Artificial intelligence-based, semi-automated segmentation for the extraction of ultrasound-derived radiomics features in breast cancer: a prospective multicenter study.

PURPOSE: To investigate the feasibility of an artificial intelligence (AI)-based semi-automated segmentation for the extraction of ultrasound (US)-derived radiomics features in the characterization of focal breast lesions (FBLs). MATERIAL AND METHODS: Two expert radiologists classified according to US BI-RADS criteria 352 FBLs detected in 352 patients (237 at Center A and 115 at Center B). An AI-based semi-automated segmentation was used to build a machine learning (ML) model on the basis of B-mode US of 237 images (center A) and then validated on an external cohort of B-mode US images of 115 patients (Center B). RESULTS: A total of 202 of 352 (57.4%) FBLs were benign, and 150 of 352 (42.6%) were malignant. The AI-based semi-automated segmentation achieved a success rate of 95.7% for one reviewer and 96% for the other, without significant difference (p = 0.839). A total of 15 (4.3%) and 14 (4%) of 352 semi-automated segmentations were not accepted due to posterior acoustic shadowing at B-Mode US and 13 and 10 of them corresponded to malignant lesions, respectively. In the validation cohort, the characterization made by the expert radiologist yielded values of sensitivity, specificity, PPV and NPV of 0.933, 0.9, 0.857, 0.955, respectively. The ML model obtained values of sensitivity, specificity, PPV and NPV of 0.544, 0.6, 0.416, 0.628, respectively. The combined assessment of radiologists and ML model yielded values of sensitivity, specificity, PPV and NPV of 0.756, 0.928, 0.872, 0.855, respectively. CONCLUSION: AI-based semi-automated segmentation is feasible, allowing an instantaneous and reproducible extraction of US-derived radiomics features of FBLs. The combination of radiomics and US BI-RADS classification led to a potential decrease of unnecessary biopsy but at the expense of a not negligible increase of potentially missed cancers.

Diagnostic performance of 2D-shear wave elastography in the diagnosis of breast cancer: a clinical appraisal of cutoff values.

PURPOSE: To assess the role of 2D-shear wave elastography (2D-SWE) in differentiating benign from malignant focal breast lesions (FBLs), providing new vendor-specific cutoff values. METHODS: 158 FBLs (size: 3.5-50 mm) detected in 151 women (age: 21-87 years) were prospectively evaluated by means 2D-SWE. For each lesion, an expert radiologist assessed US BI-RADS category and calculated the following four 2D-SWE parameters: (1) elasticity maximum (Emax); (2) mean elasticity (Emean); (3) minimum elasticity (Emin); (4) elasticity ratio (Eratio). US-guided core-biopsy was considered as standard of reference for all the FBLs classified as BI-RADS 4 or 5. For each 2D-SWE parameter, the optimal cutoff value for a diagnostic test was calculated using the Youden method. Diagnostic performance of the US BI-RADS and 2D-SWE parameters was calculated accordingly. RESULTS: 83/158 (52.5%) FBLs were benign and 75/158 (47.5%) were malignant. Statistically significant higher stiffness values were observed in malignant FBLs for all 2D-SWE parameters than in benign ones (p < 0.001). 2D-SWE cutoff values were 82.6 kPa, 66.0 kPa and 53.6 kPa, respectively, for Emax, Emean, Emin and 330.8% for Eratio. The 2D-SWE parameter showing the best diagnostic accuracy was Emax (85.44%). Considering US BI-RADS 3 (n = 60) and 4a (n = 32) FBLs, Emax and Emean showed the best diagnostic accuracy (85.87% for both), without a statistically significant decrease in sensitivity (p = 0.7003 and p = 1, respectively). CONCLUSION: Our study provides new vendor-specific cutoff values for 2D-SWE, suggesting its possible clinical use in the adjunctive assessment of category US-BI-RADS 3 and 4a breast masses.

3D DCE-MRI Radiomic Analysis for Malignant Lesion Prediction in Breast Cancer Patients.

