CT Radiomics for Predicting Pathological Complete Response of Axillary Lymph Nodes in Breast Cancer After Neoadjuvant Chemotherapy: A Prospective Study.

BACKGROUND: The diagnostic effectiveness of traditional imaging techniques is insufficient to assess the response of lymph nodes (LNs) to neoadjuvant chemotherapy (NAC), especially for pathological complete response (pCR). A radiomics model based on computed tomography (CT) could be helpful. PATIENTS AND METHODS: Prospective consecutive breast cancer patients with positive axillary LNs initially were enrolled, who received NAC prior to surgery. Chest contrast-enhanced thin-slice CT scan was performed both before and after the NAC (recorded as the first and the second CT respectively), and on both of them, the target metastatic axillary LN was identified and demarcated layer by layer. Using pyradiomics-based software that was independently created, radiomics features were retrieved. A pairwise machine learning workflow based on Sklearn (https://scikit-learn.org/) and FeAture Explorer was created to increase diagnostic effectiveness. An effective pairwise auto encoder model was developed by the improvement of data normalization, dimensionality reduction, and features screening scheme as well as the comparison of the prediction effectiveness of the various classifiers. RESULTS: A total of 138 patients were enrolled, and 77 (58.7%) in the overall group achieved pCR of LN after NAC. Nine radiomics features were finally chosen for modeling. The AUCs of the training group, validation group, and test group were 0.944 (0.919-0.965), 0.962 (0.937-0.985), and 1.000 (1.000-1.000), respectively, and the corresponding accuracies were 0.891, 0.912, and 1.000. CONCLUSION: The pCR of axillary LNs in breast cancer following NAC can be precisely predicted using thin-sliced enhanced chest CT-based radiomics.

Diffusion-weighted images (DWI) without ADC values in assessment of small focal nodules in cirrhotic liver.

OBJECTIVE: To assess if diffusion-weighted magnetic resonance (MR) imaging without apparent diffusion coefficient (ADC) values provides added diagnostic value in combination with conventional MR imaging in the detection and characterization of small nodules in cirrhotic liver. METHODS: Two observers retrospectively and independently analyzed 86 nodules (≤3 cm) certified pathologically in 33 patients with liver cirrhosis, including 48 hepatocellular carcinoma (HCC) nodules, 13 high-grade dysplastic nodules (HDN), 10 low-grade dysplastic nodules (LDNs) and 15 other benign nodules. All these focal nodules were evaluated with conventional MR images (T1-weighted, T2-weighted and dynamic gadolinium-enhanced images) and breath-hold diffusion-weighted images (DWI) (b=500 s/mm(2)). The nodules were classified by using a scale of 1-3 (1, not seen; 3, well seen) on DWI for qualitative assessment. These small nodules were characterized by two radiologists. ADC values weren't measured. The diagnostic performance of the combined DWI-conventional images and the conventional images alone was evaluated using receiver operating characteristic (ROC) curves. The area under the curves (Az), sensitivity and specificity values for characterizing different small nodules were also calculated. RESULTS: Among 48 HCC nodules, 33 (68.8%) were graded as 3 (well seen), 6 (12.5%) were graded as 2 (partially obscured), and 9 weren't seen on DWI. Among 13 HDNs, there were 3 (23.1%) and 4 (30.8%) graded as 3 and 2 respectively. Five (50%) of 10 benign nodules were partially obscured and slightly hyperintense. For 86 nodules, the average diagnostic accuracy of combined DWI-conventional images was 82.56%, which was increased significantly compared with conventional MR images with 76.17%. For HCC and HDN, the diagnostic accuracy of combined DWI-conventional images increased from 78.69% to 86.07%. CONCLUSIONS: Diffusion-weighted MR imaging does provide added diagnostic value in the detection and characterization of HDN and HCC, and it may not be helpful for LDN and regenerative nodule (RN) in cirrhotic liver.

Pairwise machine learning-based automatic diagnostic platform utilizing CT images and clinical information for predicting radiotherapy locoregional recurrence in elderly esophageal cancer patients.

