A radiomics-based interpretable model to predict the pathological grade of pancreatic neuroendocrine tumors.

OBJECTIVES: To develop a computed tomography (CT) radiomics-based interpretable machine learning (ML) model to predict the pathological grade of pancreatic neuroendocrine tumors (pNETs) in a non-invasive manner. METHODS: Patients with pNETs who underwent contrast-enhanced abdominal CT between 2010 and 2022 were included in this retrospective study. Radiomics features were extracted, and five radiomics-based ML models, namely logistic regression (LR), random forest (RF), support vector machine (SVM), XGBoost, and GaussianNB, were developed. The performance of these models was evaluated using a time-independent testing set, and metrics such as sensitivity, specificity, accuracy, and the area under the receiver operating characteristic curve (AUC) were calculated. The accuracy of the radiomics model was compared to that of needle biopsy. The Shapley Additive Explanation (SHAP) tool and the correlation between radiomics and biological features were employed to explore the interpretability of the model. RESULTS: A total of 122 patients (mean age: 50 ± 14 years; 53 male) were included in the training set, whereas 100 patients (mean age: 48 ± 13 years; 50 male) were included in the testing set. The AUCs for LR, SVM, RF, XGBoost, and GaussianNB were 0.758, 0.742, 0.779, 0.744, and 0.745, respectively, with corresponding accuracies of 73.0%, 70.0%, 77.0%, 71.9%, and 72.9%. The SHAP tool identified two features of the venous phase as the most significant, which showed significant differences among the Ki-67 index or mitotic count subgroups (p < 0.001). CONCLUSIONS: An interpretable radiomics-based RF model can effectively differentiate between G1 and G2/3 of pNETs, demonstrating favorable interpretability. CLINICAL RELEVANCE STATEMENT: The radiomics-based interpretable model developed in this study has significant clinical relevance as it offers a non-invasive method for assessing the pathological grade of pancreatic neuroendocrine tumors and holds promise as an important complementary tool to traditional tissue biopsy. KEY POINTS: • A radiomics-based interpretable model was developed to predict the pathological grade of pNETs and compared with preoperative needle biopsy in terms of accuracy. • The model, based on CT radiomics, demonstrated favorable interpretability. • The radiomics model holds potential as a valuable complementary technique to preoperative needle biopsy; however, it should not be considered a replacement for biopsy.

Clinical outcome is distinct between radiological stricture and endoscopic stricture in ileal Crohn's disease.

OBJECTIVES: Differences in clinical adverse outcomes (CAO) based on different intestinal stricturing definitions in Crohn's disease (CD) are poorly documented. This study aims to compare CAO between radiological strictures (RS) and endoscopic strictures (ES) in ileal CD and explore the significance of upstream dilatation in RS. METHODS: This retrospective double-center study included 199 patients (derivation cohort, n = 157; validation cohort, n = 42) with bowel strictures who simultaneously underwent endoscopic and radiologic examinations. RS was defined as a luminal narrowing with wall thickening relative to the normal gut on cross-sectional imaging (group 1 (G1)), which further divided into G1a (without upstream dilatation) and G1b (with upstream dilatation). ES was defined as an endoscopic non-passable stricture (group 2 (G2)). Strictures met the definitions of RS (with or without upstream dilatation) and ES were categorized as group 3 (G3). CAO referred to stricture-related surgery or penetrating disease. RESULTS: In the derivation cohort, G1b (93.3%) had the highest CAO occurrence rate, followed by G3 (32.6%), G1a (3.2%), and G2 (0%) (p < 0.0001); the same order was found in the validation cohort. The CAO-free survival time was significantly different among the four groups (p < 0.0001). Upstream dilatation (hazard ratio, 1.126) was a risk factor for predicting CAO in RS. Furthermore, when upstream dilatation was added to diagnose RS, 17.6% of high-risk strictures were neglected. CONCLUSIONS: CAO differs significantly between RS and ES, and clinicians should pay more attention to strictures in G1b and G3. Upstream dilatation has an important impact on the clinical outcome of RS but may not be an essential factor for RS diagnosis. CLINICAL RELEVANCE STATEMENT: This study explored the definition of intestinal stricture with the greatest significance for the clinical diagnosis and prognosis of patients with CD, and consequently provided effective auxiliary information for clinicians to formulate strategies for the treatment of CD intestinal strictures. KEY POINTS: • The retrospective double-center study showed that clinical adverse outcome is different between radiological strictures and endoscopic strictures in CD. • Upstream dilatation has an important impact on the clinical outcome of radiological strictures but may not be an essential factor for diagnosis of radiological strictures. • Radiological stricture with upstream dilatation and simultaneous radiological and endoscopic stricture were at increased risk for clinical adverse outcomes; thus, closer monitoring should be considered.

