An outlier removal method based on PCA-DBSCAN for blood-SERS data analysis

Surface-enhanced Raman spectroscopy (SERS) has shown promising potential in cancer screening. In practical applications, Raman spectra are often affected by deviations from the spectrometer, changes in measurement environments, and anomalies in spectrum characteristic peak intensities due to improper sample storage. Previous research has overlooked the presence of outliers in categorical data, leading to significant impacts on model learning outcomes. In this study, we propose a novel method, called Principal Component Analysis and Density Based Spatial Clustering of Applications with Noise (PCA-DBSCAN) to effectively remove outliers. This method employs dimensionality reduction and spectral data clustering to identify and remove outliers. The PCA-DBSCAN method introduces adjustable parameters (Eps and MinPts) to control the clustering effect. The effectiveness of the proposed PCA-DBSCAN method is verified through modeling on outlier-removed datasets. Further refinement of the machine learning model and PCA-DBSCAN parameters resulted in the best cancer screening model, achieving 97.41% macro-average recall and 97.74% macro-average F1-score. This paper introduces a new outlier removal method that significantly improves the performance of the SERS cancer screening model. Moreover, the proposed method serves as inspiration for outlier detection in other fields, such as biomedical research, environmental monitoring, manufacturing, quality control, and hazard prediction.

Predicting multiple linear stapler firings in double stapling technique with an MRI-based deep-learning model

Multiple linear stapler firings is a risk factor for anastomotic leakage (AL) in laparoscopic low anterior resection (LAR) using double stapling technique (DST) anastomosis. In this study, our objective was to establish the risk factors for ≥ 3 linear stapler firings, and to create and validate a predictive model for ≥ 3 linear stapler firings in laparoscopic LAR using DST anastomosis. We retrospectively enrolled 328 mid-low rectal cancer patients undergoing laparoscopic LAR using DST anastomosis. With a split ratio of 4:1, patients were randomly divided into 2 sets: the training set (n = 260) and the testing set (n = 68). A clinical predictive model of ≥ 3 linear stapler firings was constructed by binary logistic regression. Based on three-dimensional convolutional networks, we built an image model using only magnetic resonance (MR) images segmented by Mask region-based convolutional neural network, and an integrated model based on both MR images and clinical variables. Area under the curve (AUC), sensitivity, specificity, accuracy, positive predictive value (PPV), and Youden index were calculated for each model. And the three models were validated by an independent cohort of 128 patients. There were 17.7% (58/328) patients received ≥ 3 linear stapler firings. Tumor size ≥ 5 cm (odds ratio (OR) = 2.54, 95% confidence interval (CI) = 1.15-5.60, p = 0.021) and preoperative carcinoma embryonic antigen (CEA) level > 5 ng/mL [OR = 2.20, 95% CI = 1.20-4.04, p = 0.011] were independent risk factors associated with ≥ 3 linear stapler firings. The integrated model (AUC = 0.88, accuracy = 94.1%) performed better on predicting ≥ 3 linear stapler firings than the clinical model (AUC = 0.72, accuracy = 86.7%) and the image model (AUC = 0.81, accuracy = 91.2%). Similarly, in the validation set, the integrated model (AUC = 0.84, accuracy = 93.8%) performed better than the clinical model (AUC = 0.65, accuracy = 65.6%) and the image model (AUC = 0.75, accuracy = 92.1%). Our deep-learning model based on pelvic MR can help predict the high-risk population with ≥ 3 linear stapler firings in laparoscopic LAR using DST anastomosis. This model might assist in determining preoperatively the anastomotic technique for mid-low rectal cancer patients.

Establishment and clinical application value of an automatic diagnosis platform for rectal cancer T-staging based on a deep neural network

BACKGROUND

Colorectal cancer is harmful to the patient's life. The treatment of patients is determined by accurate preoperative staging. Magnetic resonance imaging (MRI) played an important role in the preoperative examination of patients with rectal cancer, and artificial intelligence (AI) in the learning of images made significant achievements in recent years. Introducing AI into MRI recognition, a stable platform for image recognition and judgment can be established in a short period. This study aimed to establish an automatic diagnostic platform for predicting preoperative T staging of rectal cancer through a deep neural network.

