Gastric adenocarcinoma is mostly locally advanced at diagnosis. An appropriate pretherapy clinical staging is essential, with assistance from contrast-enhanced CT. In primary staging, CT is the first option for tumor (T), nodal (N), and metastatic (M) staging. One of the most relevant radiologic challenges is to reach an accurate TNM staging and reduce the risk of under- or overstaging. The CT benefits for local staging are linked to an appropriate protocol, primarily based on fasting, gastric lumen distention, and administration of intravenous contrast material. It is necessary for radiologists to have a deep knowledge of the macroscopic and microscopic anatomy of the stomach to assess clinical T staging and be aware of limited CT accuracy in the differentiation between stages T1 and T2, and the mismatch between histologic and radiologic multilayer visualization. Regarding N and M staging, the relevance of CT has high variability. For N staging, standardized criteria have not been approved yet, and for M staging, the variability mostly is dependent on the localization of metastatic disease. In peritoneal involvement, laparoscopic examination currently is recommended in doubtful and high-risk patients. Upfront surgery is the first option only in stage Ia, macrosatellite instability patients, or cases of emergency. Perioperative chemotherapy currently is recommended for stage Ib or higher with microsatellite stability status. According to these therapeutic options, pretherapy clinical staging plays an important part for clinicians to manage patients properly. ©RSNA, 2025 Supplemental material is available for this article.

Clinical staging on CT has several biases, and a radiogenomics approach could be proposed. The study aimed to test the performance of a radiogenomics approach in identifying high-risk colon cancer. ATTRACT is a multicentric trial, registered in ClinicalTrials.gov (NCT06108310). Three hundred non-metastatic colon cancer patients were retrospectively enrolled and divided into two groups, high-risk and no-risk, according to the pathological staging. Radiological evaluations were performed by two abdominal radiologists. For 151 patients, we achieved genomics. The baseline CT scans were used to evaluate the radiological assessment and to perform 3D cancer segmentation. One expert radiologist used open-source software to perform the volumetric cancer segmentations on baseline CT scans in the portal phase (3DSlicer v4.10.2). Implementing the classical LASSO with a machine-learning library method was used to select the optimal features to build Model 1 (clinical-radiological plus radiomic feature, 300 patients) and Model 2 (Model 1 plus genomics, 151 patients). The performance of clinical-radiological interpretation was assessed regarding the area under the curve (AUC), sensitivity, specificity, and accuracy. The average performance of Models 1 and 2 was also calculated. In total, 262/300 were classified as high-risk and 38/300 as no-risk. Clinical-radiological interpretation by the two radiologists achieved an AUC of 0.58–0.82 (95

Magnetic resonance elastography (MRE) is an imaging technique that combines low-frequency mechanical vibrations with magnetic resonance imaging to create visual maps and quantify liver parenchyma stiffness. As in recent years, diffuse liver diseases have become highly prevalent worldwide and could lead to a chronic condition with different stages of fibrosis. There is a strong necessity for a non-invasive, highly accurate, and standardised quantitative assessment to evaluate and manage patients with different stages of fibrosis from diagnosis to follow-up, as the actual reference standard for the diagnosis and staging of liver fibrosis is biopsy, an invasive method with possible peri-procedural complications and sampling errors. MRE could quantitatively evaluate liver stiffness, as it is a rapid and repeatable method with high specificity and sensitivity. MRE is based on the propagation of mechanical shear waves through the liver tissue that are directly proportional to the organ's stiffness, expressed in kilopascals (kPa). To obtain a valid assessment of the real hepatic stiffness values, it is mandatory to obtain a high-quality examination. To understand the pearls and pitfalls of MRE, in this review, we describe our experience after one year of performing MRE from indications and patient preparation to acquisition, quality control, and image analysis.