RATIONALE AND OBJECTIVES: To develop and validate a radiomic model, with radiomic features extracted from breast Dynamic Contrast-Enhanced Magnetic Resonance Imaging (DCE-MRI) from a 1.5T scanner, for predicting the malignancy of masses with enhancement. Images were acquired using an 8-channel breast coil in the axial plane. The rationale behind this study is to show the feasibility of a radiomics-powered model that could be integrated into the clinical practice by exploiting only standard-of-care DCE-MRI with the goal of reducing the required image pre-processing (ie, normalization and quantitative imaging map generation). MATERIALS AND METHODS: 107 radiomic features were extracted from a manually annotated dataset of 111 patients, which was split into discovery and test sets. A feature calibration and pre-processing step was performed to find only robust non-redundant features. An in-depth discovery analysis was performed to define a predictive model: for this purpose, a Support Vector Machine (SVM) was trained in a nested 5-fold cross-validation scheme, by exploiting several unsupervised feature selection methods. The predictive model performance was evaluated in terms of Area Under the Receiver Operating Characteristic (AUROC), specificity, sensitivity, PPV and NPV. The test was performed on unseen held-out data. RESULTS: The model combining Unsupervised Discriminative Feature Selection (UDFS) and SVMs on average achieved the best performance on the blinded test set: AUROC = 0.725±0.091, sensitivity = 0.709±0.176, specificity = 0.741±0.114, PPV = 0.72±0.093, and NPV = 0.75±0.114. CONCLUSION: In this study, we built a radiomic predictive model based on breast DCE-MRI, using only the strongest enhancement phase, with promising results in terms of accuracy and specificity in the differentiation of malignant from benign breast lesions.

S-Detect characterization of focal solid breast lesions: a prospective analysis of inter-reader agreement for US BI-RADS descriptors.

BACKGROUND: To assess inter-reader agreement for US BI-RADS descriptors using S-Detect: a computer-guided decision-making software assisting in US morphologic analysis. METHODS: 73 solid focal breast lesions (FBLs) (mean size: 15.9 mm) in 73 consecutive women (mean age: 51 years) detected at US were randomly and independently assessed according to the BI-RADS US lexicon, without and with S-Detect, by five independent reviewers. US-guided core-biopsy and 24-month follow-up were considered as standard of reference. Kappa statistics were calculated to assess inter-operator agreement, between the baseline and after S-Detect evaluation. Agreement was graded as poor (≤ 0.20), moderate (0.21-0.40), fair (0.41-0.60), good (0.61-0.80), or very good (0.81-1.00). RESULTS: 33/73 (45.2%) FBLs were malignant and 40/73 (54.8%) FBLs were benign. A statistically significant improvement of inter-reader agreement from fair to good with the use of S-Detect was observed for shape (from 0.421 to 0.612) and orientation (from 0.417 to 0.7) (p < 0.0001) and from moderate to fair for margin (from 0.204 to 0.482) and posterior features (from 0.286 to 0.522) (p < 0.0001). At baseline analysis isoechoic (0.0485) and heterogeneous (0.1978) echo pattern, microlobulated (0.1161) angular (0.1204) and spiculated (0.1692) margins and combined pattern (0.1549) for posterior features showed the worst agreement rate (poor). After S-Detect evaluation, all variables but isoechoic pattern showed an agreement class upgrade with a statistically significant improvement of inter-reader agreement (p < 0.0001). CONCLUSIONS: S-Detect significantly improved inter-reader agreement in the assessment of FBLs according to the BI-RADS US lexicon but evaluation of margin and echo pattern needs to be further improved, particularly isoechoic pattern.

Abdominal splenosis and its differential diagnoses: What the radiologist needs to know.

Splenosis is a benign acquired condition characterized by the presence of heterotopic viable splenic tissue in other organs or within cavities such as peritoneum, retroperitoneum, or thorax after splenic trauma or surgery. Abdominal splenosis is often an incidental finding and computed tomography and magnetic resonance usually allow a confident diagnosis. The typical enhancement that parallels the spleen is a useful hallmark of splenosis. Splenic implants lack contrast uptake in the hepatobiliary phase and show high signal at high b-values on diffusion-weighted images. In some cases splenosis may mimic malignant and benign conditions in the peritoneum as well as in hollow and parenchymal abdominal organs and further investigations - including scintigraphy with Tc99m-labelled heat-denatured red blood cells or biopsy - are sometimes required in challenging cases. This pictorial essay reviews the imaging presentation and potential differential diagnosis of splenosis according to the site of implantation. A prompt and accurate radiological diagnosis of splenosis can avoid unnecessary biopsy or surgery.

