44 Analysis of abnormal chromatin configuration induced by inhibiting MEK at the 1-cell stage

After fertilisation, mammalian oocytes resume meiosis, leading to pronuclear (PN) formation. It has been known that the mitogen-activated protein kinase pathway negatively regulates PN formation. However, it remains largely unknown how the mitogen-activated protein kinase pathway is involved in the dynamic organisation of chromatin structure in PNs. Here, to address how MEK1/2 activity is involved in PN formation and transcriptional regulation, we examined whether treatment of fertilised oocytes with MEK1/2 inhibitor (PD0325901; PD) during the 1-cell stage affects their global epigenetic modifications and heterochromatin formation in mice. We performed IVF using superovulated oocytes from and ICR strain female mouse and treated IVF embryos with 1 µM PD for 10h from insemination to observe global epigenetic modifications such as H3K9me3 or H3K27me3, localization of heterochromatin protein. First, we found that PD treatment enhanced PN formation with abnormal pronuclear sizes, that is, larger female and smaller male PNs. Further, PD treatment for 10h after fertilisation caused significantly lower global H3K9me3 or H3K27me3 modifications in female PN [H3K9me3: 1.00±0.01 vs. 0.75±0.01 (N>10); P=1.0×10−2, H3K27me3: 1.00±0.01 vs. 0.55±0.01 (N>10); P=6.0×10−7], suggesting failure of establishment of normal histone modifications for heterochromatin. Consistently, we found loss of HP1b around the nucleolus precursor body, which is normally associated with pericentric heterochromatin, suggesting that MEK1/2 activity is required for normal PN and heterochromatin formation. Regardless, embryos treated with PD at the 1-cell stage recovered such global histone modifications through the pre-implantation embryos to reach the blastocyst stage. Together, our data show that MEK1/2 activity plays a pivotal role not only in regulation of PN formation but also heterochromatin formation during the 1-cell stage in mice.

Hybrid transgastric approach for the treatment of gastroesophageal junction pathologies

Flexible endoscopy has evolved to the point that it includes many endoluminal procedures that once required open or laparoscopic surgery, for instance, antireflux surgery, pyloromyotomy, mucosal and submucosal tumor resections, and even full-thickness resection. However, these procedures remain technically demanding due to flexible technology issues: difficult imaging, limited energy devices, lack of staplers, unsatisfactory suturing abilities, and so on. Transgastric laparoscopy or hybrid laparoscopy/flex endoscopy has been described for almost two decades as an alternative to a pure endoluminal approach, mainly for pancreatic pseudocyst drainage and full-thickness and mucosal resection of various lesions. The approach has never been widely adopted mostly due to cumbersome and difficult to maintain methods of gastric access. We propose to expand the indications of transgastric laparoscopy by using novel endoscopically placed ports to replicate endoscopic procedures particularly in the difficult to access proximal stomach such as endoluminal antireflux surgery. Under general anesthesia, five female pigs (mean weight: 27.6 kg) had endoscopic placement of 3, 5 mm intragastric ports (Endo-TAGSS, Leakwood KS, USA) using a technique similar to percutaneous endoscopic gastrostomy. A 5-mm laparoscope was used for visualization. Laparoendoscopic-assisted plication of the gastroesophageal junction (GEJ) was performed using 3-0 interrupted sutures (Polysorb®, Covidien, Mansfield, MA, USA). A functional lumen imagine probe (EndoFLIP®, Crospon, Inc., Galway, Ireland) was used to measure diameter, cross sectional area (CSA), distensibility, and compliance of GEJ before and after intervention. Once the TAGSS ports were removed, the gastrotomies were closed by using endoscopic over-the-scope clips. At the end of the procedure, animals were euthanized. Five laparoendoscopic-assisted endoluminal plications were performed. The mean operative time was 65.6 min (Endoscopic evaluation: 3.2 min, TAGSS Insertion: 11 min, EndoFLIP evaluation + GEJ Plication: 43.25 min, gastric wall closure: 15 minutes). In all cases, this technique was effective and allowed to achieve an adequate GEJ plication by endoscopic grading and EndoFLIP measurements. Median pre-plication GEJ diameter (D) and median pre-plication GEJ cross-sectional area (CSA) were 11.42 mm (8.6-13.6 mm) and 104.8 mm2 (58-146 mm2). After the procedure, these values were decreased to 6.14 mm (5.7-6.6 mm) and 29.8 mm2 (25-34 mm2) respectively (p = 0.0079). Median pre-plication distensibility (d) and compliance (C) were 7.87 mm2/mmHg (2.4-22.69 mm2/mmHg) and 190.56 mm3/mmHg (70.9-502.8 mm3/mmHg). After the procedure, these values decreased to 1.5 mm2/mmHg (0.7-2.2 mm2/mmHg) and 52.17 mm3/mmHg (21.9-98.7 mm3/mmHg) respectively (p = 0.0317). No intraoperative events were observed. Endoscopically, all valves were felt to be transitioned from a Hill grade 3 (normal state for the animal model) to a Hill grade 1 at the procedure completion. A hybrid laparoendoscopic approach is a feasible alternative for performing intragastric procedures with the assistance of conventional laparoscopic instruments; especially in cases where the intervention location limits the access to standard endoscopy or where endoscopic technology is inadequate. Further evaluation is planned in survival models and clinical trials.

Deep-learning-based, computer-aided classifier developed with a small dataset of clinical images surpasses board-certified dermatologists in skin tumour diagnosis

BACKGROUND

Application of deep-learning technology to skin cancer classification can potentially improve the sensitivity and specificity of skin cancer screening, but the number of training images required for such a system is thought to be extremely large.

OBJECTIVES

To determine whether deep-learning technology could be used to develop an efficient skin cancer classification system with a relatively small dataset of clinical images.

METHODS

A deep convolutional neural network (DCNN) was trained using a dataset of 4867 clinical images obtained from 1842 patients diagnosed with skin tumours at the University of Tsukuba Hospital from 2003 to 2016. The images consisted of 14 diagnoses, including both malignant and benign conditions. Its performance was tested against 13 board-certified dermatologists and nine dermatology trainees.

RESULTS

The overall classification accuracy of the trained DCNN was 76·5%. The DCNN achieved 96·3% sensitivity (correctly classified malignant as malignant) and 89·5% specificity (correctly classified benign as benign). Although the accuracy of malignant or benign classification by the board-certified dermatologists was statistically higher than that of the dermatology trainees (85·3% ± 3·7% and 74·4% ± 6·8%, P < 0·01), the DCNN achieved even greater accuracy, as high as 92·4% ± 2·1% (P < 0·001).

CONCLUSIONS

We have developed an efficient skin tumour classifier using a DCNN trained on a relatively small dataset. The DCNN classified images of skin tumours more accurately than board-certified dermatologists. Collectively, the current system may have capabilities for screening purposes in general medical practice, particularly because it requires only a single clinical image for classification.