Assessment of breast lesions by the Kaiser score for differential diagnosis on MRI: the added value of ADC and machine learning modeling.

OBJECTIVES: To evaluate the diagnostic performance of Kaiser score (KS) adjusted with the apparent diffusion coefficient (ADC) (KS+) and machine learning (ML) modeling. METHODS: A dataset of 402 malignant and 257 benign lesions was identified. Two radiologists assigned the KS. If a lesion with KS > 4 had ADC > 1.4 × 10-3 mm2/s, the KS was reduced by 4 to become KS+. In order to consider the full spectrum of ADC as a continuous variable, the KS and ADC values were used to train diagnostic models using 5 ML algorithms. The performance was evaluated using the ROC analysis, compared by the DeLong test. The sensitivity, specificity, and accuracy achieved using the threshold of KS > 4, KS+ > 4, and ADC ≤ 1.4 × 10-3 mm2/s were obtained and compared by the McNemar test. RESULTS: The ROC curves of KS, KS+, and all ML models had comparable AUC in the range of 0.883-0.921, significantly higher than that of ADC (0.837, p < 0.0001). The KS had sensitivity = 97.3% and specificity = 59.1%; and the KS+ had sensitivity = 95.5% with significantly improved specificity to 68.5% (p < 0.0001). However, when setting at the same sensitivity of 97.3%, KS+ could not improve specificity. In ML analysis, the logistic regression model had the best performance. At sensitivity = 97.3% and specificity = 65.3%, i.e., compared to KS, 16 false-positives may be avoided without affecting true cancer diagnosis (p = 0.0015). CONCLUSION: Using dichotomized ADC to modify KS to KS+ can improve specificity, but at the price of lowered sensitivity. Machine learning algorithms may be applied to consider the ADC as a continuous variable to build more accurate diagnostic models. KEY POINTS: • When using ADC to modify the Kaiser score to KS+, the diagnostic specificity according to the results of two independent readers was improved by 9.4-9.7%, at the price of slightly degraded sensitivity by 1.5-1.8%, and overall had improved accuracy by 2.6-2.9%. • When the KS and the continuous ADC values were combined to train models by machine learning algorithms, the diagnostic specificity achieved by the logistic regression model could be significantly improved from 59.1 to 65.3% (p = 0.0015), while maintaining at the high sensitivity of KS = 97.3%, and thus, the results demonstrated the potential of ML modeling to further evaluate the contribution of ADC. • When setting the sensitivity at the same levels, the modified KS+ and the original KS have comparable specificity; therefore, KS+ with consideration of ADC may not offer much practical help, and the original KS without ADC remains as an excellent robust diagnostic method.

Non-joint effusion is associated with osteoarthritis in temporomandibular joints with disk displacement.

PURPOSE: The aim of this study was to examine the possible association between different grades of joint effusion (JE) and osteoarthritis (OA) in temporomandibular joint (TMJ) anterior disk displacement without reduction (ADDwoR). MATERIAL AND METHODS: A sample of 101 female patients 20-40 years of age with unilateral TMJ ADDwoR were retrospectively reviewed. JE and OA were diagnosed with magnetic resonance imaging (MRI). JE was subdivided into three different grades: grade 0, no or minimal effusion; grade 1, moderate effusion; and grade 2, extensive effusion. Eight categories of degenerative changes were used for screening for the existence of OA. Cases with no less than one type of degenerative change were diagnosed as OA. RESULTS: In all, 71 patients (70.3%) were diagnosed as having OA in the joints with disk displacement. In the univariate analysis, the proportion of subjects with non-JE (grade 0) was higher in the OA group (p = 0.003), while the proportion of subjects with extensive effusion (grade 2) was lower in the OA group (p = 0.02). In the multivariate logistic regression analysis, non-JE was independently associated with the development of OA (odds ratio = 5.68, 95% confidence interval = 1.10-29.37, P = 0.04). CONCLUSION: The results suggested that non-JE was associated with OA in the joints with ADDwoR.

Magnetic resonance imaging assessment of temporomandibular joint soft tissue injuries of intracapsular condylar fracture.

We evaluated the soft tissue of the temporomandibular joint (TMJ) with magnetic resonance imaging (MRI) after intracapsular condylar fracture. Eighteen consecutive patients (19 TMJ) were diagnosed between 1 January 2010 and 30 October 2011. They were examined using bilateral sagittal and coronal MRI, which were obtained immediately after injury to assess the displacement of the disc, whether there was a tear in capsule or the retrodiscal tissue, and whether there was an effusion in the joint. On the affected side MRI showed disc displacement in 15 of 19, tears in the capsule in 9, and tears in the retrodiscal tissue in 16. All 19 had joint effusions. It also showed 2 joints with abnormalities on the unaffected side. We conclude that MRI is useful for diagnosis and for estimating the amount of damage to the TMJ, and is helpful in planning treatment.

In vivo tracing of superparamagnetic iron oxide-labeled bone marrow mesenchymal stem cells transplanted for traumatic brain injury by susceptibility weighted imaging in a rat model.

OBJECTIVE: To label rat bone marrow mesenchymal stem cells (BMSCs) with superparamagnetic iron oxide (SPIO) in vitro, and to monitor the survival and location of these labeled BMSCs in a rat model of traumatic brain injury (TBI) by susceptibility weighted imaging (SWI) sequence. METHODS: BMSCs were cultured in vitro and then labeled with SPIO. Totally 24 male Sprague Dawley (SD) rats weighing 200-250 g were randomly divided into 4 groups: Groups A-D (n equal to 6 for each group). Moderate TBI models of all the rats were developed in the left hemisphere following Feeney's method. Group A was the experimental group and stereotaxic transplantation of BMSCs labeled with SPIO into the region nearby the contusion was conducted in this group 24 hours after TBI modeling. The other three groups were control groups with transplantation of SPIO, unlabeled BMSCs and injection of nutrient solution respectively conducted in Groups B, C and D at the same time. Monitoring of these SPIO-labeled BMSCs by SWI was performed one day, one week and three weeks after implantation. RESULTS: Numerous BMSCs were successfully labeled with SPIO. They were positive for Prussian blue staining and intracytoplasm positive blue stained particles were found under a microscope (200). Scattered little iron particles were observed in the vesicles by electron microscopy (5000). MRI of the transplantation sites of the left hemisphere demonstrated a low signal intensity on magnitude images, phase images and SWI images for all the test rats in Group A, and the lesion in the left parietal cortex demonstrated a semicircular low intensity on SWI images, which clearly showed the distribution and migration of BMSCs in the first and third weeks. For Group B, a low signal intensity by MRI was only observed on the first day but undetected during the following examination. No signals were observed in Groups C and D at any time points. CONCLUSION: SWI sequence in vivo can consecutively and noninvasively trace and demonstrate the status and distribution of BMSCs labeled with SPIO in the brain of TBI model rats.