Development and validation of a radiomics-based nomogram for the preoperative prediction of microsatellite instability in colorectal cancer

A radiomics-boosted deep-learning for risk assessment of synchronous peritoneal metastasis in colorectal cancer

OBJECTIVES

Synchronous colorectal cancer peritoneal metastasis (CRPM) has a poor prognosis. This study aimed to create a radiomics-boosted deep learning model by PET/CT image for risk assessment of synchronous CRPM.

METHODS

A total of 220 colorectal cancer (CRC) cases were enrolled in this study. We mapped the feature maps (Radiomic feature maps (RFMs)) of radiomic features across CT and PET image patches by a 2D sliding kernel. Based on ResNet50, a radiomics-boosted deep learning model was trained using PET/CT image patches and RFMs. Besides that, we explored whether the peritumoral region contributes to the assessment of CRPM. In this study, the performance of each model was evaluated by the area under the curves (AUC).

RESULTS

The AUCs of the radiomics-boosted deep learning model in the training, internal, external, and all validation datasets were 0.926 (95% confidence interval (CI): 0.874-0.978), 0.897 (95% CI: 0.801-0.994), 0.885 (95% CI: 0.795-0.975), and 0.889 (95% CI: 0.823-0.954), respectively. This model exhibited consistency in the calibration curve, the Delong test and IDI identified it as the most predictive model.

CONCLUSIONS

The radiomics-boosted deep learning model showed superior estimated performance in preoperative prediction of synchronous CRPM from pre-treatment PET/CT, offering potential assistance in the development of more personalized treatment methods and follow-up plans.

CRITICAL RELEVANCE STATEMENT

The onset of synchronous colorectal CRPM is insidious, and using a radiomics-boosted deep learning model to assess the risk of CRPM before treatment can help make personalized clinical treatment decisions or choose more sensitive follow-up plans.

KEY POINTS

Prognosis for patients with CRPM is bleak, and early detection poses challenges. The synergy between radiomics and deep learning proves advantageous in evaluating CRPM. The radiomics-boosted deep-learning model proves valuable in tailoring treatment approaches for CRC patients.

Application research of radiomics in colorectal cancer: A bibliometric study

BACKGROUND

Radiomics has shown great potential in the clinical field of colorectal cancer (CRC). However, few bibliometric studies have systematically analyzed existing research in this field. The purpose of this study is to understand the current research status and future development directions of CRC.

METHODS

Search the English documents on the application of radiomics in the field of CRC research included in the Web of Science Core Collection from its establishment to October 2023. VOSviewer and CiteSpace software were used to conduct bibliometric and visual analysis of online publications related to countries/regions, authors, journals, references, and keywords in this field.

RESULTS

A total of 735 relevant documents published from Web of Science Core Collection to October 2023 were retrieved, and a total of 419 documents were obtained based on the screening criteria, including 376 articles and 43 reviews. The number of publications is increasing year by year. Among them, China publishes the most relevant documents (n = 238), which is much higher than Italy (n = 69) and the United States (n = 63). Tian Jie is the author with the most publications and citations (n = 17, citations = 2128), GE Healthcare is the most productive institution (n = 26), Frontiers in Oncology is the journal with the most publications (n = 60), and European Radiology is the most cited journal (n = 776). Hot spots for the application of radiomics in CRC include magnetic resonance, neoadjuvant chemoradiotherapy, survival, texture analysis, and machine learning. These directions are the current hot spots for the application of radiomics research in CRC and may be the direction of continued development in the future.

CONCLUSION

Through bibliometric analysis, the application of radiomics in CRC has been increasing year by year. The application of radiomics improves the accuracy of preoperative diagnosis, prediction, and prognosis of CRC. The results of bibliometrics analysis provide a valuable reference for the research direction of radiomics. However, radiomics still faces many challenges in the future, such as the single nature of the data source which may affect the comprehensiveness of the results. Future studies can further expand the data sources and build a multicenter public database to more comprehensively reflect the research status and development trend of CRC radiomics.

