Machine learning based prediction of recurrence after curative resection for rectal cancer.

PURPOSE: Patients with rectal cancer without distant metastases are typically treated with radical surgery. Post curative resection, several factors can affect tumor recurrence. This study aimed to analyze factors related to rectal cancer recurrence after curative resection using different machine learning techniques. METHODS: Consecutive patients who underwent curative surgery for rectal cancer between 2004 and 2018 at Gil Medical Center were included. Patients with stage IV disease, colon cancer, anal cancer, other recurrent cancer, emergency surgery, or hereditary malignancies were excluded from the study. The Synthetic Minority Oversampling Technique with Tomek link (SMOTETomek) technique was used to compensate for data imbalance between recurrent and no-recurrent groups. Four machine learning methods, logistic regression (LR), support vector machine (SVM), random forest (RF), and Extreme gradient boosting (XGBoost), were used to identify significant factors. To overfit and improve the model performance, feature importance was calculated using the permutation importance technique. RESULTS: A total of 3320 patients were included in the study. After exclusion, the total sample size of the study was 961 patients. The median follow-up period was 60.8 months (range:1.2-192.4). The recurrence rate during follow-up was 13.2% (n = 127). After applying the SMOTETomek method, the number of patients in both groups, recurrent and non-recurrent group were equalized to 667 patients. After analyzing for 16 variables, the top eight ranked variables {pathologic Tumor stage (pT), sex, concurrent chemoradiotherapy, pathologic Node stage (pN), age, postoperative chemotherapy, pathologic Tumor-Node-Metastasis stage (pTNM), and perineural invasion} were selected based on the order of permutational importance. The highest area under the curve (AUC) was for the SVM method (0.831). The sensitivity, specificity, and accuracy were found to be 0.692, 0.814, and 0.798, respectively. The lowest AUC was obtained for the XGBoost method (0.804), with a sensitivity, specificity, and accuracy of 0.308, 0.928, and 0.845, respectively. The variable with highest importance was pT as assessed through SVM, RF, and XGBoost (0.06, 0.12, and 0.13, respectively), whereas pTNM had the highest importance when assessed by LR (0.05). CONCLUSIONS: In the current study, SVM showed the best AUC, and the most influential factor across all machine learning methods except LR was found to be pT. The rectal cancer patients who have a high pT stage during postoperative follow-up are need to be more close surveillance.

Comparison of Long-Term Oncologic Outcomes Between Surgical T4 and T3 in Patients Diagnosed With Pathologic Stage IIA Right Colon Cancer.

Purpose: T stage plays an important role in the classification of subgroups in stage II colon cancer. Patients with pathologic T4 are at high risk of recurrence and it is recommended to include adjuvant chemotherapy in the treatment plan, while this is not necessary in pathologic T3. There is a discrepancy between the surgical T stage (sT), as determined by the surgeon in the operative field, and pathologic T stage (pT). The pathologic stage is considered a standard prognostic factor, but it has not been established whether the surgical stage has an oncologic impact. The aim of this study was to compare oncologic outcomes between sT4 and sT3 in pathologic stage IIA right colon cancer. Methods: Between January 2005 and December 2018, there were 354 patients who underwent right hemicolectomy performed by a single surgeon (JHB) at a tertiary hospital. The data from these patients were retrospectively collected and analyzed. Only those patients with pathologic stage IIA (pT3N0M0) right colon adenocarcinomas were included in this study. Patients with mucinous carcinoma, signet ring cell carcinoma, squamous cell carcinoma, or hereditary colon cancer, and who had emergent surgery were excluded. Finally, 86 patients were included in this study. The patients were categorized, according to their surgical records, into either the sT4 group (n=28) or the sT3 group (n=58). Results: There were no statistical differences between the two groups in terms of age, sex, body mass index, comorbidities, cancer location, histologic grade, lymphovascular invasion, perineural invasion, number of harvested lymph nodes, and adjuvant chemotherapy. The 5-year overall survival rate was significantly different between the sT4 and sT3 groups (92.6% vs. 97.7%, p=0.024). In addition, the 5-year disease-free survival rate was significantly different between the sT4 and sT3 groups (88.6% vs. 97.7%, p=0.017). In the multivariate Cox regression analysis, a classification of sT4 was a significant independent predictive factor for recurrence (p = 0.023). Conclusions: Long-term oncologic outcomes have shown significant differences between surgical T4 and T3 in pathologic stage IIA right colon cancer patients. Further large-scale, multicenter studies are required to verify the clinical impact of the surgical staging.

Low-dose radiation suppresses Pokemon expression under hypoxic conditions.

Our previous data demonstrated that CoCl2-induced hypoxia controls endoplasmic reticulum (ER) stress-associated and other intracellular factors. One of them, the transcription factor Pokemon, was differentially regulated by low-dose radiation (LDR). There are limited data regarding how this transcription factor is involved in expression of the unfolded protein response (UPR) under hypoxic conditions. The purpose of this study was to obtain clues on how Pokemon is involved in the UPR. Pokemon was selected as a differentially expressed gene under hypoxic conditions; however, its regulation was clearly repressed by LDR. It was also demonstrated that both expression of ER chaperones and ER stress sensors were affected by hypoxic conditions, and the same results were obtained when cells in which Pokemon was up- or down-regulated were used. The current state of UPR and LDR research associated with the Pokemon pathway offers an important opportunity to understand the oncogenesis, senescence, and differentiation of cells, as well as to facilitate introduction of new therapeutic radiopharmaceuticals.

Metabolic characterization of lactic acid bacterium Lactococcus garvieae sk11, capable of reducing ferric iron, nitrate, and fumarate.

A lactic acid bacterium capable of anaerobic respiration was isolated from soil with ferric iron-containing glucose basal medium and identified as L. garvieae by using 16S rDNA sequence homology. The isolate reduced ferric iron, nitrate, and fumarate to ferrous iron, nitrite, and succinate, respectively, under anaerobic N2 atmosphere. Growth of the isolate was increased about 30-39% in glucose basal medium containing nitrate and fumarate, but not in the medium containing ferric iron. Specifically, metabolic reduction of nitrate and fumarate is thought to be controlled by the specific genes fnr, encoding FNR-like protein, and nir, regulating fumarate-nitrate reductase. Reduction activity of ferric iron by the isolate was estimated physiologically, enzymologically, and electrochemically. The results obtained led us to propose that the isolate metabolized nitrate and fumarate as an electron acceptor and has specific enzymes capable of reducing ferric iron in coupling with anaerobic respiration.