CT-based peritumoral radiomics signatures to predict early recurrence in hepatocellular carcinoma after curative tumor resection or ablation.

OBJECTIVE: To construct a prediction model based on peritumoral radiomics signatures from CT images and investigate its efficiency in predicting early recurrence (ER) of hepatocellular carcinoma (HCC) after curative treatment. MATERIALS AND METHODS: In total, 156 patients with primary HCC were randomly divided into the training cohort (109 patients) and the validation cohort (47 patients). From the pretreatment CT images, we extracted 3-phase two-dimensional images from the largest cross-sectional area of the tumor. A region of interest (ROI) was manually delineated around the lesion for tumoral radiomics (T-RO) feature extraction, and another ROI was outlined with an additional 2 cm peritumoral area for peritumoral radiomics (PT-RO) feature extraction. The least absolute shrinkage and selection operator (LASSO) logistic regression model was applied for feature selection and model construction. The T-RO and PT-RO models were constructed. In the validation cohort, the prediction efficiencies of the two models and peritumoral enhancement (PT-E) were evaluated qualitatively by receiver operating characteristic (ROC) curves, calibration curves and decision curves and quantitatively by area under the curve (AUC), the category-free net reclassification index (cfNRI) and integrated discrimination improvement values (IDI). RESULTS: By comparing AUC values, the prediction accuracy in the validation cohort was good for the PT-RO model (0.80 vs. 0.79, P = 0.47) but poor for the T-RO model (0.82 vs. 0.62, P < 0.01), which was significantly overfitted. In the validation cohort, the ROC curves, calibration curves and decision curves indicated that the PT-RO model had better calibration efficiency and provided greater clinical benefits. CfNRI indicated that the PT-RO model correctly reclassified 47% of ER patients and 32% of non-ER patients compared to the T-RO model (P < 0.01); additionally, the PT-RO model correctly reclassified 24% of ER patients and 41% of non-ER patients compared to PT-E (P = 0.02). IDI indicated that the PT-RO model could improve prediction accuracy by 0.22 (P < 0.01) compared to the T-RO model and by 0.20 (P = 0.01) compared to PT-E. CONCLUSION: The CT-based PT-RO model can effectively predict the ER of HCC and is more efficient than the T-RO model and the conventional imaging feature PT-E.

Systemic Venous Drainage Is Associated with an Unfavorable Prenatal Behavior in Fetal Bronchopulmonary Sequestration.

OBJECTIVE: This study aimed to determine the significance of the venous drainage pattern of bronchopulmonary sequestration (BPS) in the prenatal course. METHODS: The venous drainage pattern of fetuses with BPS was determined with high-definition flow and confirmed by postnatal three-dimensional computed tomography angiography scan or autopsy. The volume of BPS lesions during gestation was recorded by the three-dimensional ultrasonographic Virtual Organ Computer-Aided Analysis software. The relationship between venous drainage pattern and prenatal characteristics was determined. RESULTS: Seventy-one fetuses were enrolled: 35 with systemic venous drainage (SVD) and 36 with pulmonary venous drainage (PVD). The volumes of BPS lesions significantly increased from the middle second trimester to the later second trimester in the SVD group. A marked decrease from the later second trimester to the third trimester was observed in the PVD group. The incidences of associated anomalies, hydrops, and polyhydramnios in the SVD group were 14.2, 23.3, and 33.3%, respectively, significantly higher than those in the PVD group (0, 0, and 5.6%, respectively). CONCLUSIONS: Our data indicate that SVD is correlated with a higher risk of associated anomalies and an unfavorable prenatal course in fetal BPS. Identification of the venous drainage pattern is of clinical significance in predicting the prenatal behavior of fetal BPS.