Preoperative prediction of microsatellite instability status in colorectal cancer based on a multiphasic enhanced CT radiomics nomogram model

BACKGROUND

To investigate the value of a nomogram model based on the combination of clinical-CT features and multiphasic enhanced CT radiomics for the preoperative prediction of the microsatellite instability (MSI) status in colorectal cancer (CRC) patients.

METHODS

A total of 347 patients with a pathological diagnosis of colorectal adenocarcinoma, including 276 microsatellite stabilized (MSS) patients and 71 MSI patients (243 training and 104 testing), were included. Univariate and multivariate regression analyses were used to identify the clinical-CT features of CRC patients linked with MSI status to build a clinical model. Radiomics features were extracted from arterial phase (AP), venous phase (VP), and delayed phase (DP) CT images. Different radiomics models for the single phase and multiphase (three-phase combination) were developed to determine the optimal phase. A nomogram model that combines clinical-CT features and the optimal phasic radscore was also created.

RESULTS

Platelet (PLT), systemic immune inflammation index (SII), tumour location, enhancement pattern, and AP contrast ratio (ACR) were independent predictors of MSI status in CRC patients. Among the AP, VP, DP, and three-phase combination models, the three-phase combination model was selected as the best radiomics model. The best MSI prediction efficacy was demonstrated by the nomogram model built from the combination of clinical-CT features and the three-phase combination model, with AUCs of 0.894 and 0.839 in the training and testing datasets, respectively.

CONCLUSION

The nomogram model based on the combination of clinical-CT features and three-phase combination radiomics features can be used as an auxiliary tool for the preoperative prediction of the MSI status in CRC patients.

A radiomics signature derived from CT imaging to predict MSI status and immunotherapy outcomes in gastric cancer: a multi-cohort study

BACKGROUND

Accurate microsatellite instability (MSI) testing is essential for identifying gastric cancer (GC) patients eligible for immunotherapy. We aimed to develop and validate a CT-based radiomics signature to predict MSI and immunotherapy outcomes in GC.

METHODS

This retrospective multicohort study included a total of 457 GC patients from two independent medical centers in China and The Cancer Imaging Archive (TCIA) databases. The primary cohort (n = 201, center 1, 2017-2022), was used for signature development via Least Absolute Shrinkage and Selection Operator (LASSO) and logistic regression analysis. Two independent immunotherapy cohorts, one from center 1 (n = 184, 2018-2021) and another from center 2 (n = 43, 2020-2021), were utilized to assess the signature's association with immunotherapy response and survival. Diagnostic efficiency was evaluated using the area under the receiver operating characteristic curve (AUC), and survival outcomes were analyzed via the Kaplan-Meier method. The TCIA cohort (n = 29) was included to evaluate the immune infiltration landscape of the radiomics signature subgroups using both CT images and mRNA sequencing data.

RESULTS

Nine radiomics features were identified for signature development, exhibiting excellent discriminative performance in both the training (AUC: 0.851, 95%CI: 0.782, 0.919) and validation cohorts (AUC: 0.816, 95%CI: 0.706, 0.926). The radscore, calculated using the signature, demonstrated strong predictive abilities for objective response in immunotherapy cohorts (AUC: 0.734, 95%CI: 0.662, 0.806; AUC: 0.724, 95%CI: 0.572, 0.877). Additionally, the radscore showed a significant association with PFS and OS, with GC patients with a low radscore experiencing a significant survival benefit from immunotherapy. Immune infiltration analysis revealed significantly higher levels of CD8 + T cells, activated CD4 + B cells, and TNFRSF18 expression in the low radscore group, while the high radscore group exhibited higher levels of T cells regulatory and HHLA2 expression.

CONCLUSION

This study developed a robust radiomics signature with the potential to serve as a non-invasive biomarker for GC's MSI status and immunotherapy response, demonstrating notable links to post-immunotherapy PFS and OS. Additionally, distinct immune profiles were observed between low and high radscore groups, highlighting their potential clinical implications.