The aim of this study was to compare CT radiomics and morphological features when assessing benign lymph nodes (LNs) in colon cancer (CC). This retrospective study included 100 CC patients (test cohort) who underwent a preoperative CT examination and were diagnosed as pN0 after surgery. Regional LNs were scored with a morphological Likert scale (NODE-SCORE) and divided into two groups: low likelihood (LLM: 0–2 points) and high likelihood (HLM: 3–7 points) of malignancy. The T-test and the Mann–Whitney test were used to compare 107 radiomic features extracted from the two groups. Radiomic features were also extracted from primary lesions (PLs), and the receiver operating characteristic (ROC) was used to test a LN/PL ratio when assessing the LN’s status identified with radiomics and with the NODE-SCORE. An amount of 337 LNs were divided into 167 with LLM and 170 with HLM. Radiomics showed 15/107 features, with a significant difference (p < 0.02) between the two groups. The comparison of selected features between 81 PLs and the corresponding LNs showed all significant differences (p < 0.0001). According to the LN/PL ratio, the selected features recognized a higher number of LNs than the NODE-SCORE (p < 0.001). On validation of the cohort of 20 patients (10 pN0, 10 pN2), significant ROC curves were obtained for LN/PL busyness (AUC = 0.91; 0.69–0.99; 95% C.I.; and p < 0.001) and for LN/PL dependence entropy (AUC = 0.76; 0.52–0.92; 95% C.I.; and p = 0.03). The radiomics ratio between CC and LNs is more accurate for noninvasively discriminating benign LNs compared to CT morphological features.

Background: Neuroendocrine neoplasms (NENs) are slow-growing tumors. Sarcopenia is defined as the loss of muscle mass, strength, and physical performance. First-line NEN therapy is somatostatin analogs, which could be responsible for malabsorption conditions, such as pancreatic exocrine insufficiency (EPI) with underlying sarcopenia. Aim: Evaluate the prevalence of sarcopenia in patients with NENs at diagnosis and during follow-up. Methods: A retrospective single-center study was conducted, including patients with advanced intestinal NENs G1/G2 (excluded pancreatic NENs). CT scans were analyzed at diagnosis and after 6 months of therapy, and the skeletal muscle index was assessed. Results: A total of 30 patients (F:M = 6:24) were enrolled, with the following primary tumor sites: 25 in the ileum, 1 stomach, 2 jejunum, and 2 duodenum. At diagnosis, 20 patients (66.6%) showed sarcopenic SMI values, and 10 patients (33.3%) showed non-sarcopenic SMI values. At follow-up, three more patients developed sarcopenic SMI values. Statistical significance in relation to the presence of sarcopenia was found in the group of patients with carcinoid syndrome (p = 0.0178), EPI (p = 0.0018), and weight loss (p = 0.0001). Conclusion: Sarcopenia was present in 2/3 of the patients with advanced intestinal NENs at the diagnosis and during the follow-up. It is reasonable to consider this condition to improve clinical outcomes.

To evaluate a radiomic strategy for predicting progression in advanced gastroenteropancreatic neuroendocrine tumor (GEP-NET) patients treated with somatostatin analogs (SSAs). Fifty-eight patients with GEP-NETs and liver metastases, with baseline computerized tomography (CT) scans from June 2013 to November 2020, were studied retrospectively. Data collected included progression-free survival (PFS), overall survival (OS), tumor grading, death, and Ki67 index. Patients were categorized into progressive and non-progressive groups. Two radiologists performed 3D liver segmentation on baseline CT scans using 3DSlicer v4.10.2. One hundred six radiomic features were extracted and analyzed (T-test or Mann-Whitney). Radiomic feature efficacy was evaluated via receiver operating characteristic curves, and both univariate and multivariate logistic regression were used to develop predictive models. A significance level of p < .05 was maintained. Of 55 patients, 38 were progressive (median PFS and OS: 14 and 34 months, respectively), and 17 were non-progressive (median PFS and OS: 58 months each). Six radiomic features significantly differed between groups (p < .05), with an area under the curve (AUC) range of 0.64-0.74. Ki67 was the only clinical parameter significantly associated with progression risk (odds ratio (OR) = 1.14, p < .05). The combined radiomic features and Ki67 model proved most effective, showing an AUC of 0.814 (p = .008). The radiomic model alone did not reach statistical significance (p = .07). A combined model incorporating radiomic features and the Ki67 index effectively predicts disease progression in GEP-NET patients eligible for SSA treatment.

Different conditions may underlie gastrointestinal bleeding (GIB) in children. The estimated prevalence of GIB in children is 6.4%, with spontaneous resolution in approximately 80% of cases. Nonetheless, the initial approach plays a pivotal role in determining the prognosis. The priority is the stabilization of hemodynamic status, followed by a systematic diagnostic approach. GIB can originate from either upper or lower gastrointestinal tract, leading to a broad differential diagnosis in infants and children. This includes benign and self-limiting disorders, alongside serious conditions necessitating immediate treatment. We performed a nonsystematic review of the literature, in order to describe the variety of conditions responsible for GIB in pediatric patients and to outline diagnostic pathways according to patients’ age, suspected site of bleeding and type of bleeding which can help pediatricians in clinical practice. Diagnostic modalities may include esophagogastroduodenoscopy and colonoscopy, abdominal ultrasonography or computed tomography and, when necessary, magnetic resonance imaging. In this review, we critically assess these procedures, emphasizing their respective advantages and limitations concerning specific clinical scenarios.

