Design of a Compact Hologram System Capable of 3D Lesion Diagnosis in Clinic.

MOTIVATION: This paper proposes a small-sized hologram system for the 3D imaging of lesions in a clinical environment. In a general hologram system, the distance between the beam-generating device and the screen (400 mm) and the size of the screen must be increased proportionally to obtain excellent image quality. However, in a clinical environment, the beam spread distance and screen size must be reduced. This paper proposes a method for reducing the beam divergence distance and screen size for clinical applications. METHODS: To reduce the beam spread distance and screen size, a beam prism with a 45° refractive index is used to reduce the beam spread distance by 1/3. The direction of the bent light must be adjusted such that it can reach the screen accurately. However, because the reflected light may be refracted owing to the material properties of the mirror and cause loss, this problem can be solved by using a full reflection mirror. RESULTS: The beam spread distance of the designed hologram system is 200 mm. The types of lesions obtained from the 3D images of the hologram include the lung, liver, and colon. The image resolution is 300 × 145. CONCLUSION: If the proposed method is used in a clinical environment, doctors can improve their understanding of the patient quickly and efficiently; thereby, shortening the treatment time. The proposed hologram system is expected to be useful in treatment rooms, operating rooms, and educational programs in medical schools.

Resolvin D1 suppresses inflammation-associated tumorigenesis in the colon by inhibiting IL-6-induced mitotic spindle abnormality.

While failure in resolution of inflammation is considered to increase the risk of tumorigenesis, there is paucity of experimental as well as clinical evidence supporting this association. Resolvin D1 (RvD1) is a representative pro-resolving lipid mediator that is endogenously generated from docosahexaenoic acid for the resolution of inflammation. Here, we report a decreased level of RvD1 in the blood from colorectal cancer patients and mice having inflammation-induced colon cancer, suggesting plasma RvD1 as a potential biomarker for monitoring colorectal cancer. Administration of RvD1 attenuated dextran sodium sulfate (DSS)-induced colitis and azoxymethane (AOM) plus DSS-induced colorectal carcinogenesis by suppressing the production of interleukin-6 (IL-6) and IL-6-mediated chromosomal instability. The protective effect of RvD1 against chromosomal instability is associated with downregulation of IL-6-induced Cyclin D1 expression, which appears to be mediated by blocking the Janus kinase 2 (JAK2)-signal transducer and activator of transcription 3 (STAT3) axis. RvD1 inhibited the STAT3 signaling pathway by interfering with the binding of IL-6 to its receptor (IL-6R), suggesting the novel function of RvD1 as a putative IL-6R antagonist. Together, our findings suggest that RvD1-mediated blockade of IL-6 signal transmission may contribute to inhibition of chromosomal instability and tumorigenesis.

IL-1ß induces expression of proinflammatory cytokines and migration of human colon cancer cells through upregulation of SIRT1.

SIRT1 is a mammalian NAD+-dependent deacetylase, which is known to be involved in various physiological events, such as adaptive response to environmental stresses including caloric restriction, as well as in aging and cellular senescence. However, recent studies have revealed overexpression of SIRT1 in many different types of human malignancies, particularly colon cancer. Interleukin-1ß (IL-1ß) is a proinflammatory cytokine that plays a major role in invasiveness, stemness and progression of colon cancer. However, the interaction between IL-1ß and SIRT1 in the tumor development and progression remains elusive. In this study, we found that IL-1ß induces SIRT1 protein expression in human colon cancer HCT-116 cells. IL-1ß-induced SIRT1 upregulation led to enhanced expression of mRNA transcripts of pro-inflammatory cytokines, IL-6 and IL-8 as well as that of IL-1ß. Knockdown of SIRT1 prevented IL-1ß-induced phosphorylation and nuclear accumulation of c-Jun. Furthermore, pharmacologic inhibition of SIRT1 abrogated clonogenicity and migrative capability of human colon cancer cells stimulated with IL-1ß. In summary, IL-1ß-induced SIRT1 upregulation stimulates production of proinflammatory cytokines via a nuclear accumulation of c-Jun, leadng to colon cancer growth and progression.

Effect of mistletoe extract on tumor response in neoadjuvant chemoradiotherapy for rectal cancer: a cohort study.