An integrative clinical and CT-based tumoral/peritumoral radiomics nomogram to predict the microsatellite instability in rectal carcinoma

BACKGROUND

Microsatellite instability (MSI) is detected in approximately 15% of colorectal carcinoma (CRC) patients, which has emerged as a predictor of patient response to adjuvant chemotherapy. Rectal carcinoma (RC) is the most common type of CRC. Therefore, prediction of MSI status of RC is significant for personalized medication. The purpose of this article was to develop an integrative model that combines clinical characteristics and computed tomography-based (CT-based) tumoral/peritumoral radiomics to predict the MSI status in RC.

METHODS

A cohort of 788 RCs with 97 high-MSI status (MSI-H) and 691 microsatellite stable status (MSS) were enrolled between January 2015 and January 2021 in this retrospective study. Clinical characteristics were recorded, and CT-based tumoral/peritumoral radiomic features were calculated after segmenting volume of interests. The patients were randomly divided into training and validation sets in a 7:3 proportion. Logistic models of single tumoral radiomics (LM-tRadio), peritumoral radiomics (LM-ptRadio), and combined tumoral/peritumoral radiomics (LM-Radio) were constructed to distinguish MSI-H from MSS, and a relevant radiomic score was calculated. An integrative nomogram (LM-Nomo) was developed, including significant clinical characteristics and CT-based tumoral/peritumoral radiomics. The area under receiver operator curve (AUC) was calculated to evaluate the efficacy of prediction.

RESULTS

The AUCs of LM-Radio were 0.785 (95%CI 0.732-0.837) in the training set and were 0.628 (95%CI 0.528-0.723) in the validation set, which were higher than those of LM-tRadio and LM-ptRadio. The AUCs of single LM-ptRadio were slightly higher than those of LM-tRadio (0.724 vs. 0.708 in the training set, 0.613 vs. 0.602 in the validation set). The LM-Nomo containing carcinoembryonic antigen (CEA), hypertension, and CT-based tumoral/peritumoral radiomic score showed the highest AUCs of 0.796 (95%CI 0.748-0.843) in the training set and 0.679 (95%CI 0.588-0.771) in the validation set in predicting the MSI-H status of RC.

CONCLUSION

The AUCs of LM-ptRadio were slightly higher than LM-tRadio to evaluate the MSI-H status of RC. The LM-Nomo, which includes significant clinical characteristics and CT-based tumoral/peritumoral radiomics score, demonstrated the best performance in predicting MSI-H status of RC.

Detecting microsatellite instability in colorectal cancer using Transformer-based colonoscopy image classification and retrieval

Colorectal cancer (CRC) is a major global health concern, with microsatellite instability-high (MSI-H) being a defining characteristic of hereditary nonpolyposis colorectal cancer syndrome and affecting 15% of sporadic CRCs. Tumors with MSI-H have unique features and better prognosis compared to MSI-L and microsatellite stable (MSS) tumors. This study proposed establishing a MSI prediction model using more available and low-cost colonoscopy images instead of histopathology. The experiment utilized a database of 427 MSI-H and 1590 MSS colonoscopy images and vision Transformer (ViT) with different feature training approaches to establish the MSI prediction model. The accuracy of combining pre-trained ViT features was 84% with an area under the receiver operating characteristic curve of 0.86, which was better than that of DenseNet201 (80%, 0.80) in the experiment with support vector machine. The content-based image retrieval (CBIR) approach showed that ViT features can obtain a mean average precision of 0.81 compared to 0.79 of DenseNet201. ViT reduced the issues that occur in convolutional neural networks, including limited receptive field and gradient disappearance, and may be better at interpreting diagnostic information around tumors and surrounding tissues. By using CBIR, the presentation of similar images with the same MSI status would provide more convincing deep learning suggestions for clinical use.

A computed tomography urography-based machine learning model for predicting preoperative pathological grade of upper urinary tract urothelial carcinoma

OBJECTIVES

Development and validation of a computed tomography urography (CTU)-based machine learning (ML) model for prediction of preoperative pathology grade of upper urinary tract urothelial carcinoma (UTUC).

