Pre-operative prediction of histopathological growth patterns of colorectal cancer liver metastasis using MRI-based radiomic models.

PURPOSE: Histopathological growth patterns (HGPs) of colorectal liver metastases (CRLMs) have prognostic value. However, the differentiation of HGPs relies on postoperative pathology. This study aimed to develop a magnetic resonance imaging (MRI)-based radiomic model to predict HGP pre-operatively, following the latest guidelines. METHODS: This retrospective study included 93 chemotherapy-naïve patients with CRLMs who underwent contrast-enhanced liver MRI and a partial hepatectomy between 2014 and 2022. Radiomic features were extracted from the tumor zone (RTumor), a 2-mm outer ring (RT+2), a 2-mm inner ring (RT-2), and a combined ring (R2+2) on late arterial phase MRI images. Analysis of variance method (ANOVA) and least absolute shrinkage and selection operator (LASSO) algorithms were used for feature selection. Logistic regression with five-fold cross-validation was used for model construction. Receiver operating characteristic curves, calibrated curves, and decision curve analyses were used to assess model performance. DeLong tests were used to compare different models. RESULTS: Twenty-nine desmoplastic and sixty-four non-desmoplastic CRLMs were included. The radiomic models achieved area under the curve (AUC) values of 0.736, 0.906, 0.804, and 0.794 for RTumor, RT-2, RT+2, and R2+2, respectively, in the training cohorts. The AUC values were 0.713, 0.876, 0.785, and 0.777 for RTumor, RT-2, RT+2, and R2+2, respectively, in the validation cohort. RT-2 exhibited the best performance. CONCLUSION: The MRI-based radiomic models could predict HGPs in CRLMs pre-operatively.

MRI-based tumor shrinkage patterns after early neoadjuvant therapy in breast cancer: correlation with molecular subtypes and pathological response after therapy.

BACKGROUND: MRI-based tumor shrinkage patterns (TSP) after neoadjuvant therapy (NAT) have been associated with pathological response. However, the understanding of TSP after early NAT remains limited. We aimed to analyze the relationship between TSP after early NAT and pathological response after therapy in different molecular subtypes. METHODS: We prospectively enrolled participants with invasive ductal breast cancers who received NAT and performed pretreatment DCE-MRI from September 2020 to August 2022. Early-stage MRIs were performed after the first (1st-MRI) and/or second (2nd-MRI) cycle of NAT. Tumor shrinkage patterns were categorized into four groups: concentric shrinkage, diffuse decrease (DD), decrease of intensity only (DIO), and stable disease (SD). Logistic regression analysis was performed to identify independent variables associated with pathologic complete response (pCR), and stratified analysis according to tumor hormone receptor (HR)/human epidermal growth factor receptor 2 (HER2) disease subtype. RESULTS: 344 participants (mean age: 50 years, 113/345 [33%] pCR) with 345 tumors (1 bilateral) had evaluable 1st-MRI or 2nd-MRI to comprise the primary analysis cohort, of which 244 participants with 245 tumors had evaluable 1st-MRI (82/245 [33%] pCR) and 206 participants with 207 tumors had evaluable 2nd-MRI (69/207 [33%] pCR) to comprise the 1st- and 2nd-timepoint subgroup analysis cohorts, respectively. In the primary analysis, multivariate analysis showed that early DD pattern (OR = 12.08; 95% CI 3.34-43.75; p < 0.001) predicted pCR independently of the change in tumor size (OR = 1.37; 95% CI 0.94-2.01; p = 0.106) in HR+/HER2- subtype, and the change in tumor size was a strong pCR predictor in HER2+ (OR = 1.61; 95% CI 1.22-2.13; p = 0.001) and triple-negative breast cancer (TNBC, OR = 1.61; 95% CI 1.22-2.11; p = 0.001). Compared with the change in tumor size, the SD pattern achieved a higher negative predictive value in HER2+ and TNBC. The statistical significance of complete 1st-timepoint subgroup analysis was consistent with the primary analysis. CONCLUSION: The diffuse decrease pattern in HR+/HER2- subtype and stable disease in HER2+ and TNBC after early NAT could serve as additional straightforward and comprehensible indicators of treatment response. TRIAL REGISTRATION: Trial registration at https://www.chictr.org.cn/ . REGISTRATION NUMBER: ChiCTR2000038578, registered September 24, 2020.