BACKGROUND: Mistletoe extract, used as a complementary chemotherapeutic agent for cancer patients, has anticancer effects against various malignancies. The aim of the present study was to evaluate the effect of mistletoe extract (Abnoba Viscum Q®) on tumor responses to neoadjuvant chemoradiotherapy (NCRT) for locally advanced rectal cancer. METHODS: This study included patients with rectal cancer who underwent NCRT between January 2018 and July 2020. In the mistletoe group (MG), the patients were administered Abnoba Viscum Q® subcutaneously during chemoradiotherapy-maintained just before surgery. Patient demographics, clinical outcomes, histopathological outcomes, and terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick-end labeling (TUNEL) assay results were compared between the MG and non-mistletoe group (NMG). Two rectal cancer cell lines (SNU-503 and SNU-503R80Gy) were treated with Abnoba Viscum Q® to assess its mechanistic effects in vivo. RESULTS: Overall, the study included 52 patients (MG: n = 15; NMG: n = 37). Baseline demographics between the two groups were similar, except carbohydrate antigen 19-9 levels and tumor location from the anal verge. There was no difference in the clinical stage between the two groups. A better tumor response in the MG, relative to the NMG, was observed with respect to tumor regression grade (TRG), T stage, and overall tumor-node-metastasis stage. Tumor response was significantly better in the MG than in the NMG in terms of pathologic complete response rate (53.3% vs. 21.6%, P = 0.044), good TRG response (66.7% vs. 32.4%, P = 0.024), T downstaging (86.7% vs. 43.2%, P = 0.004), and overall downstaging (86.7% vs. 56.8%, P = 0.040). The toxicities during NCRT were minimal in both groups. More apoptotic cells were noted in MG samples than in the NMG samples on TUNEL staining. Cleaved caspase-3 level following treatment with Abnoba Viscum Q® was higher in SNU-503R80Gy cells than in SNU-503 cells. CONCLUSION: Patients treated with chemoradiation combined with mistletoe extract showed better outcomes than patients not treated with mistletoe extract in terms of tumor responses. This diversity in treatment may improve the efficacy of NCRT, leading to better oncologic outcomes. Prospective and randomized studies with long-term follow-up are warranted to confirm and extend these results.

A data-driven approach to a chemotherapy recommendation model based on deep learning for patients with colorectal cancer in Korea.

BACKGROUND: Clinical Decision Support Systems (CDSSs) have recently attracted attention as a method for minimizing medical errors. Existing CDSSs are limited in that they do not reflect actual data. To overcome this limitation, we propose a CDSS based on deep learning. METHODS: We propose the Colorectal Cancer Chemotherapy Recommender (C3R), which is a deep learning-based chemotherapy recommendation model. Our model improves on existing CDSSs in which data-based decision making is not well supported. C3R is configured to study the clinical data collected at the Gachon Gil Medical Center and to recommend appropriate chemotherapy based on the data. To validate the model, we compared the treatment concordance rate with the National Comprehensive Cancer Network (NCCN) Guidelines, a representative set of cancer treatment guidelines, and with the results of the Gachon Gil Medical Center's Colorectal Cancer Treatment Protocol (GCCTP). RESULTS: For the C3R model, the treatment concordance rates with the NCCN guidelines were 70.5% for Top-1 Accuracy and 84% for Top-2 Accuracy. The treatment concordance rates with the GCCTP were 57.9% for Top-1 Accuracy and 77.8% for Top-2 Accuracy. CONCLUSIONS: This model is significant, i.e., it is the first colon cancer treatment clinical decision support system in Korea that reflects actual data. In the future, if sufficient data can be secured through cooperation among multiple organizations, more reliable results can be obtained.

Autologous blood, a novel agent for preoperative colonic localization: a safety and efficacy comparison study.