In mCRC patient with resectable metastases, the radical surgery represents the only curative treatment option, also locoregional treatments showed a survival advantage, but long-term outcomes remain disappointing. Although "pseudo-adjuvant" CT in resected mCRC is currently encouraged by the international guidelines, this treatment is still debated, and its efficacy remains controversial. We conducted a retrospective analysis of mCRC patients treated with surgery of primary tumor and surgical and/or locoregional treatment of metastases at Sant 'Andrea Hospital of Rome, from January 2002 to April 2022. The aim of the study was to evaluate a survival advantage from the addition of sequential systemic CT to surgical and/or locoregional treatment and investigate the correlation between clinical-pathological factors and treatments with survival outcomes. A total of 63 patients were evaluated, the 57% (N=36) was male and the median age was 63 years. The primary tumor site was right, transverse and left colon in 27% (N=17), 4.8% (N=3) and 36.5% (N=23) of the patients respectively, while in the 31.7% (N=20) was rectum. The sites of treated metastases were liver, lung, lymph nodes and other organs in 60.3% (N=38), 27% (N=17), 1.6% (N=1) and 6.3% (N=4) of cases, respectively. The 19% presented mucinous histology, the 86.5% mutation in RAS gene and 27.9% in BRAF. The 88.9% of patients underwent surgery on metastases, the 3.2% locoregional treatment and 7.9% both. The "pseudo-adjuvant" CT was poly-CT 5-Fluorouracil (5-FU) based in 39.6% (N=25), 5-FU single agent CT in 27% (N=17), poly-CT plus biological agent according to RAS status in 25,4% (N=16) and 5-FU plus biological agent in 8% (N=5) of cases. The univariate analysis for disease free survival (DFS) showed a statistically significant correlation with nodal disease involvement (N1) [HR=2.93 (1.28-6.68); p-value = 0.010] and with the wild-type status of KRAS [HR = 0.35 (0.16-0.79); p-value = 0.012] with a median (m) DFS in KRAS WT patients of 17 months versus 11 months for KRAS mt (p-value = 0.007). A correlation statistically significant was reported between "pseudo-adjuvant" 5-FU based poly-CT and DFS [HR = 1.35 (1.0-1.84); p-value = 0.05] with a mDFS of 16 months (CI 95% 11-20 months) in CT group versus 14 months (IC 95% 6-21 months) in CT plus biological agent group (p-value = 0.042) In terms of overall survival (OS) this correlation was confirmed [HR = 1.40 (1.01-1.96); p-value = 0.046], with a mOS of 67 months (IC 95% 56-77 months) versus 43 months (IC 95% 24-61 months) (p-value = 0.041). Our retrospective and monocentric study confirmed literature data, showing a better DFS and OS in patients treated with "pseudo-adjuvant" CT after surgical and/or locoregional treatment of mCRC, without no further benefit with the addition of biological agents. The management of patients with resected mCRC appear difficult and should always be discussed in a multidisciplinary expert team, to offer the best effective treatment and improve survival outcomes.

Objectives To perform a comprehensive within-subject image quality analysis of abdominal CT examinations reconstructed with DLIR and to evaluate diagnostic accuracy compared to the routinely applied adaptive statistical iterative reconstruction (ASiR-V) algorithm. Materials and methods Oncologic patients were prospectively enrolled and underwent contrast-enhanced CT. Images were reconstructed with DLIR with three intensity levels of reconstruction (high, medium, and low) and ASiR-V at strength levels from 10 to 100% with a 10% interval. Three radiologists characterized the lesions and two readers assessed diagnostic accuracy and calculated signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), figure of merit (FOM), and subjective image quality, the latter with a 5-point Likert scale. Results Fifty patients (mean age: 70 ± 10 years, 23 men) were enrolled and 130 liver lesions (105 benign lesions, 25 metastases) were identified. DLIR_H achieved the highest SNR and CNR, comparable to ASiR-V 100% ( p ≥ .051). DLIR_M returned the highest subjective image quality (score: 5; IQR: 4–5; p ≤ .001) and significant median increase (29%) in FOM ( p < .001). Differences in detection were identified only for lesions ≤ 0.5 cm: 32/33 lesions were detected with DLIR_M and 26 lesions were detected with ASiR-V 50% ( p = .031). Lesion accuracy of was 93.8% (95% CI: 88.1, 97.3; 122 of 130 lesions) for DLIR and 87.7% (95% CI: 80.8, 92.8; 114 of 130 lesions) for ASiR-V 50%. Conclusions DLIR yields superior image quality and provides higher diagnostic accuracy compared to ASiR-V in the assessment of hypovascular liver lesions, in particular for lesions ≤ 0.5 cm. Clinical relevance statement Deep learning image reconstruction algorithm demonstrates higher diagnostic accuracy compared to iterative reconstruction in the identification of hypovascular liver lesions, especially for lesions ≤ 0.5 cm. Key Points • Iterative reconstruction algorithm impacts image texture, with negative effects on diagnostic capabilities. • Medium-strength deep learning image reconstruction algorithm outperforms iterative reconstruction in the diagnostic accuracy of ≤ 0.5 cm hypovascular liver lesions (93.9% vs 78.8%), also granting higher objective and subjective image quality. • Deep learning image reconstruction algorithm can be safely implemented in routine abdominal CT protocols in place of iterative reconstruction.

