CT (CER on PVP) and compare its prognostic performance to prevailing grading and staging systems in pancreatic neuroendocrine neoplasms (PanNENs).

METHODS

In this retrospective study, data on 465 patients (development cohort) who underwent upfront curative-intent resection for PanNEN were used to assess the performance of CER on PVP and tumor size measured by CT (CT-Size) in predicting recurrence-free survival (RFS) using Harrell's C-index and to determine their optimal cutoffs to stratify RFS using a multi-way partitioning algorithm. External data on 184 patients (test cohort) were used to validate the performance of CER on PVP in predicting RFS and overall survival (OS) and compare its predictive performance with those of CT-Size, 2019 World Health Organization classification system (WHO), and the 8th American Joint Committee on Cancer staging system (AJCC).

RESULTS

In the test cohort, CER on PVP showed C-indexes of 0.83 (95% confidence interval [CI], 0.74-0.91) and 0.84 (95% CI, 0.73-0.95) for predicting RFS and OS, respectively, which were higher than those for the WHO (C-index: 0.73 for RFS [p = .002] and 0.72 for OS [p = .004]) and AJCC (C-index, 0.67 for RFS [p = .002] and 0.58 for OS [p = .002]). CT-Size obtained C-indexes of 0.71 for RFS and 0.61 for OS.

CONCLUSIONS

CER on PVP showed superior predictive performance on postoperative survival in PanNEN than current grading and staging systems, indicating its potential as a noninvasive preoperative prognostic tool.

KEY POINTS

• In pancreatic neuroendocrine neoplasms, the tumor-to-parenchymal enhancement ratio on portal venous-phase CT (CER on PVP) showed acceptable predictive performance of postoperative outcomes. • CER on PVP showed superior predictive performance of postoperative survival over the current WHO classification and AJCC staging system.

Pancreatic neuroendocrine tumors: tailoring imaging to specific clinical scenarios

The clinical and imaging presentation of pancreatic neuroendocrine tumors (PanNETs) is variable and depends on tumor grade, stage, and functional status. This degree of variability combined with a multitude of treatment options and imaging modalities results in complexity when choosing the most appropriate imaging studies across various clinical scenarios. While various guidelines exist in the management and evaluation of PanNETs, there is an overall lack of consensus and detail regarding optimal imaging guidelines and protocols. This manuscript aims to fill gaps where current guidelines may lack specificity regarding the choice of the most appropriate imaging study in the diagnosis, treatment planning, monitoring, and surveillance of PanNETs under various clinical scenarios.

Imaging Recommendations for Diagnosis, Staging, and Management of Pancreatic Cancer

Pancreatic cancer is the fourth most prevalent cause of cancer-related death worldwide, with a fatality rate equal to its incidence rate. Pancreatic cancer is a rare malignancy with a global incidence and death ranking of 14th and 7th, respectively. Pancreatic cancer cases are divided into three categories without metastatic disease: resectable, borderline resectable, or locally advanced disease. The category is determined by the tumor's location in the pancreas and whether it is abutting or encasing the adjacent arteries and/or vein/s.

The stage of disease and the location of the primary tumor determine the clinical presentation: the pancreatic head, neck, or uncinate process, the body or tail, or multifocal disease. Imaging plays a crucial role in the diagnosis and follow-up of pancreatic cancers. Various imaging modalities available for pancreatic imaging are ultrasonography (USG), contrast-enhanced computed tomography (CECT), magnetic resonance imaging (MRI), and 18-fluoro-deoxy glucose positron emission tomography (FDG PET).

Even though surgical resection is possible in both resectable and borderline resectable non-metastatic cases, neoadjuvant chemotherapy with or without radiotherapy has become the standard practice for borderline resectable cases as it gives a high yield of R0 resection.

Volumetric Pancreas Segmentation on Computed Tomography: Accuracy and Efficiency of a Convolutional Neural Network Versus Manual Segmentation in 3D Slicer in the Context of Interreader Variability of Expert Radiologists

PURPOSE

This study aimed to compare accuracy and efficiency of a convolutional neural network (CNN)-enhanced workflow for pancreas segmentation versus radiologists in the context of interreader reliability.

METHODS

Volumetric pancreas segmentations on a data set of 294 portal venous computed tomographies were performed by 3 radiologists (R1, R2, and R3) and by a CNN. Convolutional neural network segmentations were reviewed and, if needed, corrected ("corrected CNN [c-CNN]" segmentations) by radiologists. Ground truth was obtained from radiologists' manual segmentations using simultaneous truth and performance level estimation algorithm. Interreader reliability and model's accuracy were evaluated with Dice-Sorenson coefficient (DSC) and Jaccard coefficient (JC). Equivalence was determined using a two 1-sided test. Convolutional neural network segmentations below the 25th percentile DSC were reviewed to evaluate segmentation errors. Time for manual segmentation and c-CNN was compared.