FOXK2, regulted by miR-1271-5p, promotes cell growth and indicates unfavorable prognosis in hepatocellular carcinoma.

The forkhead transcription factor FOXK2 has been implicated in the progression of human cancers, but its role and clinical significance in hepatocellular carcinoma (HCC) have not been explored. Here we showed that FOXK2 expression was increased and associated with tumor size, TNM stage and vascular invasion. High FOXK2 expression was correlated with poor overall and disease-free survival in two independent cohorts consisting of 864 patients with HCC. The prognostic value of FOXK2 was validated by stratified survival analyses in subgroups difined by factors contributing to worse survival. Multivariate Cox regression model revealed that FOXK2 served as an independent factor for overall survival. The FOXK2 expression was reversely connected with miR-1271-5p in clinical samples. Re-introduction of miR-1271 decreased FOXK2 at both mRNA and protein levels. Luciferase assay confirmed that FOXK2 was a direct target of miR-1271 in HCC cells. Overexpression of FOXK2 enhanced the cell growth and migration, whereas FOXK2 silence resulted in the opposite phenotypes. Further studies demonstrated that FOXK2 exerted oncogenic activity via activation of PI3K/AKT signaling pathway. Collectively, our data suggest FOXK2 as an oncogene and a promising prognostic biomarker in HCC. Targeting the newly identified miR-1271/FOXK2/AKT axis may represent a potential strategy for HCC intervention.

Diagnostic and post-treatment CT appearance of biopsy proven mixed Cryptococcus and Candida cholangitis.

Cryptococcus neoformans and Candida are common opportunistic pathogens of human. There is very limited literature on Cryptococcus neoformans cholangitis or Candida cholangitis in immunocompetent patient while mixed Cryptococcus neoformans and Candida cholangitis has not been reported in the literature ever before. We hereby report the imaging findings of a case of mixed Cryptococcus neoformans and Candida cholangitis in an immunocompetent boy. The CT features included nodules in the cystic duct and common bile duct, dilatation of the intra- and extrahepatic bile ducts with mural thickening, irregular hypodense mass-like lesion extending along the bile ducts from the liver hilum to the periphery which was confirmed by the presence of enlarged and confluent lymph nodes. CT characteristics can contribute to timely diagnosis and treatment of this disease.

[Value of computed tomography features for prognosis evaluation of gastrointestinal stromal tumor].

OBJECTIVE: To analyze the association between CT features and survival rate of GIST, and to elucidate the significance of CT features for prognosis. METHODS: Clinical data of 38 patients with pathologically and immunohistochemically proven GISTs, including 11 patients at high biological risk, 13 at moderate risk, 10 at low risk and 4 at very low risk. Patients who underwent CT examination for primary tumors were included. Association between CT features and survival rate was examined. RESULTS: The mean follow-up period of 38 cases was 42.6 months and the 3-year survival rate was 86.8%. Univariate analysis revealed that tumor growth pattern, diameter, lobulated shape, irregular margin, necrosis, ulceration, adjacent invasion, and liver metastasis were associated factors of 3-year survival rate. Circumference invasion and hepatic metastases predicted poor 3-year survival rate (P<0.05). Calcification and intensity were not associated with prognosis (P>0.05). CONCLUSIONS: CT can demonstrate the tumor growth pattern, size, shape, boundary, density, necrosis, hemorrhage, calcification, ulcer, enhance features and metastasis. CT can play an important role in estimating the survival rate of GIST.

