Development of a Spine X-Ray-Based Fracture Prediction Model Using a Deep Learning Algorithm.

BACKGRUOUND: Since image-based fracture prediction models using deep learning are lacking, we aimed to develop an X-ray-based fracture prediction model using deep learning with longitudinal data. METHODS: This study included 1,595 participants aged 50 to 75 years with at least two lumbosacral radiographs without baseline fractures from 2010 to 2015 at Seoul National University Hospital. Positive and negative cases were defined according to whether vertebral fractures developed during follow-up. The cases were divided into training (n=1,416) and test (n=179) sets. A convolutional neural network (CNN)-based prediction algorithm, DeepSurv, was trained with images and baseline clinical information (age, sex, body mass index, glucocorticoid use, and secondary osteoporosis). The concordance index (C-index) was used to compare performance between DeepSurv and the Fracture Risk Assessment Tool (FRAX) and Cox proportional hazard (CoxPH) models. RESULTS: Of the total participants, 1,188 (74.4%) were women, and the mean age was 60.5 years. During a mean follow-up period of 40.7 months, vertebral fractures occurred in 7.5% (120/1,595) of participants. In the test set, when DeepSurv learned with images and clinical features, it showed higher performance than FRAX and CoxPH in terms of C-index values (DeepSurv, 0.612; 95% confidence interval [CI], 0.571 to 0.653; FRAX, 0.547; CoxPH, 0.594; 95% CI, 0.552 to 0.555). Notably, the DeepSurv method without clinical features had a higher C-index (0.614; 95% CI, 0.572 to 0.656) than that of FRAX in women. CONCLUSION: DeepSurv, a CNN-based prediction algorithm using baseline image and clinical information, outperformed the FRAX and CoxPH models in predicting osteoporotic fracture from spine radiographs in a longitudinal cohort.

Differentiation of multiple myeloma and metastases: Use of axial diffusion-weighted MR imaging in addition to standard MR imaging at 3T.

BACKGROUND: Metastasis and multiple myeloma are common malignant bone marrow lesions which may be difficult to distinguish because of similar imaging findings. The purpose of this study was to determine the value of adding diffusion-weighted imaging (DWI) to standard MR imaging to differentiate multiple myeloma from metastasis. METHODS: 25 patients with metastasis and 18 patients with multiple myeloma underwent 3T MR imaging with DWI (b = 0, 800 s/mm2) were enrolled. They all had pathologically confirmed bone lesions and were in a treatment naïve state. Two readers who were blind of final diagnosis measured the average ADC (ADCav) and minimum ADC (ADCmin) on the DWI. They then estimated the diagnosis, based on the standard MR imaging and measured ADC values. Another reader performed histogram analysis on the whole tumor volume and obtained mean ADC (ADCvol), standard deviation (SDvol), skewness, and kurtosis. Comparison of the obtained values from DWI was performed by the t-test or Mann-Whitney U test. The receiver operating characteristic (ROC) curve with areas under the curve (AUC) was used to obtain the cut off values and to evaluate the diagnostic performance of the two readers. RESULTS: ADCav, ADCmin, and ADCvol of multiple myeloma were significantly lower than those of metastasis: ADCav, 752 µm2/sec versus 1081 µm2/sec; ADCmin, 704 µm2/sec vs 835 µm2/sec; ADCvol 761 µm2/sec vs 1184 µm2/sec (p < .001). In histogram analysis, ADC values of multiple myeloma showed narrow distribution than metastasis: SDvol, 144 vs 257 (p < .001). Areas under the receiver operating characteristic curve was significantly higher with additive DWI than standard MR alone: 0.762 vs 0.953; 0.706 vs 0.950 (p < .05) for two readers. CONCLUSIONS: This study suggested that the addition of axial DWI to standard MR imaging can be helpful to diagnose multiple myeloma from metastasis at 3T.

