Artificial intelligence trained with integration of multiparametric MR-US imaging data and fusion biopsy trajectory-proven pathology data for 3D prediction of prostate cancer: A proof-of-concept study.

BACKGROUND: We aimed to develop an artificial intelligence (AI) algorithm that predicts the volume and location of clinically significant cancer (CSCa) using convolutional neural network (CNN) trained with integration of multiparametric MR-US image data and MRI-US fusion prostate biopsy (MRI-US PBx) trajectory-proven pathology data. METHODS: Twenty consecutive patients prospectively underwent MRI-US PBx, followed by robot-assisted radical prostatectomy (RARP). The AI algorithm was trained with the integration of MR-US image data with a MRI-US PBx trajectory-proven pathology. The relationship with the 3D-cancer-mapping of RARP specimens was compared between AI system-suggested 3D-CSCa mapping and an experienced radiologist's suggested 3D-CSCa mapping on MRI alone according to the Prostate Imaging Reporting and Data System (PI-RADS) version 2. The characteristics of detected and undetected tumors at AI were compared in 22,968 image data. The relationships between CSCa volumes and volumes predicted by AI as well as the radiologist's reading based on PI-RADS were analyzed. RESULTS: The concordance of the CSCa center with that in RARP specimens was significantly higher in the AI prediction than the radiologist' reading (83% vs. 54%, p = 0.036). CSCa volumes predicted with AI were more accurate (r = 0.90, p < 0.001) than the radiologist's reading. The limitations include that the elastic fusion technology has its own registration error. CONCLUSIONS: We presented a novel pilot AI algorithm for 3D prediction of PCa. AI was trained by integration of multiparametric MR-US image data and fusion biopsy trajectory-proven pathology data. This deep learning AI model may more precisely predict the 3D mapping of CSCa in its volume and center location than a radiologist's reading based on PI-RADS version 2, and has potential in the planning of focal therapy.

MRI features of the nested variant of urothelial carcinoma of the urinary bladder: report of four cases.

We report on four patients with the nested variant of urothelial carcinoma (NVUC) of the urinary bladder and focus on their magnetic resonance imaging (MRI) findings. MRI showed that all lesions had irregular wall thickening with little protrusion into the bladder lumen. All had extravesical invasion, and two had invaded other organs (uterus and seminal vesicle). On T2-weighted images, all tumors mainly showed relatively strong hypointensity similar to that of the muscularis propria, and in three cases there was also a thin hyperintense layer on the tumor surface, suggesting edematous mucosa. Diffusion-weighted images demonstrated different degrees of hyperintensity, which was faint in one case. Dynamic contrast-enhanced MRI was performed in two cases and both showed gradual contrast enhancement. It has been suggested that NVUC may produce unique MRI findings reflecting its pathological features. It would be useful for those who interpret bladder MRI to recognize this rare urothelial carcinoma variant.

MRI features of mucinous adenocarcinoma of the prostate: report of four cases.

CASE REPORTS: We report four patients with mucinous adenocarcinoma of the prostate, focusing on their magnetic resonance imaging (MRI) findings. The lesions appeared hyperintense on T2-weighted images (T2WI) in all four patients. In the two patients in whom the tumors were confined to the peripheral zone (PZ), the lesions were isointense to the surrounding normal PZ, making them difficult to identify. In all three patients who underwent diffusion-weighted imaging (b = 1000), the lesions appeared hyperintense but visually there was little or no decrease in the apparent diffusion coefficient (ADC), with ADC values of 1.02, 1.39, and 1.66 × 10-3 mm2/s, respectively. In the three patients who underwent a dynamic contrast-enhanced MR study, early enhancement was evident in two (partial in one), and gradually increasing enhancement in one. CONCLUSION: In the four cases in this instance, MRI findings of mucinous adenocarcinoma were very different from the appearance of non-mucinous adenocarcinoma. It is suggested that the conventional interpretation method of MRI for prostate cancer is not suitable for mucinous adenocarcinoma. Those who interpret prostate MRI should be aware of this rare and unique subtype of prostate cancer.

Whole-tumor histogram analysis of DWI and QSI for differentiating between meningioma and schwannoma: a pilot study.