METHODS

A total of 183 rectal cancer patients' data were collected retrospectively as research objects. Faster region-based convolutional neural networks (Faster R-CNN) were used to build the platform. And the platform was evaluated according to the receiver operating characteristic (ROC) curve.

RESULTS

An automatic diagnosis platform for T staging of rectal cancer was established through the study of MRI. The areas under the ROC curve (AUC) were 0.99 in the horizontal plane, 0.97 in the sagittal plane, and 0.98 in the coronal plane. In the horizontal plane, the AUC of T1 stage was 1, AUC of T2 stage was 1, AUC of T3 stage was 1, AUC of T4 stage was 1. In the coronal plane, AUC of T1 stage was 0.96, AUC of T2 stage was 0.97, AUC of T3 stage was 0.97, AUC of T4 stage was 0.97. In the sagittal plane, AUC of T1 stage was 0.95, AUC of T2 stage was 0.99, AUC of T3 stage was 0.96, and AUC of T4 stage was 1.00.

CONCLUSION

Faster R-CNN AI might be an effective and objective method to build the platform for predicting rectal cancer T-staging.

TRIAL REGISTRATION

chictr.org.cn: ChiCTR1900023575; http://www.chictr.org.cn/showproj.aspx?proj=39665.

Accuracy of preoperative T and N staging in colon cancer--a national population-based study

AIM

To select patients for neoadjuvant therapy in colon cancer, there is a need to improve pre-therapeutic locoregional staging. There are now data showing that the TN stage can be adequately assessed by preoperative CT in dedicated centres. In Sweden the use of preoperative CT of the abdomen for staging of the primary tumour is increasing. The aim of this study was to determine to what extent the preoperatively reported radiological TN stage correlates with the histopathological TN stage in an entire population.

METHOD

Data were collected on the preoperative cTN stage according to the radiologist and postoperative pTN stage according to the pathologist on all patients operated on for colon cancer in Sweden 2007-2010. The correlation between cTN stage and pTN stage was calculated using kappa statistics.

RESULTS

T stage was compared in 4373 patients with cT and pT stage. The correlation coefficient was 0.44, indicating fair agreement. The cN and pN correlation coefficient was 0.28, indicating a slight correlation. There was no difference in correlation related to age, gender, tumour location, body mass index or emergent vs elective surgery. A slight difference was seen between different geographical regions.

CONCLUSION

Preoperative CT in an unselected population does not result in an accurate cTN staging as previously reported from dedicated centres. To achieve adequate preoperative cTN staging nationally, the education of radiologists and optimization of the radiological method will be necessary.

EUS-FNA assessment of extramesenteric lymph node status in primary rectal cancer

BACKGROUND

Preoperative staging is an essential factor in the multidisciplinary management of rectal cancer. The accuracy of imaging alone with CT, magnetic resonance imaging, or rigid endorectal US is poor. The addition of EUS-FNA may enhance extramesenteric lymph node metastases detection (M1 disease) and overall staging accuracy.

OBJECTIVE

To evaluate the frequency of extramesenteric lymph node visualization by EUS and the rate of extramesenteric lymph node metastases by FNA. Secondary goals were to evaluate the clinical, endoscopic, and sonographic features associated with extramesenteric lymph node metastases, disease progression, and overall mortality.

DESIGN

Retrospective cohort study.

SETTINGS

Tertiary referral center.

RESULTS

Forty-one of 316 patients (13%) with primary rectal cancer over a 6-year period had M1 disease by EUS-FNA. Significant clinical, endoscopic, and sonographic features associated with extramesenteric lymph node metastases included the serum carcinoembryonic antigen level, tumor length 4 cm and longer, annularity 50% or more, sessile morphology, and lymph node size. The sensitivity and specificity of CT for extramesenteric lymph node metastases were 44% and 89%, respectively. Twenty-three of 316 rectal cancer endosonographic procedures (7.3%) were up-staged by FNA, which established extramesenteric lymph node metastases. Over a 4-year follow-up, disease progression and overall mortality of patients with extramesenteric lymph node metastases was observed in 6 patients (14.6%) and 14 patients (34%), respectively.