Origin of Satellite Ganglion Cysts with Effusion in the Flexor Hallucis Longus Tendon Sheath around the Hallux.

BACKGROUND: To describe the clinical and magnetic resonance imaging findings of ganglion cysts with effusion in the flexor hallucis longus tendon sheath around the hallux to evaluate their origin. METHODS: Patients with recurrent or painful ganglion cysts around the hallux with effusion in the flexor hallucis longus tendon sheath who underwent surgical treatment at St. Vincent's Hospital from February 2007 to August 2016 were investigated. Surgical indication was a painful or recurrent mass caused by the cystic lesions. Those without effusion of the flexor hallucis longus tendon sheath were excluded. We assessed the clinical and magnetic resonance imaging findings. RESULTS: Magnetic resonance imaging findings in all patients showed several ganglion cysts around the hallux and large fluid accumulations within the flexor hallucis longus tendon sheath. Regarding the location, six ganglion cysts were on the dorsomedial aspect, one on the plantar medial aspect, seven on the plantar lateral aspect, and one in the toe pulp. Ten patients showed joint effusions in both the metatarsophalangeal and interphalangeal joints, two in the metatarsophalangeal joints, and three in the interphalangeal joints. There were communication stalks with a tail shape or abutment between ganglion cysts with surrounding joint effusions. Intraoperatively, connections between ganglion cysts, the synovial cyst of the flexor hallucis longus tendon sheath, and surrounding joints were seen. CONCLUSIONS: Synovial fluid accumulation in the metatarsophalangeal or interphalangeal joint supplies the synovial cyst of the flexor hallucis longus tendon sheath and subsequently ganglion cysts in the hallux. In clinical practice, the surgeon should carefully check surrounding joints with tendon sheaths to prevent recurrence of the ganglion cysts around the hallux.

Differentiation of acute osteoporotic and malignant compression fractures of the spine: use of additive qualitative and quantitative axial diffusion-weighted MR imaging to conventional MR imaging at 3.0 T.

PURPOSE: To retrospectively determine the value of adding qualitative and quantitative axial diffusion-weighted (DW) imaging to standard spine magnetic resonance (MR) imaging to differentiate between acute osteoporotic and malignant compression fractures at 3.0 T. MATERIALS AND METHODS: The institutional ethics committee approved this retrospective study and waived the requirement to obtain informed consent. The authors retrospectively analyzed 3.0-T MR images, including DW images (b values: 0, 800, and 1400 sec/mm(2)), in 62 patients with acute compression fractures. Three radiologists independently interpreted MR images for the presence of malignancy by using conventional MR images alone and in combination with axial DW images with qualitative and quantitative analysis. Apparent diffusion coefficients (ADCs) were measured within solid portion with careful use of a small region of interest (ROI). The Mann-Whitney U test was performed. RESULTS: There were 30 malignant and 32 acute osteoporotic compression fractures. At qualitative analysis, hyperintensity relative to spinal cord was more frequent in malignant compression fractures than in acute osteoporotic compression fractures (87% vs 22%, respectively; P < .001). Median ADCs of malignant fractures were significantly lower than those of benign fractures (P < .001). With conventional MR imaging alone, sensitivity, specificity, and accuracy were 100%, 94%, and 97%, respectively, for reader 1; 97%, 78%, and 87% for reader 2; and 100%, 84%, and 92% for reader 3. With conventional and DW MR imaging combined, sensitivity, specificity, and accuracy were 100%, 97%, and 98% for all three readers. The addition of DW imaging led to correct changes in diagnosis: Reader 1 improved by 1.6% (one of 62 fractures), reader 2 improved by 11% (seven of 62 fractures), and reader 3 improved by 6.5% (four of 62 fractures). CONCLUSION: The addition of axial DW imaging to a conventional MR imaging protocol improved diagnostic accuracy in the differentiation of acute osteoporotic from malignant compression fractures by measuring ADCs in the solid portion with careful use of a small ROI.