RESULTS

Pancreas volumes from 3 sets of segmentations (manual, CNN, and c-CNN) were noninferior to simultaneous truth and performance level estimation-derived volumes [76.6 cm 3 (20.2 cm 3 ), P < 0.05]. Interreader reliability was high (mean [SD] DSC between R2-R1, 0.87 [0.04]; R3-R1, 0.90 [0.05]; R2-R3, 0.87 [0.04]). Convolutional neural network segmentations were highly accurate (DSC, 0.88 [0.05]; JC, 0.79 [0.07]) and required minimal-to-no corrections (c-CNN: DSC, 0.89 [0.04]; JC, 0.81 [0.06]; equivalence, P < 0.05). Undersegmentation (n = 47 [64%]) was common in the 73 CNN segmentations below 25th percentile DSC, but there were no major errors. Total inference time (minutes) for CNN was 1.2 (0.3). Average time (minutes) taken by radiologists for c-CNN (0.6 [0.97]) was substantially lower compared with manual segmentation (3.37 [1.47]; savings of 77.9%-87% [ P < 0.0001]).

CONCLUSIONS

Convolutional neural network-enhanced workflow provides high accuracy and efficiency for volumetric pancreas segmentation on computed tomography.

Application of spectral CT combined with perfusion scan in diagnosis of pancreatic neuroendocrine tumors

Background

Pancreatic neuroendocrine tumors (pNETs) are heterogeneous tumors from the pancreatic neuroendocrine system, and early diagnosis is important for tumor prognosis and treatment. In this study, we aimed to explore the diagnostic value of spectral CT combined with perfusion scanning in improving the detection rate of pNETs.

Methods

From December 2018 to December 2020, 58 patients with clinically suspected pNETs were prospectively enrolled in the study for one-stop spectral CT combined with perfusion scanning, 36 patients were confirmed with pNETs by histopathology. An independent cohort of 30 patients with pNETs who underwent routine pancreatic perfusion scanning in our hospital during the same period were retrospectively collected. The image characters of pNETs versus tumor-free pancreatic parenchymal were examined.

Results

The detection rate of spectral CT combined with perfusion was 83.1–96.2%. CT values of the pNETs lesions under each single energy in the arterial phase were statistically higher than those of the adjacent normal pancreatic parenchyma. IC, WC and NIC, in the arterial phase of pNETs lesion were all statistically higher than those of the adjacent normal pancreatic parenchyma. The perfusion parameters of pNETs including BF, BV and MSI were significantly higher than those in normal parenchyma. The average effective radiation dose during the perfusion combined energy spectrum enhanced scanning process was 17.51 ± 2.18 mSv.

Conclusion

The one-stop spectral CT combined with perfusion scan improves the detection of pNETs according to morphological features, perfusion parameters and energy spectrum characters with a relatively small radiation dose.

Mass Transport Model of Radiation Response: Calibration and Application to Chemoradiation for Pancreatic Cancer

PURPOSE

The benefit of radiation therapy for pancreatic ductal adenocarcinoma (PDAC) remains unclear. We hypothesized that a new mechanistic mathematical model of chemotherapy and radiation response could predict clinical outcomes a priori, using a previously described baseline measurement of perfusion from computed tomography scans, normalized area under the enhancement curve (nAUC).

METHODS AND MATERIALS

We simplified an existing mass transport model that predicted cancer cell death by replacing previously unknown variables with averaged direct measurements from randomly selected pathologic sections of untreated PDAC. This allowed using nAUC as the sole model input to approximate tumor perfusion. We then compared the predicted cancer cell death to the actual cell death measured from corresponding resected tumors treated with neoadjuvant chemoradiation in a calibration cohort (n = 80) and prospective cohort (n = 25). After calibration, we applied the model to 2 separate cohorts for pathologic and clinical associations: targeted therapy cohort (n = 101), cetuximab/bevacizumab + radiosensitizing chemotherapy, and standard chemoradiation cohort (n = 81), radiosensitizing chemotherapy to 50.4 Gy in 28 fractions.

RESULTS

We established the relationship between pretreatment computed v nAUC to pathologically verified blood volume fraction of the tumor (r = 0.65; P = .009) and fractional tumor cell death (r = 0.97-0.99; P < .0001) in the calibration and prospective cohorts. On multivariate analyses, accounting for traditional covariates, nAUC independently associated with overall survival in all cohorts (mean hazard ratios, 0.14-0.31). Receiver operator characteristic analyses revealed discrimination of good and bad prognostic groups in the cohorts with area under the curve values of 0.64 to 0.71.

CONCLUSIONS

This work presents a new mathematical modeling approach to predict clinical response from chemotherapy and radiation for PDAC. Our findings indicate that oxygen/drug diffusion strongly influences clinical responses and that nAUC is a potential tool to select patients with PDAC for radiation therapy.

Spectral imaging in the pediatric chest: past, present and future

Computed tomography technology continues to undergo evolution and improvement with each passing decade. From its inception in 1971, to the advent of commercially available dual-energy CT just over a decade ago, and now to the latest innovation, photon-counting detector CT, CT's utility for resolving and discriminating tissue types improves. In this review we discuss the impact of spectral imaging, including dual-energy CT and the recently available photon-counting detector CT, on the imaging of the pediatric chest. We describe the current capabilities and future directions of CT imaging, encompassing both the lungs and the surrounding tissues.

PIONEER-Panc: a platform trial for phase II randomized investigations of new and emerging therapies for localized pancreatic cancer

BACKGROUND

Personalized and effective treatments for pancreatic ductal adenocarcinoma (PDAC) continue to remain elusive. Novel clinical trial designs that enable continual and rapid evaluation of novel therapeutics are needed. Here, we describe a platform clinical trial to address this unmet need.