BACKGROUND: Preoperative localization is essential for minimally invasive colorectal surgery. However, conventional endoscopic tattooing agents such as India ink have safety issues. The availability of new endoscopic markers such as non-India-ink-based agent is limited. We assessed the efficacy and safety of preoperative endoscopic tattooing using autologous blood in colorectal surgery. METHODS: From February 2016, all patients who required localization of a target lesion before colorectal surgery underwent endoscopic tattooing using autologous blood, and the outcomes were collected prospectively. As a comparison, we retrospectively reviewed the medical records of a further 51 consecutive patients who underwent endoscopic tattooing using India ink before February 2016. A total of 102 patients who underwent endoscopic tattooing using either India ink or autologous blood were included in this study. The primary outcomes were the visibility of the tattooing in the peritoneal cavity and related adverse events. RESULTS: Endoscopic tattoos produced using India ink were visible in 49 (96.1%) patients, and tattoos created using autologous blood were visible in 47 (92.2%) patients. In the autologous blood group, the tattoo could not be identified in four patients due to excessive peritoneal fat, bleeding tendency, congenital anomaly, and suboptimal tattooing. Seven (13.7%) patients in the India ink group and three (5.9%) patients in the autologous blood group experienced endoscopic tattooing-related adverse events. CONCLUSIONS: Autologous blood is a feasible and safe tattooing agent for preoperative endoscopic localization of colorectal lesions within maximal interval of 5 days.

Risk Factors for Recurrence of Right Colonic Diverticulitis.

BACKGROUND: Right colonic diverticulitis (RCD) is more common in Asian countries than in Western countries, and the risk factors for recurrence of RCD are not fully understood. The objective of this study was to assess the risk factors for recurrence of RCD. METHODS: We analyzed 296 patients admitted for treatment of RCD in the Gachon University Gil Medical Center from December 2001 to October 2014. Gender, age, BMI, obesity, hypertension, diabetes mellitus, alcohol consumption, smoking, Hinchey classification, and hospital stay were investigated as risk factors for recurrence. RESULTS: Of the 296 patients with RCD, 31 patients recurred after conservative treatment. The median time interval between the initial episode and recurrence of diverticulitis was 10.4 months. In the univariate analysis, a high recurrence rate was observed in patients with a history of alcohol consumption, smoking, and long hospital stay. In the multivariate analysis, the recurrence rate was much higher (p < 0.001) in patients who stayed in the hospital for more than 10 days after the first attack. Smoking also elevated the recurrence rate (p = 0.011). CONCLUSION: Factors associated with recurrence of RCD may include smoking and the long hospital stay due to complexity when first diverticulitis occurs. Further prospective large-scale studies are needed to draw a definite conclusion.

In Vitro Cytotoxicity and Bioavailability of Ginsenoside-Modified Nanostructured Lipid Carrier Containing Curcumin.

Our aim was to investigate the cellular uptake, in vitro cytotoxicity and bioavailability of ginsenoside-modified nanostructured lipid carrier loaded with curcumin (G-NLC). The formulation was prepared by melt emulsification technique, in which water was added to the melted lipids and homogenized to give a uniform suspension of NLC (without ginsenoside) and G-NLC. Cellular uptake of curcumin in two colon cancer cell lines (HCT116 and HT29) was increased when administered using both NLC and G-NLC compared to control (curcumin dissolved into DMSO) as measured by fluorescence microscopy. Ginsenoside modification resulted in 2.0-fold and 1.4-fold increases in fluorescence intensity in HCT116 and HT29 cell lines, respectively, compared to plain NLC. In vitro cytotoxicity (assessed by MTT assay) had a dose-dependent relationship with curcumin concentration for both NLC and G-NLC. Although G-NLC was taken up more readily in HCT116 cells, ginsenoside modification did not produce a significant increase in cytotoxic effect; a significant increase was observed in HT29 cells. Oral administration of G-NLC in ten colon cancer patients produced an appreciable plasma level of unbound curcumin (2.9 ng/mL). In conclusion, introduction of ginsenoside into NLC enhanced the cellular uptake and cytotoxicity of curcumin as well as its oral bioavailability, and this strategy can be used to improve clinical outcomes in the treatment of colon cancer with similar genotype to HT29.

Measurement of human peritoneal surface area using artificial intelligence software in abdominal computed tomography.