Domenico De Santis, Antonella Del Gaudio, Marta Zerunian, Michela Polici, Gisella Guido, Mariarita Tarallo, Benedetta Masci, Nicolò Ubaldi, Elsa Iannicelli, Andrea Laghi, Damiano Caruso

BACKGROUND:preoperative risk assessment of gastrointestinal stromal tumors (GISTS) is required for optimal and personalized treatment planning. Radiomics features are promising tools to predict risk assessment. The purpose of this study is to develop and validate an artificial intelligence classification algorithm, based on CT features, to define GIST's prognosis as determined by the Miettinen classification.METHODS:patients with histological diagnosis of GIST and CT studies were retrospectively enrolled. Eight morphologic and 30 texture CT features were extracted from each tumor and combined to obtain three models (morphologic, texture and combined). Data were analyzed using a machine learning classification (WEKA). For each classification process, sensitivity, specificity, accuracy and area under the curve were evaluated. Inter- and intra-reader agreement were also calculated.RESULTS:52 patients were evaluated. In the validation population, highest performances were obtained by the combined model (SE 85.7%, SP 90.9%, ACC 88.8%, and AUC 0.954) followed by the morphologic (SE 66.6%, SP 81.8%, ACC 76.4%, and AUC 0.742) and texture (SE 50%, SP 72.7%, ACC 64.7%, and AUC 0.613) models. Reproducibility was high of all manual evaluations.CONCLUSIONS:the AI-based radiomics model using a CT feature demonstrates good predictive performance for preoperative risk stratification of GISTs.

Objective: The aim of this study was to develop and validate a decision support model using data mining algorithms, based on morphologic features derived from MRI images, to discriminate between complete responders (CR) and non-complete responders (NCR) patients after neoadjuvant chemoradiotherapy (CRT), in a population of patients with locally advanced rectal cancer (LARC). Methods: Two populations were retrospectively enrolled: group A (65 patients) was used to train a data mining decision tree algorithm whereas group B (30 patients) was used to validate it. All patients underwent surgery; according to the histology evaluation, patients were divided in CR and NCR. Staging and restaging MRI examinations were retrospectively analysed and seven parameters were considered for data mining classification. Five different classification methods were tested and evaluated in terms of sensitivity, specificity, accuracy and AUC in order to identify the classification model able to achieve the best performance. The best classification algorithm was subsequently applied to group B for validation: sensitivity, specificity, positive and negative predictive value, accuracy and ROC curve were calculated. Inter and intra-reader agreement were calculated. Results: Four features were selected for the development of the classification algorithm: MRI tumor regression grade (MRTRG), staging volume (SV), tumor volume reduction rate (TVRR) and signal intensity reduction rate (SIRR). The decision tree J48 showed the highest efficiency: when applied to group B, all the CR and 18/21 NCR were correctly classified (sensitivity 85.71%, specificity 100%, PPV 100%, NPV 94.2%, accuracy 95.7%, AUC 0.833). Both inter- and intra-reader evaluation showed good agreement (Kappa > 0.6). Conclusions: The proposed decision support model may help in distinguishing between CR and NCR patients with LARC after CRT.