METHODS

This is a phase II study using a Bayesian platform design to evaluate multiple experimental arms against a control arm in patients with PDAC. We first separate patients into three clinical stage groups of localized PDAC (resectable, borderline resectable, and locally advanced disease), and further divide each stage group based on treatment history (treatment naïve or previously treated). The clinical stage and treatment history therefore define 6 different cohorts, and each cohort has one control arm but may have one or more experimental arms running simultaneously. Within each cohort, adaptive randomization rules are applied and patients will be randomized to either an experimental arm or the control arm accordingly. The experimental arm(s) of each cohort are only compared to the applicable cohort specific control arm. Experimental arms may be added independently to one or more cohorts during the study. Multiple correlative studies for tissue, blood, and imaging are also incorporated.

DISCUSSION

To date, PDAC has been treated as a single disease, despite knowledge that there is substantial heterogeneity in disease presentation and biology. It is recognized that the current approach of single arm phase II trials and traditional phase III randomized studies are not well-suited for more personalized treatment strategies in PDAC. The PIONEER Panc platform clinical trial is designed to overcome these challenges and help advance our treatment strategies for this deadly disease.

TRIAL REGISTRATION

This study is approved by the Institutional Review Board (IRB) of MD Anderson Cancer Center, IRB-approved protocol 2020-0075. The PIONEER trial is registered at the US National Institutes of Health (ClinicalTrials.gov) NCT04481204 .

Diagnostic Accuracy of Transabdominal Ultrasound and Computed Tomography in Chronic Pancreatitis: A Head-to-Head Comparison

Purpose Computed tomography (CT) is the most used imaging modality for diagnosing chronic pancreatitis (CP), but advances in transabdominal ultrasound (US) technology have given US a position as a viable alternative. We aimed to evaluate the diagnostic accuracy of abdominal CT and pancreatic US compared to the reference standard, a modified Mayo score. Materials and Methods CT, US, and endoscopic ultrasound (EUS) were performed in patients referred due to suspected CP. The modified Mayo score included EUS results, clinical presentation, and results from exocrine and endocrine pancreatic function tests. We scored CT findings according to the modified Cambridge classification and US findings according to the Rosemont classification. Results In total, 73 patients were included. 53 patients (73%) were categorized as CP and 20 (27%) as non-CP. CT and US yielded similar sensitivities (68% and 64%, respectively) and specificities (75 and 85%, respectively) and similar areas under the receiver operating characteristic curves for diagnosing CP. We found no significant differences between the areas under the receiver operating characteristic curves (AUROCs) for CT (AUROC 0.75, 95% CI 0.63-0.87) and US (AUROC 0.81, 95% CI 0.71-0.91). Conclusion We conclude that CT and US had comparable, moderate accuracy in diagnosing CP. Neither modality had high enough sensitivity to exclude the diagnosis as a standalone method.

Pancreas in Hereditary Syndromes: Cross-sectional Imaging Spectrum

A wide spectrum of hereditary syndromes predispose patients to distinct pancreatic abnormalities, including cystic lesions, recurrent pancreatitis, ductal adenocarcinoma, nonductal neoplasms, and parenchymal iron deposition. While pancreatic exocrine insufficiency and recurrent pancreatitis are common manifestations in cystic fibrosis and hereditary pancreatitis, pancreatic cysts are seen in von Hippel-Lindau disease, cystic fibrosis, autosomal dominant polycystic kidney disease, and McCune-Albright syndrome. Ductal adenocarcinoma can be seen in many syndromes, including Peutz-Jeghers syndrome, familial atypical multiple mole melanoma syndrome, Lynch syndrome, hereditary breast and ovarian cancer syndrome, Li-Fraumeni syndrome, and familial pancreatic cancer syndrome. Neuroendocrine tumors are commonly seen in multiple endocrine neoplasia type 1 syndrome and von Hippel-Lindau disease. Pancreatoblastoma is an essential component of Beckwith-Wiedemann syndrome. Primary hemochromatosis is characterized by pancreatic iron deposition. Pancreatic pathologic conditions associated with genetic syndromes exhibit characteristic imaging findings. Imaging plays a pivotal role in early detection of these conditions and can positively affect the clinical outcomes of those at risk for pancreatic malignancies. Awareness of the characteristic imaging features, imaging-based screening protocols, and surveillance guidelines is crucial for radiologists to guide appropriate patient management. ^©RSNA, 2021.

Effect of deep learning image reconstruction in the prediction of resectability of pancreatic cancer: Diagnostic performance and reader confidence

OBJECTIVE

To assess the diagnostic performance and reader confidence in determining the resectability of pancreatic cancer at computed tomography (CT) using a new deep learning image reconstruction (DLIR) algorithm.

METHODS

A retrospective review was conduct of on forty-seven patients with pathologically confirmed pancreatic cancers who underwent baseline multiphasic contrast-enhanced CT scan. Image data sets were reconstructed using filtered back projection (FBP), hybrid model-based adaptive statistical iterative reconstruction (ASiR-V) 60 %, and DLIR "TrueFidelity" at low(L), medium(M), and high strength levels(H). Four board-certified abdominal radiologists reviewed the CT images and classified cancers as resectable, borderline resectable, or unresectable. Diagnostic performance and reader confidence for categorizing the resectability of pancreatic cancer were evaluated based on the reference standards, and the interreader agreement was assessed using Fleiss k statistics.

