Lung and Colon Cancer Classification Using Medical Imaging : A Feature Engineering Approach

Lung and colon cancers are the most common causes of death. Their simultaneous occurrence is uncommon, however, in the absence of early diagnosis, the metastasis of cancer cells is very high between these two organs. Currently, histopathological diagnosis and appropriate treatment are the only possibility to improve the chances of survival and reduce cancer mortality. Using artificial intelligence in the histopathological diagnosis of colon and lung cancer can provide significant help to specialists in identifying cases of colon and lung cancers with less effort, time and cost. The objective of this study is to set up a computer-aided diagnostic system that can accurately classify five types of colon and lung tissues (two classes for colon cancer and three classes for lung cancer) by analyzing their histopathological images. Using machine learning, features engineering and image processing techniques, the five models XGBoost, SVM, RF, LDA and MLP were used to perform the classification of histopathological images of lung and colon cancers that were acquired from the LC25000 dataset. The main advantage of using machine learning models is that they allow for better interpretability of the classification model since they are based on feature engineering; however, deep learning models are black box networks whose working is very difficult to understand due to the complex network design. The acquired experimental results show that machine learning models give satisfactory results and are very precise in identifying classes of lung and colon cancer subtypes. The XGBoost model gave the best performance with an accuracy of 99% and a F1-score of 98.8%. The implementation and the development of this model will help healthcare specialists identify types of colon and lung cancers. The code will be available upon request.

Segmental resection of splenic flexure colon cancers provides an adequate lymph node harvest and is a safe operative approach – an analysis of the ACS-NSQIP database

Introduction Fewer than 10% of colon cancers are found at the splenic flexure. A standard surgical approach to these cancers has not been defined. The goal of this study was to compare lymph node harvest and post-operative morbidity between segmental resection and formal left hemicolectomy for splenic flexure colon cancers. Method Patients diagnosed with a splenic flexure cancer were identified from the 2012–2018 ACS-NSQIP colectomy-targeted database. Patients were categorized based on type of surgical resection – left hemicolectomy with colorectal anastomosis or segmental colectomy with colocolonic anastomosis. Demographic, clinicopathologic, and post-operative outcomes were compared between groups. Factors independently associated with lymph node harvest, operative time, and post-operative morbidity were investigated by linear and binomial logistic regression models. Results A total of 3,049 patients underwent colectomy for a splenic flexure cancer. Of these, 83.6% had a segmental colectomy and 73% were performed by a minimally invasive approach. T- and N-stage did not differ between segmental and left hemicolectomy groups (p = 0.703 and p = 0.429, respectively). Inadequate nodal harvest (< 12 nodes) was infrequent and similar between the two procedures (7.4% vs. 9.1%, p = 0.13). Operative time was significantly shorter for segmental colectomy (213 ± 83.5 min vs. 193 ± 84.1 min, p < 0.0001) and major morbidity was similar between the two surgical techniques (8.4% vs. 8.9%, p = 0.75). After accounting for demographic, clinicopathologic, and operative factors, binomial logistic regression showed that type of procedure was not significantly associated with LN harvest (OR 0.80, 95%CI 0.54–1.17) or major morbidity (OR 1.17, 95%CI 0.36–3.81). However, on linear regression, segmental splenic flexure resection was associated with shorter operative time (estimate 20.29, 95%CI 12.61–27.97, p < 0.0001). Conclusion Splenic flexure resection for colon cancer is associated with an adequate lymph node harvest. Compared to a formal left hemicolectomy, a segmental resection also has a shorter operative time with equivalent post-operative morbidity.

