scholarly journals Cell graph neural networks enable the digital staging of tumor microenvironment and precise prediction of patient survival in gastric cancer

Author(s):  
Yanan Wang ◽  
Yu Guang Wang ◽  
Changyuan Hu ◽  
Ming Li ◽  
Yanan Fan ◽  
...  

ABSTRACTGastric cancer is one of the deadliest cancers worldwide. Accurate prognosis is essential for effective clinical assessment and treatment. Spatial patterns in the tumor microenvironment (TME) are conceptually indicative of the staging and progression of gastric cancer patients. Using spatial patterns of the TME by integrating and transforming the multiplexed immunohistochemistry (mIHC) images as Cell-Graphs, we propose a novel graph neural network-based approach, termed Cell-Graph Signature or CGSignature, powered by artificial intelligence, for digital staging of TME and precise prediction of patient survival in gastric cancer. In this study, patient survival prediction is formulated as either a binary (short-term and long-term) or ternary (short-term, medium-term, and long-term) classification task. Extensive benchmarking experiments demonstrate that the CGSignature achieves outstanding model performance, with Area Under the Receiver-Operating Characteristic curve (AUROC) of 0.960±0.01, and 0.771±0.024 to 0.904±0.012 for the binary- and ternary-classification, respectively. Moreover, Kaplan-Meier survival analysis indicates that the ‘digital-grade’ cancer staging produced by CGSignature provides a remarkable capability in discriminating both binary and ternary classes with statistical significance (p-value < 0.0001), significantly outperforming the AJCC 8th edition Tumor-Node-Metastasis staging system. Using Cell-Graphs extracted from mIHC images, CGSignature improves the assessment of the link between the TME spatial patterns and patient prognosis. Our study suggests the feasibility and benefits of such artificial intelligence-powered digital staging system in diagnostic pathology and precision oncology.

2019 ◽  
Author(s):  
Weifan Zhang ◽  
Xinhui Zhao ◽  
Zhao Liu ◽  
Hui Dang ◽  
Lei Meng ◽  
...  

Abstract Background: Few studies on the comparison among robotic, laparoscopic, and open gastrectomy had been reported in gastric cancer . The goal of this study was to evaluate the advantages of robotic-assisted gastrectomy (RAG) by comparing with laparoscopic-assisted gastrectomy(LAG) and open gastrectomy (OG). Methods: 147 gastric cancer patients who underwent gastrectomy were enrolled and retrospectively analyzed between January 2017 and July 2019. Short-term outcomes such as operation time, intraoperative estimated blood loss(EBL),number of retrieved lymph nodes, postoperative recovery, learning curve, and long-term outcome such as overall survival(OS) was compared among RAG, LAG and OG groups. Results: RAG group included 47 patients, 44 in the LAG, and 61 in the OG. Basic information such as gender, age, BMI, ASA degree were similar among three groups, and there were no statistically significances in pathological TNM staging, tumor resection extent, resection margin, methods of reconstruction( P >0.05). The cumulative sum(CUSUM) method showed that learning curve of RAG reached stability after 17 cases . For short-term outcomes, the RAG group had the shortest EBL( P =0.033), the shortest time to first flatus( P <0.001), shortest time to first intake liquid diet ( P =0.004),shortest postoperative hospital stay ( P =0.023)and the largest number of retrieved lymph nodes( P =0.044),the longest operation time( P <0.001), the most expensive treatment cost( P <0.001),however, there were no significant differences in postoperative drainage, postoperative white blood cell(WBC)count and early complications among three group( P >0.05). In addition to long-term outcome, similar OS was observed in three groups. Conclusion: Compared with LAG and OG, RAG has certain advantages in short-term outcomes and is a safe and reliable surgical method. But still need further prospective, multi-center research to confirm this.


2021 ◽  
Vol 41 (7) ◽  
pp. 3523-3534
Author(s):  
PIOTR KULIG ◽  
PRZEMYSŁAW NOWAKOWSKI ◽  
MAREK SIERZĘGA ◽  
RADOSŁAW PACH ◽  
OLIWIA MAJEWSKA ◽  
...  

Author(s):  
Hongzhi Wang ◽  
Bozhou Chen ◽  
Yueyang Xu ◽  
Kaixin Zhang ◽  
Shengwen Zheng

The major criteria to distinguish conscious Artificial Intelligence (AI) and non-conscious AI is whether the conscious is from the needs. Based on this criteria, we develop ConsciousControlFlow(CCF) to show the need-based conscious AI. The system is based on the computational model with a short-term memory (STM) and long-term memory (LTM) for consciousness and the hierarchy of needs. To generate AI based on real needs of the agent, we developed several LTMs for special functions such as feeling and sensor. Experiments have demonstrated that the agents in the proposed system behave according to the needs, which coincides with the prediction.