METHODS

A total of 140 patients with UTUC who underwent CTU examination from January 2017 to August 2023 were retrospectively enrolled. Tumor lesions on the unenhanced, medullary, and excretory periods of CTU were used to extract Features, respectively. Feature selection was screened by the Pearson and Spearman correlation analysis, least absolute shrinkage and selection operator algorithm, random forest (RF), support vector machine (SVM), and eXtreme Gradient Boosting (XGBoost). The logistic regression (LR) was used to screen for independent influencing factors of clinical baseline characteristics. Machine learning models based on different feature datasets were constructed and validated using algorithms such as LR, RF, SVM, and XGBoost. By computing the selected features, a radiomics score was generated, and a diverse feature dataset was constructed. Based on the training set, 16 ML models were created, and their performance was evaluated using the validation set for metrics including sensitivity, specificity, accuracy, area under the receiver operating characteristic curve (AUC), and others.

RESULTS

The training set consisted of 98 patients (mean age: 64.5 ± 10.5 years; 30 males), whereas the validation set consisted of 42 patients (mean age: 65.3 ± 9.78 years; 17 males). Hydronephrosis was the best independent influence factor (p < 0.05). The RF model had the best performance in predicting high-grade UTUC, with AUC of 0.914 (95% Confidence Interval [95%CI] 0.852-0.977) and 0.903 (95%CI 0.809-0.997) in the training set and validation set, and accuracy of 0.878 and 0.857, respectively.

CONCLUSIONS

An ML model based on the RF algorithm exhibits excellent predictive performance, offering a non-invasive approach for predicting preoperative high-grade UTUC.

Preoperative contrast-enhanced CT imaging and clinicopathological characteristics analysis of mismatch repair-deficient colorectal cancer

BACKGROUND

Colorectal cancer (CRC) can develop through various pathogenetic pathways, and one of the primary pathways is high microsatellite instability (MSI-H)/deficient mismatch repair (dMMR). This study investigated the correlation between preoperative contrast-enhanced CT (CECT) and clinicopathologic characteristics of colorectal cancer (CRC) according to different mismatch repair (MMR) statuses.

METHODS

From April 2021 to July 2022, a total of 281 CRC patients with preoperative CECT and available MMR status were enrolled from a single centre for this retrospective study. Preoperative CECT features and clinicopathologic characteristics were analysed. Univariate and multivariate logistic regression analyses were used for statistical analysis. A nomogram was established based on the multivariate logistic regression results. Preoperative and postoperative dynamic nomogram prediction models were established. The C-index, a calibration plot, and clinical applicability of the two models were evaluated, and internal validation was performed using three methods.

RESULTS

In total, 249 patients were enrolled in the proficient mismatch repair (pMMR) group and 32 patients in the deficient mismatch repair (dMMR) group. In multivariate analysis, tumour location (right-hemi colon vs. left-hemi colon, odds ratio (OR) = 2.90, p = .036), the hypoattenuation-within-tumour ratio (HR) (HR > 2/3 vs. HR < 1/3, OR = 36.7, p < .001; HR 1/3-2/3 vs. HR < 1/3, OR = 6.05, p = .031), the number of lymph nodes with long diameter ≥ 8 mm on CECT (OR = 1.32, p = .01), CEA status (CEA positive vs. CEA negative, OR = 0.07, p = .002) and lymph node metastasis (OR = 0.45, p = .008) were independent risk factors for dMMR. Pre- and postoperative C-statistic were 0.861 and 0.908, respectively.

CONCLUSION

The combination of pre-operative CECT and clinicopathological characteristics of CRC correlates with MMR status, providing possible non-invasive MMR prediction. Particularly for dMMR CRC, tumour-draining lymph node status should be prudently evaluated by CECT.