PURPOSE: The calculation of the intraperitoneal organ surface area is important for understanding their anatomical structure and for conducting basic and clinical studies on diseases related to the peritoneum. To measure the intraperitoneal surface area in a living body by applying artificial intelligence (AI) techniques to the abdominal cavity using computed tomography and to prepare clinical indicators for application to the abdominal cavity. METHODS: Computed tomography images of ten adult males and females with a healthy body mass index and ten adults diagnosed with colon cancer were analyzed to determine the peritoneal and intraperitoneal surface areas of the organs. The peritoneal surface was segmented and three-dimensionally modeled using AI medical imaging software. In addition to manual work, three-dimensional editing, filtering, and connectivity checks were performed to improve work efficiency and accuracy. The colon and small intestine surface areas were calculated using the mean length and diameter. The abdominal cavity surface area was defined as the sum of the intraperitoneal area and the surface areas of each organ. RESULTS: The mean peritoneal surface area of all participants was measured as 10,039 ± 241 cm2 (males 10,224 ± 171 cm2 and females 9,854 ± 134 cm2). Males had a 3.7% larger peritoneal surface area than females, with a statistically significant difference (P < 0.001). CONCLUSION: The abdominal cavity surface area can be measured using AI techniques and is expected to be used as basic data for clinical applications.

Prediction of the minimum amount of anti-adhesive agent required for entire intra-abdominal cavity using fluorescent dye.

PURPOSE: Studies on the appropriate amount of anti-adhesive agents for preventing postoperative adhesion are lacking. This animal study aimed to investigate the distribution of an anti-adhesive agent in the abdominal cavity and estimate the necessary amount to cover the entire cavity. METHODS: Fluorescent dye Flamma-552 was conjugated to Guardix-sol to create Guardix-Flamma, which was laparoscopically applied to the abdominal cavity of two 10-kg pigs in different amounts: 15 mL for G1 and 35 mL for G2. After 24 hours, the distribution of Guardix-Flamma was examined under the near-infrared mode of the laparoscope, and the thickness was measured in tissues from the omentum, small, and large intestine by immunohistochemistry. RESULTS: The average area of the abdominal cavity in 10 kg pigs was 2,755 cm2. Guardix-Flamma fluorescence was detected in the greater omentum, ascites in the pelvis, and right quadrant area in G1, whereas in G2, it was detected everywhere. On average, the total thickness of G1 and G2 were 12.68 ± 9.80 µm and 18.16 ± 15.57 µm, respectively. Guardix-Flamma thickness applied to the omentum, small, and large intestines of G2 were 1.31-, 1.45-, and 1.49-times thicker than those of G1, respectively, and were all statistically significant (P < 0.05). CONCLUSION: The entire abdominal cavity of the 10 kg pig was not evenly covered with 15 mL of Guardix. Although 35 mL of Guardix is sufficient to cover the same area with an average thickness of 18 µm, further studies should evaluate the minimum thickness required for an effective anti-adhesive function.

Single-port laparoscopic appendectomy versus conventional laparoscopic appendectomy: a prospective randomized controlled study.

OBJECTIVE: To compare surgical outcomes and quality of life between single-port laparoscopic appendectomy (SPLA) and conventional laparoscopic appendectomy (CLA) in patients with acute appendicitis. BACKGROUND: A prospective randomized single center study was performed to compare the outcome of SPLA and CLA in patients with acute appendicitis. METHODS: A total of 248 patients were randomized, but because of 18 withdrawals, the outcome of 224 is analyzed, 116 in CLA and 114 in SPLA. RESULTS: There was no significant difference in the overall complication rate (P = 0.470). There were no significant differences in infectious complications between the SPLA group and the CLA group (10.2% and 12.4%, respectively). The wound complication rate between the 2 groups was not significant (5.1% and 10.6%, respectively; P = 0.207). Cosmetic satisfaction score, 36-item short-form health survey, and postoperative pain scores were not significantly different between 2 groups. CONCLUSIONS: SPLA failed to show any advantages over CLA relative to pain and cosmesis. However, SPLA is as safe as CLA (RCT number 01348464).

Clinical impact of tumor-infiltrating lymphocytes for survival in curatively resected stage IV colon cancer with isolated liver or lung metastasis.