Intra-Tumoral Nerve-Tracing in a Novel Syngeneic Model of High-Grade Serous Ovarian Carcinoma

Dense tumor innervation is associated with enhanced cancer progression and poor prognosis. We observed innervation in breast, prostate, pancreatic, lung, liver, ovarian, and colon cancers. Defining innervation in high-grade serous ovarian carcinoma (HGSOC) was a focus since sensory innervation was observed whereas the normal tissue contains predominantly sympathetic input. The origin, specific nerve type, and the mechanisms promoting innervation and driving nerve-cancer cell communications in ovarian cancer remain largely unknown. The technique of neuro-tracing enhances the study of tumor innervation by offering a means for identification and mapping of nerve sources that may directly and indirectly affect the tumor microenvironment. Here, we establish a murine model of HGSOC and utilize image-guided microinjections of retrograde neuro-tracer to label tumor-infiltrating peripheral neurons, mapping their source and circuitry. We show that regional sensory neurons innervate HGSOC tumors. Interestingly, the axons within the tumor trace back to local dorsal root ganglia as well as jugular–nodose ganglia. Further manipulations of these tumor projecting neurons may define the neuronal contributions in tumor growth, invasion, metastasis, and responses to therapeutics.

Differential Actions of Muscarinic Receptor Subtypes in Gastric, Pancreatic, and Colon Cancer

Cancers arising from gastrointestinal epithelial cells are common, aggressive, and difficult to treat. Progress in this area resulted from recognizing that the biological behavior of these cancers is highly dependent on bioactive molecules released by neurocrine, paracrine, and autocrine mechanisms within the tumor microenvironment. For many decades after its discovery as a neurotransmitter, acetylcholine was thought to be synthesized and released uniquely from neurons and considered the sole physiological ligand for muscarinic receptor subtypes, which were believed to have similar or redundant actions. In the intervening years, we learned this former dogma is not tenable. (1) Acetylcholine is not produced and released only by neurons. The cellular machinery required to synthesize and release acetylcholine is present in immune, cancer, and other cells, as well as in lower organisms (e.g., bacteria) that inhabit the gut. (2) Acetylcholine is not the sole physiological activator of muscarinic receptors. For example, selected bile acids can modulate muscarinic receptor function. (3) Muscarinic receptor subtypes anticipated to have overlapping functions based on similar G protein coupling and downstream signaling may have unexpectedly diverse actions. Here, we review the relevant research findings supporting these conclusions and discuss how the complexity of muscarinic receptor biology impacts health and disease, focusing on their role in the initiation and progression of gastric, pancreatic, and colon cancers.

The health impact of smokeless tobacco products: a systematic review

Introduction The objective was to systematically review studies on health outcomes from smokeless tobacco (SLT) products. Methods We analysed published literature on the health outcomes from SLT use between 01/01/2015 to 01/02/2020, following Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) protocol using PubMed, Embase, Scopus, and Google Scholar. Results Of 53 studies included, six were global, 32 from Asia, Middle East and Africa (AMEA), nine from USA and six from Europe. ‘Poor’-rated studies predominated (23;43%), in particular, for global (4;66%) and AMEA (16;50%). Health outcomes differed between SLT-products and regions; those in AMEA were associated with higher mortality (overall, cancer, Coronary heart disease (CHD), respiratory but not cardiovascular disease (CVD)), and morbidity (CVD, oral and head and neck cancers), with odds ratios up to 38.7. European studies showed no excess mortality (overall, CVD, from cancers) or morbidity (ischemic heart disease (IHD), stroke, oral, head and neck, pancreatic or colon cancers) from several meta-analyses; single studies reported elevated risk of rectal cancer and respiratory disorders. Pooled study data showed protection against developing Parkinson’s disease. US studies showed mixed results for mortality (raised overall, CHD, cancer and smoking-related cancer mortality; no excess risk of respiratory or CVD mortality). Morbidity outcomes were also mixed, with some evidence of increased IHD, stroke and cancer risk (oral, head and neck). No studies reported on switching from cigarettes to SLT-products. Conclusion Our review demonstrates stark differences between different SLT-products in different regions, ranging from zero harm from European snus to greatly increased health risks in AMEA. The literature on the safety profile for SLT-products for harm reduction is incomplete and potentially misinforming policy and regulation.

miRNA-148b and its role in various cancers

miRNA-148b belongs to the family miR-148/-152, with significant differences in nonseed sequences, which can target diverse mRNA molecules. Reportedly, it may undergo deregulation in lung and ovarian cancers and downregulation in gastric, pancreatic and colon cancers. However, there is a need for further studies to better characterize its mechanism of action and in different types of cancer. In this review, we focus on the aberrant expression of miR-148b in different cancer types and highlight its main target genes and signaling pathways, as well as its pathophysiologic role and relevance to tumorigenesis in several types of cancer.