RESULTS

For prediction of margin-negative resections(ie, R0), the average area under the receiver operating characteristic curve was significantly higher with DLIR-H (0.91; 95 % confidence interval [CI]: 0.79, 0.98) than FBP (0.75; 95 % CI:0.60, 0.86) and ASiR-V (0.81; 95 % CI:0.67, 0.91) (p = 0.030 and 0.023 respectively). Reader confidence scores were significantly better using DLIR compared to FBP and ASiR-V 60 % and increased linearly with the increase of DLIR strength level (all p < 0.001). Among the image reconstructions, DLIR-H showed the highest interreader agreement in the resectability classification and lowest subject variability in the reader confidence.

CONCLUSIONS

The DLIR-H algorithm may improve the diagnostic performance and reader confidence in the CT assignment of the local resectability of pancreatic cancer while reducing the interreader variability.

Automatic Pancreas Segmentation in CT Images With Distance-Based Saliency-Aware DenseASPP Network

Pancreas identification and segmentation is an essential task in the diagnosis and prognosis of pancreas disease. Although deep neural networks have been widely applied in abdominal organ segmentation, it is still challenging for small organs (e.g. pancreas) that present low contrast, highly flexible anatomical structure and relatively small region. In recent years, coarse-to-fine methods have improved pancreas segmentation accuracy by using coarse predictions in the fine stage, but only object location is utilized and rich image context is neglected. In this paper, we propose a novel distance-based saliency-aware model, namely DSD-ASPP-Net, to fully use coarse segmentation to highlight the pancreas feature and boost accuracy in the fine segmentation stage. Specifically, a DenseASPP (Dense Atrous Spatial Pyramid Pooling) model is trained to learn the pancreas location and probability map, which is then transformed into saliency map through geodesic distance-based saliency transformation. In the fine stage, saliency-aware modules that combine saliency map and image context are introduced into DenseASPP to develop the DSD-ASPP-Net. The architecture of DenseASPP brings multi-scale feature representation and achieves larger receptive field in a denser way, which overcome the difficulties brought by variable object sizes and locations. Our method was evaluated on both public NIH pancreas dataset and local hospital dataset, and achieved an average Dice-Sørensen Coefficient (DSC) value of 85.49±4.77% on the NIH dataset, outperforming former coarse-to-fine methods.

Diagnostic Accuracy of Computed Tomography Scores in Chronic Pancreatitis

OBJECTIVES

Computed tomography (CT) is the most commonly used imaging modality when diagnosing chronic pancreatitis (CP). We aimed to evaluate the diagnostic accuracy of CT scores for diagnosing CP.

METHODS

One hundred eighteen patients were retrospectively included from an observational cohort study that comprised patients referred because of suspected CP. Patients were categorized as CP or non-CP using a modified Mayo score based on biochemistry, clinical presentation, and findings on endoscopic ultrasound and/or transabdominal ultrasound. The CT scans were scored according to the modified Cambridge classification and the unweighted CT score. Diagnostic performance indices were calculated using the modified Mayo score as reference standard.

RESULTS

Seventy-six of the 118 patients fulfilled the CP diagnostic criteria (Mayo score ≥4). The modified Cambridge classification and the unweighted CT score yielded sensitivities of 63% and 67% and specificities of 91% and 91%, respectively, and similar areas under the receiver operating characteristic curves (95% confidence interval) of 0.79 (0.71-0.88)/0.81 (0.73-0.89), respectively (P, not significant).

CONCLUSIONS

Both CT scores had similar, moderate accuracies for diagnosing CP. The limitation in diagnostic accuracy makes CT ineligible as a single method to diagnose CP, supporting that the diagnostic process for CP needs to incorporate other imaging methods and/or markers for better diagnostics.

Two-stage deep learning model for fully automated pancreas segmentation on computed tomography: Comparison with intra-reader and inter-reader reliability at full and reduced radiation dose on an external dataset

PURPOSE

To develop a two-stage three-dimensional (3D) convolutional neural networks (CNNs) for fully automated volumetric segmentation of pancreas on computed tomography (CT) and to further evaluate its performance in the context of intra-reader and inter-reader reliability at full dose and reduced radiation dose CTs on a public dataset.

METHODS

A dataset of 1994 abdomen CT scans (portal venous phase, slice thickness ≤ 3.75-mm, multiple CT vendors) was curated by two radiologists (R1 and R2) to exclude cases with pancreatic pathology, suboptimal image quality, and image artifacts (n = 77). Remaining 1917 CTs were equally allocated between R1 and R2 for volumetric pancreas segmentation [ground truth (GT)]. This internal dataset was randomly divided into training (n = 1380), validation (n = 248), and test (n = 289) sets for the development of a two-stage 3D CNN model based on a modified U-net architecture for automated volumetric pancreas segmentation. Model's performance for pancreas segmentation and the differences in model-predicted pancreatic volumes vs GT volumes were compared on the test set. Subsequently, an external dataset from The Cancer Imaging Archive (TCIA) that had CT scans acquired at standard radiation dose and same scans reconstructed at a simulated 25% radiation dose was curated (n = 41). Volumetric pancreas segmentation was done on this TCIA dataset by R1 and R2 independently on the full dose and then at the reduced radiation dose CT images. Intra-reader and inter-reader reliability, model's segmentation performance, and reliability between model-predicted pancreatic volumes at full vs reduced dose were measured. Finally, model's performance was tested on the benchmarking National Institute of Health (NIH)-Pancreas CT (PCT) dataset.