BACKGROUND: The most reliable prognostic factor to date is tumor, node, metastasis stage. However, prognostic significance of tumor-infiltrating lymphocytes (TILs) in curatively resected stage IV colon cancer with isolated liver or lung lesion has not been clarified. The aim of this study was to assess and compare the prognostic role of TILs in curatively resected stage IV colon cancers. METHODS: Immunohistochemistry was used to assess the densities of CD8(+), CD45RO(+), and FOXP3(+) according to tumor sites (primary tumor, liver, and lung) from 79 stage IV colon cancers. These were evaluated for association with histopathologic features and patients' overall survival (OS). RESULTS: Higher density of CD45RO(+) at primary and metastatic sites was associated with better patient outcomes (P = 0.009 and 0.027, respectively). The estimated 3-year OS rates for high-density CD45RO(+) at metastatic and primary sites was 82.6 and 62.4 %, respectively, compared to 60.8 and 27.1 % in low-density CD45RO(+). In multivariate analysis, CD45RO(+) at the colon primary site (P = 0.007; relative risk 0.108; 95 % confidence interval 0.021-0.546) was the strongest prognostic factor. CONCLUSIONS: High density of CD45RO(+) TILs showed independent prognostic significance for OS. This result may help to improve the prognostication of curatively resected stage IV colon cancer.

Prognostic value of circulating cytokines for stage III colon cancer.

BACKGROUND: This study was to determine preoperative serum levels of epidermal growth factor (EGF), interleukin-6 (IL-6), and C-reactive protein (CRP) in stage III colon cancer and correlate them with disease status and prognosis. The circulating EGF in correlation with primary site epidermal growth factor receptor (EGFR) was also evaluated. METHODS: Seventy-seven patients with curatively resected stage III colon cancer were selected for analysis. Enzyme-linked immunosorbent assay was used to determine EGF and IL-6 serum levels, and serum CRP levels were measured via immunoturbidimetry. EGFR expression was observed with immunohistochemical studies. RESULTS: The median levels of EGFR, IL-6, and CRP were 189.4 pg/mL, 9.09 pg/mL, and 1.4 mg/mL, respectively. The factors related to recurrence with statistical significance included positive node status (P = 0.041), lymphovascular invasion (P = 0.001), and preoperative IL-6 level ≥9 pg/mL (P = 0.020). CRP and EGF levels were not significantly associated with disease-free survival rates (P = 0.438 and P = 0.309, respectively). Multivariate analysis using Cox's proportion model revealed that lymph node status was the single independent prognostic factor for predicting time until recurrence (odds ratio, 4.99; 95% confidence interval, 1.09-22.91; P = 0.038). CONCLUSION: IL-6 expression in stage III colon cancer patients appears to be a prognostic marker of tumor behavior. No correlations between serum EGF concentrations and tumor EGFR positivity were found in this study.

The association of hospital volume with rectal cancer surgery outcomes.

PURPOSE: An association between hospital volume and postoperative mortality has been identified for several oncologic surgical procedures. Our objective was to analyze differences in surgical outcomes for patients with rectal cancer according to hospital volume in the state of California. METHODS: A cross-sectional study from 2000 to 2005 was performed using the state of California Office of Statewide Health Planning and Development database. Hospitals were categorized into low (≤30)-, medium (31-60)-, and high (>60)-volume groups based on the total number of rectal cancer operations performed during the study period. RESULTS: Overall, 7,187 rectal cancer operations were performed. Of the 321 hospitals in the study cohort, 72 % (n = 232), 20 % (n = 65), and 8 % (n = 24) were low-, medium-, and high-volume hospitals, respectively. Postoperative mortality was significantly lower- in high-volume hospitals (0.9 %) when compared to medium- (1.1 %) and low-volume hospitals (2.1 %; p < 0.001). High-volume hospitals also performed more sphincter-preserving procedures (64 %) when compared to medium- (55 %) and low-volume hospitals (51 %; p < 0.001). CONCLUSIONS: These data indicate that hospital volume correlates with improved outcomes in rectal cancer surgery. Rectal cancer patients may benefit from lower mortality and increased sphincter preservation in higher-volume centers.

Neoadjuvant treatment of mid-to-lower rectal cancer with oxaliplatin plus 5-fluorouracil and leucovorin in combination with radiotherapy: a Korean single center phase II study.

