Radiomic Features Are Predictive of Response in Rectal Cancer Undergoing Therapy.

BACKGROUND: Rectal cancer is a major mortality cause in the United States (US), and its treatment is based on individual risk factors for recurrence in each patient. In patients with rectal cancer, accurate assessment of response to chemoradiotherapy has increased in importance as the variety of treatment options has grown. In this scenario, a controversial non-operative approach may be considered in some patients for whom complete tumor regression is believed to have occurred. The recommended treatment for locally advanced rectal cancer (LARC, T3-4 ± N+) is total mesorectal excision (TME) after neoadjuvant chemoradiotherapy (nCRT). Magnetic resonance imaging (MRI) has become a standard technique for local staging of rectal cancer (tumor, lymph node, and circumferential resection margin [CRM] staging), in both the US and Europe, and it is getting widely used for restaging purposes. AIM: In our study, we aimed to use an MRI radiomic model to identify features linked to the different responses of chemoradiotherapy of rectal cancer before surgery, and whether these features are helpful to understand the effectiveness of the treatments. METHODS: We retrospectively evaluated adult patients diagnosed with LARC who were subjected to at least 2 MRI examinations in 10-12 weeks at our hospital, before and after nCRT. The MRI acquisition protocol for the 2 exams included T2 sequence and apparent diffusion coefficient (ADC) map. The patients were divided into 2 groups according to the treatment response: complete or good responders (Group 1) and incomplete or poor responders (Group 2). MRI images were segmented, and quantitative features were extracted and compared between the two groups. Features that showed significant differences (SF) were then included in a LASSO regression method to build a radiomic-based predictive model. RESULTS: We included 38 patients (26 males and 12 females), who are classified from T2 and T4 stages in the rectal cancer TNM. After the nCRT, the patients were divided into Group 1 (13 patients), complete or good responders, and Group 2 (25 patients), incomplete or poor responders. Analysis at baseline generated the following significant features for the Mann-Whitney test (out of a total of 107) for each sequence. Also, the analysis at the end of the follow-up yielded a high number of significant features for the Mann-Whitney test (out of a total of 107) for each image. Features selected by the LASSO regression method for each image analyzed; ROC curves relative to each model are represented. CONCLUSION: We developed an MRI-based radiomic model that is able to differentiate and predict between responders and non-responders who went through nCRT for rectal cancer. This approach might identify early lesions with high surgical potential from lesions potentially resolving after medical treatment.

Branch duct-intraductal papillary mucinous neoplasms (BD-IPMNs): an MRI-based radiomic model to determine the malignant degeneration potential.

BACKGROUND: Branch duct-intraductal papillary mucinous neoplasms (BD-IPMNs) are the most common pancreatic cystic tumors and have a low risk of malignant transformation. Features able to early identify high-risk BD-IPMNs are lacking, and guidelines currently rely on the occurrence of worrisome features (WF) and high-risk stigmata (HRS). AIM: In our study, we aimed to use a magnetic resonance imaging (MRI) radiomic model to identify features linked to a higher risk of malignant degeneration, and whether these appear before the occurrence of WF and HRS. METHODS: We retrospectively evaluated adult patients with a known BD-IPMN who had had at least two contrast-enhanced MRI studies at our center and a 24-month minimum follow-up time. MRI acquisition protocol for the two examinations included pre- and post-contrast phases and diffusion-weighted imaging (DWI)/apparent diffusion coefficient (ADC) map. Patients were divided into two groups according to the development of WF or HRS at the end of the follow-up (Group 0 = no WF or HRS; Group 1 = WF or HRS). We segmented the MRI images and quantitative features were extracted and compared between the two groups. Features that showed significant differences (SF) were then included in a LASSO regression method to build a radiomic-based predictive model. RESULTS: We included 50 patients: 31 in Group 0 and 19 in Group 1. No patients in this cohort developed HRS. At baseline, 47, 67, 38, and 68 SF were identified for pre-contrast T1-weighted (T1-W) sequence, post-contrast T1-W sequence, T2-weighted (T2- W) sequence, and ADC map, respectively. At the end of follow-up, we found 69, 78, 53, and 91 SF, respectively. The radiomic-based predictive model identified 16 SF: more particularly, 5 SF for pre-contrast T1-W sequence, 6 for post-contrast T1-W sequence, 3 for T2-W sequence, and 2 for ADC. CONCLUSION: We identified radiomic features that correlate significantly with WF in patients with BD-IPMNs undergoing contrast-enhanced MRI. Our MRI-based radiomic model can predict the occurrence of WF.