Background: To evaluate the diagnostic performance of a Machine Learning (ML) algorithm based on Texture Analysis (TA) parameters in the prediction of Pathological Complete Response (pCR) to Neoadjuvant Chemoradiotherapy (nChRT) in Locally Advanced Rectal Cancer (LARC) patients. Methods: LARC patients were prospectively enrolled to undergo pre- and post-nChRT 3T MRI for initial loco-regional staging. TA was performed on axial T2-Weighted Images (T2-WI) to extract specific parameters, including skewness, kurtosis, entropy, and mean of positive pixels. For the assessment of TA parameter diagnostic performance, all patients underwent complete surgical resection, which served as a reference standard. ROC curve analysis was carried out to determine the discriminatory accuracy of each quantitative TA parameter to predict pCR. A ML-based decisional tree was implemented combining all TA parameters in order to improve diagnostic accuracy. Results: Forty patients were considered for final study population. Entropy, kurtosis and MPP showed statistically significant differences before and after nChRT in patients with pCR; in particular, when patients with Pathological Partial Response (pPR) and/or Pathological Non-Response (pNR) were considered, entropy and skewness showed significant differences before and after nChRT (all p < 0.05). In terms of absolute value changes, pre- and post-nChRT entropy, and kurtosis showed significant differences (0.31 ± 0.35, in pCR, −0.02 ± 1.28 in pPR/pNR, (p = 0.04); 1.87 ± 2.19, in pCR, −0.06 ± 3.78 in pPR/pNR (p = 0.0005); 107.91 ± 274.40, in pCR, −28.33 ± 202.91 in pPR/pNR, (p = 0.004), respectively). According to ROC curve analysis, pre-treatment kurtosis with an optimal cut-off value of ≤3.29 was defined as the best discriminative parameter, resulting in a sensitivity and specificity in predicting pCR of 81.5% and 61.5%, respectively. Conclusions: TA parameters extracted from T2-WI MRI images could play a key role as imaging biomarkers in the prediction of response to nChRT in LARC patients. ML algorithms can be used to efficiently combine all TA parameters in order to improve diagnostic accuracy.

The study was aimed to develop a radiomic model able to identify high-risk colon cancer by analyzing pre-operative CT scans. The study population comprised 148 patients: 108 with non-metastatic colon cancer were retrospectively enrolled from January 2015 to June 2020, and 40 patients were used as the external validation cohort. The population was divided into two groups—High-risk and No-risk—following the presence of at least one high-risk clinical factor. All patients had baseline CT scans, and 3D cancer segmentation was performed on the portal phase by two expert radiologists using open-source software (3DSlicer v4.10.2). Among the 107 radiomic features extracted, stable features were selected to evaluate the inter-class correlation (ICC) (cut-off ICC > 0.8). Stable features were compared between the two groups (T-test or Mann–Whitney), and the significant features were selected for univariate and multivariate logistic regression to build a predictive radiomic model. The radiomic model was then validated with an external cohort. In total, 58/108 were classified as High-risk and 50/108 as No-risk. A total of 35 radiomic features were stable (0.81 ≤ ICC < 0.92). Among these, 28 features were significantly different between the two groups (p < 0.05), and only 9 features were selected to build the radiomic model. The radiomic model yielded an AUC of 0.73 in the internal cohort and 0.75 in the external cohort. In conclusion, the radiomic model could be seen as a performant, non-invasive imaging tool to properly stratify colon cancers with high-risk disease.

To test radiomic approach in patients with metastatic neuroendocrine tumors (NETs) treated with Everolimus, with the aim to predict progression-free survival (PFS) and death. Twenty-five patients with metastatic neuroendocrine tumors, 15/25 pancreatic (60%), 9/25 ileal (36%), 1/25 lung (4%), were retrospectively enrolled between August 2013 and December 2020. All patients underwent contrast-enhanced CT before starting Everolimus, histological diagnosis, tumor grading, PFS, overall survival (OS), death, and clinical data collected. Population was divided into two groups: responders (PFS ≤ 11 months) and non-responders (PFS > 11 months). 3D segmentation was performed on whole liver of naïve CT scans in arterial and venous phases, using a dedicated software (3DSlicer v4.10.2). A total of 107 radiomic features were extracted and compared between two groups (T test or Mann–Whitney), radiomics performance assessed with receiver operating characteristic curve, Kaplan–Meyer curves used for survival analysis, univariate and multivariate logistic regression performed to predict death, and interobserver variability assessed. All significant radiomic comparisons were validated by using a synthetic external cohort. P < 0.05 is considered significant. 15/25 patients were classified as responders (median PFS 25 months and OS 29 months) and 10/25 as non-responders (median PFS 4.5 months and OS 23 months). Among radiomic parameters, Correlation and Imc1 showed significant differences between two groups (P < 0.05) with the best performance (internal cohort AUC 0.86–0.84, P < 0.0001; external cohort AUC 0.84–0.90; P < 0.0001). Correlation < 0.21 resulted correlated with death at Kaplan–Meyer analysis (P = 0.02). Univariate analysis showed three radiomic features independently correlated with death, and in multivariate analysis radiomic model showed good performance with AUC 0.87, sensitivity 100%, and specificity 66.7%. Three features achieved 0.77 ≤ ICC < 0.83 and one ICC = 0.92. In patients affected by metastatic NETs eligible for Everolimus treatment, radiomics could be used as imaging biomarker able to predict PFS and death.