Quantitative assessment of perirectal tumor infiltration with dynamic contrast-enhanced multi-detector row CT in rectal cancer.

OBJECTIVE: To evaluate the efficacy of discriminant function analysis of perirectal tumor infiltration with dynamic contrast-enhanced 64-detector row CT in rectal cancer. MATERIALS AND METHODS: Forty-nine patients with rectal cancer underwent dynamic contrast-enhanced CT. A total of 96 axial CT slices containing the tumors were evaluated. The 96 images were separated into two groups with or without perirectal tumor infiltration based on pathological findings. The discriminant function was set-up using CT density differences between the mass and the adjacent perirectal tissue within 5 mm from the mass at 20 and 40 s as independent variables. The results of the discriminant function analysis were compared to those of CT morphology and pathology. RESULTS: CT morphological diagnosis was accurate on 71.9% (69/96) of the slices with 82.5% sensitivity and 64.3% specificity. Discriminant function analysis correctly identified 88.5% (85/96) of the slices with 85.0% sensitivity and 91.1% specificity. Overstaging occurred significantly more (P<0.05) on morphological analysis (20.8%, 20/96) than discriminant function analysis (5.2%, 5/96) of the CT slices. CONCLUSIONS: Discriminant function analysis of dynamic contrast-enhanced CT improves the diagnostic accuracy and specificity of perirectal tumor infiltration in rectal cancer.

[Application of 64MDCT multislice perfusion imaging in colorectal carcinoma].

OBJECTIVE: To study the correlation of time-density curves (TDC), parameters revealed by 64-multidetector-row CT (64MDCT) perfusion imaging with clinicopathological factors (staging, serosal invasion, lymph node metastasis, distant metastasis and CEA) in colorectal carcinoma (CRC). METHODS: 64 MDCT perfusion imaging was performed in 33 patients with pathologically verified CRC. TDC was created from the region of interest (ROI) drawn over the tumor, target artery and vein by 64MDCT with perfusion functional software. The parameters of individual perfusion maps included blood flow (BF), blood volume (BV), mean transit time (MTT) and permeability-surface area product (PS). Tumors were staged according to TMN classification. TDC was classified according to their shapes. The correlation between CT perfusion parameters and clinicopathological factors was studied. RESULTS: TDC of 64MDCT perfusion imaging could be classified into five types. TDC in different stages could demonstrate one or more types of the five types. There was no significant difference of CT perfusion parameters among different stages. BV and MTT were significantly higher in the patients with serosal invasion than in those without serosal invasion (t=-2.63,-2.24, P=0.0137, 0.0331, respectively). BV was significantly correlated with tumor size (r=0.41, P=0.02). BF and PS were not correlated with staging, serosal invasion, lymph node metastasis, distant place metastasis and CEA (all P>0.05). CONCLUSIONS: 64MDCT multislice perfusion imaging can reveal the blood perfusion of CRC and has potential value of clinical application.

Multiphasic MDCT in small bowel volvulus.

OBJECTIVE: Evaluate the use of MDCT with 3D CT angiography (CTA) and CT portal venography (CTPV) reconstruction for the diagnosis of small bowel volvulus (SBV). METHODS: Multiphasic MDCT findings in nine patients (seven males and two females, age range 2-70) with surgically proven SBV were retrospectively reviewed. Non-contrast and double phase contrast enhanced MDCT including 3D CTA and CTPV reconstruction were performed in all the patients. Two experienced abdominal radiologists evaluated the images and defined the location, direction and degree of SBV. RESULTS: On axial MDCT images, all cases show segmental or global dilatation of small intestine. Other findings include circumferential bowel wall thickening in eight cases, halo appearance and hyperemia in seven cases, whirl sign in six cases, beak-like appearance in six cases, closed loops in six cases and ascites in one case. CTA/CTPV showed abnormal courses involving main trunks of superior mesenteric artery (SMA) and superior mesenteric vein (SMV) in seven cases, with or without distortion of their tributaries. Normal course of SMA but abnormal course of SMV was seen in the other two cases. Of all the nine cases, whirl sign was seen in six cases and barber's pole sign in five cases. Dilated SMV was observed in eight cases and abrupt termination of SMA was found in one case. Compared with surgical findings, the location, direction and degree of SBV were correctly estimated in all cases based on CTA/CTPV. CONCLUSION: Multiphasic MDCT with CTA/CTPV reconstruction can play an important role in the diagnosis of SBV. The location, direction and degree of SBV can all be defined preoperatively using this method.