RESULTS

Three-dimensional CNN had mean (SD) Dice similarity coefficient (DSC): 0.91 (0.03) and average Hausdorff distance of 0.15 (0.09) mm on the test set. Model's performance was equivalent between males and females (P = 0.08) and across different CT slice thicknesses (P > 0.05) based on noninferiority statistical testing. There was no difference in model-predicted and GT pancreatic volumes [mean predicted volume 99 cc (31cc); GT volume 101 cc (33 cc), P = 0.33]. Mean pancreatic volume difference was -2.7 cc (percent difference: -2.4% of GT volume) with excellent correlation between model-predicted and GT volumes [concordance correlation coefficient (CCC)=0.97]. In the external TCIA dataset, the model had higher reliability than R1 and R2 on full vs reduced dose CT scans [model mean (SD) DSC: 0.96 (0.02), CCC = 0.995 vs R1 DSC: 0.83 (0.07), CCC = 0.89, and R2 DSC:0.87 (0.04), CCC = 0.97]. The DSC and volume concordance correlations for R1 vs R2 (inter-reader reliability) were 0.85 (0.07), CCC = 0.90 at full dose and 0.83 (0.07), CCC = 0.96 at reduced dose datasets. There was good reliability between model and R1 at both full and reduced dose CT [full dose: DSC: 0.81 (0.07), CCC = 0.83 and reduced dose DSC:0.81 (0.08), CCC = 0.87]. Likewise, there was good reliability between model and R2 at both full and reduced dose CT [full dose: DSC: 0.84 (0.05), CCC = 0.89 and reduced dose DSC:0.83(0.06), CCC = 0.89]. There was no difference in model-predicted and GT pancreatic volume in TCIA dataset (mean predicted volume 96 cc (33); GT pancreatic volume 89 cc (30), p = 0.31). Model had mean (SD) DSC: 0.89 (0.04) (minimum-maximum DSC: 0.79 -0.96) on the NIH-PCT dataset.

CONCLUSION

A 3D CNN developed on the largest dataset of CTs is accurate for fully automated volumetric pancreas segmentation and is generalizable across a wide range of CT slice thicknesses, radiation dose, and patient gender. This 3D CNN offers a scalable tool to leverage biomarkers from pancreas morphometrics and radiomics for pancreatic diseases including for early pancreatic cancer detection.

Evaluation of equivocal small cystic pancreatic lesions with spectral-detector computed tomography

BACKGROUND

Evaluation of small cystic lesions of the pancreas remains a challenging task, as due to their size appearance can be rather hypodense than clearly fluid-filled.

PURPOSE

To evaluate whether additional information provided by novel dual-layer spectral-detector computed tomography (SDCT) imaging can improve assessment of these lesions.

MATERIAL AND METHODS

For this retrospective study, we reviewed reports of 1192 contrast-enhanced portal-venous phase SDCT scans of the abdomen conducted between May 2017 and January 2019. On basis of the radiological report 25 small (≤1.5 cm) cystic pancreatic lesions in 22 patients were identified, in which additional short-term follow-up imaging was recommended to confirm/clarify cystic nature. Conventional images (CI) and spectral images (SI) including virtual-monoenergetic images at 40 keV (VMI), iodine-density and iodine-overlay images were reconstructed. Two readers indicated lesion conspicuity and confidence for presence of cystic nature on three-point scales. First, solely CI were evaluated, while in a second reading after a four-week interval, the combination of CI and corresponding SI were reviewed. Quantitatively, ROI-based mean attenuation was measured in CI and VMI.

RESULTS

In the subjective reading, SI significantly improved lesion conspicuity (CI 2 [1-2], SI 3 [2-3], P < 0.001) and confidence regarding presence of cystic nature (CI 2 [1-2], SI 3 [3-3], P < 0.001). Inter-observer agreement depicted by intraclass correlation coefficient improved considerably from 0.51 with only CI to 0.85 when the combination with SI was used. Further, VMI displayed significantly higher signal-to-noise (CI 1.2 ± 0.8, VMI 3.2 ± 1.8, P < 0.001) and contrast-to-noise ratios (CI 2.6 ± 0.8, VMI 4.7 ± 1.9).

CONCLUSION

Compared to CI alone, combination with SI significantly improves visualization and confidence in evaluation of small equivocal cystic pancreatic lesions.

Application of low dose pancreas perfusion CT combined with enhancement scanning in diagnosis of pancreatic neuroendocrine tumors

PURPOSE

To explore the diagnostic value of pancreatic perfusion CT combined with contrast-enhanced CT in one-time scanning (PCECT) in pancreatic neuroendocrine tumors (PNETs) and to evaluate the difference of perfusion parameters between different grades of PNETs.