Artificial intelligence to predict T4 stage of pancreatic ductal adenocarcinoma using CT imaging.

BACKGROUND: The accurate assessment of T4 stage of pancreatic ductal adenocarcinoma (PDAC) has consistently presented a considerable difficulty for radiologists. This study aimed to develop and validate an automated artificial intelligence (AI) pipeline for the prediction of T4 stage of PDAC using contrast-enhanced CT imaging. METHODS: The data were obtained retrospectively from consecutive patients with surgically resected and pathologically proved PDAC at two institutions between July 2017 and June 2022. Initially, a deep learning (DL) model was developed to segment PDAC. Subsequently, radiomics features were extracted from the automatically segmented region of interest (ROI), which encompassed both the tumor region and a 3 mm surrounding area, to construct a predictive model for determining T4 stage of PDAC. The assessment of the models' performance involved the calculation of the area under the receiver operating characteristic curve (AUC), sensitivity, and specificity. RESULTS: The study encompassed a cohort of 509 PDAC patients, with a median age of 62 years (interquartile range: 55-67). The proportion of patients in T4 stage within the model was 16.9%. The model achieved an AUC of 0.849 (95% CI: 0.753-0.940), a sensitivity of 0.875, and a specificity of 0.728 in predicting T4 stage of PDAC. The performance of the model was determined to be comparable to that of two experienced abdominal radiologists (AUCs: 0.849 vs. 0.834 and 0.857). CONCLUSION: The automated AI pipeline utilizing tumor and peritumor-related radiomics features demonstrated comparable performance to that of senior abdominal radiologists in predicting T4 stage of PDAC.

Differentiation of histopathological growth patterns of colorectal liver metastases by MRI features.

BACKGROUND: It is necessary to develop an accurate non-invasive method to determine the histopathological growth pattern (HGP) of colorectal liver metastasis (CRLM) before surgery. The present study aimed to identify various HGPs of CRLM by magnetic resonance imaging (MRI) features. METHODS: This retrospective study included 53 chemo-naïve patients with CRLM between December 2013 and September 2019. The HGPs of CRLM were assessed according to the international consensus guidelines, and were classified as either replacement HGP (rHGP) or non-rHGP. The MRI features of CRLM were retrospectively reviewed in consensus by two radiologists. The differences of MRI features between rHGP and non-rHGP tumors were compared by using Chi-square test and Student's t-test. The Spearman or Pearson correlation analysis was performed to determine the correlation between different MRI features. A receiver operating characteristic (ROC) curve was plotted to evaluate the diagnostic ability. RESULTS: Of the 53 chemo-naïve patients (mean age, 60.11±9.85 years; age range, 38-86 years), 12 were diagnosed as rHGP, while 41 were diagnosed as non-rHGP. Rim enhancement were more common in rHGP than in non-rHGP (P<0.001). Besides, the diameter difference (ΔD) between the precontrast and postcontrast images of rHGP was significantly larger than that of the non-rHGP (P=0.001). The rim width was correlated with ΔD, but not correlated with tumor size. The non-rHGP colorectal liver metastases were prone to be washed out in the delayed phases (P=0.043). The area under the curve (AUC) for the differentiation of rHGP and non-rHGP by using rim enhancement and ΔD was 0.828 (95% CI: 0.708-0.949). CONCLUSIONS: The MRI features of CRLM are characteristic and could help to differentiate rHGP and non-rHGP.

Machine learning prediction of axillary lymph node metastasis in breast cancer: 2D versus 3D radiomic features.

