Computerized Diagnosis of Liver Tumors From CT Scans Using a Deep Neural Network Approach.

The liver is a frequent site of benign and malignant, primary and metastatic tumors. Hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (ICC) are the most common primary liver cancers, and colorectal liver metastasis (CRLM) is the most common secondary liver cancer. Although the imaging characteristic of these tumors is central to optimal clinical management, it relies on imaging features that are often non-specific, overlap, and are subject to inter-observer variability. Thus, in this study, we aimed to categorize liver tumors automatically from CT scans using a deep learning approach that objectively extracts discriminating features not visible to the naked eye. Specifically, we used a modified Inception v3 network-based classification model to classify HCC, ICC, CRLM, and benign tumors from pretreatment portal venous phase computed tomography (CT) scans. Using a multi-institutional dataset of 814 patients, this method achieved an overall accuracy rate of 96%, with sensitivity rates of 96%, 94%, 99%, and 86% for HCC, ICC, CRLM, and benign tumors, respectively, using an independent dataset. These results demonstrate the feasibility of the proposed computer-assisted system as a novel non-invasive diagnostic tool to classify the most common liver tumors objectively.

MRI of the Pancreas.

MRI has played a critical role in the evaluation of patients with pancreatic pathologies, from screening of patients at high risk for pancreatic cancer to the evaluation of pancreatic cysts and indeterminate pancreatic lesions. The high mortality associated with pancreatic adenocarcinomas has spurred much interest in developing effective screening tools, with MRI using magnetic resonance cholangiopancreatography (MRCP) playing a central role in the hopes of identifying cancers at earlier stages amenable to curative resection. Ongoing efforts to improve the resolution and robustness of imaging of the pancreas using MRI may thus one day reduce the mortality of this deadly disease. However, the increasing use of cross-sectional imaging has also generated a concomitant clinical conundrum: How to manage incidental pancreatic cystic lesions that are found in over a quarter of patients who undergo MRCP. Efforts to improve the specificity of MRCP for patients with pancreatic cysts and with indeterminate pancreatic masses may be achieved with continued technical advances in MRI, including diffusion-weighted and T1 -weighted dynamic contrast-enhanced MRI. However, developments in quantitative MRI of the pancreas remain challenging, due to the small size of the pancreas and its upper abdominal location, adjacent to bowel and below the diaphragm. Further research is needed to improve MRI of the pancreas as a clinical tool, to positively affect the lives of patients with pancreatic abnormalities. This review focuses on various MR techniques such as MRCP, quantitative imaging, and dynamic contrast-enhanced imaging and their clinical applicability in the imaging of the pancreas, with an emphasis on pancreatic malignant and premalignant lesions. Level of Evidence 5 Technical Efficacy Stage 3 J. MAGN. RESON. IMAGING 2021;53:347-359.

Preoperative CT predictors of survival in patients with pancreatic ductal adenocarcinoma undergoing curative intent surgery.

PURPOSE: To evaluate the associations between computed tomography (CT) imaging features extracted from the structured American Pancreatic Association (APA)/Society of Abdominal Radiology (SAR) template and overall survival in patients with resected pancreatic ductal adenocarcinoma (PDAC). METHODS: This retrospective analysis included consecutive patients with PDAC who consented to genomic tumor testing and underwent preoperative imaging and curative intent surgical resection from December 2006 to July 2017. Two radiologists assessed preoperative CT imaging using the APA/SAR PDAC-reporting template. Univariable associations between overall survival and imaging variables were evaluated using Cox proportional hazards regression. RESULTS: The study included 168 patients (66 years ± 11; 91 women). 126/168 patients (75%) received upfront surgical resection whereas 42/168 (25%) received neoadjuvant therapy prior to surgical resection. In the entire cohort, features associated with decreased overall survival were tumor arterial contact of any kind (hazard ratio (HR) 1.89, 95% CI 1.13-3.14, p = 0.020), tumor contact with the common hepatic artery (HR 2.33, 95% CI 1.35-4.04, p = 0.009), and portal vein deformity (HR 3.22, 95% CI 1.63-6.37, p = 0.003). In the upfront surgical group, larger tumor size was associated with decreased overall survival (HR 2.30, 95% CI 1.19-4.42, p = 0.013). In the neoadjuvant therapy group, the presence of venous collaterals was the only feature associated with decreased overall survival (HR 2.28, 95% CI 1.04-4.99, p = 0.042). CONCLUSION: The application of the APA/SAR pancreatic adenocarcinoma reporting template may identify predictors of survival that can aid in preoperative stratification of patients.

Multimodal radiomics and cyst fluid inflammatory markers model to predict preoperative risk in intraductal papillary mucinous neoplasms.

Purpose: Our paper contributes to the burgeoning field of surgical data science. Specifically, multimodal integration of relevant patient data is used to determine who should undergo a complex pancreatic resection. Intraductal papillary mucinous neoplasms (IPMNs) represent cystic precursor lesions of pancreatic cancer with varying risk for malignancy. We combine previously defined individual models of radiomic analysis of diagnostic computed tomography (CT) with protein markers extracted from the cyst fluid to create a unified prediction model to identify high-risk IPMNs. Patients with high-risk IPMN would be sent for resection, whereas patients with low-risk cystic lesions would be spared an invasive procedure. Approach: Retrospective analysis of prospectively acquired cyst fluid and CT scans was undertaken for this study. A predictive model combining clinical features with a cyst fluid inflammatory marker (CFIM) was applied to patient data. Quantitative imaging (QI) features describing radiomic patterns predictive of risk were extracted from scans. The CFIM model and QI model were combined into a single predictive model. An additional model was created with tumor-associated neutrophils (TANs) assessed by a pathologist at the time of resection. Results: Thirty-three patients were analyzed (7 high risk and 26 low risk). The CFIM model yielded an area under the curve (AUC) of 0.74. Adding the QI model improved performance with an AUC of 0.88. Combining the CFIM, QI, and TAN models further increased performance to an AUC of 0.98. Conclusions: Quantitative analysis of routinely acquired CT scans combined with CFIMs provides accurate prediction of risk of pancreatic cancer progression. Although a larger cohort is needed for validation, this model represents a promising tool for preoperative assessment of IPMN.