Evaluation of angiogenesis in colorectal carcinoma with multidetector-row CT multislice perfusion imaging.

To evaluate the correlation between 64 multidetector-row CT (64MDCT) perfusion imaging in colorectal carcinoma and microvessel density (MVD) and vascular endothelial growth factor (VEGF), 64MDCT perfusion imaging was performed in 33 patients with pathologically verified colorectal carcinoma. These images were analyzed with perfusion functional software, and time-density curves (TDC) were created for the region of interest (ROI) encompassing the tumor, the target artery and vein. The individual perfusion maps generated indicated blood flow (BF), blood volume (BV), mean transit time (MTT), and permeability-surface area product (PS). MVD and VEGF were evaluated by immunohistochemical staining with anti-CD34 and anti-VEGF, respectively. Correlations between MVD or VEGF with CT perfusion parameters and clinicopathological factors (Dukes' stages, invasion depth, and lymph node and liver metastasis) were also investigated. MVD in the colorectal carcinoma was 22.61+/-9.01 per x200 field. The scores obtained for VEGF expression were 4.15+/-1.09. VEGF staining was positive in 25 of 29 tumors (86.2%). There was no significant correlation between the presence of MVD, VEGF expression and clinicopathological factors (P>0.05). There was also no correlation between MVD, VEGF expression, and any dynamic CT parameters (P>0.05). The BV and MTT were significantly higher in tumors demonstrating serous coat invasion than in those without it (t=-2.63, -2.24, P=0.0137, 0.0331, respectively). BV was also significantly correlated with tumor size (r=0.41, P=0.02). Neither BF nor PS was correlated with clinicopathological factors. In conclusion, 64MDCT perfusion imaging, MVD, and VEGF may reflect angiogenic activity, but no significant correlation among these factors.

[Evaluation of microvessel density and vascular endothelial growth factor in colorectal carcinoma with 64-multidetector-row CT perfusion imaging].

OBJECTIVE: To evaluate the correlation of 64-multidetector-row CT (64MDCT) perfusion imaging with microvessel density(MVD) and vascular endothelial growth factor(VEGF) in colorectal carcinoma. METHODS: 64MDCT perfusion imaging was performed in 33 patients with pathologically verified colorectal carcinoma. Time-density curves (TDC) were created from the region of interest (ROI) drawn over the tumor, target artery and vein by 64MDCT with perfusion functional software. The individual perfusion maps generated were for blood flow (BF), blood volume (BV), mean transit time (MTT) and permeability-surface area product (PS). MVD and VEGF expression of surgical specimens were examined by immunohistochemical staining with anti-CD34, anti-VEGF monoclonal antibody. MVD and VEGF were compared among the different types of TDC in colorectal carcinoma. The correlation of CT perfusion parameters with MVD and VEGF was also examined. RESULTS: TDC of colorectal carcinoma was divided into five types according to their shapes. MVD in the colorectal carcinoma was 22.61+/-9.01. VEGF staining was found in 25 of 29 tumors (86.2%). The score of VEGF expression was 4.15+/-1.09. No significant differences of MVD and VEGF expression among TDC types were found (F=2.59, 1.11, P>0.05). There were also no correlations of MVD and VEGF expression with any dynamic CT parameters (P>0.05). CONCLUSION: 64MDCT perfusion imaging, MVD and VEGF may reflect angiogenic activity, but no significant correlations are found among them.