Reproducibility of CT radiomic features in lung neuroendocrine tumours (NETs) patients: analysis in a heterogeneous population.

BACKGROUND: The aim is to find a correlation between texture features extracted from neuroendocrine (NET) lung cancer subtypes, both Ki-67 index and the presence of lymph-nodal mediastinal metastases detected while using different computer tomography (CT) scanners. METHODS: Sixty patients with a confirmed pulmonary NET histological diagnosis, a known Ki-67 status and metastases, were included. After subdivision of primary lesions in baseline acquisition and venous phase, 107 radiomic features of first and higher orders were extracted. Spearman's correlation matrix with Ward's hierarchical clustering was applied to confirm the absence of bias due to the database heterogeneity. Nonparametric tests were conducted to identify statistically significant features in the distinction between patient groups (Ki-67 < 3-Group 1; 3 ≤ Ki-67 ≤ 20-Group 2; and Ki-67 > 20-Group 3, and presence of metastases). RESULTS: No bias arising from sample heterogeneity was found. Regarding Ki-67 groups statistical tests, seven statistically significant features (p value < 0.05) were found in post-contrast enhanced CT; three in baseline acquisitions. In metastasis classes distinction, three features (first-order class) were statistically significant in post-contrast acquisitions and 15 features (second-order class) in baseline acquisitions, including the three features distinguishing between Ki-67 groups in baseline images (MCC, ClusterProminence and Strength). CONCLUSIONS: Some radiomic features can be used as a valid and reproducible tool for predicting Ki-67 class and hence the subtype of lung NET in baseline and post-contrast enhanced CT images. In particular, in baseline examination three features can establish both tumour class and aggressiveness.

Gastroenteropancreatic neuroendocrine neoplasms (GEP-NENs): a radiomic model to predict tumor grade.

PURPOSE: The aim of this single-center retrospective study is to assess whether contrast-enhanced computed tomography (CECT) radiomics analysis is predictive of gastroenteropancreatic neuroendocrine neoplasms (GEP-NENs) grade based on the 2019 World Health Organization (WHO) classification and to establish a tumor grade (G) prediction model. MATERIAL AND METHODS: Preoperative CECT images of 78 patients with GEP-NENs were retrospectively reviewed and divided in two groups (G1-G2 in class 0, G3-NEC in class 1). A total of 107 radiomics features were extracted from each neoplasm ROI in CT arterial and venous phases acquisitions with 3DSlicer. Mann-Whitney test and LASSO regression method were performed in R for feature selection and feature reduction, in order to build the radiomic-based predictive model. The model was developed for a training cohort (75% of the total) and validated on the independent validation cohort (25%). ROC curves and AUC values were generated on training and validation cohorts. RESULTS: 40 and 24 features, for arterial phase and venous phase, respectively, were found to be significant in class distinction. From the LASSO regression 3 and 2 features, for arterial phase and venous phase, respectively, were identified as suitable for groups classification and used to build the tumor grade radiomic-based prediction model. The prediction of the arterial model resulted in AUC values of 0.84 (95% CI 0.72-0.97) and 0.82 (95% CI 0.62-1) for the training cohort and validation cohort, respectively, while the prediction of the venous model yielded AUC values of 0.7877 (95% CI 0.6416-0.9338) and 0.6813 (95% CI 0.3933-0.9693) for the training cohort and validation cohort, respectively. CONCLUSIONS: CT-radiomics analysis may aid in differentiating the histological grade for GEP-NENs.

Radiomics in pulmonary neuroendocrine tumours (NETs).