2013 ◽  
Vol 38 (6) ◽  
pp. 1453-1460 ◽  
Author(s):  
Sohei Matsumoto ◽  
Tomoyoshi Takayama ◽  
Kohei Wakatsuki ◽  
Tetsuya Tanaka ◽  
Kazuhiro Migita ◽  
...  

2019 ◽  
Vol 37 (2) ◽  
pp. 135-144 ◽  
Author(s):  
Masahiro Sasahara ◽  
Mitsuro Kanda ◽  
Seiji Ito ◽  
Yoshinari Mochizuki ◽  
Hitoshi Teramoto ◽  
...  

Background/Aims: Identification of nutritional indicators to predict short-term and long-term outcomes is necessary to provide appropriate treatment to patients with gastric cancer. Methods: We designed an analysis of a multicenter dataset of patients with gastric cancer who underwent gastrectomy between 2010 and 2014. We enrolled 842 eligible patients who had stage II/III gastric cancer. The area under the curve (AUC) values were compared among prognostic nutritional index (PNI), calculated as 10 × albumin g/dL + 0.005 × total lymphocyte count/mm3, and its constituents, and the predictive value of preoperative PNI for postoperative short-term and long-term outcomes was evaluated. Results: Preoperative PNI exhibited higher AUC values (0.719) for 1-year survival than its constituents, and the optimal cutoff value was 47. The disease-free and overall survival of patients in the PNI-low group were significantly shorter compared with those in the PNI-high group. The prognostic difference between the PNI-high and PNI-low groups was significantly greater in the subgroup of patients who underwent total gastrectomy. Clinically relevant postoperative complications were more frequently observed in the PNI-low group. Conclusions: The preoperative PNI is a useful predictor reflecting the incidence of complications after gastrectomy and the prognosis of patients with stage II/III gastric cancer.


2017 ◽  
Vol 35 (15_suppl) ◽  
pp. e15560-e15560
Author(s):  
Ryoichi Miyamoto ◽  
Satoshi Inagawa ◽  
Naoki Sano ◽  
Sosuke Tadano ◽  
Masayoshi Yamamoto

e15560 Background: Preoperative NLR was well known as highly repeatable, cost-effective and widely available long-term postoperative prognostic marker of gastric cancer patients. However, the utility of preoperative NLR to predict short-term outcomes in gastric cancer patients remains unclear. In this study, we addressed whether the preoperative NLR is a predictive value of short-term outcome in gastric cancer patients. Methods: We retrospectively evaluated 154 consecutive gastric cancer patients. Mean NLR was calculated, and 3.5 was set as cut-off value. The patient characteristics and perioperative outcomes were respectively compared. In addition, median survival times (MSTs) were also compared. In terms of stage II/III (UICC 7th) gastric cancer patients, median disease-free survival times (MDFSTs) were compared between the two groups. Results: The patients were then divided into two groups: low-NLR group (n = 110) and high-NLR group (n = 44). Among low-NLR group and high-NLR group, significant differences were respectively observed in preoperative symptoms [56 (51%) vs. 31 (70%); p = 0.027] and perioperative outcomes including postoperative complications [3 (2.7%) vs. 5 (11.3%); p = 0.015], intraoperative blood loss (158 ± 168 g vs. 232 ± 433 g; p = 0.022), and intraoperative blood transfusion [0 vs. 3 (6.8%); p = 0.042]. MSTs and MDFSTs were significantly differed (812 vs. 594 days; p = 0.04, 848 vs. 475 days; p = 0.03, respectively). Conclusions: The present study indicated that preoperative NLR influenced not only long-term outcomes but also perioperative outcomes in gastric cancer patients. Preoperative NLR is also a useful predictive value of short-term outcomes in gastric cancer patients.


Medicine ◽  
2016 ◽  
Vol 95 (24) ◽  
pp. e3781 ◽  
Author(s):  
Mitsuro Kanda ◽  
Akira Mizuno ◽  
Chie Tanaka ◽  
Daisuke Kobayashi ◽  
Michitaka Fujiwara ◽  
...  

Sign in / Sign up

Export Citation Format

Share Document