To compare liver MRI with AIR Recon Deep Learning™(ARDL) algorithm applied and turned-off (NON-DL) with conventional high-resolution acquisition (NAÏVE) sequences, in terms of quantitative and qualitative image analysis and scanning time. This prospective study included fifty consecutive volunteers (31 female, mean age 55.5 ± 20 years) from September to November 2021. 1.5 T MRI was performed and included three sets of images: axial single-shot fast spin-echo (SSFSE) T2 images, diffusion-weighted images(DWI) and apparent diffusion coefficient(ADC) maps acquired with both ARDL and NAÏVE protocol; the NON-DL images, were also assessed. Two radiologists in consensus drew fixed regions of interest in liver parenchyma to calculate signal-to-noise-ratio (SNR) and contrast to-noise-ratio (CNR). Subjective image quality was assessed by two other radiologists independently with a five-point Likert scale. Acquisition time was recorded. SSFSE T2 objective analysis showed higher SNR and CNR for ARDL vs NAÏVE, ARDL vs NON-DL(all P < 0.013). Regarding DWI, no differences were found for SNR with ARDL vs NAÏVE and, ARDL vs NON-DL (all P > 0.2517).CNR was higher for ARDL vs NON-DL(P = 0.0170), whereas no differences were found between ARDL and NAÏVE(P = 1). No differences were observed for all three comparisons, in terms of SNR and CNR, for ADC maps (all P > 0.32). Qualitative analysis for all sequences showed better overall image quality for ARDL with lower truncation artifacts, higher sharpness and contrast (all P < 0.0070) with excellent inter-rater agreement (k ≥ 0.8143). Acquisition time was lower in ARDL sequences compared to NAÏVE (SSFSE T2 = 19.08 ± 2.5 s vs. 24.1 ± 2 s and DWI = 207.3 ± 54 s vs. 513.6 ± 98.6 s, all P < 0.0001). ARDL applied on upper abdomen showed overall better image quality and reduced scanning time compared with NAÏVE protocol.

Objective Myelofibrosis is a rare chronic myelolymphoproliferative disease and is associated with increased risk of venous thromboembolism. The objective of this study is to retrospectively evaluate patients with primary myelofibrosis who underwent abdominal US, MDCT and MRI, in order to identify the development of portal thrombosis and its correlation with portal-biliary cavernoma. Methods We evaluated 125 patients with initial diagnosis of primary myelofibrosis and nonspecific abdominal pain who had undergone US with color Doppler. In 13 patients (8 men, 5 females; age: 45–85), US detected portal thrombosis with associated portal-biliary cavernoma. All patients subsequently underwent contrast-enhanced MDCT and MRI and 4 patients MR-cholangiography. The correlation between primary myelofibrosis and portal thrombosis and cavernoma respectively was calculated using χ2 test. Results About 10% of patients with primary myelofibrosis preliminary evaluated with US had partial (8 pts) or complete (5 pts) portal thrombosis associated with portal-biliary cavernoma with a χ2 = 0. In all patients, US detected a concentric thickening of main bile duct (MBD) wall (mean value: 7 mm); color Doppler always showed dilated venous vessels within the thickened wall of the biliary tract. Contrast-enhanced CT and MRI confirmed thickening of MBD walls with their progressive enhancement and allowed better assessment of the extent of the portal system thrombosis. MR-cholangiography showed a thin appearance of the MBD lumen with evidence of ab extrinsic compression. Conclusions The evidence of portal thrombosis and portal-biliary cavernoma in 10% of the patients with primary myelofibrosis indicates a close correlation between the two diseases. In the detection of portal thrombosis and portal-biliary cavernoma, US with color Doppler is the most reliable and economical diagnostic technique while contrast-enhanced MDCT and MRI allow better assessment of the extent of the portal vein thrombosis and of the complications of myelofibrosis.