OBJECTIVE: To investigate the feasibility and accuracy of predicting locoregional recurrence (LR) in elderly patients with esophageal squamous cell cancer (ESCC) who underwent radical radiotherapy using a pairwise machine learning algorithm. METHODS: The 130 datasets enrolled were randomly divided into a training set and a testing set in a 7:3 ratio. Clinical factors were included and radiomics features were extracted from pretreatment CT scans using pyradiomics-based software, and a pairwise naive Bayes (NB) model was developed. The performance of the model was evaluated using receiver operating characteristic (ROC) curves and decision curve analysis (DCA). To facilitate practical application, we attempted to construct an automated esophageal cancer diagnosis system based on trained models. RESULTS: To the follow-up date, 64 patients (49.23%) had experienced LR. Ten radiomics features and two clinical factors were selected for modeling. The model demonstrated good prediction performance, with area under the ROC curve of 0.903 (0.829-0.958) for the training cohort and 0.944 (0.849-1.000) for the testing cohort. The corresponding accuracies were 0.852 and 0.914, respectively. Calibration curves showed good agreement, and DCA curve confirmed the clinical validity of the model. The model accurately predicted LR in elderly patients, with a positive predictive value of 85.71% for the testing cohort. CONCLUSIONS: The pairwise NB model, based on pre-treatment enhanced chest CT-based radiomics and clinical factors, can accurately predict LR in elderly patients with ESCC. The esophageal cancer automated diagnostic system embedded with the pairwise NB model holds significant potential for application in clinical practice.

Fat fraction quantification of lumbar spine: comparison of T1-weighted two-point Dixon and single-voxel magnetic resonance spectroscopy in diagnosis of multiple myeloma.

PURPOSE: We aimed to investigate the value of T1-weighted two-point Dixon technique and single-voxel magnetic resonance spectroscopy (MRS) in diagnosis of multiple myeloma (MM) through quantifying fat content of vertebral marrow. METHODS: A total of 30 MM patients and 30 healthy volunteers underwent T1-weighted two-point Dixon and single-voxel MRS imaging. The fat fraction map (FFM) was reconstructed from the Dixon images using the equation FFM = Lip/In, where Lip represents fat maps and In represents in-phase images. The fat fraction (FF) of MRS was calculated by using the integral area of Lip peak divided by the sum of integral area of Lip peak and water peak. RESULTS: FF values measured by the Dixon technique and MRS were significantly decreased in MM patients (45.99%±3.39% and 47.63%±4.38%) compared with healthy controls (64.43%±0.96% and 76.22%±1.91%) (P < 0.001 with both methods). FF values measured by Dixon technique were significantly positively correlated to those measured by MRS in MM (r = 0.837, P < 0.001) and healthy control group (r = 0.735, P < 0.001), respectively. There was no significant difference between area under the curve (AUC) obtained by the Dixon technique (0.878±0.047; range, 0.785 to 0.971; optimal cutoff, 56.35 for healthy controls vs. MM) and MRS (0.883±0.047; range, 0.791 to 0.974; optimal cutoff, 61.00 for healthy controls vs. MM). The ability of Dixon technique to differentiate MM group from healthy controls was equivalent to single-voxel MRS. CONCLUSION: Regarding detection of fat contents in vertebral bone, T1-weighted two-point Dixon technique exhibited equivalent performance to single-voxel MRS in the diagnosis of multiple myeloma. Moreover, two-point Dixon is a more convenient and stable technique for assessing bone marrow changes in MM patients than single-voxel MRS.

Intravoxel Incoherent Motion Magnetic Resonance Imaging with Integrated Slice-specific Shimming for old myocardial infarction: A Pilot Study.