OBJECTIVES: The aim of this single-centre, observational, retrospective study is to find a correlation using Radiomics between the analysis of CT texture features of primary lesion of neuroendocrine (NET) lung cancer subtypes (typical and atypical carcinoids, large and small cell neuroendocrine carcinoma), Ki-67 index and the presence of lymph nodal mediastinal metastases. METHODS: Twenty-seven patients (11 males and 16 females, aged between 48 and 81 years old-average age of 70,4 years) with histological diagnosis of pulmonary NET with known Ki-67 status and metastases who have performed pre-treatment CT in our department were included. All examinations were performed with the same CT scan (Sensation 16-slice, Siemens). The study protocol was a baseline scan followed by 70 s delay acquisition after administration of intravenous contrast medium. After segmentation of primary lesions, quantitative texture parameters of first and higher orders were extracted. Statistics nonparametric tests and linear correlation tests were conducted to evaluate the relationship between different textural characteristics and tumour subtypes. RESULTS: Statistically significant (p < 0.05) differences were seen in post-contrast enhanced CT in multiple first and higher-order extracted parameters regarding the correlation with classes of Ki-67 index values. Statistical analysis for direct acquisitions was not significant. Concerning the correlation with the presence of metastases, one histogram feature (Skewness) and one feature included in the Gray-Level Co-occurrence Matrix (ClusterShade) were significant on contrast-enhanced CT only. CONCLUSIONS: CT texture analysis may be used as a valid tool for predicting the subtype of lung NET and its aggressiveness.

FLORA software: semi-automatic LGE-CMR analysis tool for cardiac lesions identification and characterization.

PURPOSE: Today there is a growing interest in the quantification of late gadolinium enhancement (LGE) in ischemic and non-ischemic cardiac pathologies. We build an automatic self-made free software FLORA (For Late gadOlinium enhanced aReas clAssification) for the recognition, classification and quantification of LGE areas that allows to improve the observer's performances and that homogenizes the evaluations between different operators. MATERIAL AND METHODS: We have retrospectively selected 120 CMR exams: 40-ischemic with evident scar tissue on LGE sequences; 40-non-ischemic cardiomyopathy; 40-any myocardial alteration on CMR, especially on LGE sequences. FLORA's performance was compared to the radiologist's evaluation. RESULTS: FLORA identified both ischemic and non-ischemic myocardial lesions in almost all cases (80/80 and 79/80 for the double-Gaussian fit method and fixed-shift method, respectively, with sensitivity and specificity of 100%/98.8% and 55%/50%, respectively). The best results were obtained from the classification of ischemic myocardial damage, which was correctly identified in 85%-95% of cases. FLORA also increases the agreement between observers and allows a quantitative evaluation of transmurality. CONCLUSIONS: FLORA has proven to be an applicable tool that improves and facilitates the classification of LGE areas allowing their quantification.

Single Center Evaluation of Comparative Breast Radiation dose of Contrast Enhanced Digital Mammography (CEDM), Digital Mammography (DM) and Digital Breast Tomosynthesis (DBT).

RATIONALE AND OBJECTIVES: The aim of this retrospective study is to compare the radiation dose received during CEDM, short and long protocol (CEDM SP and CEDM LP), with dose received during DM and DBT on patients with varying breast thickness, age and density. MATERIALS AND METHODS: Between January 2019 and December 2019, patients having 6214 DM, 3662 DBT and 173 CEDM examinations in our department were analyzed. Protocol total single breast AGD has been evaluated for all clinical imaging protocols, extracting AGD values and exposure data from the dose DICOM Structured Report (SR) information stored in the hospital PACS system. Protocol AGD was calculated as the sum of single projection AGDs carried out in every exam for each clinical protocol. A total amount of 23,383 exams for each breast were analyzed. Protocol AGDs, stratified as a function of patient breast compression thickness, age, and breast density were assessed. RESULTS: The total protocol AGD median values for each protocol are: 2.8 mGy for DM, 3.2 mGy for DBT, 6.0 mGy for DM+DBT, 4.5 mGy for CEDM SP, 7.4 mGy for CEDM SP_DBT (CEDM SP protocol with DBT), 8.4 mGy for CEDM LP and 11.6 mGy for CEDM LP_DBT (CEDM LP protocol with DBT). CEDM SP AGD median value is 59% higher than DM AGD median value and 40% lesser than DM+DBT AGD median; this last difference was statistically confirmed with a p-value <0.001. AGD value for each standard breast CEDM SP projection results to be below 3-mGy limit. AGD value for each standard breast CEDM SP projection results to be below 3 mGy, as required by international legislation. For dense breasts, the AGD median value is 4.2 mGy, with the first and third quartile of 3.3 mGy and 6.0 mGy respectively; for non-dense breasts, the AGD median value is 4.7 mGy, with first and third quartile of 3.5 mGy and 6.3 mGy respectively. The difference between the two groups was statistically tested and confirmed, with a p-value of 0.039. CONCLUSION: CEDM SP results in higher radiation exposure compared with conventional DM and DBT but lower than the Combo mode. The dose administered during the CEDM SP is lower in patients with dense breasts regardless of their size. An interesting outcome, considering the ongoing studies on CEDM screening in patients with dense breasts.