S-Detect characterization of focal breast lesions according to the US BI RADS lexicon: a pictorial essay.

High-resolution ultrasonography (US) is a valuable tool in breast imaging. Nevertheless, US is an operator-dependent technique: to overcome this issue, the American College of Radiology (ACR) has developed the breast imaging-reporting and data system (BI-RADS) US lexicon. Despite this effort, the variability in the assessment of focal breast lesions (FBLs) with the use of BI-RADS US lexicon is still an issue. Within this framework, evidence shows that computer-aided image analysis may be effective in improving the radiologist's assessment of FBLs. In particular, S-Detect is a newly developed image-analytic computer program that provides assistance in morphologic analysis of FBLs seen on US according to the BI-RADS US lexicon. This pictorial essay describes state-of-the-art of sonographic characterization of FBLs by using S-Detect.

Impact of structured report on the quality of preoperative CT staging of pancreatic ductal adenocarcinoma: assessment of intra- and inter-reader variability.

PURPOSE: To evaluate whether a structured radiology report improves the completeness of preoperative CT staging of pancreatic ductal adenocarcinoma (PDA) compared to conventional free-text reports. METHODS: We retrospectively included 27 patients (mean age, 64 ± 11.1 years) referred for pancreatic preoperative CT scan for staging of PDA between 2015 and 2018 and in whom a diagnosis of pancreatic adenocarcinoma was ultimately confirmed. Four readers independently reported CT scans with both conventional free-text and structured reports. Differences in reported morphologic and vascular features with the two reports were assessed through McNemar Test. Intra-reader and inter-reader were calculated. RESULTS: A total of 216 reports were completed by four different readers including 108 free-text and 108 structured reports. Overall, 139 of 540 morphologic characteristics of PDA and 869 of 1188 vascular key features were only described in structured reports. Encasement of left gastric artery, gastroduodenal artery and splenic artery was described in up to 14.8% using free-text reports and in up to 29.6% using structured report, resulting in low-intra-reader agreement (k = 0.033-0.216). Inter-reader agreement improved with structured report compared to free-text one for left gastric artery (ICC = 0.844 vs. ICC = 0.493, respectively), gastroduodenal artery (ICC = 0.730 vs. ICC = 0.449, respectively), portal vein (ICC = 0.847 vs. ICC = 0.638, respectively), portal confluence (ICC = 0.848 vs. ICC = 0.422, respectively) superior mesenteric vein (ICC = 0.765 vs. ICC = 0.695, respectively), and splenic vein (ICC = 0.921 vs. ICC = 0.841, respectively). CONCLUSION: Structured reports for PDA staging significantly reduces the number of missing morphological and vascular features of PDA and improves the inter-reader agreement compared to free-text reports.

Tolerability profile of thiopurines in inflammatory bowel disease: a prospective experience.

OBJECTIVES: The occurrence of thiopurine-related adverse events (AEs) may complicate the management of patients with inflammatory bowel disease (IBD). We aimed to evaluate the tolerability of thiopurines in a current IBD setting. MATERIALS AND METHODS: All consecutive patients who started a treatment with azathioprine (AZA) from January 2010 to March 2016 were entered in a prospectively maintained database, and the AEs which led to the permanent discontinuation of the drug were reported. RESULTS: Two hundred and fifty three patients were included. Median total follow-up was 32 months (range: 0.2-75 months). At the end of the study, AZA was discontinued in 160 patients (63.2%). The main reason leading to drug withdrawal was the occurrence of AEs (109/160 patients [68.1%]; cumulative incidence among the entire cohort: 43.1%). Overall, the most frequent AEs leading to treatment withdrawal were nausea (31/253 patients, 12.3%) and subjective symptoms, i.e., poorly defined side effects such as fatigue, headache and muscle pain (20/253 patients, 7.9%). Among the 109 AZA-intolerant patients, a switch to 6-mercaptopurine (6-MP) was performed in 44 cases (40.4%). At the end of follow-up, 6-MP was discontinued in 35/44 patients (79.5%), mostly due to AEs (29/35 patients, 82.8%). Azathioprine-induced hepatic and pancreatic toxicity was associated with male gender (p = .01 and p = .03, respectively), and occurrence of nausea with Crohn's disease (p = .04). CONCLUSIONS: Our real-life prospective cohort showed the higher cumulative incidence of thiopurine withdrawal due to AEs reported to date. Switching from AZA to 6-MP was often ineffective.

