Pred-SF: A Precipitation Prediction Model Based on Deep Neural Networks

How to predict precipitation accurately and efficiently is the key and difficult problem in the field of weather forecasting. At present, we can obtain accurate meteorological data through many high-precision weather sensors and use them to forecast precipitation. However, the common numerical weather forecasting methods and radar echo extrapolation methods have insurmountable defects. Based on some common characteristics of meteorological data, this paper proposes a Pred-SF model for precipitation prediction in target areas. The model focuses on the combination of multiple meteorological modal data to carry out self-cyclic prediction and a step-by-step prediction structure. The model divides the precipitation prediction into two steps. In the first step, the spatial encoding structure and PredRNN-V2 network are used to construct the autoregressive spatio-temporal prediction network for the multi-modal data, and the preliminary predicted value of the multi-modal data is generated frame by frame. In the second step, the spatial information fusion network is used to further extract and fuse the spatial characteristics of the preliminary predicted value and, finally, output the predicted precipitation value of the target region. In this paper, ERA5 multi-meteorological mode data and GPM precipitation measurement data are used for testing to predict the continuous precipitation of a specific area for 4 h. The experimental results show that Pred-SF has strong precipitation prediction ability. Some comparative experiments were also set up for comparison to demonstrate the advantages of the combined prediction method of multi-modal data and the stepwise prediction method of Pred-SF.

Step-by-Step Method Combined with Simulation-Enhanced Learner Outcomes in Minimally Invasive Breast Surgery Education

PURPOSE

Traditional standardized training in ultrasound-guided minimally invasive breast surgery (UMIBS) focused on lecture-based learning (LBL) resulted in students' insufficient analysis, limited spatial visual conversion ability, and poor practical application. This study examined the effects of the step-by-step (SBS) method combined with a simulation model in UMIBS education.

SUBJECTS AND METHODS

A total of 84 residents participated in this study. The residents were divided into the SBS group (experience group, n=42) and the LBL group (control group, n=42), and the same teacher taught the two groups to ensure a comparable result. Based on the pork simulation model, two experts evaluated student performance scores, and the total time taken by each student was also counted. The participants were surveyed with 7 questions after the training, and each answer was assigned a score of 1, 2 or 3 to compare the participants' satisfaction.

RESULTS

The average value of the surgical skills for SBS group were significantly higher than LBS group, which was 82.8±4.4 and 72.7±4.0 (t=4.27, P<0.001), the time spend of neoplasm localization by the experience group was significantly less than the control group, which was 17.9±1.6 and 20.9±1.2 secs, (t=1.58, P<0.001), and there were significant differences in puncture accuracy and excision integrity between the two groups (P<0.05). In addition, the results of the questionnaire survey showed that learning interest, surgical ability and satisfaction were better in the SBS group than in the LBS group (P<0.05), and there were no significant differences in clinical thinking and learning pressure between the two groups.

CONCLUSION

The SBS teaching method may help to improve the surgical skills and learning interest, as well as reduce adverse reactions and cultivate clinical thinking of the students in UMIBS training. Future studies could consider multicenter clinical research to further confirm the practicality of this teaching method and reduce the risk of deviation.

Review of the Geant4-DNA Simulation Toolkit for Radiobiological Applications at the Cellular and DNA Level

Simple Summary

A brief description of the methodologies to simulate ionizing radiation transport in biologically relevant matter is presented. Emphasis is given to the physical, chemical, and biological models of Geant4-DNA that enable mechanistic radiobiological modeling at the cellular and DNA level, important to improve the efficacy of existing and novel radiotherapeutic modalities for the treatment of cancer.

Abstract

The Geant4-DNA low energy extension of the Geant4 Monte Carlo (MC) toolkit is a continuously evolving MC simulation code permitting mechanistic studies of cellular radiobiological effects. Geant4-DNA considers the physical, chemical, and biological stages of the action of ionizing radiation (in the form of x- and γ-ray photons, electrons and β^±-rays, hadrons, α-particles, and a set of heavier ions) in living cells towards a variety of applications ranging from predicting radiotherapy outcomes to radiation protection both on earth and in space. In this work, we provide a brief, yet concise, overview of the progress that has been achieved so far concerning the different physical, physicochemical, chemical, and biological models implemented into Geant4-DNA, highlighting the latest developments. Specifically, the “dnadamage1” and “molecularDNA” applications which enable, for the first time within an open-source platform, quantitative predictions of early DNA damage in terms of single-strand-breaks (SSBs), double-strand-breaks (DSBs), and more complex clustered lesions for different DNA structures ranging from the nucleotide level to the entire genome. These developments are critically presented and discussed along with key benchmarking results. The Geant4-DNA toolkit, through its different set of models and functionalities, offers unique capabilities for elucidating the problem of radiation quality or the relative biological effectiveness (RBE) of different ionizing radiations which underlines nearly the whole spectrum of radiotherapeutic modalities, from external high-energy hadron beams to internal low-energy gamma and beta emitters that are used in brachytherapy sources and radiopharmaceuticals, respectively.

Transvaginal ultrasound features of normal uterosacral ligaments

OBJECTIVE

To show a step-by-step technique to assess normal uterosacral ligaments (USLs) during transvaginal ultrasound. Uterosacral ligaments represent the most common location of deep infiltrating endometriosis (DIE) in the posterior compartment and their involvement significantly increases the risk of ureteral lesions. The ultrasonographic diagnosis of DIE involving USLs is characterized by a wide range of accuracies described between studies, probably due to variations in the examination technique, quality of ultrasound equipment, and experience of the operators. Although described as a new classification system of DIE involving USLs, the technique for visualizing normal USLs has not yet been described.

DESIGN

Stepwise demonstration of the technique with narrated video footage.

SETTING

Academic tertiary hospital.

PATIENT(S)

The video shows a 33-year-old nulliparous woman scheduled for laparoscopic removal of a para-ovarian cyst of approximately 6 cm. Procedural steps were repeated and confirmed in another four patients submitted to laparoscopy for benign diseases (Table 1). Moreover, we here show the case of a 29-year-old woman with an isolated DIE nodule of the right USL with a comparison of laparoscopic and ultrasound findings. To better point out the technique applied in a clinical setting we show also cases of infiltrated USLs (Table 2) during the ultrasound scan in women scheduled for endometriosis surgery. Informed consent was obtained from all of the patients. The study protocol was approved by the local institutional Ethics Committee (580/2018/Oss/AOUBo).

INTERVENTION(S)

To understand correctly the right position of USLs we performed an ultrasound during a surgical procedure; in this way it was possible to recognize the area of interest while the surgeon filled the pouch of Douglas with fluid and highlighted the anatomical area with pliers. Uterosacral ligaments can be seen in the mid-sagittal and transverse view of the uterus.

MAIN OUTCOME MEASURE(S)

Description of appearance of normal USLs during transvaginal ultrasound.

RESULT(S)

Uterosacral ligaments, at transvaginal ultrasound, appear as hyperechoic stripes starting from the cervix and pointing laterally in a semi-horizontal direction.

CONCLUSION(S)

We provided a step-by-step technique (Table 3) that may be a useful tool to see accurately both the USLs at transvaginal ultrasound.