MATERIALS AND METHODS

From October 2016 to December 2018, forty consecutive patients with histopathological-proven PNETs were identified retrospectively that received PCECT for the preoperative PNETs evaluation. Two board certified radiologists who were blinded to the clinical data evaluated the images independently. The image characters of PNETs vs. tumor-free pancreatic parenchymal and different grades of PNETs were analyzed.

RESULTS

One-time PCECT scanning had a detection rate of 89.1% for PNETs, which was higher than the detection accuracy of the perfusion CT only (83.6%). The perfusion parameters of PNETs including blood volume (BV), blood flow (BF), mean slope of increase (MSI), and capillary surface permeability (PS) were significantly increased than those of tumor-free pancreatic parenchyma (p < 0.05, respectively). For differential comparison between grade I (G1) and grade II (G2) tumors, the parameters of BF and impulse residue function (IRF) of tumor tissue were significantly higher in the G2 tumors (p < 0.05, for both). In this study, the total radiation dose of the whole PCECT scan was 16.241 ± 2.289 mSv.

CONCLUSION

The one-time PCECT scan may improve the detection of PNETs according to morphological features and perfusion parameters with a relative small radiation dose. The perfusion parameters of BF and IRF may be used to help distinguish G1 and G2 tumors in the preoperative evaluation.

Clinical Implementation of Dual-Energy CT for Gastrointestinal Imaging

Dual-energy CT (DECT) overcomes several limitations of conventional single-energy CT (SECT) for the evaluation of gastrointestinal diseases. This article provides an overview of practical aspects of the DECT technology and acquisition protocols, reviews existing clinical applications, discusses current challenges, and describes future directions, with a focus on gastrointestinal imaging. A head-to-head comparison of technical specifications among DECT scanner implementations is provided. Energy- and material-specific DECT image reconstructions enable retrospective (i.e., after examination acquisition) image quality adjustments that are not possible using SECT. Such adjustments may, for example, correct insufficient contrast bolus or metal artifacts, thereby potentially avoiding patient recalls. A combination of low-energy monochromatic images, iodine maps, and virtual unenhanced images can be included in protocols to improve lesion detection and disease characterization. Relevant literature is reviewed regarding use of DECT for evaluation of the liver, gallbladder, pancreas, and bowel. Challenges involving cost, workflow, body habitus, and variability in DECT measurements are considered. Artificial intelligence and machine-learning image reconstruction algorithms, PACS integration, photon-counting hardware, and novel contrast agents are expected to expand the multienergy capability of DECT and further augment its value.

New advances in CT imaging of pancreas diseases: a narrative review

Computed tomography (CT) plays a pivotal role as a diagnostic tool in many diagnostic and diffuse pancreatic diseases. One of the major limits of CT is related to the radiation exposure of young patients undergoing repeated examinations. Besides the standard CT protocol, the most recent technological advances, such as low-voltage acquisitions with high performance X-ray tubes and iterative reconstructions, allow for significant optimization of the protocol with dose reduction. The variety of CT tools are further expanded by the introduction of dual energy: the production of energy-selective images (i.e., virtual monochromatic images) improves the image contrast and lesion detection while the material-selective images (e.g., iodine maps or virtual unenhanced images) are valuable for lesion detection and dose reduction. The perfusion techniques provide diagnostic and prognostic information lesion and parenchymal vascularization and interstitium. Both dual energy and perfusion CT have the potential for pushing the limits of conventional CT from morphological evaluation to quantitative imaging applied to inflammatory and oncological diseases. Advances in post-processing of CT images, such as pancreatic volumetry, texture analysis and radiomics provide relevant information for pancreatic function but also for the diagnosis, management and prognosis of pancreatic neoplasms. Artificial intelligence is promising for optimization of the workflow in qualitative and quantitative analyses. Finally, basic concepts on the role of imaging on screening of pancreatic diseases will be provided.

Development of a volumetric pancreas segmentation CT dataset for AI applications through trained technologists: a study during the COVID 19 containment phase

PURPOSE

To evaluate the performance of trained technologists vis-à-vis radiologists for volumetric pancreas segmentation and to assess the impact of supplementary training on their performance.

METHODS

In this IRB-approved study, 22 technologists were trained in pancreas segmentation on portal venous phase CT through radiologist-led interactive videoconferencing sessions based on an image-rich curriculum. Technologists segmented pancreas in 188 CTs using freehand tools on custom image-viewing software. Subsequent supplementary training included multimedia videos focused on common errors, which were followed by second batch of 159 segmentations. Two radiologists reviewed all cases and corrected inaccurate segmentations. Technologists' segmentations were compared against radiologists' segmentations using Dice-Sorenson coefficient (DSC), Jaccard coefficient (JC), and Bland-Altman analysis.

RESULTS

Corrections were made in 71 (38%) cases from first batch [26 (37%) oversegmentations and 45 (63%) undersegmentations] and in 77 (48%) cases from second batch [12 (16%) oversegmentations and 65 (84%) undersegmentations]. DSC, JC, false positive (FP), and false negative (FN) [mean (SD)] in first versus second batches were 0.63 (0.15) versus 0.63 (0.16), 0.48 (0.15) versus 0.48 (0.15), 0.29 (0.21) versus 0.21 (0.10), and 0.36 (0.20) versus 0.43 (0.19), respectively. Differences were not significant (p > 0.05). However, range of mean pancreatic volume difference reduced in the second batch [- 2.74 cc (min - 92.96 cc, max 87.47 cc) versus - 23.57 cc (min - 77.32, max 30.19)].