Modeling the survival of colorectal cancer patients based on colonoscopic features in a feature ensemble vision transformer

The prognosis of patients with colorectal cancer (CRC) mostly relies on the classic tumor node metastasis (TNM) staging classification. A more accurate and convenient prediction model would provide a better prognosis and assist in treatment. From May 2014 to December 2017, patients who underwent an operation for CRC were enrolled. The proposed feature ensemble vision transformer (FEViT) used ensemble classifiers to benefit the combinations of relevant colonoscopy features from the pretrained vision transformer and clinical features, including sex, age, family history of CRC, and tumor location, to establish the prognostic model. A total of 1729 colonoscopy images were enrolled in the current retrospective study. For the prediction of patient survival, FEViT achieved an accuracy of 94 % with an area under the receiver operating characteristic curve of 0.93, which was better than the TNM staging classification (90 %, 0.83) in the experiment. FEViT reduced the limited receptive field and gradient disappearance in the conventional convolutional neural network and was a relatively effective and efficient procedure. The promising accuracy of FEViT in modeling survival makes the prognosis of CRC patients more predictable and practical.

CT-based deep learning model for the prediction of DNA mismatch repair deficient colorectal cancer: a diagnostic study

BACKGROUND

Stratification of DNA mismatch repair (MMR) status in patients with colorectal cancer (CRC) enables individual clinical treatment decision making. The present study aimed to develop and validate a deep learning (DL) model based on the pre-treatment CT images for predicting MMR status in CRC.

METHODS

1812 eligible participants (training cohort: n = 1124; internal validation cohort: n = 482; external validation cohort: n = 206) with CRC were enrolled from two institutions. All pretherapeutic CT images from three dimensions were trained by the ResNet101, then integrated by Gaussian process regression (GPR) to develop a full-automatic DL model for MMR status prediction. The predictive performance of the DL model was evaluated using the area under the receiver operating characteristic curve (AUC) and then tested in the internal and external validation cohorts. Additionally, the participants from institution 1 were sub-grouped by various clinical factors for subgroup analysis, then the predictive performance of the DL model for identifying MMR status between participants in different groups were compared.

RESULTS

The full-automatic DL model was established in the training cohort to stratify the MMR status, which presented promising discriminative ability with the AUCs of 0.986 (95% CI 0.971-1.000) in the internal validation cohort and 0.915 (95% CI 0.870-0.960) in the external validation cohort. In addition, the subgroup analysis based on the thickness of CT images, clinical T and N stages, gender, the longest diameter, and the location of tumors revealed that the DL model showed similar satisfying prediction performance.

CONCLUSIONS

The DL model may potentially serve as a noninvasive tool to facilitate the pre-treatment individualized prediction of MMR status in patients with CRC, which could promote the personalized clinical-making decision.

Challenges and Therapeutic Opportunities in the dMMR/MSI-H Colorectal Cancer Landscape

About 5 to 15% of all colorectal cancers harbor mismatch repair deficient/microsatellite instability-high status (dMMR/MSI-H) that associates with high tumor mutation burden and increased immunogenicity. As a result, and in contrast to other colorectal cancer phenotypes, a significant subset of dMMR/MSI-H cancer patients strongly benefit from immunotherapy. Yet, a large proportion of these tumors remain unresponsive to any immuno-modulating treatment. For this reason, current efforts are focused on the characterization of resistance mechanisms and the identification of predictive biomarkers to guide therapeutic decision-making. Here, we provide an overview on the new advances related to the diagnosis and definition of dMMR/MSI-H status and focus on the distinct clinical, functional, and molecular cues that associate with dMMR/MSI-H colorectal cancer. We review the development of novel predictive factors of response or resistance to immunotherapy and their potential application in the clinical setting. Finally, we discuss current and emerging strategies applied to the treatment of localized and metastatic dMMR/MSI-H colorectal tumors in the neoadjuvant and adjuvant setting.

