Toward an Expert Level of Lung Cancer Detection and Classification Using a Deep Convolutional Neural Network.

BACKGROUND: Computed tomography (CT) is essential for pulmonary nodule detection in diagnosing lung cancer. As deep learning algorithms have recently been regarded as a promising technique in medical fields, we attempt to integrate a well-trained deep learning algorithm to detect and classify pulmonary nodules derived from clinical CT images. MATERIALS AND METHODS: Open-source data sets and multicenter data sets have been used in this study. A three-dimensional convolutional neural network (CNN) was designed to detect pulmonary nodules and classify them into malignant or benign diseases based on pathologically and laboratory proven results. RESULTS: The sensitivity and specificity of this well-trained model were found to be 84.4% (95% confidence interval [CI], 80.5%-88.3%) and 83.0% (95% CI, 79.5%-86.5%), respectively. Subgroup analysis of smaller nodules (<10 mm) have demonstrated remarkable sensitivity and specificity, similar to that of larger nodules (10-30 mm). Additional model validation was implemented by comparing manual assessments done by different ranks of doctors with those performed by three-dimensional CNN. The results show that the performance of the CNN model was superior to manual assessment. CONCLUSION: Under the companion diagnostics, the three-dimensional CNN with a deep learning algorithm may assist radiologists in the future by providing accurate and timely information for diagnosing pulmonary nodules in regular clinical practices. IMPLICATIONS FOR PRACTICE: The three-dimensional convolutional neural network described in this article demonstrated both high sensitivity and high specificity in classifying pulmonary nodules regardless of diameters as well as superiority compared with manual assessment. Although it still warrants further improvement and validation in larger screening cohorts, its clinical application could definitely facilitate and assist doctors in clinical practice.

Uniportal video-assisted thoracoscopic segmentectomy for fetal adenocarcinoma lung cancer with severe pulmonary emphysema: a case report.

Background: Fetal adenocarcinoma is a very rare subtype of lung adenocarcinoma. Its incidence ranges from 0.1 to 0.87% among all primary lung neoplasms. Low-grade types tend to appear in the younger generation, and the age ranges from 20 to 50 years with a mean age of around 35 years. Surgical resection is currently the best way to treat fetal adenocarcinoma lung cancer without distant metastasis. Case report: This is a 56-year-old female who underwent low-dose computer tomography (LDCT) screening during the health examination. She used to be a heavy smoker for more than 30 years, and the CT images revealed severe bronchiectasis and emphysema. There is a solitary nodule with a diameter of 18.9 x 17.8mm in the central area of the left upper lobe. We decided to conduct left upper lobe S1~S3 segmentectomy under uniportal VATS. The surgery was successful, and the patient was discharged within one week and recovered well. The final diagnosis was fetal adenocarcinoma, low-grade (pT1cN0Mx, stage IA3). Conclusion: The first case reported as fetal adenocarcinoma lung cancer who underwent uniportal video-assisted thoracoscopic segmentectomy. We believe it is a safe and feasible procedure for low-grade types fetal adenocarcinoma patient with poor pulmonary function.

[Pattern of electrogastrogram in healthy neonates].

OBJECTIVE: To investigate the electrogastrogram (EGG) characteristics of healthy neonates. METHODS: Twenty healthy neonates born at 37-39 weeks of gestation (11 males and 9 females, Apagar's score 9.3 +/- 0.4) were enrolled in this study. EGG recordings were performed for half an hour pre- and postprandially at an interval of a week from birth until age 4 weeks. The EEG variables measured included the percentage of normal gastric rhythm, the percentage of tachygastria and bradygastria, the fed-to-fasting ratio of the EEG dominant power, as well as the EEG dominant frequency and its instability coefficient. The paired sample t test (95% CI) was used to compare the recordings. RESULTS: Between birth and age 28 days, the percentage of normal gastric rhythm ranged from 38.2 +/- 4.9% to 39.7 +/- 3.5% of recorded time, tachygastria was observed in the range of 23.7 +/- 5.4% to 23.5 +/- 4.3% of recorded time, and bradygastria was shown to be in the range of 38.1 +/- 5.5% to 36.8 +/- 3.9% of recorded time in the 20 neonates before meal. Statistically significant differences were not seen in neonates with different ages as well as during pre- and postprandial periods. The EEG dominant frequency of neonates before meal was 2.38 +/- 0.5, 2.43 +/- 0.2, 2.54 +/- 0.3, 2.57 +/- 0.2 and 2.59 +/- 0.1 cpm at birth and at postnatal age of 7, 14, 21 and 28 days respectively. There were no significant differences in the dominant frequency and the coefficient of instability of the dominant frequency during pre- and postprandial periods. The EEG dominant frequency at postnatal age of 14, 21 and 28 days during pre- and postprandial periods was significantly higher than that at birth and at postnatal age of 7 days (P < 0.05). The coefficient of instability of the dominant frequency at postnatal age of 21 and 28 days was significantly lower than that at birth and at postnatal age of 7 and 14 days (P < 0.05). There were no statistically significant differences in the fed-to-fasting ratio of EGG dominant power in neonates with different ages. CONCLUSIONS: The pattern of electrical activity in the normal neonatal stomach appears to be different from that demonstrated in adults and children. The percentage of normal gastric rhythm is lower, and tachygastria and bradygastria are more frequently seen. The EEG dominant frequency increases with postnatal age in neonates.

Ultrastructure of antennal sensilla of the peach aphid Myzus persicae Sulzer, 1776.

The antennal sensilla of alate Myzus persicae were mapped using transmission electron microscopy and the ultrastructure of sensilla trichoidea, coeloconica, and placoidea are described. Trichoid sensilla, located on the tip of the antennae, are innervated by 2-4 neurons, with some outer dendrites reaching the distal end of the hair. Coeloconic sensilla in primary rhinaria are of two morphological types, both equipped with two dendrites. Dendrites of Type II coeloconic sensilla are enveloped in the dendrite sheath, containing the sensillum lymph. In sensilla coeloconica of Type I, instead, dendrites are enclosed by an electron opaque solid cuticle, with no space left for the sensillum lymph. The ultrastructure of big placoid sensillum reveals the presence of three groups of neurons, with 2-3 dendrites in each neuron group, while both small placoid sensilla are equipped with a single group of neurons, consisting of three dendrites. Both large and small placoid sensilla bear multiple pores on the outer cuticle. The function of these sensilla is also discussed.