Joint segmentation and classification of breast masses based on ultrasound radio-frequency data and convolutional neural networks.

In this paper, we propose a novel deep learning method for joint classification and segmentation of breast masses based on radio-frequency (RF) ultrasound (US) data. In comparison to commonly used classification and segmentation techniques, utilizing B-mode US images, we train the network with RF data (data before envelope detection and dynamic compression), which are considered to include more information on tissue's physical properties than standard B-mode US images. Our multi-task network, based on the Y-Net architecture, can effectively process large matrices of RF data by mixing 1D and 2D convolutional filters. We use data collected from 273 breast masses to compare the performance of networks trained with RF data and US images. The multi-task model developed based on the RF data achieved good classification performance, with area under the receiver operating characteristic curve (AUC) of 0.90. The network based on the US images achieved AUC of 0.87. In the case of the segmentation, we obtained mean Dice scores of 0.64 and 0.60 for the approaches utilizing US images and RF data, respectively. Moreover, the interpretability of the networks was studied using class activation mapping technique and by filter weights visualizations.

Breast mass segmentation in ultrasound with selective kernel U-Net convolutional neural network.

In this work, we propose a deep learning method for breast mass segmentation in ultrasound (US). Variations in breast mass size and image characteristics make the automatic segmentation difficult. To address this issue, we developed a selective kernel (SK) U-Net convolutional neural network. The aim of the SKs was to adjust network's receptive fields via an attention mechanism, and fuse feature maps extracted with dilated and conventional convolutions. The proposed method was developed and evaluated using US images collected from 882 breast masses. Moreover, we used three datasets of US images collected at different medical centers for testing (893 US images). On our test set of 150 US images, the SK-U-Net achieved mean Dice score of 0.826, and outperformed regular U-Net, Dice score of 0.778. When evaluated on three separate datasets, the proposed method yielded mean Dice scores ranging from 0.646 to 0.780. Additional fine-tuning of our better-performing model with data collected at different centers improved mean Dice scores by ~6%. SK-U-Net utilized both dilated and regular convolutions to process US images. We found strong correlation, Spearman's rank coefficient of 0.7, between the utilization of dilated convolutions and breast mass size in the case of network's expansion path. Our study shows the usefulness of deep learning methods for breast mass segmentation. SK-U-Net implementation and pre-trained weights can be found at github.com/mbyr/bus_seg.

Myocardial dysfunction in children with intrauterine growth restriction: an echocardiographic study.

INTRODUCTION: The prevalence of intrauterine growth restriction (IUGR) is about 3-10% of live-born newborns and can be as high as 20% in developing countries. It may result in the occurrence of cardiovascular diseases later in life. METHODS: The aim of this study was echocardiographic evaluation, with the use of conventional and tissue Doppler parameters, of cardiac function in children born with IUGR, and comparison with healthy peers born as normally grown foetuses. RESULTS: In the IUGR group, E wave and E/A ratio were significantly lower compared to the control group. A wave, isovolumetric relaxation time, deceleration time, myocardial performance index as well as E/E' septal and E/E' lateral indices were significantly higher compared to healthy peers. CONCLUSION: Children with IUGR presented with subclinical myocardial dysfunction.

Prenatal diagnosis and follow-up of 23 cases of cardiac tumors.

OBJECTIVE: To evaluate the prenatal characteristics and postnatal outcome of cardiac tumors diagnosed at two prenatal Polish cardiology centers. METHODS: Descriptive analysis of 23 fetuses with cardiac tumors (12 multiple and 11 single) diagnosed over 16 years (from 1993 to 2009). Congestive heart failure was diagnosed when the cardiovascular profile score was seven or less. RESULTS: Associated structural congenital heart defects were present in three fetuses, extracardiac anomalies in three, and chromosomal anomalies in two. Congestive heart failure developed in five cases. Perinatal survival was not different between cases with and without cardiac failure (2/5 vs 12/18, p = 0.28). The main ultrasonographic signs observed prenatally in association with cardiac tumors were cardiomegaly, left ventricular outflow tract obstruction, pericardial effusion, and hypokinesis. A diagnosis of tuberous sclerosis was eventually made in all 12 fetuses with multiple tumors. Perinatal death occurred in 4/11 cases with single tumors and in 5/12 with multiple tumors (p = 0.57). Surgical resection of the tumor was performed in 3/11 neonates with single tumors (histopathologically: rhabdomyoma, teratoma, and fibroma) and in 2/12 with multiple tumors (both rhabdomyomas). CONCLUSIONS: Survival is not different between neonates with single and multiple tumors and between those with and without congestive heart failure.

The use of recombinant coagulation factor VIIa in uncontrolled postoperative bleeding in children undergoing cardiac surgery with cardiopulmonary bypass.

OBJECTIVE: To assess the hemostatic efficacy of recombinant coagulation factor VIIa (rFVIIa) in the management of uncontrolled bleeding in postcardiac surgery with cardiopulmonary bypass in children. DESIGN: An open-label study. SETTING: A postoperative intensive care unit. PATIENTS: Eight consecutive pediatric patients with excessive bleeding after cardiac surgery with cardiopulmonary bypass that met the criteria for reexploration and did not respond to optimal transfusions of platelets and fresh frozen plasma. INTERVENTIONS: rFVIIa 30 microg/kg was given as a bolus injection. A higher dose of 60 microg/kg was used if a patient had preoperative coagulopathy, preoperative multiple-organ failure, or indications that required an emergency operation. The same dose was repeated 15 mins after the previous injection if the bleeding had not decreased. If the bleeding had decreased but still exceeded 10 mL/hr for body weight 5 kg, the same dose was repeated 2 hrs after the previous injection. A maximum of four doses could be given before rFVIIa was considered ineffective and a reexploration was needed. MEASUREMENTS AND MAIN RESULTS: Postoperative blood loss was estimated from the volume of chest tube drainage. rFVIIa successfully controlled bleeding and prevented reexploration in all seven patients who received treatment according to the protocol. One patient who received only one dose of rFVIIa required reexploration because a second dose was not available. No adverse events related to rFVIIa were seen. CONCLUSIONS: rFVIIa may be useful in preventing reexploration in uncontrolled postoperative bleeding in children undergoing cardiac surgery with cardiopulmonary bypass. Randomized, placebo-controlled studies are needed to confirm the safety and efficacy of rFVIIa in this clinical setting.