Preoperative Prediction of Microsatellite Instability in Rectal Cancer Using Five Machine Learning Algorithms Based on Multiparametric MRI Radiomics

Objectives: To establish and verify radiomics models based on multiparametric MRI for preoperatively identifying the microsatellite instability (MSI) status of rectal cancer (RC) by comparing different machine learning algorithms. Methods: This retrospective study enrolled 383 (training set, 268; test set, 115) RC patients between January 2017 and June 2022. A total of 4148 radiomics features were extracted from multiparametric MRI, including T₂-weighted imaging, T₁-weighted imaging, apparent diffusion coefficient, and contrast-enhanced T₁-weighted imaging. The analysis of variance, correlation test, univariate logistic analysis, and a gradient-boosting decision tree were used for the dimension reduction. Logistic regression, Bayes, support vector machine (SVM), K-nearest neighbor (KNN), and tree machine learning algorithms were used to build different radiomics models. The relative standard deviation (RSD) and bootstrap method were used to quantify the stability of these five algorithms. Then, predictive performances of different models were assessed using area under curves (AUCs). The performance of the best radiomics model was evaluated using calibration and discrimination. Results: Among these 383 patients, the prevalence of MSI was 14.62% (56/383). The RSD value of logistic regression algorithm was the lowest (4.64%), followed by Bayes (5.44%) and KNN (5.45%), which was significantly better than that of SVM (19.11%) and tree (11.94%) algorithms. The radiomics model based on logistic regression algorithm performed best, with AUCs of 0.827 and 0.739 in the training and test sets, respectively. Conclusions: We developed a radiomics model based on the logistic regression algorithm, which could potentially be used to facilitate the individualized prediction of MSI status in RC patients.

Role of Machine Learning in Precision Oncology: Applications in Gastrointestinal Cancers

Gastrointestinal (GI) cancers, consisting of a wide spectrum of pathologies, have become a prominent health issue globally. Despite medical imaging playing a crucial role in the clinical workflow of cancers, standard evaluation of different imaging modalities may provide limited information. Accurate tumor detection, characterization, and monitoring remain a challenge. Progress in quantitative imaging analysis techniques resulted in "radiomics", a promising methodical tool that helps to personalize diagnosis and treatment optimization. Radiomics, a sub-field of computer vision analysis, is a bourgeoning area of interest, especially in this era of precision medicine. In the field of oncology, radiomics has been described as a tool to aid in the diagnosis, classification, and categorization of malignancies and to predict outcomes using various endpoints. In addition, machine learning is a technique for analyzing and predicting by learning from sample data, finding patterns in it, and applying it to new data. Machine learning has been increasingly applied in this field, where it is being studied in image diagnosis. This review assesses the current landscape of radiomics and methodological processes in GI cancers (including gastric, colorectal, liver, pancreatic, neuroendocrine, GI stromal, and rectal cancers). We explain in a stepwise fashion the process from data acquisition and curation to segmentation and feature extraction. Furthermore, the applications of radiomics for diagnosis, staging, assessment of tumor prognosis and treatment response according to different GI cancer types are explored. Finally, we discussed the existing challenges and limitations of radiomics in abdominal cancers and investigate future opportunities.

A nomogram model based on MRI and radiomic features developed and validated for the evaluation of lymph node metastasis in patients with rectal cancer

PURPOSE

The aim of this study was to develop and validate a nomogram model to evaluate lymph node metastasis (LNM) in patients with rectal cancer (RC).

METHODS

A total of 162 patients with RC were included in the study. The MRI reported model, the Radscore model, and the Complex model were constructed using the logistics regression (LR) algorithm. The DeLong test and decision curve analysis (DCA) were used to compare the prediction performance and clinical utility of these models. The nomogram model was constructed to visualize the prediction results of the best model. Model performance was evaluated in the training and validation groups, and the calibration curve and Hosmer-Lemeshow goodness of fit test were used to evaluate the calibration.

RESULT

All three models constructed by the LR algorithm were good at identifying LNM. The DeLong test and the DCA results showed that the Complex model outperformed the MRI reported model and the Radscore model in relation to their predictive performance and clinical utility. The nomogram of the Complex model had an area under the curve (AUC) of 0.902 (95% confidence interval (CI) 0.848-0.957) in the training group and an AUC of 0.891 (95% CI 0.799-0.983) in the validation group. Meanwhile, the nomogram showed good calibration.