Mycophenolate mofetil is a valid option in patients with inflammatory bowel disease resistant to TNF-α inhibitors and conventional immunosuppressants.

BACKGROUND: Few studies investigated the role of mycophenolate mofetil in inflammatory bowel disease, and none of them had specifically focused on patients with previous multiple intolerances and/or nonresponses to conventional immunosuppressants and biologics. AIMS: To evaluate clinical benefit and tolerability profile of mycophenolate mofetil in patients with inflammatory bowel disease and limited treatment options. METHODS: All consecutive patients with previous multiple intolerances and/or nonresponses to immunosuppressants and biologics who started an off-label treatment with mycophenolate mofetil from January 2014 to February 2016 were entered in a prospectively maintained database. RESULTS: Twenty-four patients were included. Four weeks after initiation of mycophenolate mofetil therapy, a steroid-free remission was achieved in 4 patients (16.7%), while a clinical response in 13 (54.1%). At the end of follow-up, 12 patients (50.0%) remained on mycophenolate mofetil. Six achieved and maintained steroid-free remission throughout the study period (25.0%), and a further 6 patients (25.0%) achieved a clinical response with complete discontinuation of steroids. Twelve patients (50.0%) were considered as treatment failure, and five of them underwent surgery. CONCLUSIONS: This is the first experience reporting a clinical benefit and tolerability of mycophenolate mofetil in patients with inflammatory bowel disease and multiple previous failures to other immunosuppressants and/or biologics.

Frequency of thiopurine methyltransferase mutation in patients of Mediterranean area with inflammatory bowel disease and autoimmune disorders.

BACKGROUND AND AIMS: Few studies exist on the frequency of thiopurine methyltransferase (TPMT) mutation in patients from Southern Europe. We aimed to evaluate the frequency of TPMT mutation in a homogeneous Sicilian cohort of patients with inflammatory bowel disease (IBD), autoimmune and hematological disorders, the rate of thiopurine-related adverse events, and its association with the TPMT genotype. RESULTS: Among 105 patients with IBD, 45 with autoimmune disease, and 34 with hematologic diseases, the homozygous TPMT variant genotype was found in one patient only (0.5%), while the heterozygous TPMT genotype was identified in 8 patients (4.3%). In patients with IBD, leukopenia was observed in ten patients: one had the homozygous TPMT genotype, one the heterozygous genotype, and the remaining eight the wild type genotype. CONCLUSIONS: The frequency of TPMT mutation in a Mediterranean area was low. TPMT genotyping is not a sensitive tool for predicting thiopurine-induced leukopenia.

Six year adalimumab efficacy in steroid-dependent Crohn's disease patients: A prospective single-center real life study.

BACKGROUND: Adalimumab is effective in the treatment of Crohn's disease. We have already reported data on the efficacy of adalimumab in 110 steroid-dependent patients. At the end of the study 90 patients (64.5%) maintained clinical remission. AIMS: To assess efficacy and safety of adalimumab after 6 years in patients of the original cohort who responded to treatment. METHODS: The present study is an extension of the published paper on 90/110 patients. We report results on clinical remission and safety of 6 year maintenance therapy with adalimumab. RESULTS: Of the original cohort 90 patients completed the study, 17 were lost to follow-up and 3 died. At the end of follow-up (74.16±10.3 months) 37/90 patients (41%) maintained clinical remission. Of these, 32 (86%) continued adalimumab and 5 (13%) discontinued treatment due to clinical remission and mucosal healing. Of the remaining 53/90 patients, 47 (52%) discontinued adalimumab due to clinical failure and 6 (7%) to adverse events. We obtained endoscopy data in 31/32 patients in clinical remission continuing adalimumab: 11 (36%) did not improve, 6 (19%) worsened, 14 (45%) improved. At univariable analysis no variables were related to treatment outcome. CONCLUSIONS: This "real life" prospective study shows that adalimumab is a long-term effective and safe maintenance treatment in steroid-dependent Crohn's disease patients.