PURPOSE: To evaluate the safety and efficacy of neoadjuvant chemoradiation with oxaliplatin and 5-fluorouracil (5-FU) in advanced mid-to-lower rectal cancer. METHODS: This was a single-arm, open-label phase II study conducted between August 2008 and August 2010. Thirty-one patients (n = 31) with clinical stage T3/T4 or lymph node positive rectal adenocarcinoma located in the middle or lower rectum without metastasis were enrolled onto the study. Data were analyzed according to the intention-to-treat principle. RESULTS: Thirty-one patients were enrolled into the study. Six patients (19.4%) experienced grade 3 diarrhea. Grade 2 nausea and vomiting occurred in 5 and 2 patients, respectively. Severe neurotoxicity was not observed. Grade 1 sensory neuropathy occurred in 10 patients (32.3%). Sphincter-saving surgery was performed in 29 patients (93.5%). The mean distance of the tumor from the anal verge was 4.9 cm. Anastomotic leakage occurred in 4 of 29 (13.8%) patients. The circumferential resection margin was involved in 2 patients (6.5%). Overall, 23 patients (77.4%) responded to treatment. The complete pathologic response (ypCR) rate was 12.9%. There was no death secondary to toxicity, and the mean follow-up time was 12.3 months. CONCLUSION: The overall toxicity of oxaliplatin and continuous 5-FU/leucovorin infusion in combination with radiation was well tolerated. Neoadjuvant chemoradiation for patients with locally advanced rectal cancer was associated with higher rates of sphincter preservation and downstaging, but did not significantly increase ypCR. The impact of this neoadjuvant chemoradiation regimen on survival will be determined by longer follow-up studies.

Direct comparison of Seprafilm® versus Adept ® versus no additive for reducing the risk of small-bowel obstruction in colorectal cancer surgery.

PURPOSE: To assess the effectiveness of using sodium hyaluronate-based bioresorbable membrane (Seprafilm(®)) versus 4 % icodextrin solution (Adept(®)) versus no additive intraoperatively and to prevent postoperative small-bowel obstruction in patients undergoing colorectal cancer (CRC) surgery. METHODS: The subjects of this retrospective study were 454 patients, who underwent CRC surgery between February 2007 and January 2010. Among the 454 enrolled patients, 114 patients received Seprafilm(®), 180 patients received Adept(®), and 160 patients received no additive, based on the year of their surgery. RESULTS: The overall incidences of small-bowel obstruction were 8.8, 4.3, and 6.9 %, for the Adept(®), Seprafilm(®), and no additive (control) groups, respectively. The cumulative incidence was significantly higher in the Adept(®) group than in the Seprafilm(®) and control groups (Adept(®) vs. Seprafilm(®), P = 0.043; Adept(®) vs. control group, P = 0.002). No significant difference was found between the Seprafilm group and the control group (P = 0.549). Adept(®) solution and Seprafilm(®) did not alter the liver and renal function, as assessed by blood chemistry. CONCLUSIONS: The use of Adept(®) significantly increased the incidence of small-bowel obstruction in patients undergoing CRC surgery.

Clinical manifestations of abdominal wall endometriosis: a single center experience.

OBJECTIVE: This study aimed to investigate the clinical manifestations and treatment outcomes of abdominal wall endometriosis (AWE). MATERIALS AND METHODS: Patients diagnosed with AWE at the Gil Medical Center from January 2002 to September 2010 were retrospectively reviewed. RESULTS: Thirty-seven women were treated for AWE during the study period. Median age was 34 (range 24-45) years, and median duration from last pelvic surgery until symptom onset was 30 (range 6-96) months. The most common initial symptom was a palpable mass (36, 97.2 %), followed by cyclic or spontaneous pain (21, 56.8 %). Preoperative diagnoses were accurate in 20 of 29 patients (68.9 %), who underwent a preoperative imaging study. The accuracy of abdominal US was 80 % (12/15). All patients underwent wide excision, and the median tumor size was 3.5 (range 1.0-10.0) cm. One patient experienced recurrence at 34 months postoperatively. CONCLUSIONS: Physicians should be aware of AWE in any woman presenting with palpable mass and/or pain at the abdominal wall, especially after pelvic surgery. Adequate preoperative estimation and wide excision might be essential for the treatment of AWE.

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.