CONCLUSION

The nomogram model constructed based on T2WI radiomics and MRI reported had good diagnostic efficacies for LNM in patients with RC, and provided a new auxiliary method for accurate and individualized clinical management.

Radiomics features based on internal and marginal areas of the tumor for the preoperative prediction of microsatellite instability status in colorectal cancer

Objectives

To explore whether the preoperative CT radiomics can predict the status of microsatellite instability (MSI) in colorectal cancer (CRC) patients and identify the region with the most stable and high-efficiency radiomics features.

Methods

This retrospective study involved 230 CRC patients with preoperative computed tomography scans and available MSI status between December 2019 and October 2021. Image segmentation and radiomic feature extraction were performed as follows. First, slices with the maximum tumor area (region of interest, ROI) were manually contoured. Subsequently, each ROI was shrunk inward by 1, 2, and 3 mm, respectively, where the remaining ROIs were considered as the internal region of the tumor (named as IROI1, IROI2, and IROI3), and the shrunk regions were considered as marginal regions of the tumor (named as MROI1, MROI2, and MROI3). Finally, radiomics features were extracted from each of the ROI. The intraclass correlation coefficient and least absolute shrinkage and selection operator method were used to choose the most reliable and relevant features of MSI status. Clinical, radiomics, and combined clinical radiomics models have been established. Calibration curve and decision curve analyses (DCA) were generated to explore the correction effect and assess the clinical applicability of the above models, respectively.

Results

In the testing cohort, the radiomics model based on IROI3 yielded the highest average area under the curve (AUC) value of 0.908, compared with the remaining radiomics models. Additionally, hypertension and N stage were considered as clinically independent factors of MSI status. The combined clinical radiomics model achieved excellent diagnostic efficacy (AUC: 0.928; sensitivity: 0.840; specificity: 0.867) in the testing cohort, as well as favorable calibration and clinical utility by calibration curve and DCA analyses.

Conclusions

The IROI3 model, which is based on a 3-mm shrink in the largest areas of the tumor, could noninvasively reflect the heterogeneity and genetic instability within the tumor. This suggests that it is an important biomarker for the preoperative prediction of MSI status. The model can extract more robust and effective radiomics features, which lays a foundation for the radiomics study of hollow organs, such as in CRC.

A Tumoral and Peritumoral CT-Based Radiomics and Machine Learning Approach to Predict the Microsatellite Instability of Rectal Carcinoma

Objective

To predict the status of microsatellite instability (MSI) of rectal carcinoma (RC) using different machine learning algorithms based on tumoral and peritumoral radiomics combined with clinicopathological characteristics.

Methods

There were 497 RC patients enrolled in this retrospective study. The tumoral and peritumoral CT-based radiomic features were calculated after tumor segmentation. The radiomic features from two radiologists were compared by way of inter-observer correlation coefficient (ICC). After methods of variance, correlation, and dimension reduction, six machine learning algorithms of logistic regression (LR), Bayes, support vector machine, random forest, k-nearest neighbor, and decision tree were conducted to develop models for predicting MSI status of RC. The relative standard deviation (RSD) was quantified. The radiomics and significant clinicopathological variables constituted the radiomics-clinicopathological nomogram. The receiver operator curve (ROC) was made by DeLong test, and the area under curve (AUC) with 95% confidence interval (95% CI) was calculated to evaluate the performance of the model.

Results

The venous phase of CT examination was selected for further analysis because the proportion of radiomic features with ICC greater than 0.75 was higher. The tumoral and peritumoral model by LR algorithm (M-LR) with minimal RSD showed good performance in predicting MSI status of RC with the AUCs of 0.817 and 0.726 in the training and validation set. The radiomic-clinicopathological nomogram performed better in both the training and validation set with AUCs of 0.843 and 0.737.

Conclusion

The radiomics-clinicopathological nomogram demonstrated better predictive performance in evaluating the MSI status of RC.