PURPOSE: The purpose of this study was to distinguish axillary lymph node (ALN) status using preoperative breast DCE-MRI radiomics and compare the effects of two-dimensional (2D) and three-dimensional (3D) analysis. METHODS: A retrospective study including 154 breast cancer patients all confirmed by pathology; 80 with ALN metastasis and 74 without. All MRI scans were achieved at a 3.0 Tesla scanner with 7 post-contrast MR phases sequentially acquired with a temporal resolution of 60 s. MRI radiomic features were extracted separately from a 2D single slice (i.e., the representative slice) and the 3D tumor volume. Several machine learning classifiers were built and compared using 2D or 3D analysis to distinguish positive vs negative ALN status. We performed independent test and 10-fold cross validation with multiple repetitions, and used bootstrap test, least absolute shrinkage selection operator, and receiver operating characteristic (ROC) curve analysis as statistical tests. RESULTS: The highest area under the ROC curve (AUC) was 0.81 (95% confidence intervals [CI]: 0.80-0.83) and 0.82 (95% CI: 0.81-0.82) for 2D and 3D analysis, respectively; the corresponding accuracy was 79% and 80%. The linear discriminant analysis (LDA) classifier achieved the highest classification performance. None of the AUC differences between 2D and 3D analysis was statistically significant for the several tested machine learning classifiers (all P> 0.05). CONCLUSIONS: Radiomic features from segmented tumor region in breast MRI were associated with ALN status. The separate radiomic analysis on 3D tumor volume showed a similar effect to the 2D analysis on the single representative slice in the tested machine learning classifiers.

Diffusion-weighted imaging or dynamic contrast-enhanced curve: a retrospective analysis of contrast-enhanced magnetic resonance imaging-based differential diagnoses of benign and malignant breast lesions.

OBJECTIVE: To compare the diagnostic performance of models based on a combination of contrast-enhanced (CE) magnetic resonance imaging (MRI) with diffusion-weighted imaging (DWI) or time-intensity curves (TIC) in diagnosing malignancies of breast lesions. METHODS: A double-blind retrospective study was conducted in 328 patients (254 for training and the following 74 for validation) who underwent dynamic contrast-enhanced MRI (DCE-MRI) of the breast with pathological results. Two score models, the DWI model (apparent diffusion coefficient (ADC) + morphology + enhanced information) and the TIC model (TIC + morphology + enhanced information), were established with binary logistic regression for mass and non-mass enhancements (NMEs) in the training set. The sensitivity, specificity, and area under the curve (AUC) were compared between the two models (DWI model vs. TIC model); p < 0.05 was considered as statistically different. External validation was used. RESULTS: In the training set, the sensitivities, specificities, and AUCs of the DWI/TIC model were 95.2%/95.8%, 70.8%/47.9%, and 0.932/0.891 for masses, and 94.2%/90.4%, 47.4%/47.4%, and 0.798 (95% CI, 0.686-0.884)/0.802 (95% CI, 0.691-0.887) for NMEs, respectively. The AUC of the DWI model was significantly higher than that of the TIC model (p < 0.05) for masses. In the validation set, the AUCs of the DWI/TIC model were 0.896/0.861 for masses (p < 0.05) and 0.936/0.836 for NMEs (p > 0.05). CONCLUSIONS: Combined with CE MRI, the DWI model was superior or equal to the TIC model in differentiating benign and malignant breast lesions. KEY POINTS: • Diffusion magnetic resonance imaging played an important role in the diagnosis of breast neoplasms. • On the basis of contrast-enhanced MRI, the DWI model had significantly higher diagnostic ability than the TIC model in distinguishing benign and malignant masses. • It would be reasonable to replace the time-consuming TIC with DWI for less scan time and similar diagnostic efficiency.

Utility of radiomics based on contrast-enhanced CT and clinical data in the differentiation of benign and malignant gallbladder polypoid lesions.