PURPOSE: To investigate whether whole-tumor histogram analyses of diffusivity measurements derived from q-space imaging (QSI) improves the differentiation between meningioma and schwannoma. MATERIALS AND METHODS: Fifteen extra-axial tumors (11 meningiomas and 4 schwannomas) with MR examinations from April 2011 to May 2013 were included. Three-dimensional regions of interest (ROI) encompassed the whole tumor, including cystic areas. Histogram analyses of mean displacement (MD) derived from QSI and apparent diffusion coefficient (ADC) for the ROI were performed at mean, the five percentiles of MDn and ADCn (n = 5, 25, 50, 75, 95th), kurtosis, and skewness. To determine the diagnostic ability of MDn and ADCn, we also compared the area under the curve (AUC) on receiver operating characteristic (ROC) analysis. RESULTS: Histogram analyses revealed significant differences between meningioma and schwannoma in MD75, ADC25, ADC50, ADC75, and kurtosis of ADC. The ROC analysis of kurtosis of ADC and MD75 resulted in an AUC of 1.0 and 0.96, respectively. There were no significant differences between the AUC of MD75 and that of kurtosis of ADC (p = 0.41). CONCLUSION: The histogram analyses of MD and ADC derived from QSI were both equally useful in differentiating between intracranial meningioma and schwannoma.

The split sign: The MRI equivalent of the bell clapper deformity.

METHODS:: The cases of eight patients who underwent MRI and surgery for acute scrotum between January 2010 and January 2017 were evaluated. We recorded whether hyperintense fluid on T2 weighted images existed between the posterior aspect of the epididymis and the scrotal wall ("split sign") and investigated if it correlated with BCD in surgical findings. RESULTS:: In one patient without hydrocele, readers were unable to evaluate the anatomy of the tunica vaginalis. Among seven patients with hydrocele, five had the split sign and all were surgically confirmed as BCD. In two patients with hydrocele but no split sign, one had normal scrotal anatomy and the other had a BCD with a necrotic testis adherent to the scrotal wall. CONCLUSION:: The split sign on MRI corresponded well to the lack of fixation of the epididymis to the scrotal wall and detected BCD with high sensitivity (5/6). ADVANCES IN KNOWLEDGE:: A hyperintense area on T2 weighted image between the posterior aspect of the epididymis and scrotal wall ("split sign") is a useful MRI finding for diagnosing BCD.

Can we shorten the q-space imaging to make it clinically feasible?

PURPOSE: Q-space imaging (QSI) is a novel magnetic resonance imaging (MRI) technique that enables assessment of micro-structural changes of white matter. The acquisition time, however, is comparatively long to use for routine clinical assessment. Therefore, the present study investigated the clinically feasible b value combinations to measure the water molecular displacement probability density function (PDF) in healthy subjects. METHODS: The subjects consisted of five healthy volunteers (1 female and 4 male; 40.8 ± 13.2 years). All MRIs were examined at 1.5 T. The QSI was acquired using a single-shot echo-planar imaging and Δ/δ = 142/17 ms. The magnitude of the gradients was incremented in nine steps to reach a maximal b = 10,000 s/mm2. The total acquisition time of this original QSI was 17.4 min. The feasibility of ten alternative b value combinations with the zero-filling or curve fitting technique was assessed. The mean diffusivities (MDs), kurtosis, and zero displacement probability (ZDP) were obtained, and these results were compared in manually segmented regions of interest. RESULTS: There were alternative b value combinations with a 7.5-min acquisition time and with almost the same PDF. CONCLUSION: A few alternative b value combinations with the curve fitting technique were found to be able to shorten the QSI acquisition for its clinical feasibility of MD and ZDP.

High levels of urinary midkine in various cancer patients.

Midkine (MK) is a heparin-binding growth factor, which promotes growth, migration, and survival of various cells, and MK expression is increased in many human carcinomas. We determined the urinary MK level by enzyme-linked immunoassay. Taking 311pg/mg creatinine as a cut-off level, 70% of patients with various carcinomas (n=142) gave positive values, while only 5.5% of healthy volunteers (n=330) did. In case of gastric carcinoma, 17 out of 21 patients with stage 1 tumor were positive. Urinary MK levels are expected to become a convenient marker as an aid in detection of tumors.