CONCLUSION

Trained technologists could perform volumetric pancreas segmentation with reasonable accuracy despite its complexity. Supplementary training further reduced range of volume difference in segmentations. Investment into training technologists could augment and accelerate development of body imaging datasets for AI applications.

Factors associated with missed and misinterpreted cases of pancreatic ductal adenocarcinoma

OBJECTIVES

To retrospectively examine US, CT, and MR imaging examinations of missed or misinterpreted pancreatic ductal adenocarcinoma (PDAC), and identify factors which may have confounded detection or interpretation.

METHODS

We reviewed 107 examinations in 66/257 patients (26%, mean age 73.7 years) diagnosed with PDAC in 2014 and 2015, with missed or misinterpreted imaging findings as determined by a prior study. For each patient, images and reports were independently reviewed by two radiologists, and in consensus, the following factors which may have confounded assessment were recorded: inherent tumor factors, concurrent pancreatic pathology, technical limitations, and cognitive biases. Secondary signs of PDAC associated with each examination were recorded and compared with the original report to determine which findings were missed.

RESULTS

There were 66/107 (62%) and 49/107 (46%) cases with missed and misinterpreted imaging findings, respectively. A significant number of missed tumors were < 2 cm (45/107, 42%), isoattenuating on CT (32/72, 44%) or non-contour deforming (44/107, 41%). Most (29/49, 59%) misinterpreted examinations were reported as uncomplicated pancreatitis. Almost all examinations (94/107, 88%) demonstrated secondary signs; pancreatic duct dilation was the most common (65/107, 61%) and vascular invasion was the most commonly missed 35/39 (90%). Of the CT and MRIs, 28 of 88 (32%) had suboptimal contrast dosing. Inattentional blindness was the most common cognitive bias, identified in 55/107 (51%) of the exams.

CONCLUSION

Recognizing pitfalls of PDAC detection and interpretation, including intrinsic tumor features, secondary signs, technical factors, and cognitive biases, can assist radiologists in making an early and accurate diagnosis.

KEY POINTS

• There were 66/107 (62%) and 49/107 (46%) cases with missed and misinterpreted imaging findings, respectively, with tumoral, technical, and cognitive factors leading to the misdiagnosis of pancreatic ductal adenocarcinoma. • The majority (29/49, 59%) of misinterpreted cases of pancreatic ductal adenocarcinoma were mistaken for pancreatitis, where an underlying mass or secondary signs were not appreciated due to inflammatory changes. • The most common missed secondary sign of pancreatic ductal adenocarcinoma was vascular encasement, missed in 35/39 (90%) of cases, indicating the importance of evaluating the peri-pancreatic vasculature.

Enhancement parameters of contrast-enhanced computed tomography for pancreatic ductal adenocarcinoma: Correlation with pathologic grading

BACKGROUND

Pancreatic ductal adenocarcinoma (PDA) is a malignancy with a high mortality rate and short survival time. The conventional computed tomography (CT) has been worldwide used as a modality for diagnosis of PDA, as CT enhancement pattern has been thought to be related to tumor angiogenesis and pathologic grade of PDA.

AIM

To evaluate the relationship between the pathologic grade of pancreatic ductal adenocarcinoma and the enhancement parameters of contrast-enhanced CT.

METHODS

In this retrospective study, 42 patients (Age, mean ± SD: 62.43 ± 11.42 years) with PDA who underwent surgery after preoperative CT were selected. Two radiologists evaluated the CT images and calculated the value of attenuation at the aorta in the arterial phase and the pancreatic phase (VA_arterial and VA_pancreatic) and of the tumor (VT_arterial and VT_pancreatic) by finding out four regions of interest. Ratio between the tumor and the aorta enhancement on the arterial phase and the pancreatic phase (TAR_arterial and TAR_pancreatic) was figured out through dividing VT_arterial by VA_arterial and VT_pancreatic by VA_pancreatic. Tumor-to-aortic enhancement fraction (TAF) was expressed as the ratio of the difference between attenuation of the tumor on arterial and parenchymal images to that between attenuation of the aorta on arterial and pancreatic images. The Kruskal-Wallis analysis of variance and Mann-Whitney U test for statistical analysis were used.

RESULTS

Forty-two PDAs (23 men and 19 women) were divided into three groups: Well-differentiated (n = 13), moderately differentiated (n = 21), and poorly differentiated (n = 8). TAF differed significantly between the three groups (P = 0.034) but TAR_arterial (P = 0.164) and TAR_pancreatic (P = 0.339) did not. The median value of TAF for poorly differentiated PDAs (0.1011; 95%CI: 0.01100-0.1796) was significantly higher than that for well-differentiated PDAs (0.1941; 95%CI: 0.1463-0.3194).

CONCLUSION

Calculation of TAF might be useful in predicting the pathologic grade of PDA.

Computed tomographic evaluation of pancreatic perfusion in healthy dogs

OBJECTIVE

To evaluate the feasibility of contrast-enhanced CT for assessment of pancreatic perfusion in healthy dogs.