PURPOSE: To develop and validate a novel method based on radiomics for the preoperative differentiation of benign and malignant gallbladder polypoid lesions (PLG). PATIENTS AND METHODS: A total of 145 patients with pathological proven gallbladder polypoid lesions ≥ 1 cm were included in this retrospective study. All the patients underwent abdominal contrast-enhanced computed tomography (CT) examinations 3 weeks before cholecystectomy from January 2013 to January 2019. Seventy percent of the cases were randomly selected for the training dataset, and 30% of the cases were independently used for testing. Radiomics features extracted from portal venous-phase CT of the PLG and clinical features were analyzed, and the LASSO regression algorithm was used for data dimension reduction. Multivariable logistic regression was used to generate radiomics signatures, clinical signatures, and combination signatures. The receiver operating characteristic (ROC) curve and decision curve were plotted to assess the differentiating performance of the three signatures. RESULTS: The area under the ROC curve (AUC) of the radiomics signature and clinical signature was 0.924 and 0.861 in the testing dataset, respectively. For the radiomics signature, the accuracy was 88.6%, with 88.0% specificity and 89.5% sensitivity. When combined, the AUC was 0.931, the specificity was 84.0%, and the sensitivity was 89.5%. The differences between the AUC values of the two sole models and the combination model were statistically nonsignificant. CONCLUSION: Radiomics based on CT images can be helpful to differentiate benign and malignant gallbladder polyps ≥ 1 cm in size.

Automated deep learning method for whole-breast segmentation in diffusion-weighted breast MRI.

BACKGROUND: Diffusion-weighted imaging (DWI) in MRI plays an increasingly important role in diagnostic applications and developing imaging biomarkers. Automated whole-breast segmentation is an important yet challenging step for quantitative breast imaging analysis. While methods have been developed on dynamic contrast-enhanced (DCE) MRI, automatic whole-breast segmentation in breast DWI MRI is still underdeveloped. PURPOSE: To develop a deep/transfer learning-based segmentation approach for DWI MRI scans and conduct an extensive study assessment on four imaging datasets from both internal and external sources. STUDY TYPE: Retrospective. SUBJECTS: In all, 98 patients (144 MRI scans; 11,035 slices) of four different breast MRI datasets from two different institutions. FIELD STRENGTH/SEQUENCES: 1.5T scanners with DCE sequence (Dataset 1 and Dataset 2) and DWI sequence. A 3.0T scanner with one external DWI sequence. ASSESSMENT: Deep learning models (UNet and SegNet) and transfer learning were used as segmentation approaches. The main DCE Dataset (4,251 2D slices from 39 patients) was used for pre-training and internal validation, and an unseen DCE Dataset (431 2D slices from 20 patients) was used as an independent test dataset for evaluating the pre-trained DCE models. The main DWI Dataset (6,343 2D slices from 75 MRI scans of 29 patients) was used for transfer learning and internal validation, and an unseen DWI Dataset (10 2D slices from 10 patients) was used for independent evaluation to the fine-tuned models for DWI segmentation. Manual segmentations by three radiologists (>10-year experience) were used to establish the ground truth for assessment. The segmentation performance was measured using the Dice Coefficient (DC) for the agreement between manual expert radiologist's segmentation and algorithm-generated segmentation. STATISTICAL TESTS: The mean value and standard deviation of the DCs were calculated to compare segmentation results from different deep learning models. RESULTS: For the segmentation on the DCE MRI, the average DC of the UNet was 0.92 (cross-validation on the main DCE dataset) and 0.87 (external evaluation on the unseen DCE dataset), both higher than the performance of the SegNet. When segmenting the DWI images by the fine-tuned models, the average DC of the UNet was 0.85 (cross-validation on the main DWI dataset) and 0.72 (external evaluation on the unseen DWI dataset), both outperforming the SegNet on the same datasets. DATA CONCLUSION: The internal and independent tests show that the deep/transfer learning models can achieve promising segmentation effects validated on DWI data from different institutions and scanner types. Our proposed approach may provide an automated toolkit to help computer-aided quantitative analyses of breast DWI images. LEVEL OF EVIDENCE: 3 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2020;51:635-643.

Differentiating axillary lymph node metastasis in invasive breast cancer patients: A comparison of radiomic signatures from multiparametric breast MR sequences.