Malignant Pleural Mesothelioma: Are There Imaging Characteristics Associated With Different Histologic Subtypes on Computed Tomography?

OBJECTIVES: Determine imaging characteristics specific to epithelioid (eMPM), sarcomatoid (sMPM), and biphasic (bMPM) subtypes of malignant pleural mesothelioma (MPM) on computed tomography. METHODS: Preoperative computed tomography scans of patients with MPM were retrospectively assessed for numerous features including primary affected side, volume loss, pleural thickness, pleural calcifications, pleural effusion, and lymphadenopathy. RESULTS: One hundred twenty-five patients with MPM were included. Histologic subdivision was 97 eMPM (77%), 17 bMPM (14%), and 11 sMPM (9%). Nonepithelioid MPM (bMPM and sMPM) was more likely than eMPM to have calcified pleural plaques (P = 0.035). Analyzed separately, bMPM and sMPM each demonstrated calcified plaques more frequently than eMPM, and sMPM more often had internal mammary nodes; however, P values did not reach significance (P = 0.075 and 0.071, respectively). CONCLUSIONS: Calcified plaques are significantly more common in nonepithelioid subtypes compared with eMPM. Given the different prognoses and management of MPM subtypes, accurate noninvasive subtype classification is clinically vital.

Computed tomographic evaluation of the thymus-does obesity affect thymic fatty involution in a healthy young adult population?

OBJECTIVE: To determine a relationship between increased body mass index (BMI) and fatty involution of the thymus in subjects aged between 20 and 30 years. METHODS: CT images of 94 patients aged between 20 and 30 years were reviewed. Quantitative thymic mean attenuation was recorded and qualitative thymic attenuation was assigned to 1 of 4 possible grades. BMI and subcutaneous fat thickness were documented. Correlations between thymic attenuation, and BMI and subcutaneous fat thickness were assessed using linear regression models. Differences in thymic attenuation in overweight vs normal weight patients were assessed using t-test and Pearson Χ2 analysis. RESULTS: Low mean thymic attenuation values were associated with higher patient BMI (p = 0.024). Normal weight patients had a mean quantitative thymic attenuation of 15.5 Hounsfield unit and overweight patients had a mean quantitative thymic attenuation of -16.4 Hounsfield unit (p = 0.0218). There was a significant association between increasing subcutaneous fat thickness and reduced mean quantitative thymic attenuation (p < 0.0001). There was also a significant difference in subcutaneous fat thickness when comparing qualitatively assessed thymic Grade 0 with grades 2 and 3 (p = 0.027 and 0.001 respectively); and Grade 1 with Grade 3 (p = 0.001). CONCLUSION: In patients between 20 and 30 years old, the degree of thymic fatty infiltration is related to BMI. Advances in knowledge: Multidetector CT can assess fatty involution of the thymus gland. This retrospective study demonstrates a relationship between BMI and thymus gland fatty involution. Subjects with increased subcutaneous fat have decreased mean thymus gland attenuation.

Systemic chemokine levels in breast cancer patients and their relationship with circulating menstrual hormones.

INTRODUCTION: The chemokines Stromal Cell-Derived Factor-1alpha (SDF-1alpha/CXCL12) and Monocyte Chemotactic Protein-1 (MCP-1/CCL2) have been implicated in breast cancer progression. We recently reported elevated systemic MCP-1 in breast cancer patients. This study investigated circulating levels of SDF-1alpha in breast cancer patients, and addressed potential hormonal regulation of these two potent chemokines. METHODS: SDF-1alpha levels were determined by ELISA in 114 breast cancer patients and 85 controls, and correlated with clinical data. Blood samples were collected from 36 healthy premenopausal volunteers weekly for four weeks to measure Luteinising Hormone (LH), Follicular Stimulating Hormone (FSH), Oestradiol and Progesterone using a Bayer ADVIA Centaur Immunoassay system, in parallel with SDF-1alpha and MCP-1. CXCL12 expression was determined using RQ-PCR in primary tumour stromal cells (n = 16) harvested at surgery. RESULTS: Plasma SDF-1alpha was significantly higher in breast cancer patients than age-matched controls and had a significant correlation with tumour grade and epithelial subtype. Investigation of menstrual variations of these chemokines revealed lower SDF-1alpha levels in the mid-luteal phase of the menstrual cycle and a significant positive correlation with circulating Oestradiol. MCP-1 levels showed no correlation with menstrual hormones. There was a trend towards increased CXCL12 expression in tumour compared to normal stromal cells. CONCLUSIONS: The elevated level of SDF-1alpha detected in breast cancer patients, and it's correlation with prognostic indicators, highlights the importance of this chemokine in disease progression. Elucidation of factors influencing chemokine secretion supports clarification of their role in tumourigenesis.