CONCLUSIONS

Preoperative EUS-FNA identification of extramesenteric lymph node metastases outside of standard radiation fields or total mesorectal excision resection margins could affect medical and surgical planning.

Role of tumor size in the pre-operative management of rectal cancer patients

Background

Clinical management of rectal cancer patients relies on pre-operative staging. Studies however continue to report moderate degrees of over/understaging as well as inter-observer variability. The aim of this study was to determine the sensitivity, specificity and accuracy of tumor size for predicting T and N stages in pre-operatively untreated rectal cancers.

Methods

We examined a test cohort of 418 well-documented patients with pre-operatively untreated rectal cancer admitted to the University Hospital of Basel between 1987 and 1996. Classification and regression tree (CART) and logistic regression analysis were carried out to determine the ability of tumor size to discriminate between early (pT1-2) and late (pT3-4) T stages and between node-negative (pN0) and node-positive (pN1-2) patients. Results were validated by an external patient cohort (n = 28).

Results

A tumor diameter threshold of 34 mm was identified from the test cohort resulting in a sensitivity and specificity for late T stage of 76.3%, and 67.4%, respectively and an odds ratio (OR) of 6.67 (95%CI:3.4-12.9). At a threshold value of 29 mm, sensitivity and specificity for node-positive disease were 94% and 15.5%, respectively with an OR of 3.02 (95%CI:1.5-6.1). Applying these threshold values to the validation cohort, sensitivity and specificity for T stage were 73.7% and 77.8% and for N stage 50% and 75%, respectively.

Conclusions

Tumor size at a threshold value of 34 mm is a reproducible predictive factor for late T stage in rectal cancers. Tumor size may help to complement clinical staging and further optimize the pre-operative management of patients with rectal cancer.

Multimodal preoperative evaluation system in surgical decision making for rectal cancer: a randomized controlled trial

PURPOSE

Multimodal preoperative evaluation (MPE) is a novel strategy for surgical decision making, incorporating the transrectal ultrasound (TRUS), 64 multi-slice spiral computer tomography (MSCT), and serum amyloid A protein (SAA) for rectal cancer. This trial aims to determine the accuracy of MPE in preoperative staging and its role in surgical decision making for rectal cancer.

METHODS

Two hundred twenty-five participants with histologically proven rectal cancer with tumor height less than 10 cm were randomly assigned into three arms in the ratio 1:1:1. Arm A (MPE) was multimodal staged by the combination of MSCT, TRUS, and SAA. Arm B (MSCT+SAA) was staged by MSCT and SAA. Arm C (MSCT) was staged only by MSCT. The primary endpoints were the accuracy of preoperative staging and expected surgical procedures. This study is registered as an International Standard Randomised Controlled Trial, number ChiCTR-DT-00000409.

RESULTS

The analysis showed statistical difference in the accuracy of T staging between arm A and B (94.6% vs. 77.8%, P=0.003) and arm A and C (94.6% vs. 80.6%, P=0.010). Statistical difference was also observed between the accuracies of preoperative N staging between arm A and C (85.1% vs. 69.4%, P=0.023) and arm A and B (85.1% vs. 84.7%, P=0.029). Surgical decision making in arm A was more accurate than that in arm C (95.9% vs. 80.6%, P=0.001). Pathological T stage (P<0.001), N stage (P<0.001), tumor node metastasis stage (P<0.001), serum level of SAA (P=0.002), and tumor height (P=0.030) were significantly associated with final surgical procedures.

CONCLUSION

MPE is an effective strategy in preoperative staging and more accurate than other available strategies in surgical decision making for rectal cancer.