Currently, little is known regarding the value of quantitative parameters derived from the intravoxel incoherent motion (IVIM) magnetic resonance imaging (MRI) with integrated slice-specific shimming (iShim) sequence in detecting old myocardial infarction and myocardial fibrosis. This study was to investigate the value of IVIM-MRI with iShim sequence in diagnosing old myocardial infarction and fibrosis. Thirty-five patients with both old myocardial infarction and myocardial fibrosis and 12 healthy volunteers were prospectively enrolled to undergo cardiac diffusion-weighted imaging (DWI) using seven b-values (0, 20, 60, 80, 120, 200 and 600 s/mm2). The iShim sequence was used for IVIM data acquisition, and the diffusion parameters, D, D* and f values for IVIM, and conventional apparent diffusion coefficient (ADC) were evaluated on the anterior, posterior and lateral walls of the ventricular septum using the short axis of the heart. Significant differences were found in the D, D* and f values between healthy subjects and patients with old myocardial infarction and myocardial fibrosis (P = 0.000), with the median value of the D and f significantly smaller in the myocardial infarction and fibrosis than in the normal control but the median value of D* significantly greater in the myocardial infarction and fibrosis than in the normal control. In the receiver operating curve analysis, the areas under the curve were 0.939, 0.988 and 0.959 for the D, D* and f values, respectively. The sensitivities and specificities were 84.6% and 94.4% for D, 88.9% and 84.6% for D* and 100% and 93.1% for the f values, respectively. In conclusion, the IVIM-derived parameters (D, D* and f) obtained using the iShim DWI technique showed high capacity in diagnosing old myocardial infarction and myocardial fibrosis by providing diffusion and perfusion information, which may have great importance in future clinical practice.

Differences in brain pathological changes between rotenone and 6-hydroxydopamine Parkinson's disease models.

Rotenone and 6-hydroxydopamine are two drugs commonly used to generate Parkinson's disease animal models. They not only achieve degenerative changes of dopaminergic neurons in the substantia nigra, but also satisfy the requirements for iron deposition. However, few studies have compared the characteristics of these two models by magnetic resonance imaging. In this study, rat models of Parkinson's disease were generated by injection of 3 µg rotenone or 10 µg 6-hydroxydopamine into the right substantia nigra. At 1, 2, 4, and 6 weeks after injection, coronal whole-brain T2-weighted imaging, transverse whole-brain T2-weighted imaging, and coronal diffusion tensor weighted imaging were conducted to measure fractional anisotropy and T2* values at the injury site. The fractional anisotropy value on the right side of the substantia nigra was remarkably lower at 6 weeks than at other time points in the rotenone group. In the 6-hydroxydopamine group, the fractional anisotropy value was decreased, but T2* values were increased on the right side of the substantia nigra at 1 week. Our findings confirm that the 6-hydroxydopamine-induced model is suitable for studying dopaminergic neurons over short periods, while the rotenone-induced model may be appropriate for studying the pathological and physiological processes of Parkinson's disease over long periods.

Detecting dopaminergic neuronal degeneration using diffusion tensor imaging in a rotenone-induced rat model of Parkinson's disease: fractional anisotropy and mean diffusivity values.

Dopamine content in the basal ganglia is strongly associated with the degree of dopaminergic neuron loss in the substantia nigra pars compacta. Symptoms of Parkinson's disease might not arise until more than 50% of the substantia nigra pars compacta is lost and the dopamine content in the basal ganglia is reduced by more than 80%. Greater diagnostic sensitivity and specificity would allow earlier detection of Parkinson's disease. Diffusion tensor imaging is a recently developed magnetic resonance imaging technique that measures mean diffusivity and fractional anisotropy, and responds to changes in brain microstructure. When the microscopic barrier (including cell membranes, microtubules and other structures that interfere with the free diffusion of water) is destroyed and extracellular fluid volume accumulates, the mean diffusivity value increases; when the integrity of the microstructure (such as myelin) is destroyed, fractional anisotropy value decreases. However, there is no consensus as to whether these changes can reflect the early pathological alterations in Parkinson's disease. Here, we established a rat model of Parkinson's disease by injecting rotenone (or sunflower oil in controls) into the right substantia nigra. Diffusion tensor imaging results revealed that in the stages of disease, at 1, 2, 4, and 6 weeks after rotenone injection, fractional anisotropy value decreased, but mean diffusivity values increased in the right substantia nigra in the experimental group. Fractional anisotropy values were lower at 4 weeks than at 6 weeks in the right substantia nigra of rats from the experimental group. Mean diffusivity values were markedly greater at 1 week than at 6 weeks in the right corpus striatum of rats from the experimental group. These findings suggest that mean diffusivity and fractional anisotropy values in the brain of rat models of Parkinson's disease 4 weeks after model establishment can reflect early degeneration of dopaminergic neurons. The change in fractional anisotropy values after destruction of myelin integrity is likely to be of greater early diagnostic significance than the change in mean diffusivity values.