Long-term Oncological Outcomes of Totally Laparoscopic Colectomy With Intracorporeal Anastomosis for Colon Cancer: Propensity Score Matching Analysis

Background: This retrospective study aimed to compare long-term oncological outcomes between laparoscopic-assisted colectomy (LAC) with extracorporeal anastomosis (EA) and totally laparoscopic colectomy (TLC) with intracorporeal anastomosis (IA) for colon cancers, including right- and left-sided colon cancers.Methods: Patients with stage I–III colon cancers who underwent elective laparoscopic colectomy between January 2013 and December 2017 were analyzed retrospectively. Patients converted from laparoscopic to open surgery and R1/R2 resection were excluded. Propensity score matching (PSM) analysis (1:1) was performed to overcome patient selection bias.Results: A total of 388 patients were reviewed. After PSM, 83 patients in the EA group and 83 patients in the IA group were compared. Median follow-up was 56.5 months in the EA group and 55.5 months in the IA group. Estimated 3-year overall survival (OS) did not differ significantly between the EA group (86.6%; 95% confidence interval (CI), 77.4–92.4%) and IA group (84.8%; 95%CI, 75.0–91.1%; P = 0.68). Estimated 3-year disease-free survival (DFS) likewise did not differ significantly between the EA group (76.4%; 95%CI, 65.9–84.4%) and IA group (81.0%; 95%CI, 70.1–88.2%; P = 0.12).Conclusion: TLC with IA was comparable to LAC with EA in terms of 3-year OS and DFS. TLC with IA thus appears to offer an oncologically feasible procedure.

Deciphering Promoter Hypermethylation of Genes Encoding for RASSF/Hippo Pathway Reveals the Poor Prognostic Factor of RASSF2 Gene Silencing in Colon Cancers

The aims of this study were to assess the frequency of promoter hypermethylation of the genes encoding the Ras associated domain family (RASSF)/Hippo pathway, as well as the impact on overall (OS) and progression-free survival (PFS) in a single-center retrospective cohort of 229 patients operated on for colon cancers. Hypermethylation status was investigated by methylation-specific PCR on the promoters of the RASSF1/2, STK4/3 (encoding Mammalian Ste20-like protein 1 and 2 (MST1 and 2), respectively), and LATS1/2 genes. Clinicopathological characteristics, recurrence-free survival, and overall survival were analysed. We found the RASSF/Hippo pathway to be highly silenced in colon cancer, and particularly RASSF2 (86%). The other promoters were hypermethylated with a lesser frequency of 16, 3, 1, 10 and 6%, respectively for RASSF1, STK4, STK3, LATS1, and LATS2 genes. As the hypermethylation of one RASSF/Hippo family member was by no means exclusive from the others, 27% of colon cancers displayed the hypermethylation of at least two RASSF/Hippo member promotors. The median overall survival of the cohort was 60.2 months, and the median recurrence-free survival was 46.9 months. Survival analyses showed a significantly poorer overall survival of patients when the RASSF2 promoter was hypermethylated (p = 0.03). The median OS was 53.5 months for patients with colon cancer with a hypermethylated RASSF2 promoter versus still not reached after 80 months follow-up for other patients, upon univariate analysis (HR = 1.86, [95% CI: 1.05–3.3], p < 0.03). Such difference was not significant for relapse-free survival as in multivariate analysis. A logistic regression model showed that RASSF2 hypermethylation was an independent factor. In conclusion, RASSF2 hypermethylation is a frequent event and an independent poor prognostic factor in colon cancer. This biomarker could be investigated in clinical practice.