BACKGROUND: The axillary lymph node status is critical for breast cancer staging and individualized treatment planning. PURPOSE: To assess the effect of determining axillary lymph node (ALN) metastasis by breast MRI-derived radiomic signatures, and compare the discriminating abilities of different MR sequences. STUDY TYPE: Retrospective. POPULATION: In all, 120 breast cancer patients, 59 with ALN metastasis and 61 without metastasis, all confirmed by pathology. FIELD STRENGTH/SEQUENCE: 3 .0T scanner with T1 -weighted imaging, T2 -weighted imaging, diffusion-weighted imaging, and dynamic contrast-enhanced (DCE) sequences. ASSESSMENT: Typical morphological and texture features of the segmented tumor were extracted from four sequences, ie, T1 WI, T2 WI, DWI, and the second postcontrast phase (CE2) of the dynamic contrast-enhanced sequences. Additional contrast enhancement kinetic features were extracted from all DCE sequences (one pre- and seven postcontrast phases). Linear discriminant analysis classifiers were built and compared when using features from an individual sequence or the combination of the sequences in differentiating the ALN metastasis status. STATISTICAL TESTS: Mann-Whitney U-test, Fisher's exact test, least absolute shrinkage selection operator (LASSO) regression, and receiver operating characteristic analysis were performed. RESULTS: The accuracy/AUC of the four sequences was 79%/0.87, 77%/0.85, 74%/0.79, and 79%/0.85 for the T1 WI, CE2, T2 WI, and DWI, respectively. When CE2 was augmented by adding kinetic features, the model achieved the highest performance (accuracy = 0.86 and AUC = 0.91). When all features from the four sequences and the kinetics were combined, it did not lead to a further increase in the performance (P = 0.48). DATA CONCLUSION: Breast tumor's radiomic signatures from preoperative breast MRI sequences are associated with the ALN metastasis status, where CE2 phase and the contrast enhancement kinetic features lead to the highest classification effect. Level of Evidence 3 Technical Efficacy Stage 2 J. Magn. Reson. Imaging 2019;50:1125-1132.

Preoperative Prediction of Axillary Lymph Node Metastasis in Breast Cancer using Radiomics Features of DCE-MRI.

The accurate and noninvasive preoperative prediction of the state of the axillary lymph nodes is significant for breast cancer staging, therapy and the prognosis of patients. In this study, we analyzed the possibility of axillary lymph node metastasis directly based on Magnetic Resonance Imaging (MRI) of the breast in cancer patients. After mass segmentation and feature analysis, the SVM, KNN, and LDA three classifiers were used to distinguish the axillary lymph node state in 5-fold cross-validation. The results showed that the effect of the SVM classifier in predicting breast axillary lymph node metastasis was significantly higher than that of the KNN classifier and LDA classifier. The SVM classifier performed best, with the highest accuracy of 89.54%, and obtained an AUC of 0.8615 for identifying the lymph node status. Each feature was analyzed separately and the results showed that the effect of feature combination was obviously better than that of any individual feature on its own.

A radiomic nomogram based on an apparent diffusion coefficient map for differential diagnosis of suspicious breast findings.

OBJECTIVE: To develop and validate a radiomic nomogram based on an apparent diffusion coefficient (ADC) map for differentiating benign and malignant lesions in suspicious breast findings classified as Breast Imaging Reporting and Data System (BI-RADS) category 4 on breast magnetic resonance imaging (MRI). METHODS: Eighty-eight patients diagnosed with BI-RADS 4 findings on breast MRI in the First Affiliated Hospital of China Medical University from December 2014 to December 2015 were retrospectively analyzed in this study. Sixty-three were randomized electronically to establish forecasting models, and the other 25 were used for validation. Radiomic features based on the ADC map were generated automatically by Artificial Intelligence Kit software (A.K. software; GE Healthcare, China). Feature reduction was conducted using the Mann-Whitney test and Spearman correlation after pre-treatment. A prediction model of ADC radiomics was established by logistic linear regression and cross-validation. A nomogram was established based on ADC radiomic features, pharmacokinetics and clinical features, including the morphology and ADC value for breast BI-RADS 4 lesions on MRI. RESULTS: A total of 396 radiomic features were extracted automatically by the A.K. software. Five features were selected after pre-processing, Mann-Whitney tests and Spearman correlation analysis. The area under the ROC curve of the prediction model comprising ADC radiomic features was 0.79 when the cutoff value was 0.45, and the accuracy, sensitivity and specificity were 80.0%, 0.813 and 0.778, respectively. A visualized differential nomogram based on the radiomic score, pharmacokinetics and clinical features was established. The decision curve showed good consistency. CONCLUSIONS: ADC radiomic features could provide an important reference for differential diagnosis between benign and malignant lesions in suspicious BI-RADS 4 lesions. The visualized nomogram based on ADC radiomic features, pharmacokinetics and clinical features may have good prospects for clinical application.