Correlation of CT radiomic features for GISTs with pathological classification and molecular subtypes: preliminary and monocentric experience.

PURPOSE: Our primary purpose was to search for computed tomography (CT) radiomic features of gastrointestinal stromal tumors (GISTs) that could potentially correlate with the risk class according to the Miettinen classification. Subsequently, assess the existence of features with possible predictive value in differentiating responder from non-responder patients to first-line therapy with Imatinib. METHODS: A retrospective study design was carried out using data from June 2009 to December 2020. We analyzed all the preoperative CTs of patients undergoing surgery for GISTs. We segmented non-contrast-enhanced CT (NCECT) and contrast-enhanced venous CT (CECT) images obtained either on three different CT scans (heterogeneous cohort) or on a single CT scan (homogeneous cohort). We then divided the patients into two groups according to Miettinen classification criteria and based on the predictive value of response to first-line therapy with Imatinib. RESULTS: We examined 54 patients with pathological confirmation of GISTs. For the heterogeneous cohort, we found a statistically significant relationship between 57 radiomic features for NCECT and 56 radiomic features for CECT using the Miettinen risk classification. In the homogeneous cohort, we found the same relationship between 8 features for the NCECT and 5 features for CECT, all included in the heterogeneous cohort. The various radiomic features are distributed with different values in the two risk stratification groups according to the Miettinen classification. We also found some features for groups predictive of response to first-line therapy with Imatinib. CONCLUSIONS: We found radiomic features that correlate with statistical significance for both the Miettinen risk classification and the molecular subtypes of response. All features found in the homogeneous study cohort were also found in the heterogeneous cohort. CT radiomic features may be useful in assessing the risk class and prognosis of GISTs.

On the dependence of quantitative diffusion-weighted imaging on scanner system characteristics and acquisition parameters: A large multicenter and multiparametric phantom study with unsupervised clustering analysis.

PURPOSE: The purpose of this multicenter phantom study was to exploit an innovative approach, based on an extensive acquisition protocol and unsupervised clustering analysis, in order to assess any potential bias in apparent diffusion coefficient (ADC) estimation due to different scanner characteristics. Moreover, we aimed at assessing, for the first time, any effect of acquisition plan/phase encoding direction on ADC estimation. METHODS: Water phantom acquisitions were carried out on 39 scanners. DWI acquisitions (b-value = 0-200-400-600-800-1000 s/mm2) with different acquisition plans (axial, coronal, sagittal) and phase encoding directions (anterior/posterior and right/left, for the axial acquisition plan), for 3 orthogonal diffusion weighting gradient directions, were performed. For each acquisition setup, ADC values were measured in-center and off-center (6 different positions), resulting in an entire dataset of 84 × 39 = 3276 ADC values. Spatial uniformity of ADC maps was assessed by means of the percentage difference between off-center and in-center ADC values (Δ). RESULTS: No significant dependence of in-center ADC values on acquisition plan/phase encoding direction was found. Ward unsupervised clustering analysis showed 3 distinct clusters of scanners and an association between Δ-values and manufacturer/model, whereas no association between Δ-values and maximum gradient strength, slew rate or static magnetic field strength was revealed. Several acquisition setups showed significant differences among groups, indicating the introduction of different biases in ADC estimation. CONCLUSIONS: Unsupervised clustering analysis of DWI data, obtained from several scanners using an extensive acquisition protocol, allows to reveal an association between measured ADC values and manufacturer/model of scanner, as well as to identify suboptimal DWI acquisition setups for accurate ADC estimation.