ANIMALS

6 healthy purpose-bred female Treeing Walker Coonhounds.

PROCEDURES

Contrast-enhanced CT of the cranial part of the abdomen was performed with 3-mm slice thickness. Postprocessing computer software designed for evaluation of human patients was used to calculate perfusion data for the pancreas and liver by use of 3-mm and reformatted 6-mm slices. Differences in perfusion variables between the pancreas and liver and differences in liver-specific data of interest were evaluated with the Friedman test.

RESULTS

Multiple pancreatic perfusion variables were determined, including perfusion, peak enhancement index, time to peak enhancement, and blood volume. The same variables as well as arterial, portal, and total perfusion and hepatic perfusion index were determined for the liver. Values for 6-mm slices appeared similar to those for 3-mm slices. The liver had significantly greater median perfusion and peak enhancement index, compared with the pancreas.

CONCLUSIONS AND CLINICAL RELEVANCE

Measurement of pancreatic perfusion with contrast-enhanced CT was feasible in this group of dogs. Hepatic arterial and pancreatic perfusion values were similar to previously published findings for dogs, but hepatic portal and hepatic total perfusion measurements were not. These discrepancies might have been attributable to physiologic differences between dogs and people and related limitations of the CT software intended for evaluation of human patients. Further research is warranted to assess reliability of perfusion variables and applicability of the method for assessment of canine patients with pancreatic abnormalities.

Imaging and Management of Pancreatic Cancer

Pancreatic cancer is an aggressive disease with rising incidence and high mortality despite advances in imaging and therapeutic options. Surgical resection is currently the only curative treatment, with expanding roles for adjuvant and neoadjuvant chemoradiation. Accurate detection, staging, and post-treatment monitoring of pancreatic cancer are critical to improving survival and imaging plays a central role in the multidisciplinary approach to this disease. This article will provide a broad overview of the imaging and management of pancreatic cancer with a focus on diagnosis and staging, operative and nonoperative treatments, and post-therapeutic appearances after surgery and chemoradiation therapy.

Advanced imaging techniques for chronic pancreatitis

MRI and MRCP play an important role in the diagnosis of chronic pancreatitis (CP) by imaging pancreatic parenchyma and ducts. MRI/MRCP is more widely used than computed tomography (CT) for mild to moderate CP due to its increased sensitivity for pancreatic ductal and gland changes; however, it does not detect the calcifications seen in advanced CP. Quantitative MR imaging offers potential advantages over conventional qualitative imaging, including simplicity of analysis, quantitative and population-based comparisons, and more direct interpretation of detected changes. These techniques may provide quantitative metrics for determining the presence and severity of acinar cell loss and aid in the diagnosis of chronic pancreatitis. Given the fact that the parenchymal changes of CP precede the ductal involvement, there would be a significant benefit from developing MRI/MRCP-based, more robust diagnostic criteria combining ductal and parenchymal findings. Among cross-sectional imaging modalities, multi-detector CT (MDCT) has been a cornerstone for evaluating chronic pancreatitis (CP) since it is ubiquitous, assesses primary disease process, identifies complications like pseudocyst or vascular thrombosis with high sensitivity and specificity, guides therapeutic management decisions, and provides images with isotropic resolution within seconds. Conventional MDCT has certain limitations and is reserved to provide predominantly morphological (e.g., calcifications, organ size) rather than functional information. The emerging applications of radiomics and artificial intelligence are poised to extend the current capabilities of MDCT. In this review article, we will review advanced imaging techniques by MRI, MRCP, CT, and ultrasound.

The North American Neuroendocrine Tumor Society Consensus Paper on the Surgical Management of Pancreatic Neuroendocrine Tumors

This manuscript is the result of the North American Neuroendocrine Tumor Society consensus conference on the surgical management of pancreatic neuroendocrine tumors from July 19 to 20, 2018. The group reviewed a series of questions of specific interest to surgeons taking care of patients with pancreatic neuroendocrine tumors, and for each, the available literature was reviewed. What follows are these reviews for each question followed by recommendations of the panel.

Duodeno-pancreatic and extrahepatic biliary tree trauma: WSES-AAST guidelines

Duodeno-pancreatic and extrahepatic biliary tree injuries are rare in both adult and pediatric trauma patients, and due to their anatomical location, associated injuries are very common. Mortality is primarily related to associated injuries, but morbidity remains high even in isolated injuries. Optimal management of duodeno-bilio-pancreatic injuries is dictated primarily by hemodynamic stability, clinical presentation, and grade of injury. Endoscopic and percutaneous interventions have increased the ability to non-operatively manage these injuries. Late diagnosis and treatment are both associated to increased morbidity and mortality. Sequelae of late presentations of pancreatic injury and complications of severe pancreatic trauma are also increasingly addressed endoscopically and with interventional radiology procedures. However, for moderate and severe extrahepatic biliary and severe duodeno-pancreatic injuries, immediate operative intervention is preferred as associated injuries are frequent and commonly present with hemodynamic instability or peritonitis. The aim of this paper is to present the World Society of Emergency Surgery (WSES) and American Association for the Surgery of Trauma (AAST) duodenal, pancreatic, and extrahepatic biliary tree trauma management guidelines.