T and N Staging of Gastric Cancer Using Dual-Source Computed Tomography.

AIM: This study is aimed at comparing gastric cancer T and N staging between virtual monochromatic energy images and fusion images generated by dual-source computed tomography (DSCT) dual-energy mode data acquisition prospectively while measuring the iodine concentration of gastric cancer and lymph nodes at different T and N stages from iodine map retrospectively. METHODS: A total of 71 patients (50 males and 21 females; mean age: 59 ± 11 years) confirmed with gastric cancer by endoscopic biopsy with no neoadjuvant chemotherapy were enrolled for the CT examination before surgeries. The preoperative T and N staging results were compared between groups with pathological results as the gold standard. The iodine concentrations of the gastric lesions and LNs were measured on the iodine-based material decomposition images. All iodine concentration values were normalized against those in the abdominal aorta and defined as normalized iodine concentration (nIC) values. The short axis length of LNs and nIC values were statistically analyzed. RESULTS: Group A was better than group B for T3 and T4 staging. No statistically significant difference in the overall accuracies for N staging was found between groups. For the late arterial and delayed phases, T3 and T4 nIC values of the extraserosal adipose tissue showed statistically significant differences. The nIC values between N0 and Nm (N1-N3) showed statistically significant differences in the portal phase only. CONCLUSIONS: T3 and T4 nIC values of the extraserosal adipose tissue showed statistically significant differences. Hence, dual-source CT may be helpful in the differential diagnosis between T3 and T4.

The primary study of low-dose pancreas perfusion by 640- slice helical CT: a whole-organ perfusion.

To discuss the feasibility of low-dose whole-pancreas imaging utilizing 640-slice dynamic volume CT.80 patients (40 cases of normal pancreas and 40 patients supposed of having pancreatic carcinoma or focal pancreatic space-occupying lesions were mainly refered) referred for CT pancreas perfusion were enrolled in the study. 80 patients randomly assigned to 3 groups: Group ① (whole sequence). Group ② (odd number sequence). Group ③ (even number group)(Compared to ①, the scanning times and effective radiate dose of ② and ③ decreased about 50% respectively). The head, body, tail of each normal pancreas without any pancreatic disease, lesion and lesion-surrounding areas of each pancreatic cancer were selected as ROI, and tissue peak, blood flow are measured.According to pathology and clinical materials, 27 patients were diagnosed as pancreatic cancer; 40 patients were diagnosed as normal pancreas. The tissue peak and blood flow of the head, body, tail of normal pancreas without any pancreatic disease are 109.63 ± 16.60 and 131.90 ± 41.61, 104.38 ± 19.39 and 127.78 ± 42.52, 104.55 ± 15. 44 and 123.50 ± 33.44 respectively. The tissue peak and blood flow of pancreatic cancer is 59.59 ± 18.20 and 60.00 ± 15.36. For and between each group, there is no significant statistical difference for the tissue peak and blood flow of normal areas of the head, body, tail of normal pancreas. There is statistical difference for the tissue peak and blood flow of lesion and lesion-surrounding areas of pancreatic cancer in each group. However, there is no statistical difference for the tissue peak and blood flow of normal and diseasing areas between 3 groups.Low-dose whole-pancreas perfusion with 640-slice dynamic volume CT is feasible.